| Home |
| Search |
| Today's Posts |
|
#1
|
|||
|
|||
Revox Phase Inverter Circuit?
Hi All, I was surfing the web, and happened across the schematic for an interesting Revox Hi-Fi amplifier, you can find the schematic at this URL: http://www.audioantik.de/Stereo/revox.pdf The amplifier uses a pair of EL-34s as output tubes, a pentode input stage, and an ECC83 for the phase inverter. So far so good, sounds a little like the Mullard 5-20 circuit, but upon closer examination the ECC83 is not being used as a long tailed pair phase inverter, instead one section of the ECC83 is used as a cathode follower to drive the grid of one output tube, and the input to a floating paraphase phase inverter which drives the grid of the second output tube. Did any other manufacturer use this circuit, or did only Revox use it? What advantages might it posses that lead Revox to use it? Regards, John Byrns Surf my web pages at, http://users.rcn.com/jbyrns/ |
|
#2
|
|||
|
|||
|
"John Byrns" wrote in message ... Hi All, I was surfing the web, and happened across the schematic for an interesting Revox Hi-Fi amplifier, you can find the schematic at this URL: http://www.audioantik.de/Stereo/revox.pdf The amplifier uses a pair of EL-34s as output tubes, a pentode input stage, and an ECC83 for the phase inverter. So far so good, sounds a little like the Mullard 5-20 circuit, but upon closer examination the ECC83 is not being used as a long tailed pair phase inverter, instead one section of the ECC83 is used as a cathode follower to drive the grid of one output tube, and the input to a floating paraphase phase inverter which drives the grid of the second output tube. Did any other manufacturer use this circuit, or did only Revox use it? What advantages might it posses that lead Revox to use it? ** The phase inverter has its gain set by *resistors* at effectively unity. It also has a wide output voltage swing available - unlike a concertina circuit. Unlike a long tailed pair - gain matching of the two halves of the 12AX7 is non critical. Enough good things to appeal to some "Eric the Engineer" at Revox don't you think ?? ........... Phil |
|
#3
|
|||
|
|||
|
John Byrns wrote:
Hi All, I was surfing the web, and happened across the schematic for an interesting Revox Hi-Fi amplifier, you can find the schematic at this URL: http://www.audioantik.de/Stereo/revox.pdf The amplifier uses a pair of EL-34s as output tubes, a pentode input stage, and an ECC83 for the phase inverter. So far so good, sounds a little like the Mullard 5-20 circuit, but upon closer examination the ECC83 is not being used as a long tailed pair phase inverter, instead one section of the ECC83 is used as a cathode follower to drive the grid of one output tube, and the input to a floating paraphase phase inverter which drives the grid of the second output tube. Did any other manufacturer use this circuit, or did only Revox use it? What advantages might it posses that lead Revox to use it? Regards, John Byrns Surf my web pages at, http://users.rcn.com/jbyrns/ Don't know who else may have used it but gain of the phase inverter as a whole is only one. That may not be a problem here since a pentode with a rather high Rl is used as the input stage. Output capability of the entire phase inverter is double that of a concertina which it should be since there are double the cathodes at work. Output impedance of both phases is low & would be an advantage. Also noticed several phase compensation networks used. To me that is an indication of too much NFB resulting in stability problems. Would I use it? Maybe, but I can't think of where!!! Cheers, John Stewart |
|
#4
|
|||
|
|||
|
John Stewart wrote: John Byrns wrote: Hi All, I was surfing the web, and happened across the schematic for an interesting Revox Hi-Fi amplifier, you can find the schematic at this URL: http://www.audioantik.de/Stereo/revox.pdf The amplifier uses a pair of EL-34s as output tubes, a pentode input stage, and an ECC83 for the phase inverter. So far so good, sounds a little like the Mullard 5-20 circuit, but upon closer examination the ECC83 is not being used as a long tailed pair phase inverter, instead one section of the ECC83 is used as a cathode follower to drive the grid of one output tube, and the input to a floating paraphase phase inverter which drives the grid of the second output tube. Did any other manufacturer use this circuit, or did only Revox use it? What advantages might it posses that lead Revox to use it? Regards, John Byrns Surf my web pages at, http://users.rcn.com/jbyrns/ Don't know who else may have used it but gain of the phase inverter as a whole is only one. That may not be a problem here since a pentode with a rather high Rl is used as the input stage. Output capability of the entire phase inverter is double that of a concertina which it should be since there are double the cathodes at work. Output impedance of both phases is low & would be an advantage. Also noticed several phase compensation networks used. To me that is an indication of too much NFB resulting in stability problems. Would I use it? Maybe, but I can't think of where!!! Cheers, John Stewart I tried the audioantic site, and couldn't raise it, also via http://www.one-electron.com/FC_Consumer.html And couldn't raise the link. Maybe you could post it at abse, or direct. Patrick Turner. |
|
#5
|
|||
|
|||
|
In article , "Phil
Allison" wrote: "John Byrns" wrote in message ... Hi All, I was surfing the web, and happened across the schematic for an interesting Revox Hi-Fi amplifier, you can find the schematic at this URL: http://www.audioantik.de/Stereo/revox.pdf The amplifier uses a pair of EL-34s as output tubes, a pentode input stage, and an ECC83 for the phase inverter. So far so good, sounds a little like the Mullard 5-20 circuit, but upon closer examination the ECC83 is not being used as a long tailed pair phase inverter, instead one section of the ECC83 is used as a cathode follower to drive the grid of one output tube, and the input to a floating paraphase phase inverter which drives the grid of the second output tube. Did any other manufacturer use this circuit, or did only Revox use it? What advantages might it posses that lead Revox to use it? ** The phase inverter has its gain set by *resistors* at effectively unity. It also has a wide output voltage swing available - unlike a concertina circuit. Unlike a long tailed pair - gain matching of the two halves of the 12AX7 is non critical. Enough good things to appeal to some "Eric the Engineer" at Revox don't you think ?? No, that's not enough, Eric is smarter than that, there is one more important reason. Regards, John Byrns Surf my web pages at, http://users.rcn.com/jbyrns/ |
|
#7
|
|||
|
|||
|
In article , Patrick Turner
wrote: John Stewart wrote: John Byrns wrote: Hi All, I was surfing the web, and happened across the schematic for an interesting Revox Hi-Fi amplifier, you can find the schematic at this URL: http://www.audioantik.de/Stereo/revox.pdf The amplifier uses a pair of EL-34s as output tubes, a pentode input stage, and an ECC83 for the phase inverter. So far so good, sounds a little like the Mullard 5-20 circuit, but upon closer examination the ECC83 is not being used as a long tailed pair phase inverter, instead one section of the ECC83 is used as a cathode follower to drive the grid of one output tube, and the input to a floating paraphase phase inverter which drives the grid of the second output tube. Did any other manufacturer use this circuit, or did only Revox use it? What advantages might it posses that lead Revox to use it? Regards, John Byrns Surf my web pages at, http://users.rcn.com/jbyrns/ Don't know who else may have used it but gain of the phase inverter as a whole is only one. That may not be a problem here since a pentode with a rather high Rl is used as the input stage. Output capability of the entire phase inverter is double that of a concertina which it should be since there are double the cathodes at work. Output impedance of both phases is low & would be an advantage. Also noticed several phase compensation networks used. To me that is an indication of too much NFB resulting in stability problems. Would I use it? Maybe, but I can't think of where!!! Cheers, John Stewart I tried the audioantic site, and couldn't raise it, also via http://www.one-electron.com/FC_Consumer.html And couldn't raise the link. Maybe you could post it at abse, or direct. I don't know how to post at "abse", so I put a temporary copy he http://users.rcn.com/jbyrns/revox.pdf Regards, John Byrns Surf my web pages at, http://users.rcn.com/jbyrns/ |
|
#8
|
|||
|
|||
|
"John Byrns" wrote in message ... In article , "Phil Allison" wrote: ** The phase inverter has its gain set by *resistors* at effectively unity. It also has a wide output voltage swing available - unlike a concertina circuit. Unlike a long tailed pair - gain matching of the two halves of the 12AX7 is non critical. Enough good things to appeal to some "Eric the Engineer" at Revox don't you think ?? No, that's not enough, Eric is smarter than that, there is one more important reason. ** Ok - there is also very high impedance input from the cathode follower ( 20 Mohms or more ) suiting the pentode first stage and low impedance drive to both EL34 grids. BTW You did not say this was a quiz - that is devious. ......... Phil |
|
#9
|
|||
|
|||
|
Only thing I can think of is that the CF has an extremely high input
impedance, due to low Miller effect and the bootstrapped grid return. Lets you run the input pentode into a high anode load with recuded phase/frequency response issues. Note the 15pF capacitor from the anode of the first stage to ground! I also noted that grid current in the output tubes drives the phase inverter toward cutoff, but that would affect only the upper EL34. This amp might have odd overload characteristics. "John Byrns" wrote in message ... Hi All, I was surfing the web, and happened across the schematic for an interesting Revox Hi-Fi amplifier, you can find the schematic at this URL: http://www.audioantik.de/Stereo/revox.pdf The amplifier uses a pair of EL-34s as output tubes, a pentode input stage, and an ECC83 for the phase inverter. So far so good, sounds a little like the Mullard 5-20 circuit, but upon closer examination the ECC83 is not being used as a long tailed pair phase inverter, instead one section of the ECC83 is used as a cathode follower to drive the grid of one output tube, and the input to a floating paraphase phase inverter which drives the grid of the second output tube. Did any other manufacturer use this circuit, or did only Revox use it? What advantages might it posses that lead Revox to use it? Regards, John Byrns Surf my web pages at, http://users.rcn.com/jbyrns/ |
|
#10
|
|||
|
|||
|
In article , "Phil
Allison" wrote: "John Byrns" wrote in message ... In article , "Phil Allison" wrote: ** The phase inverter has its gain set by *resistors* at effectively unity. It also has a wide output voltage swing available - unlike a concertina circuit. Unlike a long tailed pair - gain matching of the two halves of the 12AX7 is non critical. Enough good things to appeal to some "Eric the Engineer" at Revox don't you think ?? No, that's not enough, Eric is smarter than that, there is one more important reason. ** Ok - there is also very high impedance input from the cathode follower ( 20 Mohms or more ) suiting the pentode first stage and low impedance drive to both EL34 grids. BTW You did not say this was a quiz - that is devious. It wasn't a quiz, but your list didn't seem like enough to get Eric to use that circuit, there had to be more, thanks for identifying it, now I know why Eric used it. Regards, John Byrns Surf my web pages at, http://users.rcn.com/jbyrns/ |
|
#11
|
|||
|
|||
|
John Byrns wrote: In article , Patrick Turner wrote: John Stewart wrote: John Byrns wrote: Hi All, I was surfing the web, and happened across the schematic for an interesting Revox Hi-Fi amplifier, you can find the schematic at this URL: http://www.audioantik.de/Stereo/revox.pdf The amplifier uses a pair of EL-34s as output tubes, a pentode input stage, and an ECC83 for the phase inverter. So far so good, sounds a little like the Mullard 5-20 circuit, but upon closer examination the ECC83 is not being used as a long tailed pair phase inverter, instead one section of the ECC83 is used as a cathode follower to drive the grid of one output tube, and the input to a floating paraphase phase inverter which drives the grid of the second output tube. Did any other manufacturer use this circuit, or did only Revox use it? What advantages might it posses that lead Revox to use it? Regards, John Byrns Surf my web pages at, http://users.rcn.com/jbyrns/ Don't know who else may have used it but gain of the phase inverter as a whole is only one. That may not be a problem here since a pentode with a rather high Rl is used as the input stage. Output capability of the entire phase inverter is double that of a concertina which it should be since there are double the cathodes at work. Output impedance of both phases is low & would be an advantage. Also noticed several phase compensation networks used. To me that is an indication of too much NFB resulting in stability problems. Would I use it? Maybe, but I can't think of where!!! Cheers, John Stewart I tried the audioantic site, and couldn't raise it, also via http://www.one-electron.com/FC_Consumer.html And couldn't raise the link. Maybe you could post it at abse, or direct. I don't know how to post at "abse", so I put a temporary copy he http://users.rcn.com/jbyrns/revox.pdf Regards, John Byrns Surf my web pages at, http://users.rcn.com/jbyrns/ Thankyou John, I am now able to see this circuit. There are no more phase correctiionn networks on this amp than what you may find elswhere, but they would be needed, because the pentode output stage has no R+C networks to limit the HF gain of the output tubes. The bottom opv is driven from a CF, but the top opv is driven from the anode of a floating parafeed triode, a 1/2 12AX7. The drive to the 12AX7 grid is around 33 k and the miller effect would cause a slightly early open loop roll off in drive to one, but not both the opvs. Cin of the opvs is low, so not much attentuation will happen in the opvs. The overload character would cause serious DV imbalances, since the drive to each opv comes from very different value impeadances, one from a CF of say 1k, and the other from an anode of 12AX7, of over 50 kohms. I prefer to have the driver Z equal to that of a medium U triode, and balanced for each opv. This is not of huge importance, since nothing is ever overloaded. I think this is a circuit was designed mainly to be original, rather than copy the established genre of Mullard and Williamson topologies, or be a favourable development from circuits of the day, and the Revox conveys no special benefits in engineering terms. The HF/LF stability would be mainly dependant on the quality of the OPT. We don't know what the quality is, ie, what the value of leakage inductance and stray shunt capacitance is, and if the OPTs are poor, many old designs like this oscillate well with loads of 0.22 uF. I think the design at http://www.turneraudio.com.au/htmlwe...0ulabinteg.htm addresses stability issues far better, and provide better symetrical drive. Note the networks which step the open loop gain at both extremes of the bandwidth. The zobel RC network offers mind damping at above 200 kHz, and I sometimes use say 0.002 plus 2k7 across the 1/2 primaries of the OPT, but since my OPTs have such low leakage inductance, I found this to be unnecessary. The same circuit can be used in principle with any sort of output tubes. Compensation networks must be carefully chosen to suit the OPT. Patrick Turner. |
|
#12
|
|||
|
|||
|
"Patrick Turner" Thankyou John, I am now able to see this circuit. The overload character would cause serious DV imbalances, since the drive to each opv comes from very different value impeadances, one from a CF of say 1k, and the other from an anode of 12AX7, of over 50 kohms. ** Hey - Turneriod !! You forgot to allow for the nearly 50 times NFB applied in the phase inverter - making its output Z the same as the CF. I think this is a circuit was designed mainly to be original, rather than copy the established genre of Mullard and Williamson topologies, or be a favourable development from circuits of the day, and the Revox conveys no special benefits in engineering terms. ** None that a mentally crippled autodidact can see - anyhow. ............. Phil |
|
#13
|
|||
|
|||
|
John Byrns wrote: In article , "Phil Allison" wrote: "John Byrns" wrote in message ... In article , "Phil Allison" wrote: ** The phase inverter has its gain set by *resistors* at effectively unity. It also has a wide output voltage swing available - unlike a concertina circuit. Unlike a long tailed pair - gain matching of the two halves of the 12AX7 is non critical. Enough good things to appeal to some "Eric the Engineer" at Revox don't you think ?? No, that's not enough, Eric is smarter than that, there is one more important reason. ** Ok - there is also very high impedance input from the cathode follower ( 20 Mohms or more ) suiting the pentode first stage and low impedance drive to both EL34 grids. BTW You did not say this was a quiz - that is devious. It wasn't a quiz, but your list didn't seem like enough to get Eric to use that circuit, there had to be more, thanks for identifying it, now I know why Eric used it. Regards, John Byrns Surf my web pages at, http://users.rcn.com/jbyrns/ In addition to what I said earlier about the Revox, I have more to add. Looking at the phase inverter triode, it is set up as an anode follower, but the two drive signals to the opv have a cascaded source. But both drive signals thus come from low impedance sources. I still think the overload would cause some asymetrical DV gymnastics within the amp. The circuit shown allows the pentode input stage to see low miller C, and uses two cascaded follower stages to produce drive signals. The bandwidth should be quite good. But a concertina would also work fine, and although the Ro from a concertina measures differently between anode and cathode. But because the loads on a concertina remain similar, including the opv miller C, the drive voltages are similar, so there is no need to have the extra stage of the Revox circuit. The open loop gain of the driver amp is only that of the pentode, and so it has to make the 15 volts drive for the opv grids , and the thd of the this pentode will be higher than if we used a medium U triode input tube, followed by an LTP using medium U triodes. The EF80 is a 6BX6 which has a Gm of up to 7mA/V at 10ma Iaq, but a lot lower gm at the 0.7 mA at which it operates here, so its gain could be somewhere between 100 and 200, since its load is just under 200k. Looking at the amp overall, its able to make around 25 watts into 7.5 ohms, which is 13.7 volts of output, which is divided down to make 2.28 volts of feedback. The open loop input voltage will be perhaps be around 0.1 volts if the pentode input tube has a gain of 150. Therefore around 2.4 volts are needed at the input for full power with FB connected. Thus the amount of global FB is around 27 dB. Such a large amount of NFB would require decent output trannies, but give a pretty decent reduction of the open loop thd and Ro of amp. Because there are two cascaded high NFB stages enclosed by the loop of global FB, ie, the CF and AF driver stages, HF instablity is of concern, and presumably the 15 pF between the EF80 anode and ground was found to limit HF gain, and the total C to 0V including stray C would be around 25pF, with Rout of the EF80 anode being about 180k, the HF pole of the pentode input stage would be 35 kHz, with a roll off at 6dB/octave. Ths would be better than a Mullard 520, where an EF86 with high Ro has to power a 12AX7 LTP with a miller C of about 50 pF. The Williamson driver stages have a BW of 400+kHz using 6SN7, and my gain + LTP around 200 kHz, using 6CG7. Modern practice more commonly uses 9 dB less FB, but up to around 30 dB of FB was common with pentode amps. The amount of FB applied increases as the load does, since pentode gain is determined by A = U x RL / ( Ra + RL ), and U is around 120, and when RL is high without any load connected, A = U, and this amp has far more than 27 dB effectively applied, because the open loop gain is 23 dB above the gain with 7.5 ohms at the output. I still cannot see any great superiority in the Revox circuit. Whether any superiority is to be heard is a moot point. And I have my reasons for prefering UL, CFB, or triode output stages, with medium mu triode drivers like 6SN7/6CG7. Revox probably bought a pile of EF80 cheaply, and of course 12AX7 were cheaper than 6SN7/6CG7. The Revox is nevertheless a clever circuit. Dynaco ST70 had no trouble using a pentode to drive a concertina PI connected to EL34. BTW, a poster said Unlike a long tailed pair - gain matching of the two halves of the 12AX7 is non critical. The gain matching of the two triodes of an LTP is not critical, as long as the tail is a CCS, and the anode load RLs are equal. With the Revox, the values of the bias resistors R41 and R51 which also form a shunt FB network ARE critical, to make sure the two drive signals are equal. The R values would need to be adjusted for balance. Patrick Turner. |
|
#14
|
|||
|
|||
|
"Patrick Turner" In addition to what I said earlier about the Revox, I have more to add. Looking at the phase inverter triode, it is set up as an anode follower, but the two drive signals to the opv have a cascaded source. But both drive signals thus come from low impedance sources. ** ****ing gobbledegook - the Turneroid's forte. But a concertina would also work fine, and although the Ro from a concertina measures differently between anode and cathode. ** Not when similarly loaded - which they normally re. I still cannot see any great superiority in the Revox circuit. ** As I listed: Immunity from mismatch in the 12AX7. Larger clean output swing hence THD. Lower drive impedance. Whether any superiority is to be heard is a moot point. ** The Turneroid's posturing stock answer to every damn thing he no clue about. Revox probably bought a pile of EF80 cheaply, and of course 12AX7 were cheaper than 6SN7/6CG7. The Revox is nevertheless a clever circuit. Dynaco ST70 had no trouble using a pentode to drive a concertina PI connected to EL34. ** More Turneroid drivel - all utterly uninformative. BTW, a poster said Unlike a long tailed pair - gain matching of the two halves of the 12AX7 is non critical. The gain matching of the two triodes of an LTP is not critical, as long as the tail is a CCS, and the anode load RLs are equal. ** More Turneroid bull****. With the Revox, the values of the bias resistors R41 and R51 which also form a shunt FB network ARE critical, ** So ****ing what - arsehole ??? They are not dependant on tube gain tolerance. to make sure the two drive signals are equal. The R values would need to be adjusted for balance. ** More Turneroid stinking lies. ................ Phil |
|
#15
|
|||
|
|||
|
There's no need for me to comment on Phil Allison's
foul mouthed & insulting reply to my post. He shows a complete lack of communication ability, and a poor knowledge of tube craft. Patrick Turner. Phil Allison wrote: "Patrick Turner" In addition to what I said earlier about the Revox, I have more to add. Looking at the phase inverter triode, it is set up as an anode follower, but the two drive signals to the opv have a cascaded source. But both drive signals thus come from low impedance sources. ** ****ing gobbledegook - the Turneroid's forte. But a concertina would also work fine, and although the Ro from a concertina measures differently between anode and cathode. ** Not when similarly loaded - which they normally re. I still cannot see any great superiority in the Revox circuit. ** As I listed: Immunity from mismatch in the 12AX7. Larger clean output swing hence THD. Lower drive impedance. Whether any superiority is to be heard is a moot point. ** The Turneroid's posturing stock answer to every damn thing he no clue about. Revox probably bought a pile of EF80 cheaply, and of course 12AX7 were cheaper than 6SN7/6CG7. The Revox is nevertheless a clever circuit. Dynaco ST70 had no trouble using a pentode to drive a concertina PI connected to EL34. ** More Turneroid drivel - all utterly uninformative. BTW, a poster said Unlike a long tailed pair - gain matching of the two halves of the 12AX7 is non critical. The gain matching of the two triodes of an LTP is not critical, as long as the tail is a CCS, and the anode load RLs are equal. ** More Turneroid bull****. With the Revox, the values of the bias resistors R41 and R51 which also form a shunt FB network ARE critical, ** So ****ing what - arsehole ??? They are not dependant on tube gain tolerance. to make sure the two drive signals are equal. The R values would need to be adjusted for balance. ** More Turneroid stinking lies. ............... Phil |
|
#16
|
|||
|
|||
|
John Byrns wrote:
Hi All, I was surfing the web, and happened across the schematic for an interesting Revox Hi-Fi amplifier, you can find the schematic at this U= RL: http://www.audioantik.de/Stereo/revox.pdf The amplifier uses a pair of EL-34s as output tubes, a pentode input stage, and an ECC83 for the phase inverter. So far so good, sounds a little like the Mullard 5-20 circuit, but upon closer examination the ECC83 is not being used as a long tailed pair phase inverter, instead o= ne section of the ECC83 is used as a cathode follower to drive the grid of= one output tube, and the input to a floating paraphase phase inverter which drives the grid of the second output tube. Did any other manufacturer use this circuit, or did only Revox use it? What advantages might it posses that lead Revox to use it? Regards, John Byrns Surf my web pages at, http://users.rcn.com/jbyrns/ Since there appears to be some difference of opinion within the NG as to = the strengths & weaknesses of the Revox circuit, here are my observations for= what they are worth. Please keep in mind these are my opinions only. Yours may= differ. If so, lets all hear why. =46rom my view of the Revox circuit I see the following. I count three HF phase correction circuits as follows- C14, R21??R14=85=85..This is the proper place to use phase correction. C15, R14=85=85..Another step in the voltage amp response C61, R61=85=85=85C61 should not be required There should be no RC time constants connected across the OP transformer= At the LF end I see C12, R15 & C13, R12 which hopefully have been selecte= d to stabilize NFB which includes the LF response of the OP transformer. Other= OP transformers would need a different combination of these to attain stabil= ity. Both OP tube grids are driven by a low impedance source, one by a CF, whi= le the other is driven by an anode follower whose gain is one. Unfortunately= , the anode follower includes an extra LF time constant. While the input impedance of the CF is high, it is driven by a 200K sourc= e so has minimal advantage in this respect over a concertina (split load) inve= rter. Most but not all concertinas are DC coupled to the previous stage, thus eliminating another LF time constant. Although the phase splitter arrangement has double the voltage output capability of the concertina most of that advantage has been given away through their use of a high mu triode 12AX7/ECC83 with it=92s limited cur= rent handling capability. Perhaps this arrangement was used to get around some= ones patent. RDH4 will tell you that there should be one only dominant phase correctio= n at each of the LF & HF ends of the amplifier spectrum, in a section on what = they refer too as =91staggering=92 the response. Norman Crowhurst, a well recognized guru of the art tells us the same thi= ng several times in many of his articles, one of which I posted on ABSE a wh= ile ago. If anyone is interested, I will post it again. For those who are una= ble to get ABSE, email me direct for a copy. As a result some amplifiers have what we call =93Conditional Stability=94= =2E In these situations the amplifier may have bursts of ultrasonic oscillation = & sometimes sub-audible =91breathing=92 These may not show up in ordinary t= esting since for the most part tests are done into a fixed resistive load. In th= e real world the load is complex, that being a loudspeaker in it=92s box. P= entodes & Beam Power tubes especially don=92t like complex loads. All of this is = said to result in listener fatigue. I built myself a loudspeaker simulator in ord= er to more fully test any of my designs. My view, Anyway!!!!!!!! Cheers, John Stewart |
|
#17
|
|||
|
|||
|
Correction- Since I prepared my note in MS Word 97, at least one of the s=
ymbols I used was not recognized by Netscape. The two question marks in one of the= lines should each be the symbol for a 'pipe', which would indicate a parallel c= onnexion of the resistors R21 & R14. Refered to the following line. C14, R21??R14=85=85..means R21 in parallel with R14. Another version of the subject circuit was published by E. Rodenhuis of P= hilips originally in 1957. Rodenhuis left out the CF & went straight to the OP s= tage. Phase inversion is accomplished by the triode connected pentode section o= f an ECL80. The included triode is connected as a 2MHZ oscillator providing bi= as for the output & driver stage, something we don't often see. Cheers, John Stewart |
|
#18
|
|||
|
|||
|
Having spent quite some time dealing with European engineering talent ( in
the electrical power industry), I feel it fair to say that the engineering discipline has a much higher "artistic" value in Europe than it does in the US. Doesn't necessarily mean that their approach is better or worse, just different and perhaps less cost-constrained. It happens on occasion, however, that the form impairs the function! Somewhere around here I have a schematic of a German-made FM receiver that I repaired once. You know about reflex amplifiers? This thing takes the cake. A pentagrid converter, but the LO runs at, let us say, 30 MHz for a received frequency of 100.7 MHz. They pick that up, run it through a crystal diode (!) tripler to 90 MHz and back into the bottle it goes. The plate circuit has coils and/or transformers tuned to 30, 90 and 10.7 MHz, along with a 60 MHz trap for undesired mixing products. The 60 and 90 MHz circuits are broadbanded. The image rejection is not perfect. "John Stewart" wrote in message ... Correction- Since I prepared my note in MS Word 97, at least one of the symbols I used was not recognized by Netscape. The two question marks in one of the lines should each be the symbol for a 'pipe', which would indicate a parallel connexion of the resistors R21 & R14. Refered to the following line. C14, R21??R14....means R21 in parallel with R14. Another version of the subject circuit was published by E. Rodenhuis of Philips originally in 1957. Rodenhuis left out the CF & went straight to the OP stage. Phase inversion is accomplished by the triode connected pentode section of an ECL80. The included triode is connected as a 2MHZ oscillator providing bias for the output & driver stage, something we don't often see. Cheers, John Stewart |
|
#19
|
|||
|
|||
|
In article , wrote:
Another version of the subject circuit was published by E. Rodenhuis of P= hilips originally in 1957. Rodenhuis left out the CF & went straight to the OP s= tage. Phase inversion is accomplished by the triode connected pentode section o= f an ECL80. The included triode is connected as a 2MHZ oscillator providing bi= as for the output & driver stage, something we don't often see. The circuit without the "CF" is common, for example H.H.Scott used a similar circuit in their stereo amplifiers through about 1961 when they switched to a concertina phase inverter. Without the "CF" it is just a plain old floating paraphase phase inverter driven by a pentode gain stage. It was the use of the "CF" that I thought was out of the ordinary in the Revox design. What are all the "=" signs in your posts recently about? They are most annoying as they seem to split words at the end of lines, sort of like a hyphen. Regards, John Byrns Surf my web pages at, http://users.rcn.com/jbyrns/ |
|
#20
|
|||
|
|||
|
John Byrns wrote:
In article , wrote: Another version of the subject circuit was published by E. Rodenhuis of P= hilips originally in 1957. Rodenhuis left out the CF & went straight to the OP s= tage. Phase inversion is accomplished by the triode connected pentode section o= f an ECL80. The included triode is connected as a 2MHZ oscillator providing bi= as for the output & driver stage, something we don't often see. The circuit without the "CF" is common, for example H.H.Scott used a similar circuit in their stereo amplifiers through about 1961 when they switched to a concertina phase inverter. Without the "CF" it is just a plain old floating paraphase phase inverter driven by a pentode gain stage. It was the use of the "CF" that I thought was out of the ordinary in the Revox design. What are all the "=" signs in your posts recently about? They are most annoying as they seem to split words at the end of lines, sort of like a hyphen. Regards, John Byrns Surf my web pages at, http://users.rcn.com/jbyrns/ Hi John- Not sure why the "=" sign shows up at your end. Since it is the end of a line perhaps it is a CR symbol not properly ID'ed. Just had another look here & perhaps it has to do with max line length set in your reader. My reader doesn't seem to care, that is no = sign shows up anywhere while reading from the NG's, my posts included. I'm using very old Netscape 4.7. Does your respond OK to the very long line lengths as posted by AJ? I will repost the above text again shorter & we will see what happens. Another version of the subject circuit was published by E. Rodenhuis of Philips originally in 1957. Rodenhuis left out the CF & went straight to the OP stage. Phase inversion is accomplished by the triode connected pentode section of an ECL80. The included triode is connected as a 2MHZ oscillator providing bias for the output & driver stage, something we don't often see. Cheers, John Stewart |
|
#21
|
|||
|
|||
|
In article , wrote:
John Byrns wrote: Hi All, I was surfing the web, and happened across the schematic for an interesting Revox Hi-Fi amplifier, you can find the schematic at this U= RL: http://www.audioantik.de/Stereo/revox.pdf The amplifier uses a pair of EL-34s as output tubes, a pentode input stage, and an ECC83 for the phase inverter. So far so good, sounds a little like the Mullard 5-20 circuit, but upon closer examination the ECC83 is not being used as a long tailed pair phase inverter, instead o= ne section of the ECC83 is used as a cathode follower to drive the grid of= one output tube, and the input to a floating paraphase phase inverter which drives the grid of the second output tube. Did any other manufacturer use this circuit, or did only Revox use it? What advantages might it posses that lead Revox to use it? Regards, John Byrns Surf my web pages at, http://users.rcn.com/jbyrns/ Since there appears to be some difference of opinion within the NG as to = the strengths & weaknesses of the Revox circuit, here are my observations for= what they are worth. Please keep in mind these are my opinions only. Yours may= differ. If so, lets all hear why. =46rom my view of the Revox circuit I see the following. I count three HF phase correction circuits as follows- C14, R21??R14=85=85..This is the proper place to use phase correction. C15, R14=85=85..Another step in the voltage amp response C61, R61=85=85=85C61 should not be required There should be no RC time constants connected across the OP transformer= At the LF end I see C12, R15 & C13, R12 which hopefully have been selecte= d to stabilize NFB which includes the LF response of the OP transformer. Other= OP transformers would need a different combination of these to attain stabil= ity. Both OP tube grids are driven by a low impedance source, one by a CF, whi= le the other is driven by an anode follower whose gain is one. Unfortunately= , the anode follower includes an extra LF time constant. While the input impedance of the CF is high, it is driven by a 200K sourc= e so has minimal advantage in this respect over a concertina (split load) inve= rter. Most but not all concertinas are DC coupled to the previous stage, thus eliminating another LF time constant. Although the phase splitter arrangement has double the voltage output capability of the concertina most of that advantage has been given away through their use of a high mu triode 12AX7/ECC83 with it=92s limited cur= rent handling capability. Perhaps this arrangement was used to get around some= ones patent. RDH4 will tell you that there should be one only dominant phase correctio= n at each of the LF & HF ends of the amplifier spectrum, in a section on what = they refer too as =91staggering=92 the response. Norman Crowhurst, a well recognized guru of the art tells us the same thi= ng several times in many of his articles, one of which I posted on ABSE a wh= ile ago. If anyone is interested, I will post it again. For those who are una= ble to get ABSE, email me direct for a copy. As a result some amplifiers have what we call =93Conditional Stability=94= =2E In these situations the amplifier may have bursts of ultrasonic oscillation = & sometimes sub-audible =91breathing=92 These may not show up in ordinary t= esting since for the most part tests are done into a fixed resistive load. In th= e real world the load is complex, that being a loudspeaker in it=92s box. P= entodes & Beam Power tubes especially don=92t like complex loads. All of this is = said to result in listener fatigue. I built myself a loudspeaker simulator in ord= er to more fully test any of my designs. My view, Anyway!!!!!!!! Cheers, John Stewart The Dynaco amps also use three high frequency compensation networks, two are the same as two of the Revox networks, with a third from the primary of the OPT back to somewhere I can't remember. Why do you say that "R61|||C61 should not be required", this seems to be the most prevalent of all forms of HF compensation used in tube Hi-Fi Amps? Presumably it is at least helpful. I guess it all depends on whether you are a leader or a laggard. You missed one of the LF networks, R45-C45. Regards, John Byrns Surf my web pages at, http://users.rcn.com/jbyrns/ |
|
#22
|
|||
|
|||
|
In article , wrote:
John Byrns wrote: In article , wrote: Another version of the subject circuit was published by E. Rodenhuis of P= hilips originally in 1957. Rodenhuis left out the CF & went straight to the OP s= tage. Phase inversion is accomplished by the triode connected pentode section o= f an ECL80. The included triode is connected as a 2MHZ oscillator providing bi= as for the output & driver stage, something we don't often see. The circuit without the "CF" is common, for example H.H.Scott used a similar circuit in their stereo amplifiers through about 1961 when they switched to a concertina phase inverter. Without the "CF" it is just a plain old floating paraphase phase inverter driven by a pentode gain stage. It was the use of the "CF" that I thought was out of the ordinary in the Revox design. What are all the "=" signs in your posts recently about? They are most annoying as they seem to split words at the end of lines, sort of like a hyphen. Regards, John Byrns Surf my web pages at, http://users.rcn.com/jbyrns/ Hi John- Not sure why the "=" sign shows up at your end. Since it is the end of a line perhaps it is a CR symbol not properly ID'ed. Just had another look here & perhaps it has to do with max line length set in your reader. My reader doesn't seem to care, that is no = sign shows up anywhere while reading from the NG's, my posts included. I'm using very old Netscape 4.7. Does your respond OK to the very long line lengths as posted by AJ? I will repost the above text again shorter & we will see what happens. Another version of the subject circuit was published by E. Rodenhuis of Philips originally in 1957. Rodenhuis left out the CF & went straight to the OP stage. Phase inversion is accomplished by the triode connected pentode section of an ECL80. The included triode is connected as a 2MHZ oscillator providing bias for the output & driver stage, something we don't often see. My antique News Reader, over 10 years old, responds perfectly to AJ's posts, I have line wrapping turned on. I do get the long lines if I save one of AJ's posts as a text file, and then look at it. As far as the equal signs in your posts, I guess that must be something my News Reader is adding when it doesn't understand a character, in addition to the equal signs at the ends of the lines, there also seem to be a number of other characters it doesn't understand, and has replaced by an equal sign followed by a numeric value for the character. I looked up the same post on Google, and all the strange characters were replaced by question marks rather than =nn as my News Reader is doing. Google has terminated the lines at the same places, in the middle of words, as my News Reader has, it just hasn't added the equal signs at the ends of the lines. All very strange, I have seen this on other posts from time to time before, but haven't the slightest idea what causes it. Looking at the Google version, I would say my News Reader isn't the problem, just that my News Reader reports more detailed information on the strange characters than other programs do, for example Google seems to just substitute question marks for the strange characters. Regards, John Byrns Surf my web pages at, http://users.rcn.com/jbyrns/ |
|
#23
|
|||
|
|||
|
John Stewart wrote: John Byrns wrote: Hi All, I was surfing the web, and happened across the schematic for an interesting Revox Hi-Fi amplifier, you can find the schematic at this URL: http://www.audioantik.de/Stereo/revox.pdf The amplifier uses a pair of EL-34s as output tubes, a pentode input stage, and an ECC83 for the phase inverter. So far so good, sounds a little like the Mullard 5-20 circuit, but upon closer examination the ECC83 is not being used as a long tailed pair phase inverter, instead one section of the ECC83 is used as a cathode follower to drive the grid of one output tube, and the input to a floating paraphase phase inverter which drives the grid of the second output tube. Did any other manufacturer use this circuit, or did only Revox use it? What advantages might it posses that lead Revox to use it? Regards, John Byrns Surf my web pages at, http://users.rcn.com/jbyrns/ Since there appears to be some difference of opinion within the NG as to the strengths & weaknesses of the Revox circuit, here are my observations for what they are worth. Please keep in mind these are my opinions only. Yours may differ. If so, lets all hear why. From my view of the Revox circuit I see the following. I count three HF phase correction circuits as follows- C14, R21??R14……..This is the proper place to use phase correction. C15, R14……..Another step in the voltage amp response C61, R61………C61 should not be required There should be no RC time constants connected across the OP transformer At the LF end I see C12, R15 & C13, R12 which hopefully have been selected to stabilize NFB which includes the LF response of the OP transformer. The time constants involved with these components would not be small enough to provide a stepped LF response beginning at say 20 Hz in the open loop response. I have such typical purpose made stepped response neworks in the coupling between V1 and the LTP in my amp at http://www.turneraudio.com.au/htmlwe...0ulabinteg.htm Other OP transformers would need a different combination of these to attain stability. Both OP tube grids are driven by a low impedance source, one by a CF, while the other is driven by an anode follower whose gain is one. Unfortunately, the anode follower includes an extra LF time constant. Its not a big deal, but increasing the time constants of bypassing used in the screen and catode circuits, with say 4 uF screen bypass, and 1,000 uF cathode bypass, with the network I use placed somewhere in the coupling will work better with an OPT with a marginal amount of Lp. Even with oodles of Lp, the LF gain stepper makes recovery after overload far less prone to oscillations. The stepped open loop response means the open loop LF phase shift at 10 Hz is reduced, along with the gain, so less FB is applied at F below 20 Hz, where FB is just not needed so much. There are HF compo elements in the amp with C14,R21,C15, and C61. With so much FB used, these values are critical, and I wouldn't mind betting the amp would oscillate with no RL, and a 0.22 uF across the output terminals. While the input impedance of the CF is high, it is driven by a 200K source so has minimal advantage in this respect over a concertina (split load) inverter. Most but not all concertinas are DC coupled to the previous stage, thus eliminating another LF time constant. Agreed. Although the phase splitter arrangement has double the voltage output capability of the concertina most of that advantage has been given away through their use of a high mu triode 12AX7/ECC83 with it’s limited current handling capability. Perhaps this arrangement was used to get around some ones patent. The CPI has slightly less voltage ability than the stages shown, but the output stage would be easy to drive, requiring only a max of 15v to each outout grid, and the CPI would have little more thd than the Revox set up. I'd prefer the CPI, anyday, and a 12AT7 would allow higher idle current. RDH4 will tell you that there should be one only dominant phase correction at each of the LF & HF ends of the amplifier spectrum, in a section on what they refer too as ‘staggering’ the response. Application of this is in all my amps. Norman Crowhurst, a well recognized guru of the art tells us the same thing several times in many of his articles, one of which I posted on ABSE a while ago. If anyone is interested, I will post it again. For those who are unable to get ABSE, email me direct for a copy. As a result some amplifiers have what we call “Conditional Stability”. In these situations the amplifier may have bursts of ultrasonic oscillation & sometimes sub-audible ‘breathing’ These may not show up in ordinary testing since for the most part tests are done into a fixed resistive load. In the real world the load is complex, that being a loudspeaker in it’s box. Pentodes & Beam Power tubes especially don’t like complex loads. All of this is said to result in listener fatigue. I built myself a loudspeaker simulator in order to more fully test any of my designs. The worst case loads are purely L or C. If the amp is stable into any L or C load, then its stable. Patrick Turner. My view, Anyway!!!!!!!! Cheers, John Stewart |
|
#24
|
|||
|
|||
|
John Byrns wrote:
The Dynaco amps also use three high frequency compensation networks, two are the same as two of the Revox networks, with a third from the primary of the OPT back to somewhere I can't remember. Sometimes a problem results from an imbalance of capacitance of the primary ends of a PP OP transformer to ground or to the secondary, even in a well made OPT. One of the fixes is as you have described, usually with a rather small capacitance. I remember actually managing to fix a problem like that with a twisted wire gimmick NFB. Why do you say that "R61|||C61 should not be required", this seems to be the most prevalent of all forms of HF compensation used in tube Hi-Fi Amps? Yes, that is how many try to stabilize the FB connexion, including myself for many years. However, after reading some of Crowhurst's stuff, I realized I was usually headed in the wrong direction. The cap should not be used as shown at C61. Since changing, my results have been much better. I will email you direct a copy of one of Crowhurst's articles which addresses the subject very well. Presumably it is at least helpful. I guess it all depends on whether you are a leader or a laggard. You missed one of the LF networks, R45-C45. Yes, you are correct. That one needs to be rather large since the amp may go Class AB. However, the frequency it sees is double that of the program material, which helps. Cheers, John Stewart Regards, John Byrns Surf my web pages at, http://users.rcn.com/jbyrns/ |
|
#25
|
|||
|
|||
|
"Patrick Turner" There's no need for me to comment on Phil Allison's foul mouthed & insulting reply to my post. ** You aint seen nothing yet - arsehole. ........... Phil |
|
#26
|
|||
|
|||
|
Patrick Turner wrote:
John Stewart wrote: John Byrns wrote: Hi All, I was surfing the web, and happened across the schematic for an interesting Revox Hi-Fi amplifier, you can find the schematic at th= is URL: http://www.audioantik.de/Stereo/revox.pdf The amplifier uses a pair of EL-34s as output tubes, a pentode inpu= t stage, and an ECC83 for the phase inverter. So far so good, sounds= a little like the Mullard 5-20 circuit, but upon closer examination t= he ECC83 is not being used as a long tailed pair phase inverter, inste= ad one section of the ECC83 is used as a cathode follower to drive the gri= d of one output tube, and the input to a floating paraphase phase invert= er which drives the grid of the second output tube. Did any other manufacturer use this circuit, or did only Revox use = it? What advantages might it posses that lead Revox to use it? Regards, John Byrns Surf my web pages at, http://users.rcn.com/jbyrns/ Since there appears to be some difference of opinion within the NG as= to the strengths & weaknesses of the Revox circuit, here are my observations= for what they are worth. Please keep in mind these are my opinions only. Yours= may differ. If so, lets all hear why. From my view of the Revox circuit I see the following. I count three HF phase correction circuits as follows- C14, R21??R14=85=85..This is the proper place to use phase correctio= n. C15, R14=85=85..Another step in the voltage amp response C61, R61=85=85=85C61 should not be required There should be no RC time constants connected across the OP transfo= rmer At the LF end I see C12, R15 & C13, R12 which hopefully have been sel= ected to stabilize NFB which includes the LF response of the OP transformer. The time constants involved with these components would not be small en= ough to provide a stepped LF response beginning at say 20 Hz in the open loo= p response. I have such typical purpose made stepped response neworks in the coupli= ng between V1 and the LTP in my amp at http://www.turneraudio.com.au/htmlwe...0ulabinteg.htm Other OP transformers would need a different combination of these to attain st= ability. Both OP tube grids are driven by a low impedance source, one by a CF,= while the other is driven by an anode follower whose gain is one. Unfortuna= tely, the anode follower includes an extra LF time constant. Its not a big deal, but increasing the time constants of bypassing used= in the screen and catode circuits, with say 4 uF screen bypass, and 1,000 uF cathode bypass, with the network I use placed somewhere in the coupling will work better with an OPT with a marginal amount of = Lp. Even with oodles of Lp, the LF gain stepper makes recovery after overlo= ad far less prone to oscillations. The stepped open loop response means the open loop LF phase shift at 10 Hz is reduced, along with the gain, so less FB is applied at F below 20 Hz, where FB is just not needed so much. There are HF compo elements in the amp with C14,R21,C15, and C61. With so much FB used, these values are critical, and I wouldn't mind betting the amp would oscillate with no RL, and a 0.22 uF across the output terminals. While the input impedance of the CF is high, it is driven by a 200K s= ource so has minimal advantage in this respect over a concertina (split load) = inverter. Most but not all concertinas are DC coupled to the previous stage, th= us eliminating another LF time constant. Agreed. Although the phase splitter arrangement has double the voltage output= capability of the concertina most of that advantage has been given aw= ay through their use of a high mu triode 12AX7/ECC83 with it=92s limited= current handling capability. Perhaps this arrangement was used to get around = some ones patent. The CPI has slightly less voltage ability than the stages shown, but the output stage would be easy to drive, requiring only a max of 15v to each outout grid, and the CPI would have little more thd than th= e Revox set up. I'd prefer the CPI, anyday, and a 12AT7 would allow higher idle current= =2E RDH4 will tell you that there should be one only dominant phase corre= ction at each of the LF & HF ends of the amplifier spectrum, in a section on w= hat they refer too as =91staggering=92 the response. Application of this is in all my amps. Norman Crowhurst, a well recognized guru of the art tells us the same= thing several times in many of his articles, one of which I posted on ABSE = a while ago. If anyone is interested, I will post it again. For those who are= unable to get ABSE, email me direct for a copy. As a result some amplifiers have what we call =93Conditional Stabilit= y=94. In these situations the amplifier may have bursts of ultrasonic oscillat= ion & sometimes sub-audible =91breathing=92 These may not show up in ordina= ry testing since for the most part tests are done into a fixed resistive load. I= n the real world the load is complex, that being a loudspeaker in it=92s bo= x. Pentodes & Beam Power tubes especially don=92t like complex loads. All of this= is said to result in listener fatigue. I built myself a loudspeaker simulator in= order to more fully test any of my designs. The worst case loads are purely L or C. If the amp is stable into any L or C load, then its stable. Patrick Turner. My view, Anyway!!!!!!!! Cheers, John Stewart Now Patrick, please pay attention. When John B posted the Revox schematic= , he asked for our view of the circuit, not your view of my circuit critiqu= e as above or your view of the opinions of others. I have already stated in my opening that my opinion is my own & possibly not that of others. You are of course entitled to your own thoughts in re= spect to the circuit & that is the information that John suggested to be posted= , not your opinion of what others may have in mind. In your various posts you appear from my vantage point to be in many ways defending your own circuits, which is a waste of space since it is unlikely that any of us will ever buy one of your amps. Most of us are quite able to build our own amp & probably have done just that on many occasions. The NG is not a good place to market your amps. I read your posts now only occasionally since I find them far to long. If you are to hold the interest of others you will need to get their atte= ntion quickly & then have something of interest to say in order to maintain tha= t. In most cases when you are making a point, better leave all the arithmeti= c for another time. Perhaps you should consider writing a book. This post is not meant to insult you but rather it is a wakeup call, something like you might get from your manager if you were not self employed as you are. Just some thoughts, John Stewart |
|
#27
|
|||
|
|||
|
My view, Anyway!!!!!!!! Cheers, John Stewart Now Patrick, please pay attention. When John B posted the Revox schematic, he asked for our view of the circuit, not your view of my circuit critique as above or your view of the opinions of others. I am paying attention, and what you see is what you get, and I can only see things, ie, have a view, expressed in terms of comparisons between what is presented for comment, and my own creations, which include the product of my opinions, to give an explanation of why things are the way they are. I do sometimes have to round down the comments of folks like Allison and Jute because they post so much abuse, and I am a juicy target. I have already stated in my opening that my opinion is my own & possibly not that of others. You are of course entitled to your own thoughts in respect to the circuit & that is the information that John suggested to be posted, not your opinion of what others may have in mind. In your various posts you appear from my vantage point to be in many ways defending your own circuits, which is a waste of space since it is unlikely that any of us will ever buy one of your amps. I don't expect folks here to buy my product, and none have. private emails suggest that plenty have built my designs, and profited with the great sound they got, and the clear guidance contained in power supply and transformer info. I don't consider my attitude a waste of time. I am here as a maker of fine audio amps, and there is a lot of thought behind what I do, which I recommend folks follow, should they wish to build something for themselves, or to sell to others. I "defend" my circuits. Why would I not? Nobody has convinced me that there is anything sub-standard in my technical approach. I have gradually learned a lot by being here during the last 3 years. Most of us are quite able to build our own amp & probably have done just that on many occasions. The NG is not a good place to market your amps. I read your posts now only occasionally since I find them far to long. I don't worry if folks find the length of the posts too long. I try to make the posts contain the detail I think is needed. If I have but one person reading something I wrote, its enough. What folks build is their business, but I hope they benefit from my ideas. If you are to hold the interest of others you will need to get their attention quickly & then have something of interest to say in order to maintain that. In most cases when you are making a point, better leave all the arithmetic for another time. Perhaps you should consider writing a book. One reason I established the website, which is a sort of book online, is to provide answers for those who ask all these questions here, and to save me the work of repeating myself, and wearing myself out. I rarely dismiss someone's idea about doing something and I often just try to help them understand what they are doing and why. There are quite a few other splendid websites which perform the same purpose, and of course some fine books, like RDH4, Tremain's Audio Encyclopedia, Morgan Jones, Ben Duncan's book on SS amps, etc, etc. This post is not meant to insult you but rather it is a wakeup call, something like you might get from your manager if you were not self employed as you are. One minute ppl are complaining that RAT has gone to the dogs because its dominated with the neurotic rantings of Allison and Jute, and just a few others like myself who want to abuse nobody and just talk tubes. Now, when I do start saying quite a lot about tubes, with the Revox amp being the vehicle for discussing analysis, I get told I am saying too much. Its almost like I can't win. My manager is that part of me which controls the rest of me, and he does take some getting use to.... Now if we had 20 ppl like yourself all posting long detailed discussions, and even promoting their favourite circuits, and comparing them amoung themselves, analysing without fear or catowing to anyone, or calling each other autistics et all, we'd have a real discussion group. Just some thoughts, John Stewart I welcome them, Patrick Turner. |
|
#28
|
|||
|
|||
|
John Stewart wrote:
: John Byrns wrote: : What are all the "=" signs in your posts recently about? They are most : annoying as they seem to split words at the end of lines, sort of like a : hyphen. : Hi John- Not sure why the "=" sign shows up at your end. Since it : is the end of a line perhaps it is a CR symbol not properly ID'ed. Header line from your "problem" posts: Content-Type: text/plain; charset=iso-8859-1 Header line from your "normal" posts: Content-Type: text/plain; charset=us-ascii The problem may be related to your preparing that note in MS Word 97. Perhaps you pasted directly from it into NS news, and the charset changed? If you save it as a text file and copy/paste that, it might not occur. That's a guess, of course, but the differences in headers are clear. I don't know the NS news options, but it should be & stay set to us-ascii. : =46rom my view of the Revox circuit I see the following. : : I count three HF phase correction circuits as follows- : C14, R21??R14=85=85..This is the proper place to use phase correction. : C15, R14=85=85..Another step in the voltage amp response : C61, R61=85=85=85C61 should not be required : There should be no RC time constants connected across the OP transformer= etc. is the way it looks here in Agent when sent with charset=iso-8859-1 Google doesn't see the =85 or EOL = characters in your originals, but others do, as you can see from some of the followups, and it does render the ?? marks.The EOL = seems like a soft CR but it breaks the word up. Hope that helps. (The keyboard "|" pipe symbol still works in us-ascii:-) Otherwise, it's been an interesting discussion. Thanks. Ross |
|
#29
|
|||
|
|||
|
Patrick Turner said:
The HF/LF stability would be mainly dependant on the quality of the OPT. We don't know what the quality is, ie, what the value of leakage inductance and stray shunt capacitance is, and if the OPTs are poor, many old designs like this oscillate well with loads of 0.22 uF. The Revox amps I've seen (PP EL84) , used very good quality OPTs with a C-core, comparable to the Philips AD 9058 tranny. This Philips transformer allows up to 30 dB of NFB in a standard Mullard (= Philips!) application with 2 x EL84, so this says something about the winding quality. Not that I would use 30 dBs of NFB these days :-) -- Sander deWaal Vacuum Audio Consultancy |
|
#30
|
|||
|
|||
|
In article , wrote:
Now Patrick, please pay attention. When John B posted the Revox schematic= , he asked for our view of the circuit, not your view of my circuit critiqu= e as above or your view of the opinions of others. Well actually while I didn't explicitly state it, I also implicitly wanted any relevant views on the opinions of others, as I hoped to gain some factual understanding of the circuit, and unsupported, undebated opinions aren't always very useful. I have already stated in my opening that my opinion is my own & possibly not that of others. You are of course entitled to your own thoughts in re= spect to the circuit & that is the information that John suggested to be posted= , not your opinion of what others may have in mind. On the contrary, relevant comments on the opinions of others can be quite helpful in gaining a fuller understanding. In your various posts you appear from my vantage point to be in many ways defending your own circuits, which is a waste of space since it is unlikely that any of us will ever buy one of your amps. Most of us are quite able to build our own amp & probably have done just that on many occasions. The NG is not a good place to market your amps. Oh, Oh, it sounds like you have thrown in with the AJ anti Turner campaign, which I still don't understand. In any case why shouldn't he defend his circuits? It helps in understanding the operation of the circuits, and where Patrick is coming from, even if I don't always agree with his "opinions". Also there are really two issues with peoples so called opinions, first is it technically correct, or not, which is ultimately a mostly black or white issue, and second does one agree subjectively with the results achieved by the circuit/idea. This last is purely a matter of personal preference and values, and as such is not an absolute. I read your posts now only occasionally since I find them far to long. If you are to hold the interest of others you will need to get their atte= ntion quickly & then have something of interest to say in order to maintain tha= t. In most cases when you are making a point, better leave all the arithmeti= c for another time. Perhaps you should consider writing a book. The arithmetic is the best part of Patrick's posts! I haven't been paying attention to the exact form of Patrick's posts, but your suggestions for how to write them are good, except I would add one additional element. After he has first gotten their attention, and then said something of interest, that holds their attention, he should finish up with a final section that covers the arithmetic, and this last section is very important, although by keeping it a separate section those who aren't interested can stop reading after the interest maintaining part. You wouldn't want him to write like Crowhurst in that paper you emailed me, shoveling out a lot of unsubstantiated opinions that may or may not be correct, and in any case most assuredly don't give the whole story. But more on that in another post. This post is not meant to insult you but rather it is a wakeup call, something like you might get from your manager if you were not self employed as you are. I would think his manager is talking to him most everyday. Just some thoughts, John Stewart I thought it was your premise not to give "thoughts" on other peoples thoughts? Regards, John Byrns Surf my web pages at, http://users.rcn.com/jbyrns/ |
|
#31
|
|||
|
|||
|
Looks to me that the biggest problem would be the uneven impedance
presented to the output tubes. One triode drives as a cathode follower, the other as what amounts to an anode follower. Jack John Byrns wrote: Hi All, I was surfing the web, and happened across the schematic for an interesting Revox Hi-Fi amplifier, you can find the schematic at this URL: http://www.audioantik.de/Stereo/revox.pdf The amplifier uses a pair of EL-34s as output tubes, a pentode input stage, and an ECC83 for the phase inverter. So far so good, sounds a little like the Mullard 5-20 circuit, but upon closer examination the ECC83 is not being used as a long tailed pair phase inverter, instead one section of the ECC83 is used as a cathode follower to drive the grid of one output tube, and the input to a floating paraphase phase inverter which drives the grid of the second output tube. Did any other manufacturer use this circuit, or did only Revox use it? What advantages might it posses that lead Revox to use it? Regards, John Byrns Surf my web pages at, http://users.rcn.com/jbyrns/ |
|
#32
|
|||
|
|||
|
"Jack Crenshaw" wrote in message ink.net... Looks to me that the biggest problem would be the uneven impedance presented to the output tubes. One triode drives as a cathode follower, the other as what amounts to an anode follower. ** You have made the same dumb error the Turneroid charlatan did. What about the effect of negative feedback on that second triode???? It operates at unity gain. .............. Phil |
|
#33
|
|||
|
|||
|
I thought the impedance of an anode follower was of similar magnitude to a cathode follower, they are both "followers" after all. Regards, John Byrns In article . net, Jack Crenshaw wrote: Looks to me that the biggest problem would be the uneven impedance presented to the output tubes. One triode drives as a cathode follower, the other as what amounts to an anode follower. Jack John Byrns wrote: Hi All, I was surfing the web, and happened across the schematic for an interesting Revox Hi-Fi amplifier, you can find the schematic at this URL: http://www.audioantik.de/Stereo/revox.pdf The amplifier uses a pair of EL-34s as output tubes, a pentode input stage, and an ECC83 for the phase inverter. So far so good, sounds a little like the Mullard 5-20 circuit, but upon closer examination the ECC83 is not being used as a long tailed pair phase inverter, instead one section of the ECC83 is used as a cathode follower to drive the grid of one output tube, and the input to a floating paraphase phase inverter which drives the grid of the second output tube. Did any other manufacturer use this circuit, or did only Revox use it? What advantages might it posses that lead Revox to use it? Regards, John Byrns Surf my web pages at, http://users.rcn.com/jbyrns/ Surf my web pages at, http://users.rcn.com/jbyrns/ |
|
#34
|
|||
|
|||
|
John Byrns wrote: I thought the impedance of an anode follower was of similar magnitude to a cathode follower, they are both "followers" after all. Regards, The Ro from a CF is approximately 1/Gm, so 12AX7 at 0.7mA idle current has Ra effectively at 666 ohms. See page 318, RDH4. The cathode load RL is so large it won't affect this figure. The anode follower has its gain reduced to near unity, and its Rout is also reduced to a low figure, but it isn't quite as low as the CF, but low enough to give little practical difference to the operation of the circuit in this case. grid current will cause flattening of the positve going voltage peaks on the CF side, since the 12AX7 hasn't got much ability to work into grid current, and the coupling cap will charge up to increase the bias during overload; its the rectification effect. The negative going signal is not flattened, but applied to the anode follower, and this inverts the phase, and a similar change to bias occurs at grid current, but but one that is symetrical. Its not important with a hi-fi amp. With EL34 pentode operation, rail clipping occurs well before grid current. Like the Dynaco, the Revox circuit has to make all the driver gain in the pentode input stage, but as I explained in an earlier post, the EF80/6BX6 should manage to make a lot of gain, and as much as a pair of cascaded 6SN7/12AU7, but with less bandwidth and more distortion. The Dynaco bootstraps the load seen by the pentode input tube to increase its gain, and the Revox just uses a higher Gm pentode. Quad II use a pair of pentodes, the rather low Gm EF86, and have about 1/2 the FB in the output stage, and 1/2 as global, thus using even more drive to the output tubes from a high impedance source, but it still works OK. IN effect the output tube driver source impedance is reduced by the global FB, because if you were to restrict the voltage input to one of the opv grids, the FB would try to cause the distorted signal so that it is corrected, which is the same if the driver impedance had been reduced. The Revox uses a lot of global FB applied, for a tube amp, to get Ro of this pentode amp to a low level of less than 1 ohm. Patrick Turner. John Byrns In article . net, Jack Crenshaw wrote: Looks to me that the biggest problem would be the uneven impedance presented to the output tubes. One triode drives as a cathode follower, the other as what amounts to an anode follower. Jack John Byrns wrote: Hi All, I was surfing the web, and happened across the schematic for an interesting Revox Hi-Fi amplifier, you can find the schematic at this URL: http://www.audioantik.de/Stereo/revox.pdf The amplifier uses a pair of EL-34s as output tubes, a pentode input stage, and an ECC83 for the phase inverter. So far so good, sounds a little like the Mullard 5-20 circuit, but upon closer examination the ECC83 is not being used as a long tailed pair phase inverter, instead one section of the ECC83 is used as a cathode follower to drive the grid of one output tube, and the input to a floating paraphase phase inverter which drives the grid of the second output tube. Did any other manufacturer use this circuit, or did only Revox use it? What advantages might it posses that lead Revox to use it? Regards, John Byrns Surf my web pages at, http://users.rcn.com/jbyrns/ Surf my web pages at, http://users.rcn.com/jbyrns/ |
|
#35
|
|||
|
|||
|
"Patrick Turner" wrote in message ... John Byrns wrote: I thought the impedance of an anode follower was of similar magnitude to a cathode follower, they are both "followers" after all. The anode follower has its gain reduced to near unity, and its Rout is also reduced to a low figure, ** Now that Phil Allison pointed this out the stinking Turneroid adopts it as his own a few days later. Typical deceitful Turneroid SOTE stunt. Quad II use a pair of pentodes, the rather low Gm EF86, and have about 1/2 the FB in the output stage, and 1/2 as global, thus using even more drive to the output tubes from a high impedance source, but it still works OK. IN effect the output tube driver source impedance is reduced by the global FB, because if you were to restrict the voltage input to one of the opv grids, the FB would try to cause the distorted signal so that it is corrected, which is the same if the driver impedance had been reduced. ** Insane Turneroid autistic drivel. Low drive source impedance extends the amp's open loop bandwidth - something NFB **cannot** do. ............ Phil |
|
#36
|
|||
|
|||
|
As I recall, the output impedance of the anode follower is twice that of
the cathode follower. Still low, but different. But I think the two circuits would react quite differently when feeding an output tube that's drawing significant grid current. In short, I think the problem would be what the circuit sounds like when driven into overload. Jack John Byrns wrote: I thought the impedance of an anode follower was of similar magnitude to a cathode follower, they are both "followers" after all. Regards, John Byrns In article . net, Jack Crenshaw wrote: Looks to me that the biggest problem would be the uneven impedance presented to the output tubes. One triode drives as a cathode follower, the other as what amounts to an anode follower. Jack John Byrns wrote: Hi All, I was surfing the web, and happened across the schematic for an interesting Revox Hi-Fi amplifier, you can find the schematic at this URL: http://www.audioantik.de/Stereo/revox.pdf The amplifier uses a pair of EL-34s as output tubes, a pentode input stage, and an ECC83 for the phase inverter. So far so good, sounds a little like the Mullard 5-20 circuit, but upon closer examination the ECC83 is not being used as a long tailed pair phase inverter, instead one section of the ECC83 is used as a cathode follower to drive the grid of one output tube, and the input to a floating paraphase phase inverter which drives the grid of the second output tube. Did any other manufacturer use this circuit, or did only Revox use it? What advantages might it posses that lead Revox to use it? Regards, John Byrns Surf my web pages at, http://users.rcn.com/jbyrns/ Surf my web pages at, http://users.rcn.com/jbyrns/ |
|
#37
|
|||
|
|||
|
"Jack Crenshaw" wrote in message hlink.net... As I recall, the output impedance of the anode follower is twice that of the cathode follower. Still low, but different. But I think the two circuits would react quite differently when feeding an output tube that's drawing significant grid current. In short, I think the problem would be what the circuit sounds like when driven into overload. ** A hi-fi amp should not be driven into overload - excepting maybe on brief signal peaks. More than that produces very audible distortion - which makes the sound no longer hi-fi. Maybe you were thinking of guitar amps where overdriving is the norm. .......... Phil |
|
#38
|
|||
|
|||
|
HI
most amps have a push pull driver after the Phase splitter? Any using triodes do as the drive voltage is high. El34's as triodes maybe don't need them. Phil Lawrence "Phil Allison" wrote in message u... "Jack Crenshaw" wrote in message hlink.net... As I recall, the output impedance of the anode follower is twice that of the cathode follower. Still low, but different. But I think the two circuits would react quite differently when feeding an output tube that's drawing significant grid current. In short, I think the problem would be what the circuit sounds like when driven into overload. ** A hi-fi amp should not be driven into overload - excepting maybe on brief signal peaks. More than that produces very audible distortion - which makes the sound no longer hi-fi. Maybe you were thinking of guitar amps where overdriving is the norm. ......... Phil |
|
#39
|
|||
|
|||
|
Phil Lawrence wrote: HI most amps have a push pull driver after the Phase splitter? Any using triodes do as the drive voltage is high. El34's as triodes maybe don't need them. Phil Lawrence About 10 years ago, a gentleman by the name of Clive Locke introduced me to the idea of triodes without feedback, and one of his favourites was a quad of EL34 in triode, with a paralleled 6DJ8 used as the concertina PI as a driver, and another 6DJ8 paralleled and used as an input gain stage with a CCS used as the RL, to minimise V1 thd. The amp without FB was sensitive to about a volt. The 30 watts of class A was very listenable, and with a 5k : 5ohm OPT, Ro was about 1.4 ohms. I didn't stay with that idea for long, as I found careful application of FB with UL circuits worked and sounded probably better, used with simple 6CG7 input triode, and an LTP as the driver, again with 6CG7. The best 6CG7 have the same internals as a 6SN7. They usually sound warmer than 6DJ8, don't ask me why. Its easy to get 80vrms in a normal plate loaded gain stage with 6SN7, 6DJ8, 12AU7, 12BH7, 6CG7, etc, with a load of say 50k, and with a sufficiently high supply voltage. Splitting the load evenly into the cathode and anode circuits means that you can get 40 vrms from anode and cathode, and there is plenty to drive EL34, which might be biased at -28v, so 20vrms drive is plenty, since EL34 don't like AB2 operation. A pair of 300B in the output PP stage would need lots more drive, because the U is only 4, not 8 to10, like EL34. Patrick Turner. "Phil Allison" wrote in message u... "Jack Crenshaw" wrote in message hlink.net... As I recall, the output impedance of the anode follower is twice that of the cathode follower. Still low, but different. But I think the two circuits would react quite differently when feeding an output tube that's drawing significant grid current. In short, I think the problem would be what the circuit sounds like when driven into overload. ** A hi-fi amp should not be driven into overload - excepting maybe on brief signal peaks. More than that produces very audible distortion - which makes the sound no longer hi-fi. Maybe you were thinking of guitar amps where overdriving is the norm. ......... Phil |
| Reply |
| Thread Tools | |
| Display Modes | |
|
|
Similar Threads
|
||||
| Thread | Forum | |||
| Doppler Distortion - Fact or Fiction | Pro Audio | |||
| Positive Feedback in P-P 6V6 Amp | Vacuum Tubes | |||
| Transient response of actively filtered speakers | Tech | |||
| Blindtest question | High End Audio | |||
| Negative/Positive Phase Shift in a Transformer | Pro Audio | |||
Linear Mode
