Home |
Search |
Today's Posts |
#161
Posted to rec.audio.tubes
|
|||
|
|||
On the origin and workings of "Fast" Power supplies
On 12 Dec 2006 19:21:20 -0800, "Bret Ludwig"
wrote: Use the relay to short the output to ground, rather than open it. The power amp should have circuitry that protects it from running into a dead short. If THAT fails, better to blow up the output transistors than a potentially $10K+ speaker system Crowbars, including solid-state ones using SCR's, have been used since the 1970's, for example, the McIntosh MA6100. They do have their limitations, as you've noted, and their advantages. To take advantage of crowbars, the designer must integrate them with a slow-start regimen to prevent turn-on transients (under *all* conditions) and with a soft turn-off (under *all* conditions). Do-able, obviously, but not trivial. For critical use, the contact resistance of a relay idling for years(?) is too problematical, so it's never done. All good fortune, Chris Hornbeck "Too soon oldt; too late schmardt." |
#162
Posted to rec.audio.tubes
|
|||
|
|||
"Fast" Power supplies
"flipper" wrote in message
On Tue, 12 Dec 2006 18:18:54 -0500, "Arny Krueger" wrote: "flipper" wrote in message A passive component can cause nonlinear distortion - a good example would be a diode. Of course. So can a transformer and, as that article shows, a capacitor. However, if the distortion only arises under operational conditions that never arise in typical use, then it is irrelevant. It was relevant to the topic, as was explained in what you snipped out as a prerequisite to making that supercilious comment. My comment was precipitated by the fact that you tried to introduce irrelevant information, and then declined to take responsbility for that act. |
#163
Posted to rec.audio.tubes
|
|||
|
|||
On the origin and workings of "Fast" Power supplies
"Chris Hornbeck" wrote in
message news On 12 Dec 2006 19:21:20 -0800, "Bret Ludwig" wrote: Use the relay to short the output to ground, rather than open it. The power amp should have circuitry that protects it from running into a dead short. If THAT fails, better to blow up the output transistors than a potentially $10K+ speaker system Crowbars, including solid-state ones using SCR's, have been used since the 1970's, for example, the McIntosh MA6100. They do have their limitations, as you've noted, and their advantages. To take advantage of crowbars, the designer must integrate them with a slow-start regimen to prevent turn-on transients (under *all* conditions) and with a soft turn-off (under *all* conditions). Do-able, obviously, but not trivial. For critical use, the contact resistance of a relay idling for years(?) is too problematical, so it's never done. Dry-contact relays with sealed contacts (e.g. reed relays) or current-handling relays with self-wiping contacts made of reasonable materials, are extremely reliable. They just take a little care to properly specify, and a little money to obtain. |
#164
Posted to rec.audio.tubes
|
|||
|
|||
"Fast" Power supplies
"Eeyore" wrote in message ... : : : Ruud Broens wrote: : : "Eeyore" wrote in message : : Ruud Broens wrote: : : : : : : ...opinions you get to have :-) : : i've kept an eye on B&W over the years, as they've repeatedly used the same : : design considerations i used in my diy speakers - that inspires confidence : :-) : : alas, haven't yet heard the 802D's, yet : : : : Do they use driver impedance correction ? : : : : Graham : : : nah, they bought some guitar speakers from a surplus outlet, : glued 'm into a cardboard box, threw in some parts from the : junk bin for good measure and : that's about it : : Ah - you're a funny man too. : : Let me guess.... You really don't know but felt obliged to say something daft : anyway ? just to give you an idea of how silly your line of questioning is, consider eg. DIY magazines like Hobby Hifi, a german bimonthly, where the Kit Pro 17.5 2 way is presented ...Audax Aerogel low/mid, Bohlender magnetostat tweeter the developer, Rainer Kroenke, has one of the largest 'quiet rooms' in industry, he used B&Kj equipment, MLSSA , the works final result: 50Hz - 15kHz on axis virtually flat fr, measurement data on the Bohlender-Graebener Neo 3 PDR-W/FPtweeter MLSSA waterfall diagram over nearly all of 400 Hz/40 kHz dead as a doormat in 1 ms *some worst case 2ms around minor resonances in the tweeter 2-3 kHz, formidable indeed data, measured for the Audax HM 170 MNO very good cascading waterfall spectrum, matching the tweeter at around 2.5 kHz, fortunately enabling crossover freq there crossover fixes some minor details and uses Mundorf caps oohh, impedance correction was implemented in the filter of this 250 euro/speaker project, as well :-) Rudy somewhat better informed |
#165
Posted to rec.audio.tubes
|
|||
|
|||
"Fast" Power supplies
"Ruud Broens" wrote in message ... : : "Eeyore" wrote in message : ... : : ...DIY magazines like Hobby Hifi, a german bimonthly, : where the Kit Pro 17.5 2 way is presented : ..Audax Aerogel low/mid, Bohlender magnetostat tweeter : the developer, Rainer Kroenke, has one of the largest : 'quiet rooms' in industry, he used B&Kj equipment, : MLSSA , the works : final result: : 50Hz - 15kHz on axis virtually flat fr, : : measurement data on the Bohlender-Graebener Neo 3 PDR-W/FPtweeter : MLSSA waterfall diagram over nearly all of 400 Hz/40 kHz : dead as a doormat in 1 ms *some worst case 2ms around : minor resonances in the tweeter 2-3 kHz, formidable indeed : : data, measured for the Audax HM 170 MNO : very good cascading waterfall spectrum, matching : the tweeter at around 2.5 kHz, fortunately enabling : crossover freq there : : crossover fixes some minor details and uses : Mundorf caps : : oohh, impedance correction was implemented in the filter : of this 250 euro/speaker project, as well :-) : : Rudy : somewhat better informed : see he http://www.hobbyhifi.de/Aktuell/Aktu...e_ausgabe.html (if you're serious about speakers, learn to read German! the above lines being a condensation of a 9 page article, lot's of graphs, construction plan & notes ...the usual German gruendlichkeit, eh :-) R. |
#166
Posted to rec.audio.tubes
|
|||
|
|||
On the origin and workings of "Fast" Power supplies
Eeyore wrote: "Bob H." wrote: No , they don't need a speaker *DC* protection. IME, the relays needed for speaker protection of direct-coupled amplifiers, have, over the service life of good equipment, *more* of a sonic penalty and much more of a reliability penalty than a good transformer. And that's saying a lot. How about just using a high grade audio cap instead of the relay? If you want to get serious, a large film cap? The whole raison d'etre of DC coupled amps is to get rid of that cap. 2000uf polypropylene btw ? Graham You say that, but I do have one 470uF polypropylene. Four of those? You may ask if they are big........... |
#167
Posted to rec.audio.tubes
|
|||
|
|||
On the origin and workings of "Fast" Power supplies
Bob H. wrote: Eeyore wrote: 2000uf polypropylene btw ? Graham MY Harmon Kardon uses a 100uf cap, and it's the only one in the signal chain. Since it's my only SS amp, it's what I'm familiar with. And I will be replacing them with at least Silmacs, and probably foil caps when I get around to it. Actually, I think it's 1000uf, and yes it's the coupling cap to the speakers. I remember a factor of 10. I'll check the schematic again when I get home. But how about 1 or 2kuf silmac or similar, bypassed by a 100 or 50uf film? Does anyone build low voltage foil caps? (lower than speaker caps). They could be wound with thinner film, for smaller size, and for SS, much lower voltages are needed. Bob H. |
#168
Posted to rec.audio.tubes
|
|||
|
|||
On the origin and workings of "Fast" Power supplies
"Bob H." wrote: Bob H. wrote: Eeyore wrote: 2000uf polypropylene btw ? Graham MY Harmon Kardon uses a 100uf cap, and it's the only one in the signal chain. Since it's my only SS amp, it's what I'm familiar with. And I will be replacing them with at least Silmacs, and probably foil caps when I get around to it. Actually, I think it's 1000uf, and yes it's the coupling cap to the speakers. I remember a factor of 10. I'll check the schematic again when I get home. But how about 1 or 2kuf silmac or similar, bypassed by a 100 or 50uf film? Is this a 'single supply' amplifier design ? The idea of bypassing is yet more unfounded nonsense btw. I'm not sure what the great idea's supposed to be with silk included in the dielectric. I'd use a modern low-ESR design. It'll help to use a larger value too like the 2000 uF I mentioned. Does anyone build low voltage foil caps? (lower than speaker caps). They could be wound with thinner film, for smaller size, and for SS, much lower voltages are needed. The lowest voltage film caps I've normally seen are 50V. There's a limit to how thin the film can be made reliably. Graham |
#169
Posted to rec.audio.tubes
|
|||
|
|||
On the origin and workings of "Fast" Power supplies
"Eeyore" wrote in message ... : : : "Bob H." wrote: : : Bob H. wrote: : Eeyore wrote: : : 2000uf polypropylene btw ? : : Graham : : MY Harmon Kardon uses a 100uf cap, and it's the only one in the signal : chain. Since it's my only SS amp, it's what I'm familiar with. And I : will be replacing them with at least Silmacs, and probably foil caps : when I get around to it. : : Actually, I think it's 1000uf, and yes it's the coupling cap to the : speakers. I remember a factor of 10. I'll check the schematic again : when I get home. But how about 1 or 2kuf silmac or similar, bypassed : by a 100 or 50uf film? : : Is this a 'single supply' amplifier design ? : : The idea of bypassing is yet more unfounded nonsense btw. : : I'm not sure what the great idea's supposed to be with silk included in the : dielectric. I'd use a modern low-ESR design. It'll help to use a larger value : too like the 2000 uF I mentioned. : : : Does anyone build low voltage foil caps? (lower than speaker caps). : They could be wound with thinner film, for smaller size, and for SS, : much lower voltages are needed. : : The lowest voltage film caps I've normally seen are 50V. There's a limit to how : thin the film can be made reliably. : : Graham a modern stacked polyprop can easily have some 2500 layers. as the thickness of the metallic layer can be neglected, a 10 mm tall stack makes for a 4 um dielectric thickness. the odd and even layers are end connected by what is called the Schoops process. quite advanced manufacturing at work :-) 50V across 4 um is like 12 KV/mm, so limits are imposed by the dielectric's punch through capability Rudy |
#170
Posted to rec.audio.tubes
|
|||
|
|||
On the origin and workings of "Fast" Power supplies
MY Harmon Kardon uses a 100uf cap, and it's the only one in the signal chain. Since it's my only SS amp, it's what I'm familiar with. And I will be replacing them with at least Silmacs, and probably foil caps when I get around to it. My mistake. The 100uf caps are at the input of the amp section. The ouputs are direct coupled. It's been a couple of years since I looked at the schematic, and just made a mental note to look into alternatives for those caps. Bob H. |
#171
Posted to rec.audio.tubes
|
|||
|
|||
On the origin and workings of "Fast" Power supplies
"Bob H." wrote: MY Harmon Kardon uses a 100uf cap, and it's the only one in the signal chain. Since it's my only SS amp, it's what I'm familiar with. And I will be replacing them with at least Silmacs, and probably foil caps when I get around to it. My mistake. The 100uf caps are at the input of the amp section. The ouputs are direct coupled. It's been a couple of years since I looked at the schematic, and just made a mental note to look into alternatives for those caps. Is it in the feedback loop by any chance ? Graham |
#172
Posted to rec.audio.tubes
|
|||
|
|||
On the origin and workings of "Fast" Power supplies
Eeyore said:
MY Harmon Kardon uses a 100uf cap, and it's the only one in the signal chain. Since it's my only SS amp, it's what I'm familiar with. And I will be replacing them with at least Silmacs, and probably foil caps when I get around to it. My mistake. The 100uf caps are at the input of the amp section. The ouputs are direct coupled. It's been a couple of years since I looked at the schematic, and just made a mental note to look into alternatives for those caps. Is it in the feedback loop by any chance ? A DC-blocking cap at the input? Not likely. -- - Ever seen someone with 5.1 ears? So, what does that tell you? - |
#173
Posted to rec.audio.tubes
|
|||
|
|||
On the origin and workings of "Fast" Power supplies
Sander deWaal wrote: A DC-blocking cap at the input? Not likely. It's between the high cut filter and the input to the amp stage, and it's definatley a signal cap. Looking throught the tone control section, I see another 33uf cap if the tone controls are bypassed, a 220u if not. Also, a 10u after the volume pot to the stage prior to the tone controls. So I guess there are at least 3 caps in the signal stage. |
#174
Posted to rec.audio.tubes
|
|||
|
|||
On the origin and workings of "Fast" Power supplies
"Bob H." said:
A DC-blocking cap at the input? Not likely. It's between the high cut filter and the input to the amp stage, and it's definatley a signal cap. My comment was directed at Graham, he asked if the cap was somewhere in the feedback loop. That could only be possible, if the feedback resistor from the base of the inverting input of the differential pair is connected to ground with a cap. IMO, that's not a harmful place ;-) Looking throught the tone control section, I see another 33uf cap if the tone controls are bypassed, a 220u if not. Also, a 10u after the volume pot to the stage prior to the tone controls. So I guess there are at least 3 caps in the signal stage. The reason for the 100 uF being that large, is that in effect, all caps are in series, with some resistors and pots inbetween. This gives several different corner frequencies. -- - Ever seen someone with 5.1 ears? So, what does that tell you? - |
#175
Posted to rec.audio.tubes
|
|||
|
|||
On the origin and workings of "Fast" Power supplies
Sander deWaal wrote: "Bob H." said: A DC-blocking cap at the input? Not likely. It's between the high cut filter and the input to the amp stage, and it's definatley a signal cap. My comment was directed at Graham, he asked if the cap was somewhere in the feedback loop. That could only be possible, if the feedback resistor from the base of the inverting input of the differential pair is connected to ground with a cap. IMO, that's not a harmful place ;-) It's potentially as equally harmful as the input cap actually. Looking throught the tone control section, I see another 33uf cap if the tone controls are bypassed, a 220u if not. Also, a 10u after the volume pot to the stage prior to the tone controls. So I guess there are at least 3 caps in the signal stage. The reason for the 100 uF being that large, is that in effect, all caps are in series, with some resistors and pots inbetween. This gives several different corner frequencies. If someone cares to offer me the working voltages at those point and the input impedances I can easily work out if the use of zero-bias electrolytics is 'harmful' or not. Graham |
#176
Posted to rec.audio.tubes
|
|||
|
|||
On the origin and workings of "Fast" Power supplies
Eeyore said:
That could only be possible, if the feedback resistor from the base of the inverting input of the differential pair is connected to ground with a cap. IMO, that's not a harmful place ;-) It's potentially as equally harmful as the input cap actually. I disagree. The signal amplitude at the inverting input is attenuated by the feedback voltage divider, and there's a fairly constant DC offset voltage across it. The input cap has to deal with an unknown DC polarity at the input and higher signal voltages, hence the need for a relatively large value and a bipolar electrolytic. I use Nichicon Muse ES caps for this application. -- - Ever seen someone with 5.1 ears? So, what does that tell you? - |
#177
Posted to rec.audio.tubes
|
|||
|
|||
"Fast" Power supplies
4 Graham:
: : "Eeyore" wrote in message http://www.bwspeakers.com/downloadFi...opment_ Paper "What is not so clear-cut is why different capacitors, with ostensibly the same specification, can sound so different from one another. The difficulty in mapping physical properties to the perceived performance characteristics further compounds the problem." R. extra bulletin |
#178
Posted to rec.audio.tubes
|
|||
|
|||
"Fast" Power supplies
"Ruud Broens" wrote in message
4 Graham: "Eeyore" wrote in message http://www.bwspeakers.com/downloadFi...opment_ Paper "What is not so clear-cut is why different capacitors, with ostensibly the same specification, can sound so different from one another. The difficulty in mapping physical properties to the perceived performance characteristics further compounds the problem." Just another example of the deep pit that people can put themselves in with sighted, non-level-matched, non-time-synched listening evaluations. |
#179
Posted to rec.audio.tubes
|
|||
|
|||
"Fast" Power supplies
One of my concerns about massive C's in the power supply chain comes from
the fact that the HV supply to ground ends up participating in the signal path. Hagerman uses 470nF bypasses on his 47uF PSU caps to overcome this problem--and if one uses larger caps, then I would suspect a double by-pass is needed to get the same result. I prefer the choke input filters--as less C is needed to get equivalent ripple. The other issue of choke input filters is that the output of the choke is a pure DC offest sin wave--and it gets around the HF noise that comes from a C input PSU. Of course a larger (and more expensive) choke is needed for choke input filters. Another plus: if you are using a FW rectifier the choke input effectively knocks out rectifer switching noise as well. The question of "fast" is something I have also wondered about--and one of these days I'll do a real world test of it. The point is that you want your HT not to sag on high current peaks. And you don't want it to ring on recovery from peaks. Theoretically, an infinitely large reservoir would fit the bill--thus favoring the long C chain. But if your reservoir is not approximately infinite then the large time constant is going to work against recovery from long peak draw periods. Thus, a fast recovery system (small C's) can work as well if properly designed to eliminate ringing. The other element that helps a fast recovery system is obviously a power transformer that is overspec--and thus doesn't contribute to the recovery time via saturation (during peaks). In practice I've noticed that commercially designed PSU on moderate priced products tend to be running at close to the limits--thus saving cost. These amps, in my experience, benefit from a beefer PSU of either variety. YMMV--have fun and try a few different approaches. "Iain Churches" wrote in message ... There seem to be two schools of thought regarding linear PSUs in valve tube/amps. Some favour a long chain with high capacitance electrolytics. Others favour a simple C-L-C pi filter with caps of 47uF or so, saying that a "fast" psu sounds better. With valve rectifiers, one is limited in the value of the first electrolytic, but with FW bridges, these limitations no longer apply. What type of supply do RATs prefer? Does a "fast" supply sound different/better? If so, how does this improvement manifest itself? regards to all. Iain |
#180
Posted to rec.audio.tubes
|
|||
|
|||
"Fast" Power supplies
In article ,
"coffeedj" wrote: One of my concerns about massive C's in the power supply chain comes from the fact that the HV supply to ground ends up participating in the signal path. Hagerman uses 470nF bypasses on his 47uF PSU caps to overcome this problem--and if one uses larger caps, then I would suspect a double by-pass is needed to get the same result. I prefer the choke input filters--as less C is needed to get equivalent ripple. The other issue of choke input filters is that the output of the choke is a pure DC offest sin wave--and it gets around the HF noise that comes from a C input PSU. Are you sure it is really a "pure DC offest sin wave"? I was under the impression that the output of a choke input PSU also included at least the odd harmonics of the sine wave? Of course a larger (and more expensive) choke is needed for choke input filters. Another plus: if you are using a FW rectifier the choke input effectively knocks out rectifer switching noise as well. The question of "fast" is something I have also wondered about--and one of these days I'll do a real world test of it. The point is that you want your HT not to sag on high current peaks. And you don't want it to ring on recovery from peaks. Theoretically, an infinitely large reservoir would fit the bill--thus favoring the long C chain. But if your reservoir is not approximately infinite then the large time constant is going to work against recovery from long peak draw periods. Thus, a fast recovery system (small C's) can work as well if properly designed to eliminate ringing. The other element that helps a fast recovery system is obviously a power transformer that is overspec--and thus doesn't contribute to the recovery time via saturation (during peaks). "Peaks" on the load side of a power transformer don't cause saturation, if anything they tend to reduce the possibility of transformer saturation. Regards, John Byrns -- Surf my web pages at, http://fmamradios.com/ |
#181
Posted to rec.audio.tubes
|
|||
|
|||
"Fast" Power supplies
That is what one of the texts say--and it appeared to be true on the DSO.
I'll have to take a second look though to be sure. "John Byrns" wrote in message ... In article , "coffeedj" wrote: One of my concerns about massive C's in the power supply chain comes from the fact that the HV supply to ground ends up participating in the signal path. Hagerman uses 470nF bypasses on his 47uF PSU caps to overcome this problem--and if one uses larger caps, then I would suspect a double by-pass is needed to get the same result. I prefer the choke input filters--as less C is needed to get equivalent ripple. The other issue of choke input filters is that the output of the choke is a pure DC offest sin wave--and it gets around the HF noise that comes from a C input PSU. Are you sure it is really a "pure DC offest sin wave"? I was under the impression that the output of a choke input PSU also included at least the odd harmonics of the sine wave? Of course a larger (and more expensive) choke is needed for choke input filters. Another plus: if you are using a FW rectifier the choke input effectively knocks out rectifer switching noise as well. The question of "fast" is something I have also wondered about--and one of these days I'll do a real world test of it. The point is that you want your HT not to sag on high current peaks. And you don't want it to ring on recovery from peaks. Theoretically, an infinitely large reservoir would fit the bill--thus favoring the long C chain. But if your reservoir is not approximately infinite then the large time constant is going to work against recovery from long peak draw periods. Thus, a fast recovery system (small C's) can work as well if properly designed to eliminate ringing. The other element that helps a fast recovery system is obviously a power transformer that is overspec--and thus doesn't contribute to the recovery time via saturation (during peaks). "Peaks" on the load side of a power transformer don't cause saturation, if anything they tend to reduce the possibility of transformer saturation. Regards, John Byrns -- Surf my web pages at, http://fmamradios.com/ |
Reply |
Thread Tools | |
Display Modes | |
|
|
Similar Threads | ||||
Thread | Forum | |||
FA: Vintage Neve Console Modules, Panels, Power Supplies + much more | Pro Audio | |||
here is how firewire ports fail | Pro Audio | |||
List of NOS mostly tubes | Vacuum Tubes | |||
"The Audibility of Power Supplies" | Vacuum Tubes | |||
FS: SOUNDSTREAM CLOSEOUTS AND MORE!! | Car Audio |