Home |
Search |
Today's Posts |
#1
|
|||
|
|||
Tube Amp Power Supply
Hello Gurus!
I've completed my amplifier testing (EL34 based) and noted that the voltage supply drops from 480VDC to 440VDC over a period of 1.5hours. Of course the power transformer was taking quite long to get hot (still touchable) and after 1.5hours, the supply voltage remains at 440VDC. When the supply voltage drops, the bias voltage across the cathod resistor for the EL34 drops from 0.75VDC to 0.65VDC and stays there. Is this normal for a tube amplifier to be behaving this way when the output tranformer gets hot? Is it possible to keep everything stable using a LM317 floating voltage regulator in place of a choke? Regards, Mark |
#2
|
|||
|
|||
F.H.Chan wrote:
after 1.5hours, the supply voltage remains at 440VDC. When the supply voltage drops, the bias voltage across the cathod resistor for the EL34 drops from 0.75VDC to 0.65VDC and stays there. Hi, quick questions. o What is the topology of your output circuit, single, push-pull, triode-connection or pentode-connection? o What is the exact biasing method for the EL34? o Is your cathode register very hot? I was kinda confused for the cathode resistor voltage of 0.75-0.65[V]. If this is the actual bias voltage, it is too low for EL34 with plate supply voltage of over 400[V]. So I assume that you are also applying grid bias voltage somehow, but I don't know for sure. Atsunori |
#3
|
|||
|
|||
Hi Atsunori,
Hi, quick questions. o What is the topology of your output circuit, single, push-pull, triode-connection or pentode-connection? Its a push-pull amplifier o What is the exact biasing method for the EL34? o Is your cathode register very hot? I was kinda confused for the cathode resistor voltage of 0.75-0.65[V]. If this is the actual bias voltage, it is too low for EL34 with plate supply voltage of over 400[V]. So I assume that you are also applying grid bias voltage somehow, but I don't know for sure. Yes there's a grid bias (negative) voltage applied. The cathod resistor is connected to ground and it was biased at 0.75V (to determine current through the tube). Nope the resistor doesn't get hot at all. Regards, Mark |
#4
|
|||
|
|||
Mark wrote:
Its a push-pull amplifier I see. So the cathodes of two EL34's are connected together, and is grounded through a resistor, over which you notice the change of voltage from 0.75[V] to 0.65[V] after a while. Yes there's a grid bias (negative) voltage applied. The cathod resistor is connected to ground and it was biased at 0.75V (to determine current through the tube). Nope the resistor doesn't get hot at all. Sounds like the operating current for the output stage actually decreased after 1.5 hours, mean while the supply voltage also dropped from 480 to 440v. I would then check the followings over the 1.5 hours of operation as the plate supply voltage changes. o If the primary line voltage(AC 110v or whatever) measured at the power transformer changes. o If the secondary high voltage from the power transformer changes. o If you are using rectifier tube(s), see if the voltage drop(s) across your rectifier tube(s) change. Atsunori |
#5
|
|||
|
|||
"F.H.Chan" wrote in message
... Hello Gurus! I've completed my amplifier testing (EL34 based) and noted that the voltage supply drops from 480VDC to 440VDC over a period of 1.5hours. Of course the power transformer was taking quite long to get hot (still touchable) and after 1.5hours, the supply voltage remains at 440VDC. When the supply voltage drops, the bias voltage across the cathod resistor for the EL34 drops from 0.75VDC to 0.65VDC and stays there. Is this normal for a tube amplifier to be behaving this way when the output tranformer gets hot? Is it possible to keep everything stable using a LM317 floating voltage regulator in place of a choke? Regards, Mark Long term drift problems can be difficult to troubleshoot. Perhaps this thought is radical (maybe Atsunori can express his opinion on this), but can the output stage be disconnected and a large resistor (dummy load) be hung in its place and then monitor the PS to see if there is still drift? Much can then be eliminated as the culprit. Back to earth. For long-term temperature drift type problems, I try to hasten the time by alternating a heat gun and cold (Freon type) spray. If you then narrowed the problem to a particular area, the long nozzle tube can be directed to individual components. Food for thought (may give you indigestion though). Cordially, west |
#6
|
|||
|
|||
west wrote: "F.H.Chan" wrote in message ... Hello Gurus! I've completed my amplifier testing (EL34 based) and noted that the voltage supply drops from 480VDC to 440VDC over a period of 1.5hours. Of course the power transformer was taking quite long to get hot (still touchable) and after 1.5hours, the supply voltage remains at 440VDC. When the supply voltage drops, the bias voltage across the cathod resistor for the EL34 drops from 0.75VDC to 0.65VDC and stays there. Is this normal for a tube amplifier to be behaving this way when the output tranformer gets hot? Is it possible to keep everything stable using a LM317 floating voltage regulator in place of a choke? Regards, Mark Long term drift problems can be difficult to troubleshoot. Perhaps this thought is radical (maybe Atsunori can express his opinion on this), but can the output stage be disconnected and a large resistor (dummy load) be hung in its place and then monitor the PS to see if there is still drift? Much can then be eliminated as the culprit. Back to earth. For long-term temperature drift type problems, I try to hasten the time by alternating a heat gun and cold (Freon type) spray. If you then narrowed the problem to a particular area, the long nozzle tube can be directed to individual components. Food for thought (may give you indigestion though). Cordially, west There is more analysis than you suggest required for finding out why the B+ sags by 40V after 1.5hrs. Just how hot is that power tranny getting? A hot B+ winding on the brink of destroying its insulation has more DC resistance than when cold after being on for only 2 minutes. Presumably the resistor at the cathodes is 10ohms, so hence the 0.75V volts and 0.65 volts measured; it means the Ik is reducing from 75mA to 65 mA due to the reduction of the B+ by 40V. But something is grossly WRONG! Just what is a mystery. There is something sucking power and causing the B+ to fall so much. Is it leaky electros? one would expect they are also unduly hot to be able to dissipate the power needed to make the B+ sag by the 40V, even though when that happens the tubes draw less current! Patrick Turner. |
#7
|
|||
|
|||
"Patrick Turner" wrote in message ... There is more analysis than you suggest required for finding out why the B+ sags by 40V after 1.5hrs. Just how hot is that power tranny getting? A hot B+ winding on the brink of destroying its insulation has more DC resistance than when cold after being on for only 2 minutes. Presumably the resistor at the cathodes is 10ohms, so hence the 0.75V volts and 0.65 volts measured; it means the Ik is reducing from 75mA to 65 mA due to the reduction of the B+ by 40V. But something is grossly WRONG! Just what is a mystery. There is something sucking power and causing the B+ to fall so much. Is it leaky electros? one would expect they are also unduly hot to be able to dissipate the power needed to make the B+ sag by the 40V, even though when that happens the tubes draw less current! Patrick Turner. The Power Tranny is only suppling a 5V and 440-0-440 @ 750mA (as per written on the tranny). Its hot but I still can place my hands on it for a whole 1 minute yet I don't have to worry about getting burned. The Cathode resistor is a 15ohms, hence at 0.75V = 50mA While using the amp, I had a 500mA meter attached across the B+ and the amp. It initially went to 190mA and later dropped to 170~160mA. At this point, the voltage stays at 440V ~ 450V. I left it for more that 2 hours and it stayed there. All of the caps at the power supply is cold. None appears to be warm when touched. Warm regards, Mark |
#9
|
|||
|
|||
Mark wrote: "Patrick Turner" wrote in message ... There is more analysis than you suggest required for finding out why the B+ sags by 40V after 1.5hrs. Just how hot is that power tranny getting? A hot B+ winding on the brink of destroying its insulation has more DC resistance than when cold after being on for only 2 minutes. Presumably the resistor at the cathodes is 10ohms, so hence the 0.75V volts and 0.65 volts measured; it means the Ik is reducing from 75mA to 65 mA due to the reduction of the B+ by 40V. But something is grossly WRONG! Just what is a mystery. There is something sucking power and causing the B+ to fall so much. Is it leaky electros? one would expect they are also unduly hot to be able to dissipate the power needed to make the B+ sag by the 40V, even though when that happens the tubes draw less current! Patrick Turner. The Power Tranny is only suppling a 5V and 440-0-440 @ 750mA (as per written on the tranny). Its hot but I still can place my hands on it for a whole 1 minute yet I don't have to worry about getting burned. Well that suggests the T rise is letting the tranny rise to about 40C. The Cathode resistor is a 15ohms, hence at 0.75V = 50mA So the current is even les than thought. While using the amp, I had a 500mA meter attached across the B+ and the amp. It initially went to 190mA and later dropped to 170~160mA. At this point, the voltage stays at 440V ~ 450V. I left it for more that 2 hours and it stayed there. All of the caps at the power supply is cold. None appears to be warm when touched. When you turn on a tube amp, the B+ often goes rapidly up to a maximum voltage, and then it falls to a lower value when the output tubes begin to conduct and the bias is properly established. A 440v-0-440v winding on a transformer which measures this Vrms voltage with no load should give you a maximum of +616V DC with no load. Then when the output tubes conduct, this voltage should fall as you say it does to a stable value which you say is about +450V DC. I would say you are using a tube rectifier because with silicon diodes the DC voltage does not fall as much from the maximum DC voltage with no load; ie, if maximum is 616Vdc, then it would fall to maybe 560Vdc with silicon diodes. But the tube rectifier would have considerable effective series resistance so this is why you get only +450V from a tranny winding of 440Vrms, which is 616Vpeak voltage swing. The initial current required to charge up the PS capacitors may reach a peak value for a short time and then reduce until the output tubes conduct when the current increases to a steady value. All this should happen happen in the first 20 seconds of operation. There should be no DC current flow into the CT on the OPT for about 8 seconds after turn on when the amp is cold So I don't know why you are measuring 190mA flowing into the CT at switch on and then seeing it drop back to 170mA. You may not be measuring the current at the right point. There may not be anything wrong with your amp. You say there is B+ drop from 480V to 440V over 1.5 hrs. But does this drop occur after 20 seconds? if so, the amp is OK, but if the B+ starts at +480V at after say 30 seconds, then very slowly falls 40V after 1.5 hrs, something is causing it. It's never happened in amps I have built or serviced. Patrick Turner. Warm regards, Mark |
#10
|
|||
|
|||
Mark wrote:
While using the amp, I had a 500mA meter attached across the B+ and the amp. It initially went to 190mA and later dropped to 170~160mA. At this point, the voltage stays at 440V ~ 450V. I left it for more that 2 hours and it stayed there. In any event, Mark, I think it's better to check the power transformer's secondary AC voltage at two time points. - When the amplifier just started and the power transformer is still cool. - After 1.5 hours, where the B+ dropped to 440v and stable. If the secondary AC voltage does not change, then other part of the amplifier is causing this phenomenon, possibly your rectifier circuit for one of many candidates. But since I don't know what rectification circuit is used, I can't go further. If the secondary AC voltage changes, there are also many possible scenarios. But I think it's worth checking out the presence of transformer's secondary AC voltage change. Atsunori |
#11
|
|||
|
|||
"Jon Yaeger" wrote in message ... Mark, Question: Does this amp have a C-L-C filter? Jon Oh yes - sorry forgot to mention that earlier |
#12
|
|||
|
|||
"Patrick Turner" wrote in message ... There should be no DC current flow into the CT on the OPT for about 8 seconds after turn on when the amp is cold So I don't know why you are measuring 190mA flowing into the CT at switch on and then seeing it drop back to 170mA. You may not be measuring the current at the right point. There may not be anything wrong with your amp. You say there is B+ drop from 480V to 440V over 1.5 hrs. But does this drop occur after 20 seconds? if so, the amp is OK, but if the B+ starts at +480V at after say 30 seconds, then very slowly falls 40V after 1.5 hrs, something is causing it. It's never happened in amps I have built or serviced. During initial turn-on, the current meter shows 0 as you mentioned. It will slowly reach 190mA after 30 seconds or so. At this point, the voltage measured is 480VDC. The drop occurs gradually over a period of 1 hour till 1.5 hours. The Amp meter is fitted between the output of the power supply and the input supply to the amp. Regards, Mark |
#13
|
|||
|
|||
|
#14
|
|||
|
|||
Mark wrote: "Patrick Turner" wrote in message ... There should be no DC current flow into the CT on the OPT for about 8 seconds after turn on when the amp is cold So I don't know why you are measuring 190mA flowing into the CT at switch on and then seeing it drop back to 170mA. You may not be measuring the current at the right point. There may not be anything wrong with your amp. You say there is B+ drop from 480V to 440V over 1.5 hrs. But does this drop occur after 20 seconds? if so, the amp is OK, but if the B+ starts at +480V at after say 30 seconds, then very slowly falls 40V after 1.5 hrs, something is causing it. It's never happened in amps I have built or serviced. During initial turn-on, the current meter shows 0 as you mentioned. It will slowly reach 190mA after 30 seconds or so. At this point, the voltage measured is 480VDC. The drop occurs gradually over a period of 1 hour till 1.5 hours. The Amp meter is fitted between the output of the power supply and the input supply to the amp. It seems that after the CLC you have the CT of the OPT and the B+ supply to the rest of the amp. What are the actual loads connected? Are they constant? you say not only does the B+ drop, but the output tube current reduces. So where is the mystery current going? So what rectifier tube do you have? have you tried replacing it with another? Its possible the HT winding has high DC resistance and which is getting very hot. what is the DCR of the whole of the 440-0-440 winding? Patrick Turner. Regards, Mark |
#15
|
|||
|
|||
Jon Yaeger wrote: in article , Mark at wrote on 9/2/05 12:00 PM: "Jon Yaeger" wrote in message ... Mark, Question: Does this amp have a C-L-C filter? Jon Oh yes - sorry forgot to mention that earlier Mark, Just for fun, disconnect the 1st section of the electrolytic and bypass it with a new axial cap of similar value. Measure the B+ over time. Reasoning: I have seen a number of amps have electros that fail OPEN. If it is the 1st cap section, it will drive the voltage down significantly . . . but typically a lot more than you experienced. The PS becomes a L-C instead of a C-L-C. Just suppose, for a minute, that you have a cap section that decreases in value or there is an increasing equivalent series resistance. It could drop the voltage like you are experiencing. Of course, I may be quite wrong, but it's easy to try & rule out. Jon Yes, its a possibility the caps are crook. The ripple voltage at C1 should be = 1,800 x Idc / C uF for 60Hz mains or 2,200 x Idc / Cuf if mains f = 50 Hz, So if the Idc is known, and the capacitor value is know from what is on the can, then the measured ripple voltage at C1 should be what one calculates, so if C1 = 47uF, then Vripple at 170mA shoulod be with 60Hz mains = 1,800 x 0.17 / 47 = 6.5Vrms . If the ripple measures say 30Vrms, then C must be a lot lower than it should be. But it don't explain the slow fall in B+ over 1.5hrs; this is more liklely to the tied to the temperature of the HT winding wire. Patrick Turner. |
#16
|
|||
|
|||
I vote for Jon Yaeger's diagnosis. Cheers, John Stewart
|
#17
|
|||
|
|||
"Patrick Turner" wrote in message ... It seems that after the CLC you have the CT of the OPT and the B+ supply to the rest of the amp. What are the actual loads connected? Are they constant? you say not only does the B+ drop, but the output tube current reduces. So where is the mystery current going? So what rectifier tube do you have? have you tried replacing it with another? Its possible the HT winding has high DC resistance and which is getting very hot. what is the DCR of the whole of the 440-0-440 winding? Hello Guys! Today's Sunday, so finnally got some time to do some measurements on the amp. I got the following results today. The previous was normally measured in the night on weekdays, but today I did in during the day. This time I moved the amp meter to right after the 5AR4 rec tube and before the caps and the rest. Current moved up slowly to 190mA ~ 195mA and stopped. Time - Secondary VAC 0:00 - 448VAC 0:30 - 441VAC 0:45 - 440VAC 1:00 - 436VAC 1:30 - 433VAC 1:45 - 434VAC 2:00 - 432VAC The B+ today measured from 480VDC and slowly went down to 476VDC Current drop a little to around 185mA after 2 hours. Do you think its something to do with the mains voltage dropping during the night? Todays measurements seems to be okay, not too far off as earlier. Would a computer AVR be useful to prevent too much drops in the mains? Or is it bad for audio use? Regards, Mark |
#18
|
|||
|
|||
|
#19
|
|||
|
|||
Jon Yaeger wrote: in article , Mark at wrote on 9/4/05 12:07 PM: "Patrick Turner" wrote in message ... It seems that after the CLC you have the CT of the OPT and the B+ supply to the rest of the amp. What are the actual loads connected? Are they constant? you say not only does the B+ drop, but the output tube current reduces. So where is the mystery current going? So what rectifier tube do you have? have you tried replacing it with another? Its possible the HT winding has high DC resistance and which is getting very hot. what is the DCR of the whole of the 440-0-440 winding? Hello Guys! Today's Sunday, so finnally got some time to do some measurements on the amp. I got the following results today. The previous was normally measured in the night on weekdays, but today I did in during the day. This time I moved the amp meter to right after the 5AR4 rec tube and before the caps and the rest. Current moved up slowly to 190mA ~ 195mA and stopped. Time - Secondary VAC 0:00 - 448VAC 0:30 - 441VAC 0:45 - 440VAC 1:00 - 436VAC 1:30 - 433VAC 1:45 - 434VAC 2:00 - 432VAC The B+ today measured from 480VDC and slowly went down to 476VDC Current drop a little to around 185mA after 2 hours. Do you think its something to do with the mains voltage dropping during the night? Todays measurements seems to be okay, not too far off as earlier. Would a computer AVR be useful to prevent too much drops in the mains? Or is it bad for audio use? Regards, Mark Why not measure mains & B+ concurrently? A 1V change in mains voltage will give you about a 4V change in B+. That much? Maybe in the US with 120V. Here in Oz I often see a range of mains from 235V to 250V, so that if there is +425V from 235~ then I would get +451V from 250~. Some chinese amps built to run on 220V such as Jolida and designed to have B+ = +450V will then have +511V with 250~, and then some of the electros blow up, and if ppl bias the amps as the instructions state the tubes are too hot. I had a biasing problem when i recently serviced a Stingray amp which had been bought in europe with 230V mains. On many days i measure 250~ here. But our original poster said he had a drop of B+ from 480V to 440V. That's about 8% change. Perhaps indeed he has an 8% variation in the mains voltage. Patrick Turner. A 16V delta corresponds to a 4V mains drop. Nothing to lose sleep over. An AVR may or may not degrade the sound. My adjustable AC power supply makes some amps sound awful. Jon |
#20
|
|||
|
|||
Another possibility is to check for the "aw ****" test.
Swap out the rectifier tube and see what happens. "John Stewart" wrote in message ... I vote for Jon Yaeger's diagnosis. Cheers, John Stewart |
#21
|
|||
|
|||
Patrick Turner wrote:
Jon Yaeger wrote: in article , Mark at wrote on 9/2/05 12:00 PM: "Jon Yaeger" wrote in message ... Mark, Question: Does this amp have a C-L-C filter? Jon Oh yes - sorry forgot to mention that earlier Mark, Just for fun, disconnect the 1st section of the electrolytic and bypass it with a new axial cap of similar value. Measure the B+ over time. Reasoning: I have seen a number of amps have electros that fail OPEN. If it is the 1st cap section, it will drive the voltage down significantly . . . but typically a lot more than you experienced. The PS becomes a L-C instead of a C-L-C. Just suppose, for a minute, that you have a cap section that decreases in value or there is an increasing equivalent series resistance. It could drop the voltage like you are experiencing. Of course, I may be quite wrong, but it's easy to try & rule out. Jon Yes, its a possibility the caps are crook. The ripple voltage at C1 should be = 1,800 x Idc / C uF for 60Hz mains or 2,200 x Idc / Cuf if mains f = 50 Hz, So if the Idc is known, and the capacitor value is know from what is on the can, then the measured ripple voltage at C1 should be what one calculates, so if C1 = 47uF, then Vripple at 170mA shoulod be with 60Hz mains = 1,800 x 0.17 / 47 = 6.5Vrms . If the ripple measures say 30Vrms, then C must be a lot lower than it should be. But it don't explain the slow fall in B+ over 1.5hrs; this is more liklely to the tied to the temperature of the HT winding wire. Patrick Turner. He is too far from the 'Line Drop Compensator' on the distribution system. Sometimes called a 'Line Drop Constipator' by the unwashed among us! A magical, high power auto transformer on a pole near your house whose switch is automatically adjusted by a control gadget so that the voltage at the load remains constant under various conditions. Sometimes you will see three of them on succeeding poles doing their work in a 3-phase system. Hard to believe, but this very old technology really works. So there! That is my contribution for this week. What? No tubes? You will have to read the latest issues of AudioXpress to see what I've really been doing. Cheers, John Stewart |
Reply |
Thread Tools | |
Display Modes | |
|
|
Similar Threads | ||||
Thread | Forum | |||
World Tube Audio Newsletter 06/05 | Vacuum Tubes | |||
KISS 113 by Andre Jute | Vacuum Tubes | |||
One for the Tube Grabbers | Vacuum Tubes | |||
FS: SOUNDSTREAM CLOSEOUTS AND MORE!! | Car Audio | |||
EASY way to power tube audio projects | Vacuum Tubes |