Just an update. I tried to diconnect the balance pot and gnd the grid and the voltage at pentode plate goes up to 45v from 18v. And vg2 to around 35v. With supply of 157. Tried replacing the 50 k dropping resistor to 20k and voltages rise up but still below the 70 v plate. Change the two multiple filter cap and voltages has improved. Tomorrow i will replace the doubler filter caps. The miltiple caps has really dried up. Thanks gentlemen for your kind suggestions.
One more question is how come that the grid got that 1.3v via balance pot. As per schema it was cap coupled via filters. Im planning on replacing the balance pot with a new one but this time put a coupling cap to grid and put a grid leak resistor as well as series grid resistor.
Is 6an8 work better with low plate voltage?
One more question is how come that the grid got that 1.3v via balance pot. As per schema it was cap coupled via filters. Im planning on replacing the balance pot with a new one but this time put a coupling cap to grid and put a grid leak resistor as well as series grid resistor.
Is 6an8 work better with low plate voltage?
Definitely replace all the electrolytic PS caps.
The voltage on the grid and the low voltages elsewhere would make me wonder if there is DC voltage getting to the chassis somehow. If there is voltage on the chassis your readings will be lower, assuming that you are measuring between the tube pins and the chassis. Most vintage amps didn't use a 3 wire, grounded, power cord. So make sure the chassis is grounded and see how that affects your voltage readings.
The voltage on the grid and the low voltages elsewhere would make me wonder if there is DC voltage getting to the chassis somehow. If there is voltage on the chassis your readings will be lower, assuming that you are measuring between the tube pins and the chassis. Most vintage amps didn't use a 3 wire, grounded, power cord. So make sure the chassis is grounded and see how that affects your voltage readings.
Junm,
Respectfully, kindly again check posts #27, #28, #33 (end), #37, #39 and #40.
All are kindly asking you one way or another to, before you do anything else, short pins 8 of the 6AN8s to verified earth and then tell us what you measure at the various points, including B+ voltages.
While all other advice is valuable, matters will stay muddled until you perform this step. It must be determined what is happening at those pins 8 and why before we can further help you or you can help yourself successfully. There should be zero d.c. voltage there, and if not the reason must be found. Until then, this will upset things irrespective of however many components you change or what you do. Some of us have been there; we are asking this for your own good!
Hope this is clear?
Respectfully, kindly again check posts #27, #28, #33 (end), #37, #39 and #40.
All are kindly asking you one way or another to, before you do anything else, short pins 8 of the 6AN8s to verified earth and then tell us what you measure at the various points, including B+ voltages.
While all other advice is valuable, matters will stay muddled until you perform this step. It must be determined what is happening at those pins 8 and why before we can further help you or you can help yourself successfully. There should be zero d.c. voltage there, and if not the reason must be found. Until then, this will upset things irrespective of however many components you change or what you do. Some of us have been there; we are asking this for your own good!
Hope this is clear?
ERROR!
Junm,
My face is red! You can in turn direct me to your last post! I apologise, somehow I misinterpreted that: YOU DID TELL WHAT HAPPENED WHEN SHORTING PINS 8!
Again so sorry. As you may notice from the hour, I ahve been up for the last 30 hours, not an excuse but perhaps an explanation.
I must go to bed now but will work through the new measurements with perhaps greater awareness than now. Likely one of the other honourable members would have replied by then!
Junm,
My face is red! You can in turn direct me to your last post! I apologise, somehow I misinterpreted that: YOU DID TELL WHAT HAPPENED WHEN SHORTING PINS 8!Again so sorry. As you may notice from the hour, I ahve been up for the last 30 hours, not an excuse but perhaps an explanation.
I must go to bed now but will work through the new measurements with perhaps greater awareness than now. Likely one of the other honourable members would have replied by then!
Junm,
Still not going to bed: If I dare make further suggestions(!) you can perhaps then remove (disconnect a lead) the capacitors 30µf 6V over the cathode resistor as well as the 0,2µF screen capacitor. Keeping pins 8 shorted to earth and everything else as is, if you can then give resultant electrode voltages ....
Now off with me!
Still not going to bed: If I dare make further suggestions(!) you can perhaps then remove (disconnect a lead) the capacitors 30µf 6V over the cathode resistor as well as the 0,2µF screen capacitor. Keeping pins 8 shorted to earth and everything else as is, if you can then give resultant electrode voltages ....
Now off with me!
Hi johan i just finished recapping it today. I already got the voltage og pentode plate at very close to the indicated readings by sansui. Hoever the screen voltage still at 50v lesser than 55 as indicated. Replaced all caps including film types. I put a 680k on grid of pentode to gnd. And temporarily put coupling cap from balance pot. Check bias voltages for 7591 and at -19v. I did testing and it sings however i test only one channel since i only have pair of 7591. Not fully satisfied since its not to loud. Maybe around 10w only.
I rewire the pins for 6p3s since i have lot of these.. adjust screen bias to -24v at 420v b+. The 6p3s red plate. I cant adjust the since it already atmaximum. Put 470 ohm cathode res for 2 tube and cathode reads 16v. The 6p3s did not red plate at 40 ma .
Question? Is it good to add cathode resistor . Fixed bias gives -24v and plus cathode bias reads 16v.. if i remove the cathode resistor and fixed bias it to -40v will it be the same?
I rewire the pins for 6p3s since i have lot of these.. adjust screen bias to -24v at 420v b+. The 6p3s red plate. I cant adjust the since it already atmaximum. Put 470 ohm cathode res for 2 tube and cathode reads 16v. The 6p3s did not red plate at 40 ma .
Question? Is it good to add cathode resistor . Fixed bias gives -24v and plus cathode bias reads 16v.. if i remove the cathode resistor and fixed bias it to -40v will it be the same?
In my opinion, finding the fault should be limited to checking all the components and NOT changing the design. This amplifier worked just fine at some point in time, unless the schematic doesn't reflect the actual circuit. If you want to creat some kind of mutant, then go ahead, change the circuit, use different tubes. But if you want to retain the Sansui design, then find the real fault. After checking all involved components, use a new, tested 6AN8 tube, same brand if needed to be. Just my humble opinion.
A first step of finding a bad component that I am using is comparing measurements on both channels, one good, one bad. Disconnect power and discharge filter caps. Remove both 6AN8 tubes. Measure resistance on each tube pin to ground and then each tube pin to B+, say at the 30K resistor. Write down each value. Do the same on the good channel. Discrepancies of 5%-10% should be normal I think. If you find a spot with bigger differences, there should be your further investigation going.
Thank gentlemen with your good input. My line voltage is 218v at the test time. The power trafo is rated 232v fora 180v secondary and im only getting 172 v then double rectified to 425vdc.
Yes i had do comparison with each channel thats why i foun that problem is common on them so iconcentrated on ths common items such as the psu. The other channel reads little lower than the other but in minimim i think.
@ arto. Thanks men i will going to put maybe arond 100 uf to 220 uf. Do i still need to increase the feedback resistor to have more gain for 6p3s and put a series resistor fo the screen? It has direct connection at 400v. Little bit lower than plate of 420 v.
Yes i had do comparison with each channel thats why i foun that problem is common on them so iconcentrated on ths common items such as the psu. The other channel reads little lower than the other but in minimim i think.
@ arto. Thanks men i will going to put maybe arond 100 uf to 220 uf. Do i still need to increase the feedback resistor to have more gain for 6p3s and put a series resistor fo the screen? It has direct connection at 400v. Little bit lower than plate of 420 v.
What i wiil be doing next is finding out excessive high freq from the preamp section which has 12ax7 with tone switches. When turning the treble pot isee some flickering on the output valves. What was the indication of blinking of tube when treble pot is rotated .Scrathes is heard on speakers?
Last edited:
Junm,
Your questions seem to have been addressed in previous posts. While all posts are good advice, I would single out post #48 from Electrix. This because a fundamental basic rule is indicated, of importance to all.
It is essential to determine the cause of any fault before fiddling about with component values, as Electrix said, even if you dislike the original design. Otherwise one can forever go on changing component values and get nowhere because a undetected fault condition keeps upsetting your expectations. This is a basic rule of electronic work! (Most of us learnt this the hard way!)
This can be as a result of overloading. By advancing any control beyond the original flat frequency response, overloading will occur when going near the amp's max capability. Typically the 'flickering' might look like a 'sparkling' g2 - symptoms vary. (Technically, overloading means that a tube has run out of 'headroom'. Normal gain characteristics cease to exist, feedback if any tries to compensate and cuts out, and the whole amplifier goes into a state of momentary instability. Everything is then out of control until conditions go back to normality.)
Tubes are not instantly damaged by this, but matters should go back to normal as quickly as possible. One gets the same result in the Sansui AU111, where the main amplifier GFB loop is used for tone control purposes (Presence). (I am tempted to state: Never do that! but that is OT here. Still ....)
I am hunting here for a 6AN8 to try hook up your circuit and do a few experiments myself. (Using a cathodine phase inverter direct coupled to an input tube can become critical when high output is desired, as a result of tube spec spread.) - but don't wait for that! You might be almost done!
Your questions seem to have been addressed in previous posts. While all posts are good advice, I would single out post #48 from Electrix. This because a fundamental basic rule is indicated, of importance to all.
It is essential to determine the cause of any fault before fiddling about with component values, as Electrix said, even if you dislike the original design. Otherwise one can forever go on changing component values and get nowhere because a undetected fault condition keeps upsetting your expectations. This is a basic rule of electronic work! (Most of us learnt this the hard way!)
This can be as a result of overloading. By advancing any control beyond the original flat frequency response, overloading will occur when going near the amp's max capability. Typically the 'flickering' might look like a 'sparkling' g2 - symptoms vary. (Technically, overloading means that a tube has run out of 'headroom'. Normal gain characteristics cease to exist, feedback if any tries to compensate and cuts out, and the whole amplifier goes into a state of momentary instability. Everything is then out of control until conditions go back to normality.)
Tubes are not instantly damaged by this, but matters should go back to normal as quickly as possible. One gets the same result in the Sansui AU111, where the main amplifier GFB loop is used for tone control purposes (Presence). (I am tempted to state: Never do that! but that is OT here. Still ....)
I am hunting here for a 6AN8 to try hook up your circuit and do a few experiments myself. (Using a cathodine phase inverter direct coupled to an input tube can become critical when high output is desired, as a result of tube spec spread.) - but don't wait for that! You might be almost done!
Yes - will have to check that. Pentodes can work quite satisfactory at low Va, as long as they are not overloaded. For power amp input stages the required output is usually low ...
BUT for this circuit the pentode must deliver the full EL34 drive (the phase splitter has no gain). For a power tube bias of 19V, that signal will need to be 38 Vpp, which might leave little headroom. I will look at the pentode load line and determine if all is kosher there.
BUT for this circuit the pentode must deliver the full EL34 drive (the phase splitter has no gain). For a power tube bias of 19V, that signal will need to be 38 Vpp, which might leave little headroom. I will look at the pentode load line and determine if all is kosher there.
Junm,
I am going to violate one of my own advices and comment on the circuit as is! I have fetched the necessary 6AN8 graphs and done some computing. Although the graphs are small, I believe my findings are close to practice.
But let me at least honour the advice against starting to change component values before a fault is found. What I am presenting here does not take the place of fault-finding; it comes afterwards.
Meaning that when one computes, that triode phase splitter barely makes it to drive the 7591s to full output, i.e. some 20Vp on the control grids, without overloading, also taking tolerances into account.
Thus, I would like you to really get +235V h.t. where indicated, for the 6AN8 triode not to run out of steam. [Brief comment about driving some power tubes with a concertina: In this case the effective signal between phase inverter anode and cathode turns out to be 80Vpp. Thats pushing it with only some +235V effective h.t.] To fit in the signal requirement comfortably plus allowing some for tolerances, the pentode anode (pin 6) should not really be higher than +60V . That will place the triode anode (pin 1) at some 170V as indicated. I must caution again against influencing voltage measurements by connecting a meter, even DVM, to the circuit. A 10 meg input-R meter is really essential.
As I calculated, the triode bias should be at some -4,1V; best measured directly beteen pins 2 and 3, so as not to load the circuit appreciably. Your best low impedance check is from pin 3 to earth, preferably below 70V.
Not to burden the thread with arithmetic; should you want me to take you through the arithmetical spadework from graphs, we can perhaps take this to PM.
Last point for the present: To finally get the desired pin 6 potential, one can most conveniently adjust the pentode cathode resistor of 1K5 ohm until the pin-3 voltage is about 65V or slightly below. (One temporarily replace it by a say 5K pot for adjustment and measure the resistance afterwards.)
I repeat, this is to get the phase inverter to a better work point where it will not risk overloading at max. output. It has nothing to do with your fault-finding; that comes first as mentioned before.
This is gettting long so I will stop for now.
I am going to violate one of my own advices and comment on the circuit as is! I have fetched the necessary 6AN8 graphs and done some computing. Although the graphs are small, I believe my findings are close to practice.
But let me at least honour the advice against starting to change component values before a fault is found. What I am presenting here does not take the place of fault-finding; it comes afterwards.
Meaning that when one computes, that triode phase splitter barely makes it to drive the 7591s to full output, i.e. some 20Vp on the control grids, without overloading, also taking tolerances into account.
Thus, I would like you to really get +235V h.t. where indicated, for the 6AN8 triode not to run out of steam. [Brief comment about driving some power tubes with a concertina: In this case the effective signal between phase inverter anode and cathode turns out to be 80Vpp. Thats pushing it with only some +235V effective h.t.] To fit in the signal requirement comfortably plus allowing some for tolerances, the pentode anode (pin 6) should not really be higher than +60V . That will place the triode anode (pin 1) at some 170V as indicated. I must caution again against influencing voltage measurements by connecting a meter, even DVM, to the circuit. A 10 meg input-R meter is really essential.
As I calculated, the triode bias should be at some -4,1V; best measured directly beteen pins 2 and 3, so as not to load the circuit appreciably. Your best low impedance check is from pin 3 to earth, preferably below 70V.
Not to burden the thread with arithmetic; should you want me to take you through the arithmetical spadework from graphs, we can perhaps take this to PM.
Last point for the present: To finally get the desired pin 6 potential, one can most conveniently adjust the pentode cathode resistor of 1K5 ohm until the pin-3 voltage is about 65V or slightly below. (One temporarily replace it by a say 5K pot for adjustment and measure the resistance afterwards.)
I repeat, this is to get the phase inverter to a better work point where it will not risk overloading at max. output. It has nothing to do with your fault-finding; that comes first as mentioned before.
This is gettting long so I will stop for now.
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.