Increase of 299C power

Thx. I reflowed some questionable solder joints on the CL80s and hope that was the issue. But need a set of new 7591s anyway. Its my understanding that only TungSol makes one that is the same physical size in bottle and pins (7591a).
I have noticed that factory matched quads have plate current measurements from 80-100 marked on the boxes (based on unknown test parameters). Would you be hesitant about buying a set on the lower end (80ma)?
 
Last edited:
I've got an early version 208 here with the same problem. The schematic for my unit is more realistic calling for 470 vdc to the OPT CT, but I'm getting just shy of 500 volts.
Dave
FYI
While troubleshooting my problem, I found that this unit has the amp of the 208 model with a preamp in front (almost identical to schematic of 208). That would make all my voltages on the money and the transformer correct for a 208 and for this unit.

http://hhscott.com/pdf/fs/HHS_208_SD_web.jpg
 
Last edited:
I believe the original question was if you "had" to replace the Se rectifier. And the answer of course is no. If the amp is working within its specified operating conditions then there is no REAL need to replace anything. I have gone through around a hundred tube amps now, and truthfully, i have not seen a failed Se rectifier yet. When they are working at their design parameters they will do their job without problems. But when they are abused and over voltaged or drawing too many amps like in the video you showed the outcome is a spectacular failure. How convenient that he had a video camera to film the gory details.
I am not saying that i don't change them. Changing them out is a perfect excuse for doing a procedure an earning some money. You might even say that as a repair tech one should for OSHA rationale, change out such a rectifier. But there will always be collectors who only care that the piece is 100% original, down to the Se rectifier.
That's fine but sel. rects. voltage drop increases with age. Why have to monitor the rectifier when you can replace it for pennies and some labor? Plus, if the voltage drop rises on the rect, the bias voltage slowly falls over time, wouldn't that cause more and more bias current to flow?
 
That's fine but sel. rects. voltage drop increases with age. Why have to monitor the rectifier when you can replace it for pennies and some labor? Plus, if the voltage drop rises on the rect, the bias voltage slowly falls over time, wouldn't that cause more and more bias current to flow?
Components have specs that show the expected lifespan and i am at all familiar with what that is for the Selinium rectifier in the unit concerened. If we know that info then one can determine how the expected reduced performance of the part will affect the bias voltage, and if that is a real concern. Without that info we are talking in generalities and thus, in the original query the need to change out the SE rectifier is not an absolute. But, i as a matter of course i change them out for a smaller and less problematic silicon diode.
 
Old -- I would doubt that the markings you are referring to are true mA readings, as the 7591 is only capable of a maximum plate current of 80 mA to begin with, so the numbers likely represent a grade or percentage grade for matching purposes. As a result, I wouldn't be hesitant to use tubes with lower numbers as long as they're all close in number.

Glad you found the problem!

Dave
 
.

Here is a short YT video of what happens when one flames out.. and remember they produce toxic fumes.
Seriously, you believe this guy actually had all his video gear set up and coincidentally filmed a Se Rectifier blowing up? C'mon the guy blew it up to make a video. Probably has a repair shop and want to scare up some business.
 
Unfortunately, none of the schematics fit my unit. For reference, I use pin6 on the power tubes to see if any schematics match up (other than power supply which is a 299C).

I believe my channel problem is a signal issue inside the unit as the right speaker makes the same angry hum just as the left speaker when testing pin 2 of the corresponding 12ax7s (V2 and V102). Otherwise dead silence on right channel.
Hopefully voltages will work out once both channels are amplifying signal. For now, running it at 97 vac keeps the voltages in line.
Also noticed that after heating up for awhile, the balance on the bias of the right side tubes was way off (-7 and -19) and had to be rebalanced.

Edit: Also, would it be advisable to add screen resistors? Will be in the area adding some test jacks anyway.
View attachment 984658
Oldman - Is there about 425v on each Pin 6 of the power tubes of the dead channel? And yes, you should add screen grid resistors. I used 470 k Ohm, 1/2w in my similar LK-72 (aka299c).
I misspoke, its a TR13-2-1. Shows how bad my memory is from the time it takes to flip it over to the time it takes typing on the computer.

Not even sure yet what the bias should be as I dont have schematics to match this unit. I've seen 299C shematics from -18.5 to -20.5. And then there are some that dont have the bias voltage. This unit most closely resembles a 72A with a 470 ohm in lieu of 270 ohm for R40 though I havent checked the complete schematic.

Will figure the it out with adding the cathode resistor I guess. Had to order some 470 ohm resistors as I was one shy, maybe I should have just put in some 270s.
Oldman - I wouldn't be too concerned about the right negative bias voltage, it's the product of the current and the plate voltage (power) that is what's most important. Since you have installed 10 Ohm cathode resistors to use as monitors, what I do is find out what the max. plate dissipation is for said tube. In this case, the 7591, which has a max. Pd of 19 Watts. So, at idle, I read the plate voltage (with the power line voltage 'of your choice'). Say it's 430v. Then I adjust the bias to say 30 ma. The product of which is 12.9w, but you have to subtract the screen grid dissipation, which I estimate at 5%. 12.9w x 0.95 = 12.25w. That's the plate dissipation. Divide 12.25 by 19 and get 64.5%, which is a bit cold according to the "70% dissipation rule". So increase the bias current a little and re-calculate the Pd. until you get 70%. I prefer closer to 75% sound wise, a bit warm, full.
Therefore, bias voltage is simply an average of a sampling of amps. that were produced and is a quicker (but inaccurate) way to set the proper bias level.
 
Oldman - Is there about 425v on each Pin 6 of the power tubes of the dead channel? And yes, you should add screen grid resistors. I used 470 k Ohm, 1/2w in my similar LK-72 (aka299c).

Oldman - I wouldn't be too concerned about the right negative bias voltage, it's the product of the current and the plate voltage (power) that is what's most important. Since you have installed 10 Ohm cathode resistors to use as monitors, what I do is find out what the max. plate dissipation is for said tube. In this case, the 7591, which has a max. Pd of 19 Watts. So, at idle, I read the plate voltage (with the power line voltage 'of your choice'). Say it's 430v. Then I adjust the bias to say 30 ma. The product of which is 12.9w, but you have to subtract the screen grid dissipation, which I estimate at 5%. 12.9w x 0.95 = 12.25w. That's the plate dissipation. Divide 12.25 by 19 and get 64.5%, which is a bit cold according to the "70% dissipation rule". So increase the bias current a little and re-calculate the Pd. until you get 70%. I prefer closer to 75% sound wise, a bit warm, full.
Therefore, bias voltage is simply an average of a sampling of amps. that were produced and is a quicker (but inaccurate) way to set the proper bias level.
I adjusted it to 33ma with 456v plate which would give me 75%. However I had to adj bias to -19.5 due to weak tubes. But it has considerable distortion and clips at a low volume. There has to be a point where you have to bias so much to get your power up that it starts this distortion. I cannot find that info on the tube charts. In fact, the 456v with 33 ma is off the charts; http://www.tubezone.net/pdf/7591.pdf and the 456v exceeds max for this tube even though the amp design is for 460v. Unless I am reading this wrong which may very well be the case.
 
470K for a screen resistor is way too large. You'd get a lot of voltage drop across that and the performance would tank. 470 ohm you could do though. I usually use 100 but I seem to recall that basically anything in the 100 - 1000 ohm range will get it done without really impacting performance any.
 
470K for a screen resistor is way too large. You'd get a lot of voltage drop across that and the performance would tank. 470 ohm you could do though. I usually use 100 but I seem to recall that basically anything in the 100 - 1000 ohm range will get it done without really impacting performance any.
If I stated that then I misspoke. If I added a screen resistor, I would have put in a 150 ohm resistor as you had suggested on the Sherwood I was working on at the time.
old
 
I adjusted it to 33ma with 456v plate which would give me 75%. However I had to adj bias to -19.5 due to weak tubes. But it has considerable distortion and clips at a low volume. There has to be a point where you have to bias so much to get your power up that it starts this distortion. I cannot find that info on the tube charts. In fact, the 456v with 33 ma is off the charts; http://www.tubezone.net/pdf/7591.pdf and the 456v exceeds max for this tube even though the amp design is for 460v. Unless I am reading this wrong which may very well be the case.
Any comment/advice on this post?
thx
 
Oldman - Is there about 425v on each Pin 6 of the power tubes of the dead channel? And yes, you should add screen grid resistors. I used 470 k Ohm, 1/2w in my similar LK-72 (aka299c).

Oldman - I wouldn't be too concerned about the right negative bias voltage, it's the product of the current and the plate voltage (power) that is what's most important. Since you have installed 10 Ohm cathode resistors to use as monitors, what I do is find out what the max. plate dissipation is for said tube. In this case, the 7591, which has a max. Pd of 19 Watts. So, at idle, I read the plate voltage (with the power line voltage 'of your choice'). Say it's 430v. Then I adjust the bias to say 30 ma. The product of which is 12.9w, but you have to subtract the screen grid dissipation, which I estimate at 5%. 12.9w x 0.95 = 12.25w. That's the plate dissipation. Divide 12.25 by 19 and get 64.5%, which is a bit cold according to the "70% dissipation rule". So increase the bias current a little and re-calculate the Pd. until you get 70%. I prefer closer to 75% sound wise, a bit warm, full.
Therefore, bias voltage is simply an average of a sampling of amps. that were produced and is a quicker (but inaccurate) way to set the proper bias level.
And yes, you should add screen grid resistors. I used 470 k Ohm, 1/2w in my similar LK-72 (aka299c). Sorry fellas', my mistake, 470 OHM resistors.
 
Oldman - If you are biased at -19.5v, does the driver put out at least double that or 39 volts pk-pk? Weak drive may be causing your distorted output.
 
Oldman - If you are biased at -19.5v, does the driver put out at least double that or 39 volts pk-pk? Weak drive may be causing your distorted output.
You may have nailed it. One channel was 65v and other 40v (design is 85v). The tubes were from a "nos" sleeve. Found two that met spec and now have 80v and 104v. I questioned why one tab of each box in the sleeve was cut. Now I know.
sleeve 6GH8A.jpg
And emission readings have no meaning as they all tested strong.
Strangely, plate current is actually down a little. But bias is running between -19.5 and -21.5.
 
You may have nailed it. One channel was 65v and other 40v (design is 85v). The tubes were from a "nos" sleeve. Found two that met spec and now have 80v and 104v. I questioned why one tab of each box in the sleeve was cut. Now I know.
View attachment 1026750
And emission readings have no meaning as they all tested strong.
Strangely, plate current is actually down a little. But bias is running between -19.5 and -21.5.
Thanks. 6GH8s are very squirrely as far as consistency goes. I had to choose from 6 pieces to find a good pair. I chose Sylvania. Remember, the plate voltage of the Pentode driver is crucial to bias the phase inverter properly. Get it as close to the schematic as possible for maximum power output. It affects the symmetry of the drive waveform.
 
if that is direct coupled, you'll sometimes get variations in the pentode section that messes with how the triode side works. It changes how the triode biases so the drive doesn't end up being equal. You can meddle with the screen resistor to fix it, but then you have to do it next time you change the tube. Easier to swap tubes until you get a pair that work right. Luckily they're pretty cheap.
 
Thanks. 6GH8s are very squirrely as far as consistency goes. I had to choose from 6 pieces to find a good pair. I chose Sylvania. Remember, the plate voltage of the Pentode driver is crucial to bias the phase inverter properly. Get it as close to the schematic as possible for maximum power output. It affects the symmetry of the drive waveform.
How do I adjust it? Seems like the tubes really make the difference. On the HV, I am at 183vdc and 218vdc on the drivers.
Edit: 203 schematic I am now using.
208.jpg
Transformers are hot.
 
Last edited:
Back
Top Bottom