Blowing rectifier tubes

Hi folks
I've been blowing rectifier tubes like crazy in my st-70 since I put in the VTA mod w/ the power supply cap upgrade 6 months ago.
At the same time I put in a new 80-40-30-20 quad cap with the 80 uf section oriented towards the back of the amp.
I've also done the diode mod with in 4007's and that hasn't helped.
Mostly there have been sparks at the bottom of the tubes after warm up.
Two have glowed bright red, also after warm up.
Voltages at pins 7&5, before the diodes are 380ac.
And the voltage between 8&2 is 5.6 ac.
Does anybody have any ideas?
Seems like a lot of people think that modern gz34's are pretty unreliable.
I did have a PM that lasted 5 months.
Thanks.
Bob

are you running the correct amount of bias?

Hi Krebob

I built two Mark III kits with the same VTA boards... I recall that I managed to "pop" one of the rectifier tubes by incorrect wiring between the quad cap (also an 80-40-30-20) and the VTA board (when building the 2nd amplifier - lost concentration I guess). Check the wiring VERY carefully, unfortunately I can't remember how / what I managed to mess up - but I DID mess up something really simple for sure, and it did take out the rectifier tube.

FWIW, I work from home and my Mark IIIs are running almost constantly every day for going on two years now - probably 1000s of hours of operation - and I have never lost a rectifier tube again (I actually "cycle" between my Mark IIIs and my ST-70, just to spread the "load" a bit).


-- JJ

Hello,
Yes, I've double checked the bias and the wiring around the quad cap and rectifier tube.
I will go through it all again.
Also, I did just put in the triode mod.
I'll go through that wiring again too.
Thanks for your ideas.
Bob

Bob (krebob),

Look at the sockets that sparked under the tubes for carbon tracks. You may need a strong light source to tell if there are any, but it is possible the socket(s) may have a conductive carbon track to the chassis. I found (caused ) a carbon track on my dad's EL84SE stereo amp by accidentally feeding a signal without a speaker connected. That caused an arc to ground, and the arc caused a carbon track to form and it was severely loading the rectifier tube and power transformer.

Fastest socket replacement I ever did in my life. He still doesn't know about it nearly 20 years later.

Tony

what bias are you using?

Hi Folks
The sparks I'm getting are on the inside of the rectifier tubes, towards the bottom.
It's happened to several of them.
The output tubes are fairly new winged c's
The bias I'm using is 0.40 vdc.
Thanks,
Bob

Hi Folks
I'm still trying to understand why I'm going thru all these rectifier tubes.
Any ideas would be greatly appreciated.
Thanks,
Bob

Thanks for your reply Bob
My DMM does have a capacitance test on it.
I took the cap out of the unit and each section measures very close to what it's rated.
I haven't been blowing fuses, just the rectifier tube.
Is it still possible that I have a bad output tube?
Unfortunately, I don't have any gz34's in stock that don't spark right now.
Thanks,
Bob

There are a few 5AR4's to stay away from. The cheap Chinese 5AR4's will always go up in a display of sparks. Another is the JJ tube as these don't last long either. a suitable replacement is the Sovtek 5AR4 rectifier. There is absolutely no need to put out your hard earned bucks on a Mullard or GZ37 from across the pond.

Have to ask this....which cap section follows the 5AR4? About 40-45mfd is all that one can safely use after the 5AR4. Larger first cap than that usually ends up bad.

Hello
So far I've had 2 JJ's, 2 sovetek's go up in sparks.
I had a pm that lasted for 5 months before it started glowing bright red.
I do have the 40uf section of the quad cap following the 5ar4.
I do appreciate the question because I have made plenty of those kinds of mistakes.
I should have a TAD 5ar4 in the mail today.
Thanks for your reply.
Bob

How old is your cap? Is it a NOS? Where did you purchase it? Brand?

It's 6 or 7 months old from Dynakit parts.
My DMM has a capacitance test and shows it to be okay.
Is there any way to test it under full voltage?

Thanks,
Bob

I have had good success with cheap Shuguang and Sovtek rectifiers. I don't think the GZ34/5AR4 are your problem. "Pre-rectifying" with silicon diodes should be fine, but I would avoid bypassing your 5AR4 with diodes until you root cause the failures. Are you still seeing glowing red plates in the output pentodes? You may need to darken the room to see this. By any chance would anyone be cycling your amp on/off without waiting a minute or two between turn-on events?

If you have had good luck with the Shuguang tubes you are in the minority because most do NOT survive past 6 months tops. The Sovtek on the other hand have had a good reliability record and the tube is known on several other forums for reliability. Note* I actually liked the sound of that rectifier up until the sparks flew.

A good post has been made here... what is your bias point 40mA per output tube? or higher? Individual tube bias? Roys board? A bit of personal experience here... had one of Roys boards in the past and the individual 50K bias pots are tricky to adjust and generally do not stay put regardless of if its new tubes or old. The 50K pots just make bias adjustment finiky in my opinion but are still better than the stock boards.

The power supply mod (added diodes) is meant to take the strain off the 5Ar4 because they do the rectifying but the 5Ar4 still allows for that nice slow easy B+ ramp up. If it were me I would lower the bias to 40mA per output tube. Then systematically check the driver board stages for proper voltage and operation. Check and record all voltages and put them up against the normal voltages for this particular driver board.

Hello, and thanks for your interest,
I haven't seen any glowing plates on the output tubes. Only on a couple of the 5ar4's.
I do have one of Roy's VTA boards with individual bias pots set at .40 VDC as per Roy's instructions.
The pots are sensitive, but I haven't had any problem with them.
I should get new rectifier tubes on Thursday.
Is there any other tests I can do without a tube?

Also, can someone tell me what exactly is happening when the 5ar4 starts sparking?
Is there too much current going through the tube?

Thanks again,
Bob

One thing I did when I was having quad cap problems was I bought a single polarized 10-20uF, 600V electrolytic cap from a local guitar amp shop. I then used that cap as a surrogate for the each segment in the quad cap--one at a time--to see which segment was bad. Like you, my quad cap measured okay at the low voltage used by the multimeter. Of course you might have multiple bad segments.

quote
Also, can someone tell me what exactly is happening when the 5ar4 starts sparking?
Is there too much current going through the tube?


Yes, the 5AR4 starts to spark or ark because of too much current. Usually this condition is caused by either a bad section of the Quad cap or a first section that is larger than the recommended value of 40mfd. This is exactly why I am a fan of the add on power supply boards that are available as an aftermarket thought. Once a tube does this the condition only gets worse in my experience. With respect to the bias setting pots... my experience again... even with new matched tubes the bias setting drifts a little. Have had the same experience with 50K pots in my Dynaco's and in my Quicksilver mono block amps. The problem seems to go away with 10K pots. On the positive side the 50K pots allow for a broader array of tubes. My dynaco's will allow the usage of EL34's,6BG6GA's 6L6GC's, and using a switch to modify the circuit I can use KT88's and 6550's.

Quote:
t's 6 or 7 months old from Dynakit parts.
My DMM has a capacitance test and shows it to be okay.
Is there any way to test it under full voltage?



If it were mine I would construct a test circuit that woul allow me to power up the cap from a rectified variac and monitor the current draw on the cap. If the cap were to draw too much then the cap is defective. This is the problem with purchasing caps like the Quad cap off a suppliers shelf. You simply don't know how old it is. I personally have had several of the "new" quad caps fail in the past and whipe out a good rectifier. While not a fan of the Mullard or other expensive rectifiers I do like something along the lines of a Sylvania or Ge rectifier with some age on it.

Your meter will show an obvious shorted condition but not necessarily a cap that is drawing too much on the get go.

Hi Folks
So I did figure out why I was going through all those rectifier tubes.
I had just put in the pentode- triode switches and used the grounded lug on the terminal strip for pin 1.
I must have looked at it a couple dozen times before I "saw" it.
Thanks to all of you for your ideas and support.
I'm learning a lot these days, though I'm often learning the hard way.
I'll be coming back soon with my latest fiasco.
Be Well,
Bob