ST70 Bias Unstable, Oscillation

I have a factory assembled 64 ST70 I’ve used for several years. Some years ago I had the can cap changed, bias caps and input jacks upgraded with Dynakit parts. I also installed a NOS quad of RFT EL34s. Amp played great for some time.

Recently I went through the amp as it had some long term low end hum in the R channel. Tested all tubes, replaced any out of spec resistors, replaced caps on the board with Solens, Micas, reflowed all solder, replaced selenium with a Diode, replaced bias pots with parts from Dynakit, cleaned all tube sockets, etc.

The issue now is when I attempt to bias the amp. Without any input connected, the amp bias sets fine and is stable on both sides. When I connect an RCA cable with a source, the bias on the R swings dramatically up to 3.5v or so, L swings down a bit. If I attempt to correct the bias back down to 1.3v when I get close, the bias goes wild and the amp starts to oscillate with a loud low end thumping. I cannot get the right side then to adjust past .5v without the swing and oscillation. L side can get somewhat stable but also swings.

I’ve tried swapping tubes from side to side, problem stays in R channel. I’ve tried other known good GZ34, EL34 and 7199 and the problem is still present. I’ve been through the amp several times, rebuilt the RCA inputs and nothing changes the problem. Transformers, read proper ohms.

I’ve also tried a direct RCA to iPad and a PAS 3, alternate cables and all create the issue. Hopefully someone else has had the same issue. I would love to put what is a great amp back into rotation.

Images

Correct me if I'm wrong, but if I'm understanding you correctly you are attempting to bias the amp with a signal passing through it.
A tube amp should always have it's bias set at idle and will give the erroneous readings you describe when done so with music playing.

Mike

mijohn wrote:Correct me if I'm wrong, but if I'm understanding you correctly you are attempting to bias the amp with a signal passing through it.
A tube amp should always have it's bias set at idle and will give the erroneous readings you describe when done so with music playing.

Mike

Thanks and I would assume the same, but I am not running a signal. This is only from having connection at the RCA jack with either a iPad or the Pre-Amp.
If I bias without any input but on and warmed up, I can set it fine. When device is connected bias goes instable. I know that a signal through any amp with cause bias to rise and fall but not at all this much.

NIBW wrote:

mijohn wrote:Correct me if I'm wrong, but if I'm understanding you correctly you are attempting to bias the amp with a signal passing through it.
A tube amp should always have it's bias set at idle and will give the erroneous readings you describe when done so with music playing.

Mike

Thanks and I would assume the same, but I am not running a signal. This is only from having connection at the RCA jack with either a iPad or the Pre-Amp.
If I bias without any input but on and warmed up, I can set it fine. When device is connected bias goes instable. I know that a signal through any amp with cause bias to rise and fall but not at all this much.

Will it bias properly with the preamp connected but turned off? That's the way I bias my amp. If The ST-70 sounds fine when playing maybe there isn't a problem. Why do you need to re-bias after the preamp is connected?

mijohn wrote:

NIBW wrote:

mijohn wrote:Correct me if I'm wrong, but if I'm understanding you correctly you are attempting to bias the amp with a signal passing through it.
A tube amp should always have it's bias set at idle and will give the erroneous readings you describe when done so with music playing.

Mike

Thanks and I would assume the same, but I am not running a signal. This is only from having connection at the RCA jack with either a iPad or the Pre-Amp.
If I bias without any input but on and warmed up, I can set it fine. When device is connected bias goes instable. I know that a signal through any amp with cause bias to rise and fall but not at all this much.

Will it bias properly with the preamp connected but turned off? That's the way I bias my amp. If The ST-70 sounds fine when playing maybe there isn't a problem. Why do you need to re-bias after the preamp is connected?

After further thought I realize what I wrote previously was wrong and there is a problem. As you noted one channel is biasing normally. If the problem surfaced after your upgrades a detailed comparison between the good and bad channels should be done, but you seem have done all the right things there. Do the input devices need to be powered up to cause the instability or does just plugging them in unpowered have the same result?

PAS emits a small DC component at it's output. The non-X models about 0.6V, this can upset
the first tube in ST70 as teh amp has no input capacitor.

Perhaps check the bias pot. Maybe pot is bad, wiper is dirty ? And short the inputs for good measure.

mijohn wrote:

mijohn wrote:

NIBW wrote:

mijohn wrote:Correct me if I'm wrong, but if I'm understanding you correctly you are attempting to bias the amp with a signal passing through it.
A tube amp should always have it's bias set at idle and will give the erroneous readings you describe when done so with music playing.

Mike

Thanks and I would assume the same, but I am not running a signal. This is only from having connection at the RCA jack with either a iPad or the Pre-Amp.
If I bias without any input but on and warmed up, I can set it fine. When device is connected bias goes instable. I know that a signal through any amp with cause bias to rise and fall but not at all this much.

Will it bias properly with the preamp connected but turned off? That's the way I bias my amp. If The ST-70 sounds fine when playing maybe there isn't a problem. Why do you need to re-bias after the preamp is connected?

After further thought I realize what I wrote previously was wrong and there is a problem. As you noted one channel is biasing normally. If the problem surfaced after your upgrades a detailed comparison between the good and bad channels should be done, but you seem have done all the right things there. Do the input devices need to be powered up to cause the instability or does just plugging them in unpowered have the same result?

I’ll have to try it with only device plugged in but not powered up. I do know that if a cord is plugged in it is fine, it’s only when I connect to a device.

I have PAS 3x but in the past I was able to bias the amp without this issue.

peterh wrote:PAS emits a small DC component at it's output. The non-X models about 0.6V, this can upset
the first tube in ST70 as teh amp has no input capacitor.

Rec wrote:Perhaps check the bias pot. Maybe pot is bad, wiper is dirty ? And short the inputs for good measure.

Original bias pots were cleaned and ultimately replaced with new.

Tried again to clean and rebuild the input RCAs as it almost appears as a ground issue. Cleaned up the board again and bias was the same, although now with the bias pot turned all the way down, the lowest I can get the voltage is 3.6v or so. Very odd.

Measured voltage across the board, the only flag was the 2 1.5M resistors. On the L channel there is about 39v coming off the 1.5m, on the R it’s about 19v. Not sure this is the issue but I’ll certainly replace those as they are original.

Had a mind to look into the transformers and behold! The R channel was a disaster of a repair, more importantly the primary leads were wrong. I cleaned it up some added new leads in close to correct wire color and reflowed some more solder.
A few tests and the Bias was stable as it could be. Been working on tracking this down since Sept! Very well worth the effort and happy to rescue the original iron.

Well done, you should be quite pleased with yourself!

It's amazing what those little gremlins can do.  I had a 12kv power supply that would short the moment it was energized, popping the huge main breaker.  A few of our techs worked on it and couldn't figure it out.  Then I was called in.  Turned out to be a bad lead to the primary coil that's supposed to pull the grounding contacts up, allowing the large capacitors to charge.  Resoldered the lead to the coil, the grounding contacts pulled up, and no more scary, loud bang, lights out, pants changing incidents.  Oh, and the transmitter powered up also.

Nice save! .....well done!

1] Make sure that your AC V mains input voltage is not more than 117 (120VAC, worst case scenerio).
2] Clean your valve sockets.
3] Test AC input voltage at full wave 5AR4. If it is over 540 VAC, then your mains are delivering voltage that is too high. (A common problem in the USA). The stock ST70 was made to run on 117 VAC and not over 120 VAC. The 5AR4 valve is being pushed to it's limit.
4] You can try to re-solder the pins on the power valves. Do this fast and precisely with a med-hot regulated iron. No promises; one or more EL34s are likely shot from main voltage that is too high.
5] Check the voltage in and out of your bias diodes. Also check the B+ from the output trannies.
6] Basically, thoroughly check every POWER SUPPLY voltage point in the original Dyna manual. Get to know this manual. You will be surprised.
7] You need to properly rebuild this unit. go to Triodel and get their Cap Board, a new preamp / phase spliter board (the EF86 board is best). Make sure ALL resistors are 1% Metal film. The phase splitter resistors MUST be 1% metal film. USE 1/2 WATT resistors. Yes, 1/4 Watt will work...for a while. Replace your bias resistors with two 1 Watt metal film resistors (somewhere around a 42 and 45 ohm in parallel, —if my memory serves. Bias VDC need not be exactly 1.56 DCV. Bias voltage is non critical to a couple 1/10ths of a volt. 1.56 DCV was originally used because most amateurs had only a basic analogue VOM, NOT a VTVM nor a modern digital meter. Old VOMs had a low impedance input that made then less than accurate. Plus, there was resonance and/or induction in most of them. 1.56 VDC allowed calibration with a dry cell battery.)
All signal caps should be poly-film. Orange drops are fine. Replace your octal valve sockets. Ceramic are best and cheapest. Be careful when mounting. Do NOT over torque. Get fibre washers for mounting. If there are rivets than use a dremmel grinding tool to remove the rivets.
8] A run of NEW replacement power x-formers was made at the old Thoradson factory in Chicago. These supply 1/3d more current! —A great upgrade and you can run almost any power valve with only a minor bias reset. If they still exist, Triodel will be able to get these as well. Last I saw, they were about $250. Money well spent.
I use a VARIAC to make sure that I ALWAYS feed my valve gear exactly 117 VAC. You can also buy an APC 500 Watt regulator and set it on 110 VAC (Mexican). Reset your bias and your amp will run fine. You valves will also last MANY times longer.

The APC 500 Watt voltage regulator has 3 settings. REMEMBER, choose 110 VAC. They sell for under $100. Also, GREAT for PCs as they will keep your PC up during brown outs less than one second long. They also protect against spikes. Ordinary power strips (just Varistors) are worthless. Save your money.

THE most important part of ANY Williamson, class A/B power amp is the POWER SUPPLY. This is where most failures occur. As with ANY amp, remember that the music you listen to is created by the power supply. —I.e., you are LISTENING to the power supply! The cleaner and more accurate the power supply, the better the fidelity.

Good luck

PS: Get a digital VOM if you do not have one (it's far better than a VTVM) . Get a 35 - 40 Watt REGULATED soldering iron. It makes a BIG diff. See AES in Tempe AZ, as well as TRIODEL

I like what bixlives! says above, and completely agree... I'd add one thing to address his concern about the voltage... Put a NTC in series on the power cord. You can use a CL-80 or a pair of CCL-90s in parallel, this will do two things. First, it will drop the voltage a couple of volts bringing it closer to the 117 VAC that Dynaco originally wanted. Secondly and perhaps just as important, the NTC creates a slow turn on, this especially affects the tube heaters. According to one 1950's expert, it increases tube heater life expectancy by 200-300 percent.

A cheap improvement that really pays off.