• WARNING: Tube/Valve amplifiers use potentially LETHAL HIGH VOLTAGES.
    Building, troubleshooting and testing of these amplifiers should only be
    performed by someone who is thoroughly familiar with
    the safety precautions around high voltages.

Help with ST-70 variant hum in both channels

The red/yellow CT wire should ideally connect as near as possible to the negative of the first filter cap in the power supply. This is the highest ripple current connection in the amp and you don’t want that ripple causing noise in other parts of the circuit.
 
  • Like
Reactions: rayma
I guess my current questions would be: Is there any ground connection (other than the power cord ground) that should connect at the chassis rather than to the ground-loop breaker circuit?

How/where do I test ripple? (I have a scope and know how to put it in a mode to remove DC from displayed AC wave and vice-versa.)

If I'm going to try physical shielding with anti-EMI/RF film or tape, where do you recommend I apply it?

Do you think the negative feedback selector boards could be a source of noise? (there are three different pairs of caps and resistors with a jumper used to select which pair is in the feedback path, depending on whether you're using the 4-ohm, 8-ohm, or 16-ohm impedance taps.) I only have 8-ohm taps connected so don't really need the boards but put them in for future flexibility. There are a couple of small (2-inch) wires hanging off of them which would be the outputs to currently non-existent speaker outputs.

Is there any point in twisting high-voltage DC wires together for the short distance that they run near each other (and run near the wires for the choke)?

Thanks to everyone who's trying to help me with this exceptionally frustrating situation.
 
Is there any ground connection (other than the power cord ground) that should connect at the chassis rather than to the ground-loop breaker circuit?
The frames of transformers and chokes should be electrically connected to chassis; scratch off paint and/or anodization and connect with star washers. This is to provide a safe path in case of insulation failure from windings to framework - you wouldn't want the framework to be sitting at B+ relative to the chassis and inches away - and this failure mode definitely sometimes happens.

Tracking down hum issues is only difficult because we believe schematics. They invariably pretend that something called "ground" exists, and that it is a magical source/sink of electrons/charge. If we instead think about how current flows and remember that voltage is generated across even small impedances like "grounds" we don't fool ourselves quite as badly. The best rule of thumb is to not let noisy currents travel though any significant impedance in a way that can be added to signal. For example, a DC power supply should have a path, preferably a twisted pair, of B+ and B- (not! "ground") that makes its way towards the amplifier. Then, the amplifier should have the same consideration: from the output stage back towards the input each stage's B+ should be bypassed to its local B- at a location where its residual noise doesn't impinge on signal (normally at the bottom of the cathode resistor).

Golden Age equipment usually makes a nightmare of this, with multiple chassis "grounds" and other loops, but the "one and only one" rule always applies. Multi-section can capacitors are a common problem, with a bunch of B-'s tied together including the first/worst one off the rectifier, very non-ideal, so much so that I'd really recommend forgoing the ground loop breaker (there're plenty of other loops) unless it's unavoidable. Plug everything into the same power strip, which keeps the ground loops small and local, and it's very rare to have any issues with the PE ground loop.

In summary, think about how charging pulses generate noise voltages across small (but not small enough to forget) impedances, and you can solve this puzzle yourself. It's very much just Ohm's Law stuff.

All good fortune,
Chris
 
Thanks, @Chris Hornbeck. I do have a lot to learn and am enjoying learning it even while being frustrated at the difficulty of finding the issue.

I have read in some places about B- but that term is not referenced in this amplifier so I'm not sure what it means. I mean, I'd think it would mean negative DC voltage but I don't know if I have that here. (And I don't know if I MUST have it if I have B+.)

Since this kit has been successfully built as "dead quiet" by quite a few people, and since it's unlikely that brand new parts have failed, I must assume that I've done something wrong. Small enough to be very hard to find, but wrong.

I notice that the LEDs for the auto-bias board are very slightly dimmer (but enough to notice, obviously) on the right side than on the left, that suggests to me that I have an improper connection between something and ground. Does that logic seem sensible?

And since the ground-loop breaker board didn't end the hum, it seems that the hum must be within the circuitry that's connected to that board, which may be why the hum doesn't go away even when I use a cheater plug on the power outlet.

What I'm not sure of is what to measure. So far, using fairly precise "4-wire resistance" functionality, I can't find anything on the right channel that tests as if it has lower resistance to ground than does the equivalent point in the left channel.

I'm thinking I should desolder all the wires to the right channel power tubes and then reattach them.

I'm also going to take out the negative feedback selector boards even though I'd be surprised if they're part of the problem.

BTW, as you probably noticed already, I don't have a can capacitor in this configuration...

Thanks again!
 
Since we have no idea what the auto-bias board or the negative feedback board are like, no one can comment specifically to help you. Be sure to thank the originator of this kit, who's Secret Sauce is so secret and ground-breaking that it can't be exposed to public scrutiny.

This is diyAudio, where ideas are not secret. All good fortune,
Chris
 
  • Like
Reactions: rayma
So I think the measurement I'm supposed to get related to that board is around 400 mV. And I'm getting 402 or 403 which is less than one percent off so I think it is OK. At least that part is OK. I don't know if anything else could be an issue though.

I tried to do a little bit of testing looking for too much or two little resistance to Ground at various points, but I couldn't find any of that kind of error.
 
Just an aside re: the bias supply. It is supplied to the bias board so if it has much ripple that is transmitted to the grids of the output tubes. If they are reasonably well matched then most of it should cancel in the output transformer.
 
What seems out of whack is that your B+ was below 400V (if I'm remembering right). That's why I would check that you have the expected roughly 50mA per output tube. Presumably there is a sense resistor in the cathode of the output tube(s) the board is getting it's measurement from that could be checked. The other likely culprit is the power transformer.
 
  • Like
Reactions: Francois G
I'm supposed to get related to that board is around 400 mV. And I'm getting 402 or 403 which is less than one percent off so I think it is OK.
I assume you measure across each of the 10 Ohm resistors. That represents .400/10 = 0.040 amps through each EL34 cathode which is fine.

Could we recap what you have done so far? And the outcome. Here is what I recall and with the help of your post #111:

1. Replaced the rectifier diode. No change
2. Tightly wisted and moved wiring around. Resulted in slight improvement of the hum.
3. Installed the ground loop breaker and move all circuit grounds, including the red/yellow center tap connects to ground of the ground loop breaker, not to chassis. No change observed, right?
4. Replace 6sn7s with a fresh set en swapped the EL34s from channel to channel. No change.
5. You disconnected the feedback wire by removing the jumper on you seletable feedback boards. This increased the hum. You then replaced the feedback wires and used shielded wire for feedback. Did that make any difference? In removing the jumper originally I realize that the long wire to the PCB input could still pick up induced AC along the way.
6. You shorted the input signal to the first triode directly at the PCB. No difference.
7. You checked the B+ voltage after changing B+ resistor and it was slight higher that the Miller manual listed. You were planning to get in between values to get it just right, but did not see any difference in hum level. Here is where the power supply gets murky for me. The resistor value is given in the Miller manual (page 10) as 2.2k to 4.3k and the ideal B+ as 370-390 Vdc. Such a high value resistor can not be controlling the actual B+ that goes to the OT centertaps, as @jgf implied above. My guess is that the 370-390 V is really the supply to the phase splitter. I still would like to see a diagram of the PS, but perhaps you could start with a chart of voltages you now measure at each pin on the EL34 sockets.

The white wire is the ground connection of what I think are filter caps. (Two caps with the outer leg of each going to green and brown transformer wires, respectively, and inner leg of each tied together and to ground (to ground loop breaker.)
I’m not sure what you meant here. What are the brown and green transformer wires? There should not be transformer wires directly going to filter capacitors. I guess Miller is using lower voltage series pairs to make up the individual HV filter caps.

In post #111 could you explain exactly what is the graph you posted and how did you get it? That seems very noisy and messy. Now that we basically have eliminated almost everything else as the source of the hum we should make sure the PS works correctly. This messy graph does not look right to me.

When I run the transformer noise through a spectrum analyzer it seems, oddly, to have a spike at 90Hz (and a much bigger one at 120Hz).
 
Last edited:
Where exactly are you measuring the current mentioned - “amps drop from .335 to .312” . Is it across R B+? What is its value? The 0.312 amp current you measured is much higher than I would expect for a Dynaco ST70 at idle. And the relatively high value (2k-4k) of R B+ makes me think that it is only serving the PCB and NOT the power tubes. (For other readers- R B+ is shown upper left of the PCB photo on page 11 of the Mille Audio Manual previously referenced)
Has this been addressed? Do we know the current the amplifier(s) draw is as expected when idling? I still would like an answer to this question.
 
You say that you removed the wiring from the rca Jack's to the driver board? But does that mean the input circuit is open circuit? You should reconnect the rca's and make some shorting plugs from some spare input jackc or use an old cable and cut the ends off. Turn on and see if it hums then?
 
Has OP @Rossputin given up on the amplifier? If not we would love to have an update.
It seems to me without complete circuit diagrams this is a case of "terminal complexity".
A situation where due to the complexity of the technologies used issues increase at a faster rate than a user's trouble shooting can ever comprehend.

The only way to move forward from "terminal complexity" is to simplify by disabling almost every function and then start by resorting one function at a time until the noise issue presents it's self or you end up at a fully functioning quiet amplifier.
However a complete set of circuit diagrams is needed to start.

The hum could be due to many many things, excess bias current in the output tubes-imbalance in the output tube bias current setting-instability in the output tube bias level (all auto bias issues), imbalance in the tubes themselves (matching or defect), ground loop issues, heater noise pickup in the input tubes (imbalance in grounding or lack of DC heater offset voltage), power supply ripple, rectifier transformer secondary ringing (lack of snubbers) and on and on.
Some of these may well be design issues in the mod circuit boards provided and with no circuit diagrams...

Just my opinion however I personally will not purchase "black box" products that sellers will not provide full service information on.
It has always ended over the long run the same.
Later for some reason or other the manufacture drops support or charges punitive costs to repair, then the product malfunctions, now becoming just more for the land fill.