ng5921's DIY Amplifiers

[L0rdGwyn]

Hey all!

Got a couple updates. I've been working to track down this infuriating hum in the amplifier, and I could use input from the other DIY gurus on here. @LobalWarming, @g0ldl10n, @funch, @L0rdGwyn, hope ya'll don't mind me pinging you guys.

So from what I can tell, it is a 60hz hum, and I initially suspected a ground loop because the hum intensity is constant regardless of the amplifier volume. I am convinced it is not a 120hz hum, because it does not change intensity with increasing/decreasing the first filter cap in the power supply. I am also reassured it is not noise/interference because there is no change in intensity with volume changes.

Things I have tried without any improvement in the hum:

• Using a star-of-stars grounding scheme, which carefully separates the signal grounds grounds, power grounds, and cathode grounds, and all meet at the chassis star
• Using a ground loop breaker circuit
• Increasing capacitance of first filter cap in the power supply (it is currently 8uF)
• Grounding the shielding of the Mogwai input signal wires from the RCA sockets only at the RCA socket (which is then tied to the signal ground)
• Plugged the amp into different electrical circuits in my apartment

Here is a drawing showing the grounding schema:

(The ground loop breaker cap is actually a film cap, I drew it wrong).


I am starting to suspect the power transformer may be at fault. I am using a Hammond 272FX. I am looking at the Lundahl LL1683 as a replacement.


My other ideas for addressing the hum are to add a high/lo gain switch on the headphone output, and to negative feedback on a toggle switch. There is no hum with lower sensitivity cans.

Could use any and all help, at this point I am out of ideas! Pics of the amp internals are attached for reference.

A few thoughts. Noise not being variable with the turning of the potentiometer just means that it isn't coming from the amplifier inputs, you can have constant noise that is being picked up further downstream that will not change when turning the pot. 120Hz noise can enter the circuit through means other than power supply ripple. Noise can be injected into the signal path from 120Hz ripple currents from the rectifier pulses charging the reservoir cap, so it is important to distinguish whether this is 60Hz or 120Hz noise. This is why it's advisable to keep the transformer - rectifier - reservoir capacitor loop away from the signal path. 120Hz noise tends to sound more like a buzz than a hum given the sawtooth waveform. If it's a 60Hz hum, it could be radiation from the transformer. Try putting a steel pot over the transformer and see if it reduces the noise. If so, switching to the Lundahl may help. Using their branded steel enclosures will reduce radiation further. If this is an OTL circuit and it's using 6922, the circuit may also just have an overabundance of gain which is going to raise the noise floor. You could try switching to a lower gain input tube, which may bring the hum down low enough to be inaudible.

 

[ng5921]

A few thoughts. Noise not being variable with the turning of the potentiometer just means that it isn't coming from the amplifier inputs, you can have constant noise that is being picked up further downstream that will not change when turning the pot. 120Hz noise can enter the circuit through means other than power supply ripple. Noise can be injected into the single path from 120Hz ripple currents from the rectifier pulses charging the reservoir cap, so it is important to distinguish whether this is 60Hz or 120Hz noise. This is why it's advisable to keep the transformer - rectifier - reservoir capacitor loop away from the signal path. 120Hz noise tends to sound more like a buzz than a hum given the sawtooth waveform. If it's a 60Hz hum, it could be radiation from the transformer. Try putting a steel pot over the transformer and see if it reduces the noise. If so, switching to the Lundahl may help. Using their branded steel enclosures will reduce radiation further. If this is an OTL circuit and it's using 6922, the circuit may also just have an overabundance of gain which is going to raise the noise floor. You could try switching to a lower gain input tube, which may bring the hum down low enough to be inaudible.

Ah I didn't consider that the noise could be injected after the amplifier stage, I appreciate that insight. I will try and hook up the amplifier to my Analog Discovery 2 this weekend and measure the frequency of the hum since I may be misinterpreting the sound. Will give the steel pot a try tomorrow too and look into some lower gain tubes.

Thanks L0rdGwyn!

 

[L0rdGwyn]

Also if you're using AC heaters, make sure they are center tapped and ground referenced. If they already are and nothing else is working, you could also try DC elevating the heater supply, this can solve some heater to cathode leakage noise issues. Sorry for the quick notes, have to get up early for work, off to bed.

 

[g0ldl10n]

Hey all!

Got a couple updates. I've been working to track down this infuriating hum in the amplifier, and I could use input from the other DIY gurus on here. @g0ldl10n, @funch, @L0rdGwyn, hope ya'll don't mind me pinging you guys.

So from what I can tell, it is a 60hz hum, and I initially suspected a ground loop because the hum intensity is constant regardless of the amplifier volume. I am convinced it is not a 120hz hum from insufficient filtering by the power supply because it does not change intensity with increasing/decreasing the first filter cap in the power supply.

Things I have tried without any improvement in the hum:

• Using a star-of-stars grounding scheme, which carefully separates the signal grounds grounds, power grounds, and cathode grounds, and all meet at the chassis star
• Using a ground loop breaker circuit
• Increasing capacitance of first filter cap in the power supply (it is currently 8uF)
• Grounding the shielding of the Mogwai input signal wires from the RCA sockets only at the RCA socket (which is then tied to the signal ground)
• Plugged the amp into different electrical circuits in my apartment

Here is a drawing showing the grounding schema:

(The ground loop breaker cap is actually a film cap, I drew it wrong).


I am starting to suspect the power transformer may be at fault. I am using a Hammond 272FX. I am looking at the Lundahl LL1683 as a replacement.


My other ideas for addressing the hum are to add a high/lo gain switch on the headphone output, and to negative feedback on a toggle switch. There is no hum with lower sensitivity cans.

Could use any and all help, at this point I am out of ideas! Pics of the amp internals are attached for reference.

Whenever I have encountered constant hum/noise (although, it would be important to validate the frequency of the hum you are hearing through measurements) where the volume pot did not have any effect on it was when magnetic coupling occurred from poorly placed wire runs, and or transformer placement. The other situation I have had constant noise was bad solder joints, but this was probably not 60 or 120Hz hence why its important to validate the frequency of the noise/hum you are hearing.

Also, I am going to assume this is in both channels?

 

[ng5921]

Yes the hum is equal in both channels. I also just tried the steel pot and covered various parts of the amplifier, no reduction in the hum unfortunately, making EMI from the transformer to the tubes unlikely. Could definitely still be a wire inside the amp that is causing interference with something.

I might have been wrong about the frequency of the hum fellas. It does not sound exactly like a 60hz, and not exactly like 120hz, so I used my phone and a frequency measuring app. I shoved the phone mic into the headphone cup, and measured a peak around 120-140hz, so maybe this is a 120hz hum and I did not hear it correctly.

So my initial assumptions would have been wrong, and I'll try to confirm this finding with my Analog Discovery 2 over the weekend.

I will also poke around the amp with a chopstick while it is powered on and see if the hum changes at all.

 

[g0ldl10n]

Yes the hum is equal in both channels. I also just tried the steel pot and covered various parts of the amplifier, no reduction in the hum unfortunately, making EMI from the transformer to the tubes unlikely. Could definitely still be a wire inside the amp that is causing interference with something.

I might have been wrong about the frequency of the hum fellas. It does not sound exactly like a 60hz, and not exactly like 120hz, so I used my phone and a frequency measuring app. I shoved the phone mic into the headphone cup, and measured a peak around 120-140hz, so maybe this is a 120hz hum and I did not hear it correctly.

So my initial assumptions would have been wrong, and I'll try to confirm this finding with my Analog Discovery 2 over the weekend.

I will also poke around the amp with a chopstick while it is powered on and see if the hum changes at all.

I have a feeling it's your grounding bar scheme - this same sort of deployment caused me many headaches in my own build with 120Hz hum until I changed it. Essentially, I removed all the grounding bar except a single piece/wire ran down the center, which connected to the chassis at a single point, then tied each channel/stages grid, cathode, and filter caps grounds to the exact same point on this bar, one behind the other. Once I did this, all my 60Hz and 120Hz nightmares disappeared (and only having the CT of the PT tied to the same grounding point on this as the 1st filter cap, as I think your drawing eluded to). Originally, I was placing the grounds as conveniently as possible to these grounding bars sent to each location/stage/channel - maybe I did something wrong originally, but once I changed it to what I just described, extremely quiet amp now.

 

[ng5921]

I have a feeling it's your grounding bar scheme - this same sort of deployment caused me many headaches in my own build with 120Hz hum until I changed it. Essentially, I removed all the grounding bar except a single piece/wire ran down the center, which connected to the chassis at a single point, then tied each channel/stages grid, cathode, and filter caps grounds to the exact same point on this bar, one behind the other. Once I did this, all my 60Hz and 120Hz nightmares disappeared (and only having the CT of the PT tied to the same grounding point on this as the 1st filter cap, as I think your drawing eluded to). Originally, I was placing the grounds as conveniently as possible to these grounding bars sent to each location/stage/channel - maybe I did something wrong originally, but once I changed it to what I just described, extremely quiet amp now.

I've been wondering about that too. I'm running the grounding bar between the ground terminals of the two blue capacitors, which are secondary filter caps in the power supply, specifically for the power tubes. This might not be the best spot to have that grounding bus bar run.

The simplest option might to just be abandoning the grounding bar, and routing all these grounds to a star ground and then tying that to the chassis star.

 

[g0ldl10n]

I've been wondering about that too. I'm running the grounding bar between the ground terminals of the two blue capacitors, which are secondary filter caps in the power supply, specifically for the power tubes. This might not be the best spot to have that grounding bus bar run.

The simplest option might to just be abandoning the grounding bar, and routing all these grounds to a star ground and then tying that to the chassis star.

Yes, and I know that connecting each of the stages grid, cathode and filter caps grounds together (or exactly at the same spot), but "behind" one another was also crucial - essentially, you don't want grounding/ripple currents from the noisier stages like the 1st PS filter cap flowing over the other grounding points. I am pretty sure Merlin's grounding pdf goes over this in more a more detailed sense.

https://www.valvewizard.co.uk/Grounding.pdf

 

[ng5921]

Ripped out the bus bar and replaced everything with star grounding, and the hum is still there

Next step, poke around inside the amp with a chopstick and see if moving any wires makes the hum change.

 

[g0ldl10n]

Ripped out the bus bar and replaced everything with star grounding, and the hum is still there

You got pictures of this?

@ng5921 oh, and forgot to ask, since something similar effected my amp - you're sure its not the rectifier tube, and or voltage regulator tube emitting its EMI/EMF effecting the power, and or input tubes? Have you tried shielding either of those? I see the voltage regulator sits quite close to the input tube, although, I am unsure how much EMI/EMF it emits externally, but id try shielding them to ses (excuse me if you've already tried this).

 

[ng5921]

@g0ldl10n Here are some pics! The power supply grounds are all using green wires. They tie to 2 separate stars, which then route to the chassis star ground.

I also poked around in the amp with a chopstick while it was powered on, no changes in hum intensity/frequency despite moving a bunch of wires around.

I don't think it is EMI/EMF - earlier today I used a steel pot (per l0rdgwyn's suggestion) and covered up various elements in the amp and there was no change in the hum. I've also tried different rectifiers, different voltage regulators, different input and power tubes..

Also, not sure if this is of any significance, but when listening to music if I turn off the amp the sound will continue to play for maybe 20 seconds, during which there is no hum at all.

 

[Swagsupply]

@g0ldl10n Here are some pics! The power supply grounds are all using green wires. They tie to 2 separate stars, which then route to the chassis star ground.

I also poked around in the amp with a chopstick while it was powered on, no changes in hum intensity/frequency despite moving a bunch of wires around.

I don't think it is EMI/EMF - earlier today I used a steel pot (per l0rdgwyn's suggestion) and covered up various elements in the amp and there was no change in the hum. I've also tried different rectifiers, different voltage regulators, different input and power tubes..

Also, not sure if this is of any significance, but when listening to music if I turn off the amp the sound will continue to play for maybe 20 seconds, during which there is no hum at all.

Sounds like it might be transformer hum

 

[ng5921]

Not purchasing a transformer yet (still need to verify if this is 60hz vs 120hz on the AD2), but wanted to think about options for the transformer (ideally 115/120V 60hz, secondary 300-0-300V CT @>100mA, 5V CT @3A, 6.3V CT @5A)

1) Get another Hammond 272FX ($120ish) and hope it does not hum. Benefit is it will mount without modifying chassis

2) Go with Lundahl LL1683 (~$200), 3D print a bracket and find a transformer cover

3) other transformer manufacturer I should consider?

4) custom transformer?

Problem with (2) is that I can't mount the lunadhl metal housing on my existing chassis in the standard orientation (Lundahl logo horizontal), and if I turned it 90 degrees it would end up orienting the transformer magnetic field in line with the rectifier tube which wouldn't be good. So I would 3D print a mounting bracket that orients it properly, and find a metal transformer cover that will fit or 3D print one (but then there's no magnetic shielding).

 

[Swagsupply]

I've got a couple options to consider for transformers.

1) Get another Hammond 272FX ($120ish) and hope it does not hum. Benefit is it will mount without modifying chassis

2) Go with Lundahl LL1683 (~$200), 3D print a bracket and find a transformer cover

3) other transformer manufacturer I should consider?

4) custom transformer?

Problem with (2) is that I can't mount the lunadhl metal housing on my existing chassis in the standard orientation (Lundahl logo horizontal), and if I turned it 90 degrees it would end up orienting the transformer magnetic field in line with the rectifier tube which wouldn't be good. So I would 3D print a mounting bracket that orients it properly, and find a metal transformer cover that will fit or 3D print one (but then there's no magnetic shielding).

Maybe try an edcor

 

[g0ldl10n]

Not purchasing a transformer yet (still need to verify if this is 60hz vs 120hz on the AD2), but wanted to think about options for the transformer (ideally 115/120V 60hz, secondary 300-0-300V CT @>100mA, 5V CT @3A, 6.3V CT @5A)

1) Get another Hammond 272FX ($120ish) and hope it does not hum. Benefit is it will mount without modifying chassis

2) Go with Lundahl LL1683 (~$200), 3D print a bracket and find a transformer cover

3) other transformer manufacturer I should consider?

4) custom transformer?

Problem with (2) is that I can't mount the lunadhl metal housing on my existing chassis in the standard orientation (Lundahl logo horizontal), and if I turned it 90 degrees it would end up orienting the transformer magnetic field in line with the rectifier tube which wouldn't be good. So I would 3D print a mounting bracket that orients it properly, and find a metal transformer cover that will fit or 3D print one (but then there's no magnetic shielding).

You dont have any other PT on hand that you could sub in at all, that may not be the exact voltages you need to verify/isolate it to transformer humming? Maybe even just for the input and output heaters, and or possibly the rectifier heaters? Possibly just external PSU for the input and output heaters?

And if not, do you have a full SS bridge on hand you can swap in just to make sure the noise isn't coming from the current rectification/power supply setup? I know you said you increased the capacitance of the resevoir cap, but I guess I am just one that would want to be absolutely positive before dropping money on a new mains PT, only to get it in and install it, only for the hum to still be there.

And btw, did you ever use the analog discovery to verify the frequency of the hum?