EL34, Claus Byrith amplifier, mains noise problem

[MadsKaizer]

Hey All

I built a amplifier after the schematics revised by Claus Byrith and is now testing and adjusting it. I have after his instructions managed to set DC/AC balance and bias point to within the specified limits.

But I have a mains clicking/hum that I do not seem to be able to get rid off.

I have grounded downstream this way "mono block -> power supply -> star point on chassis"

Tripled filtering capacity of the power supply without any noticeable change.

Tried 1nF decoupling on the power supply

Moved power transformer around in different alignments to the output transformer, no change either.


Heater supply is AC, but run as twisted wires and as much as possible runs perpendicular to signal wires.

I hope that some of you experienced tube experts can take a look at this video and maybe hear what my problem is, I am still learning

The sound in the video is a sine wave input from 100Hz to 5kHz, you might have to turn up your sound to hear the 50/100Hz noise...

MVI_5140.MOV - YouTube

[nigelwright7557]

Sounds like just mains hum.

Try removing earth to zero volts connection.
Or try adding earth to zero volts connection to see if it changes.

[DualTriode]

Hello,
It looks like you are having fun, that is the important part.
You have it all hanging out bare naked on the bench. I do not see it in the video but I bet it is there, your soldering station. With your soldering station turned on pick it up and move it around your operating amplifier. It will surprise you how sensitive all that open circuitry is to Electro Magnetic Radiation. That soldering is a multitool.
Try putting the amplifier in a temporary shielded, grounded metal box.
DT
All just for fun!

[MadsKaizer]

Originally Posted by nigelwright7557

Sounds like just mains hum.
Try removing earth to zero volts connection.
Or try adding earth to zero volts connection to see if it changes.

Grouding it all, all the way to the wall socket dampens the output to about 10%, both noise and signal. I am not all too sure about this.

Originally Posted by DualTriode

Hello,
It looks like you are having fun, that is the important part.
You have it all hanging out bare naked on the bench. I do not see it in the video but I bet it is there, your soldering station. With your soldering station turned on pick it up and move it around your operating amplifier. It will surprise you how sensitive all that open circuitry is to Electro Magnetic Radiation. That soldering is a multitool.
Try putting the amplifier in a temporary shielded, grounded metal box.
DT
All just for fun!

The soldering iron has no influence on the amplifier in anyway, I have tried waving it everywhere without any change.

I measured 8 to 12 Volt ripple, with a differential probe on my scope, in the 420VDC in the power supply, I tried adding a large 3300uF capacitor without any change either.

I also tried a line in filter from a old PSU, no change either.

After I added the 3300uF capacitor in the power supply it have enough energy stored to continue operation after the AC line power is removed, here there is no problem at all with noise, its crystal clear sound from the signal generator. Not a surprising discovery, but shows its something in the amplifier and not a external source.

[DF96]

What grounding arrangements have you used, especially around the PSU? Did you use his PCBs? You may be injecting charging pulses into your signal ground.

[MadsKaizer]

Originally Posted by DF96

What grounding arrangements have you used, especially around the PSU? Did you use his PCBs? You may be injecting charging pulses into your signal ground.

As I described in the first post, its grounded down stream.

I made my own PCBs, images are PCB layouts and the bmp files are the schematics.

Do you mean switching noise from the bridge rectifier? It puzzles me that 3300uF above the original 200uF did not make any difference in ripple voltage.

[DF96]

No, not switching noise but charging pulses. Your PSU PCB attaches both the heater supply CT and the bias supply to potentially noisy points on the ground trace. In addition, the bias supply has its ground attached to the input instead of the output. PSU ground connections should be clean, not dirty, so they establish potential and conduct DC rather than AC. Changes I would suggest are:

- connect junction of R4/R5 to Ground Out, instead of current location.
- remove connection of D6/D7 to ground trace.
- add new bias ground connection from C8 + to C6 -.

You can probably make these changes temporarily by cutting traces and adding wire links. Adding extra capacitance to a poorly grounded PSU can make things worse.

The amp PCB looks OK. Make sure the OPT secondary is grounded to the amp PCB input, not output. Otherwise any voltage drops across the PCB ground trace will get added into the feedback voltage, which may create distortion.

[MadsKaizer]

Originally Posted by DF96

No, not switching noise but charging pulses. Your PSU PCB attaches both the heater supply CT and the bias supply to potentially noisy points on the ground trace. In addition, the bias supply has its ground attached to the input instead of the output. PSU ground connections should be clean, not dirty, so they establish potential and conduct DC rather than AC. Changes I would suggest are:

- connect junction of R4/R5 to Ground Out, instead of current location.
- remove connection of D6/D7 to ground trace.
- add new bias ground connection from C8 + to C6 -.

You can probably make these changes temporarily by cutting traces and adding wire links. Adding extra capacitance to a poorly grounded PSU can make things worse.

The amp PCB looks OK. Make sure the OPT secondary is grounded to the amp PCB input, not output. Otherwise any voltage drops across the PCB ground trace will get added into the feedback voltage, which may create distortion.

I made the PCBs two years ago, I have learned much since, but I must admit I never gave it a thought that it is the outputs that needs to have the shortest path to the ground, its obvious and I should have known

I made the changes you suggested and it damped the noise somewhat, but I have now removed it completely.... and here is the embarrassing part. The noise was injected from my signal generator

The OPT is connected at the ground.

The amp is now perfectly playing Queens of the Stone Age - Mosquito song, what a relief, but its playing it from my phone, so I guess there is some input grounding issues.

[DF96]

Sorry, we should have asked you to disconnect everything else and see what that does to the problem. Remote diagnosis/debugging always involves assumptions.

It is likely that the sig gen itself is not to blame, but by connecting it you created a ground loop. At least by correcting the PSU PCB you have reduced internally created hum.

[MadsKaizer]

Thanks for the help and here is a short video of it working

It sure ain't easy doing remote fault finding

MVI_5141.MOV - YouTube