I will go through Bonsai’s tips. Here is a photo of the ground wire connections…
I don’t have a true RMS meter. I do have a couple of oscilloscopes, but haven’t learned how to use them yet. It’s on the long list of things I need to learn. If only both channels were like the right channel…
It’s not easy to see in the photo, but there is a wire loop in the middle between GND2.2 and GND3.2 that I could connect all the ground wires to rather than at GND2 and GND3. My work on ground needs to be cleaned up as the wires are longer than needed just in case I need to add the additional LC. I will work through Bonsai’s material before I do anything else. I agree that if I have a good and a bad channel, the problem is unlikely to be the common power supply.
I think there's an opportunity lurking here. You'll be able to see the ripple... and measure it as well - very accurately. You'll also be able to check for oscillations and trace the input signal as it gets amplified... which may help you identify the root cause of your problem.
There is no doubt that you are correct, but I don't know the first thing about how to use an oscilloscope. I have two that are supposed to be fully operational, but I bought them and never tried using them. Is trouble shooting this circuit with an oscilloscope something that can be learned quickly?
For the most part, I build tube amps. I have built an ACA from a kit, and like it, but prefer the DIY tube amps that are my daily drivers. I have always wanted to hear a First Watt amp in my system to see if it could displace my tube amps, but building my first First Watt clone has already been a large investment in time, and a steep learning curve. To be fair, the residual noise I am getting in my left channel probably wouldn't be audible through most speakers, but my tube amps do not have this noise, so it has been very disappointing to not be able to debug this amp. I will carefully go through the Bonsai article next.
PROBLEM SOLVED!
It seems that the either the transformer or the rectifiers are too close to the circuit PCBs in my amp. Placing the metal sheet between the transformer and rectifier eliminated the buzz, and made the left channel as quiet as the right one. I think that because the PCBs are not mirrored (ideally they would be), different parts of the circuit are exposed to the radiated fields of the transformer/rectifiers, and some parts of the circuit are more sensitive to this radiated energy than others. Part of the problem is that when I layed out the amp, I did so with the intention of leaving enough room on the top part of the heatsink for the 2017 Sony VFET amp that I also have a kit for. I figured I could put one power supply and two amps into one chassis since the heatsinks are so tall. This put the Aleph J PCB closer to the transformer than it would have been had I built for just one amp. My chassis is also on the narrow side.
I am going to cut the sheet in the photograph in half and install one half per side as a shield between the transformers and the PCB. Thank you all for your help troubleshooting. Hopefully, by tomorrow night I will be listening to my completed and noise free Aleph J!
It seems that the either the transformer or the rectifiers are too close to the circuit PCBs in my amp. Placing the metal sheet between the transformer and rectifier eliminated the buzz, and made the left channel as quiet as the right one. I think that because the PCBs are not mirrored (ideally they would be), different parts of the circuit are exposed to the radiated fields of the transformer/rectifiers, and some parts of the circuit are more sensitive to this radiated energy than others. Part of the problem is that when I layed out the amp, I did so with the intention of leaving enough room on the top part of the heatsink for the 2017 Sony VFET amp that I also have a kit for. I figured I could put one power supply and two amps into one chassis since the heatsinks are so tall. This put the Aleph J PCB closer to the transformer than it would have been had I built for just one amp. My chassis is also on the narrow side.
I am going to cut the sheet in the photograph in half and install one half per side as a shield between the transformers and the PCB. Thank you all for your help troubleshooting. Hopefully, by tomorrow night I will be listening to my completed and noise free Aleph J!
I'm lucky to have custom made Donuts on disposal, so I'm always getting them made with lower magnetic flux (more wire turns per Volt), with static shield (so called blind winding) and magnetic shield ( in this case several turns of magnetic strip same as core is made of)
some ppl around (on DiyA) are ordering fancy ones from Toroidy Poland, but I don't like several complaints of their working temperature ........ something is fishy having that
some ppl around (on DiyA) are ordering fancy ones from Toroidy Poland, but I don't like several complaints of their working temperature ........ something is fishy having that
My toroid is from Antek, and does not have a shield. I should have built a separate power supply chassis. The amp would have been quiet from the get go if I had. Or I could have bought a better toroidal transformer.
Just as an additional data point:
I have by now bought at least 6 different transformers from Toroidy (in most cases their middle quality, "Audio Grade", TTSA...), and never had a problem with high working temperature ... but in most cases, I put about 1/3 to 1/2 of their rated VA as a load on them. They only tend to get mildly warm, if at all.
Regards, Claas
Yes, I see that the non-mirrored boards are an issue here. The left channel has the signal carrying output mosfets right next to the transformer whereas the right channel has the output mosfets at the back of the case, away from the transformer.
So shielding is a solution. Another solution is to rotate the left channel board 180 degrees.
So shielding is a solution. Another solution is to rotate the left channel board 180 degrees.
For some reason, Antek doesn’t manufacture transformers larger than 400va with a shield.
Yes. I will try shielding the sensitive side of the board that was a problem on the left board, on both boards, to see if I can get the amp even quieter. I only have the one metal sheet now, and need to cut it in half. The shields are going to require removing in order to adjust the bias and offset on the amp. How often are these adjustments typically needed?
Yes. I was looking online to see if I could order a 500VA replacement, and they don’t have one. Would a 400VA make much difference. The Antek toroids are reasonable, so it wouldn’t be very expensive, and I can apply the one I have to a separate external power supply for my next amp.
Last edited:
The amp sounded great even with the buzz. I have a purpose built listening room that is very quiet. In this room any added noise from the electronics adds noticeably to the noise floor. This has made me a stickler for silent amps. A little while back, the equally helpful folks here on the tube forum helped me work my way through a hum/buzz problem with a Type 30 and EL84 amp I had trouble with. My situation is not a normal one for most. My guess is that the amp would be perfectly acceptable for most users without the shielding. Hopefully this will add a helpful data point for people in a similar situation.
You may not need to shield the whole left channel board. Try shielding only between the transformer and output mosfets.
If you haven't done the Bonsai debugging tips, I suggest you do to confirm that there are no other issues.
If shielding of the right channel does not reduce the right channel noise further, then it is likely that a reduction of the power supply ripple is needed to further reduce the noise. The easiest way to do that is to replace the resistors in the CRC filter with two 154B per rail. Or with more work, add LC to the rails.
Rotating the left channel board 180 degrees is also an alternative to shielding if you find that shielding of the right channel does not reduce the right channel noise.
Alternatively, you can shield the transformer with a circular shield.
If you haven't done the Bonsai debugging tips, I suggest you do to confirm that there are no other issues.
If shielding of the right channel does not reduce the right channel noise further, then it is likely that a reduction of the power supply ripple is needed to further reduce the noise. The easiest way to do that is to replace the resistors in the CRC filter with two 154B per rail. Or with more work, add LC to the rails.
Rotating the left channel board 180 degrees is also an alternative to shielding if you find that shielding of the right channel does not reduce the right channel noise.
Alternatively, you can shield the transformer with a circular shield.
Lucky Mighty bstrd. If only we all were as lucky as you!
Yes, Toroidy is very good, but I have a very hot and buzzing audio grade I wanna throw out the window.
Luckily, the world is small, and mighty’s friends are soft. There is supposedly 6 kilos of custom iron waiting for me in my shed
