Just follow Ben’s advice. It is not essential that you make a plate, or that you cut the other junctions. Several examples of people using several junctions, happily. Me too, see the ground planes: junctions from the first cap.
With obly one junction, the equalizing current in fact increases and so you may have to establish takeoff further from it. So perhaps more are merry, see PL’s biguns: planes all the way, takeoffs here and there too.
Ben’s advice is to just make the Tee using wires, and I second that: simple and probably just as good or better than a plate. And you can always make a plate at some point anyways if you so desire
PS: Ben is the king of quiet amps, not me.
I was in the process of reworking the amp tonight to place the power transformer vertically in the front, put the rectifiers close underneath it, and move the PSU board closer to the front of the amp in order to make room for a few chokes and capacitors, when I noticed that one of the V- wires from the power supply board to the right channel circuit PCB was making contact with one of the bolts through the wood plate that the power supply board is mounted to. This means that there was a connection, albeit a poor one, between V- and the chassis. My guess is that this is the source of the noise issue that I am having. Is this a fair guess? I don’t know if I should bother adding the extra chokes and capacitors given this discovery.
The PS PCB mounting holes are isolated from the PCB tracks. I don't believe there was any harm caused. The best way to make sure is to:
Power up the amplifier with the AMP PCBs disconnected. Are you getting around + and - 25V DC (unloaded)?
If YES and everything looks good, connect the AMP PCBs and confirm that the bias across the 0.47ohm resistors is around 400mV, and that the V+ and V- are now sitting at around 22-23V DC(this will depend on the bias)
Since you like to play a lot, I'd suggest that you:
1. Get the variac, two 8-ohm 100W (or 50W) resistors, a tone generator and an oscilloscope
2. Connect the amp via variac to the mains supply, apply something like 0.25V 1kHz sinewave signal to amp inputs, DO NOT load the amp outputs, and observe the output on the oscilloscope while you slowly increase the mains voltage to your amp using that variac. Eventually, you will have a full, complete, nice-looking amplified sinewave on the oscilloscope - once the variac reaches your nominal mains voltage (110-120V AC??)
3. Do the same as 2. but this time with the load connected to the amp output.
I do not like the chokes' sound... but that is me.
Aleph J is a very high input impedance / wide bandwidth amplifier. If used in a single-ended mode, it will feed the ground noise (remember, you connected -IN and ground together) to one side of the input differential (JFETs) pair - hence the differential pair can not reject the ground noise (the other input is connected to the HOT of your source)... there is no common mode/out-of-phase noise present with single-ended input - hence the noise can not be rejected.
Aleph J should be used in a balanced configuration; where it will exhibit 0 mains' frequency (x2) hum/buzz noise, AND 0 hiss (if no other crazy-amount-of-gain devices are connected to its input). This is how I use my Aleph J now (I started with a single-ended configuration as well...same as you).
Power up the amplifier with the AMP PCBs disconnected. Are you getting around + and - 25V DC (unloaded)?
If YES and everything looks good, connect the AMP PCBs and confirm that the bias across the 0.47ohm resistors is around 400mV, and that the V+ and V- are now sitting at around 22-23V DC(this will depend on the bias)
Since you like to play a lot, I'd suggest that you:
1. Get the variac, two 8-ohm 100W (or 50W) resistors, a tone generator and an oscilloscope
2. Connect the amp via variac to the mains supply, apply something like 0.25V 1kHz sinewave signal to amp inputs, DO NOT load the amp outputs, and observe the output on the oscilloscope while you slowly increase the mains voltage to your amp using that variac. Eventually, you will have a full, complete, nice-looking amplified sinewave on the oscilloscope - once the variac reaches your nominal mains voltage (110-120V AC??)
3. Do the same as 2. but this time with the load connected to the amp output.
I do not like the chokes' sound... but that is me.
Aleph J is a very high input impedance / wide bandwidth amplifier. If used in a single-ended mode, it will feed the ground noise (remember, you connected -IN and ground together) to one side of the input differential (JFETs) pair - hence the differential pair can not reject the ground noise (the other input is connected to the HOT of your source)... there is no common mode/out-of-phase noise present with single-ended input - hence the noise can not be rejected.
Aleph J should be used in a balanced configuration; where it will exhibit 0 mains' frequency (x2) hum/buzz noise, AND 0 hiss (if no other crazy-amount-of-gain devices are connected to its input). This is how I use my Aleph J now (I started with a single-ended configuration as well...same as you).
The V- wire was touching a bolt used to secure the wood that the power supply PCB is attached to. That bolt was attached to the chassis. I believe that it was in that state for the entire trouble shooting process that you have all very generously helped me with. I’m going to put it back together without the additional caps and chokes to see if that fixes things. Here are a couple of photos of the new cleaner layout….
you need greater area for Donut support
either change to wider L bracket, or find another hat/circle for Donut ( as you have on other side) and mount in between L bracket and Donut
think of pressure on just few wires, versus pressure on many wires, when having proper support area
edit: also be sure that you have some thermal goop under the bridges and always split washer at bridge screw; thermal cycling is sure thing to have screw unbolted, and bridge is Dodo
either change to wider L bracket, or find another hat/circle for Donut ( as you have on other side) and mount in between L bracket and Donut
think of pressure on just few wires, versus pressure on many wires, when having proper support area
edit: also be sure that you have some thermal goop under the bridges and always split washer at bridge screw; thermal cycling is sure thing to have screw unbolted, and bridge is Dodo
Yes. The contact area on the L bracket seems problematic. I couldn’t find a wider L, but I will find a very large flat washer for the side of the transformer that only has the L bracket. On the rectifiers I used a flat black washer against the rectifier, and a star lock washer between the flat black washer and the bolt. I can change to a split ring if that works better, and I will add thermal compound as well. Thanks again to you all for your help. It is much appreciated.
In the process of moving things around, one of the leads on one of the rectifiers snapped off, so I had to go buy two more and replace them. There is still a faint hum in the left channel. The right channel is very close to silent, with only the slightest hiss right next to the speaker. The left channel is the one that was problematic previously. I am going to measure AC on the rails and report back. I haven’t added the additional LC.
I added the thermal compound under the new bridges, and a circular metal plate over the L bracket to distribute the contact area on the transformer.
I noticed something on shutting down the amp that may tie in with the residual hum on the left channel. These photos show that the LEDs on the power supply board dim unevenly, with the left channel power supply LED dimming more quickly than the right.
1. This is with the amp on. LEDs evenly lit.
2. 10 seconds after amp is turned off. Left LED on power supply PCB is noticeably dimmer.
3. About 90 seconds after amp is shut off. Left LED is off. Right is still on.
Could this be related to the noise issue?
Will do. I should have left more wire, but I should be able to spin it a bit. I'll try and report back after I measure the AC on the rails.
These are the results of measuring the rails…
The positive rail, as measured between ground and V+ on the power supply PCB.
The negative rail, as measured between ground and V- on the power supply PCB.
I rotated the transformer to put the leads on the top, with no change in these figures. I will try attaching speakers again to find out if it reduced the audible hum in the left channel and report back.
I am not sure whether the meter's AC measurement is accurate since 1mV of ripple seems too low for a CRC power supply.
Since the power supply powers both channels, the difference in noise between the channels is most likely not caused by the power supply.
Have you tried Bonsai's Debugging Amplifier Noise tips (see post #8,768)?
Are the left channel V+, V-, and Gnd wires connections to the power supply board made at exactly the same locations as the right channel wires?
Since the power supply powers both channels, the difference in noise between the channels is most likely not caused by the power supply.
Have you tried Bonsai's Debugging Amplifier Noise tips (see post #8,768)?
Are the left channel V+, V-, and Gnd wires connections to the power supply board made at exactly the same locations as the right channel wires?
Last edited: