Yaqin MC-5881A amplifier improvements

[DanDini]

One other thing I didn't note before (and should have). The noise is present from switch on - and doesn't get any worse (or less) when the valves warm up. In other words, it must be some ground loop hum or similar from the supply I guess?..

[Salas]

Originally posted by DanDini


OK, I've done that and get 41.7Vdc down to 3.9mV on one side and 40.8Vdc down to 4.1mV on the other.

Any idea where this voltage inconsistency comes from as the transformers appear to deliver an even supply. Is it enough to be an issue?

---- No its not an issue, its just that one pair draws just a little differently because of valve tolerances. In fact its a triumph that we got practically on target for heater voltage and bias mods just from watching a photo! Now each valve draws 41mA and your heaters are averaging 6.3VAC. Heat is normal and bias too. Super!

Originally posted by DanDini
One other thing I didn't note before (and should have). The noise is present from switch on - and doesn't get any worse (or less) when the valves warm up. In other words, it must be some ground loop hum or similar from the supply I guess?..

Its a ground loop by any bookmaker's favorite bet, and I would experiment braking it using a 10R 5-10W resistor. I would leave a cable go to chassis from mains ground terminal on the IEC socket, but I would insert the resistor between the ground wires common point and mains ground terminal. So mains ground directly to chassis, but ground cables common point is broken by the resistor to chassis. Keep in mind that chasing a ground loop for a given construction, layout and ground scheme from a pic its just betting on issues and it will be a miracle if we solve it or just attenuate it. If I had the amp before me, tweaking it, I could be much more positive.

But we can try ideas one after another. Just do that with the resistor for now so we see what it does.

P.S. As you practically saw, such cheap amps look good but they are not only using 3rd class components but they can be so badly cathode heated and biased out of spec that they can enter fire hazard zone. Having a strong ground loop in a finished commercial product is childish at least.

The chosen circuit is not bad at all though, and you can benefit from Auricaps, =C= 6L6GC output valves, and a small PSU choke if you can fit it. But small chokes have a lot of resistance and if common for channels it will eat much B+. Another idea is to use an extra 100uF 450V cap before the 470uF that is already there, and link them with a 10R 10W. CRC will cut somehow on your PSU ripple, and you will lose just 10V B+. A no space for 2 good chokes situation can benefit somehow from the CRC alternative suggestion. Appreciably better than the single cap you have there now.

[Salas]

Here is what the simulator shows for your single 470uF / 450V cap.

[Salas]

And here is what it shows for the 100uF - 10R - 470uF CRC. Notice the leaner horizontal line. A good reduction for ripple given the tight space that may not allow for chokes. Economic and straightforward too.

[DanDini]

Right, after some fiddling with the grounds I discovered that most of the noise is coming from the input select and volume module. When this is disconnected from earth - it pretty much shuts up. You can see in the picture that there is (was) a wire soldered to this module and screwed to the chassis, but it is also earthed through the bodies of the switch and pot where they physically touch the chassis.



If you look at the scope pic with the earth connected:



...and without:



..you'll see the difference. There is still some noise (as shown in the pic!) but a LOT less.

I think I'll try it like this for a while and consider the other mods as a next step. Overall it really doens't sound bad and now runs much cooler. It's interesting that the manufacturers specifically earthed this if it creates this problem.

Thanks for all the help!

[Salas]

Bravo Dan! They have created a loop within the GND circuit and you have broken it. If you try breaking with a resistor in other places like the one I mentioned earlier, you may probably kill hum even more. But you don't run extra sensitive speakers so maybe its futile. I am happy that we did a fruitful step by step problem solving with no back and forth. Well worthed the effort.

Regards

Salas.

[DanDini]

The big problem now is isolating the thing. Unfortunately the body of the input switcher is grounded and as the chassis is metal it's hard to get it rights. Is there not any way of resolving these things without breaking the connection to earth? Can you not 'equalise' the gnds so no loops occur?

Thanks,

Dan

[DanDini]

Another thing regarding the current we adjusted by lifting one of the 500ohm resistors... according to this data sheet, we are safe with 88ma, didn't I only have 70 or so per tube in the first place?

http://www.retrovox.com.au/tsol5881.pdf

Thanks,

Dan

[Salas]

1. Yes, to 'equalize' the GND currents is what a competent layout in any commercial amp should have done in the first place. See about that 10R break suggestion I have proposed earlier.

2. That data sheet says 'values are for two tubes'. Was overheated, wasn't it? Each tube was at max dissipation before. Devouring itself.

[DanDini]

1. Yes, to 'equalize' the GND currents is what a competent layout in any commercial amp should have done in the first place. See about that 10R break suggestion I have proposed earlier.

I'm looking into this next

2. That data sheet says 'values are for two tubes'. Was overheated, wasn't it? Each tube was at max dissipation before. Devouring itself

Sorry - it was a misreading. I thought it was for one valve. My mistake.

Dan