Aleph-X builder's thread.

[wuffwaff]

well,

it depends a bit on how your power supply is build but if you´ve got abs dc offset then one half of the power supply is using more power than the other half (because of the resistors to ground at the output) causing the two secondaries to be loaded unequally wich can cause hum.

William

[Eric]

Got it, William! Thanks!

I spent the day yesterday tweaking and measuring voltages and currents and adjusting the trim pots. First, I changed R12 and R34 back to 1k3 and was able to measure AC Current Gain at 50.5% on each side of the amp. I then changed the McMillan Resistors from 3k3 to 10k on each side of the amp. I also moved Q6 from the back panel of the amp to the heat sink directly over one of the output transistors. I double checked to make sure it was electrically isolated from the chassis when I moved it. Absolute DC Offset now starts at 14v. With Q6 on the heatsink vs on the back panel of the chassis, it fell to near zero volts more quickly. Within 5 mins it fell to 6.2v, by 10 mins it was at 3.2v, and it reach 0.41v at 20 mins.

The amp ran all day (largely with the cover on) and abolute DC Offset dropped to about -2v, relative DC offset stayed < .01v, and bias current continued to fall as the amp heated up. After a few hours, I readjusted Absolute DC offset to near 0v, and bumped the bias back up to about 7.5A. The power supply is providing about 20.4v at 7.5A, and measurements of AC Current Gain stayed at 50%.

I then drove the amp with my preamp and a 70Hz tone for a while using my dummy load. On the 8-ohm setting, those resistors really got hot this time! With my preamp delivering 9.1v to the input, I was able to swing 27.8v into 4 ohms (just about 96w) and 38.3v into 8 ohms (just about (91w). I made these measurements with my DMM. Putting the preamp at its max volume, output topped out at 100w into both 4 and 8 ohms. The resistors got really hot and the heatsinks ran noticeably cooler as power went out through the speaker terminals.

Playing some music, it sounded very nice (though its hard to evaluate with only a single speaker).

Before I shut it down, I re-measured everything and it was spot-on. Power supply voltage was holding at 20.4 or so, bias current was holding steady at about 7.5A per transisor bank, AC Current Gain still measured 50%.

Then I powered down and let it cool. Now when I turn it on, it behaves strangely with respect to temperature. The heat sinks on the output side of the PCB (Q2 and Q11) run extra hot - I'm measuring 66c on the screw the holds the transistors to the sink. These sinks heat up VERY quickly. On the Current Source Side (Q1 and Q10), they take MUCH longer to heat up than the output side. After 5 minutes, Q1 and Q10 are still pretty cool, but Q2 and Q11 are running hot. Absolute DC Offset now takes much longer to come down to zero v, but still starts at 14v when cold.

After an hour of running, Q1 and Q10 run about 58c (just about the same as before), but now Q2 and Q11 are running about 66c (hotter than before). Absolute DC Offset is running about 0.5v by this time.

I never noticed the two side of the amp running at different temperatures or warming at obviously different rates before. AC Current Gain is still 50% and all transistor banks are measuring the same 0.420v across the source resistors (about 7.5A at 20.4v). All of the measurements look great except the temperature differential. Playing music still sounds good, too.

Is this normal behavior? Any ideas what is going on? Did something get damaged while I was measuring output power into my dummy load?

Thanks!

[wuffwaff]

Hi,

if you´ve got a positive absolute DC Offsett then the positive side will dissapate Ibias x (Vpower supply / 2 -V abs dc) and the other side will dissipate I bias x (V supply /2 - V abs dc offset) wich is a very big difference at 14V startup voltage. 6 V x 7.5A = 60 watts, 34Volts x 7.5A = a lot more.

So it is logical that the positive side heats up a lot slower. 14V is a lot.
What resistor did you use between output and ground?

Maybe putting Q6 on the other (colder?) heatsink would be better.

William

[Eric]

Hi William,

Thanks, this is making me feel much better! I was wondering if this was connect to absolute DC offset - I didn't fully understand what it meant before.

The output to ground resistor (R1/R4 on the PCB) is 78-ohms made from 5 x 390ohm 1% 1w all in parallel. Perhaps I'll move Q6 to the other side and/or reduce the 10k value for the McMillan resistors to something closer to 6k.

Thanks, this makes me feel MUCH better!
Eric

[wuffwaff]

Hi,

something else occured to me. I´ve got two heatsinks per mono with one current source and one output per heatsink. This way they both get to the same temperature or at least nearly the same.

William

[Eric]

Funny, I was thinking about this same arrangement. The way its wired now is really just an artifact of the way I decided to orient the PCB within the chassis. If I rotate the board 90 degrees, then I'll have the same arrangement that you do.

From what I have observed, it looks like the McMillan resistors control both the initial DC offset (high resistance leads to higher initial values of absolute DC offset) and the rate at which the absolute DC offset declines. Using 3k3, the DC offset fell pretty quickly - within 20 mins, it was at zero. Using 10k, the DC offset falls much more slowly.

I think I'll stack two 3k3 in series for a total value of 6k6. If this does not work, perhaps I'll explore moving Q6 to a small, external heatsink in an effort to keep it thermally isolated from the changing temperatures inside the chassis.

[wuffwaff]

Hi,

I think keeping it as cool as possible would be a good idea as most of the change in abs. dc offset comes from the temperature rise in Q6. It must stay coupled to the case otherwise your abs offset will depend on ambient temperature......

William

[Eric]

I'll see how it behaves with a value of 6k6 and having Q6 on the main heat sink first. If DC offset falls to zero within 15-20 mins, I'm not going to worry about it any more. Its rather fascinating to study how the behavior of this amp changes (voltage, current, temperature) and how each change affects the other settings...

I don't think coupling it to ambient temperature in this case is such a bad thing. Its for my basement theater, which keeps a pretty consistent temperature all year round (maybe a cold as 65-66F in winter and as high as 69-70F in summer).

[Eric]

On another note, I have the opportunity to bring my amp into the EE lab on campus this week and hook it to a function generator and a scope. Other than using a sine wave as input and looking for symmetric waveform and clipping behavior on the output, is there anything else that I can easily look for while I'm there?

Eric

[wuffwaff]

Look at square waves......1kHz, 10kHz.

This helps you choose the right caps for the feedback loops and the current sources.

William