Hartke HS1200 - what keeps the bias V in the middle ?

[chazpope]

I guess the 'other thread' is now so deep and old that nobody bothers to look anymore So I am cross posting this as a fresh problem, well it really is


Hi Guys


I am back onto trying to fix this HS1200, and I have started pulling out transistors to find what's wrong. It generally behaves as if one side is not working - I guess what I 'hear' is half the signal distorted at low and higher volume. The problem is apparent in a static state, but I can't yet figure out what causes it.

Without load, the base of the Q210 driver sits at .9V and the base of Q211 is -1.2V. If I plug headphones the sound is OK.

But when I plug the speaker the base of Q210 goes to -5V and the -side goes to -7V and so the whole thing is offset by about -6 volts!

I have taken the heat sink off - and when I play the amp in that state the negative side outputs do not get warm (only the driver does). Conversely the + side does get warm while its driver doesn't.

I have pulled out and checked Q210, 205, 206, 207, 208. I also removed the caps of the limiter just to eliminate them for now. The Mid point seems to be at DC 0 and the pot seems to operate predictably, also the biasing pot works and modifies the voltage difference between the two bases, but not where that is with regards to 0.

So now I wonder what in this circuit makes the biasing voltage stay in the middle rather than go off like I get ?

Thanks

[chazpope]

Well nobody answered so I carried on pulling stuff out and, after checking every other transistor I found that all the B688s have their B-Es blown, two are showing nothing, the other one .9V

This amp has been played weekly but only at very low volume say at 15W max and has never heated up during play. I wonder why this has happened and how to prevent it from happening again.

The drivers seem to be OK. Now I have ordered new outputs but they will take 3 weeks to come, in the meantime I am tempted to try a TIP35/36C pair only they are rated to 100V and the rails are 65V, so I would expect the B688s to be rated at above 130V but they are 120V , now isn't that too close for comfort.

[Steve Conner]

What makes the bias voltage stay in the middle is the global negative feedback. It forces the output voltage to zero, so the bias stacks up evenly either side of zero.

The MJL3281, MJL1302, MJW21194 etc. are all pretty good output transistors. They come in a few different package styles with slightly different part numbers.

[chazpope]

What makes the bias voltage stay in the middle is the global negative feedback. It forces the output voltage to zero, so the bias stacks up evenly either side of zero.

Well in my case the output DC was 0 but the voltage at the driver bases was way off the middle ? I know I should have pulled the output transistors first but the fact that there was no DC on the output made me think they are OK, I had not seen this kind of burn before, usually I see a EC short. I wonder what caused it, the drivers seem OK

[Steve Conner]

For me the most likely cause of a B-E short would be the transistor losing the main supply to its collector, so that the driver stage tries to supply all of the speaker current through the output device's B-E junction. This would also explain the blown drivers as they are not up to driving the speaker by themselves.

Check the supply rails to the output stage, solder joints etc.

[chazpope]

well it is rather a BE 'open' than a short - one is showing open circuit the other .9V. The drivers seem OK, I thought I could try it just on the drivers but then I thought not cause their emitters do not go to the speaker. What is the point in this floating arrangement is it to save a resistor? Could it be that somehow the outputs got too much BE Voltage before burning, esp that all of them are burned the same way. Funny that the amp was played for an hour in this distorted state before anyone noticed

[rf7]

What are you using for the output devices? I can't find 2SD718's or 2SB688's anywhere. Do you have a good source for them? Working on a KM-200 though, not a HS-1200.

[chazpope]

well yeah - ebay !

[Jazz P Bass]

The drivers seem OK, I thought I could try it just on the drivers but then I thought not cause their emitters do not go to the speaker.

It is possible, for testing the circuit, to run a 1 watt, 1K resistor from the output transistor base pad to the emitter pad. Each side.
This is for testing the preoutput stage.
I would not run the amp into a load.

[Jazz P Bass]

Please let us know how you make out with the transistors.

[chazpope]

FIXED! So I got new B688/D718 transistors, tried on pair on and it worked. I had also bought new drivers but they haven't arrived so I left the old ones in. I then put in the remaining 2 output pairs and closed it shut. I set the bias pot at about 5% off its lowest point. Next I played it in the pub last night for about 3h and it worked ok.

Again, in the hot pub and in the kickback position I found it had gotten really hot to the touch, despite the fact that it was never more than 9 o'clock turned up. With 3 sets of outputs it feels as if there is no heatsinking going on whatsoever. I will try it with a fan next time but this must be bad design - can't think of what will happen if I play it the nominal 120W, surely it will cook itself to death!

The question remains about the voltage rating of the stock outputs - does that explain the way they had failed their BE junction. Also I wonder if I should have replaced the drivers - although they did check out ok and the amp works fine.

Finally I should have checked the outputs first as I usually do to diag this fault, but I guess too much reading of forums and the odd way they had burned made me waste time looking for other faults

Also this amp has always had a nasty DC pop about 2 secs after switching off, looks like as the voltage drops something suddenly gets out of balance and the speaker is pushed out wildly. Not a big deal but it makes me wonder if it will turn on next time ..

[Enzo]

You need to properly set the bias for the amp. Setting it by which way the control points is not right. If the design were bad, then ALL of them would be sitting there burning themselves up. AMps that get hot just idling are not properly adjusted.

Drivers? I'd replace them. Sure the amp is working, and you checked the transistors. I bet you checked the parts with a hand meter. A very low current test at a very low voltage. That does not remotely tell you if the internal structure of the transistor is cracked. For a couple of dollars, is it really economy to take the chance?

[Jazz P Bass]

The EBay transistor supplier was located where?
All of the China, Thailand, Hong Cong parts are remakes.
If you want to dive in again, I would measure the idle voltage across each 0.47 ballast resistor.
That will give you an idea how much current each transistor is trying to pull.
Another item you can check is the wattage that the amp is drawing from the mains, at idle.
50 watts is a rough estimate.
If it is drawing more than that, the bias is way hot.

[chazpope]

Transistors came from China - (somehow within 2 days).

Now the amp is definitely not getting hot at idle - I checked that first. I did not measure it but it can't have been drawing anywhere near 50W -at idle the limiter lamp was barely showing a colour. Longevity was my only criteria when setting the bias - and not knowing what it is supposed to be emitter-current-wise - I set it straight to the lowest possible point. There was no audible distortion and this is good enough for pub bass, so I set it at just above that.

I did try what happens when I turn the pot the other way and at almost full you can barely see a difference in the lamp (40W) limiter, I did not stay there for long though.

Enzo this is exactly what I was not sure of - can you have this sort of half-damage on the drivers. I will change them as soon as they arrive (from china

I guess I should test the thing in upright and non upright state and see what the T difference is - but I do not yet have a consistent test signal, I did buy a cheap scope but the chinese manuals were unbelievable so I returned it straight away and I will look for some other scope solution

[Jazz P Bass]

How about measurements.
The lamp limiter is 'limiting' the amp.
That tells very little, as far as what the amp is actually doing.
The measurements across the .047 resistors is a must do.
If you have one 'hogging' current, that may be a problem.