How to adjust (bias) solid state amplifiers ?

[Tilman]

Hi,

I have to admit, I am a tube guy. My solid state knowledge is not that deep, although I think I have the basics covered.
Unfortunately, most amps I get for repair are SS - not tubes. 60% of them have shorted power amp MOSFETs or bipolar transistors.
Provided I was able to find suitable parts, I can replace them and most amps work again.

My questions are:

1.There is a trim pot on the power amp PCB most of times.
When I have a service manual (I almost never have) things are clear. The manual says where to measure what and how to adjust that trim pot.
If I do not have service manual - I don`t know what do do with this pot.

I have tried to bring the amp to full power and see if I can get a cross over notch on the scope and adjust the pot until it disappears. The problem is : There is no crossover notch.

I measured DC at the output and tried to adjust the pot to read zero. No DC can be adjusted.

So : How to figure out what to do with this pot ?
Example: Right now I have a power mixer with blown ECX10P20/ECX10N20 MOSFETs and an Gallien Krüger bass amp with TIP35+36 BJTs.
Both have trim pots.

Is there a logical way to handle these trim pots ? I hope so!

2. Sometimes (not very often) the solid state amp works after I have replaced blown BJTs for a short while and than blows the BJTs again. It could be that I just did not adjust the trim pot correctly.
It could be that pairs (PNP,NPN or P-channel, N-channel) BJTs or MOSFETs have to me paired / matched to work stable for a long time ?
I really don't know!

Thanks for your help!
Tilman

[Steve Conner]

If I do not have service manual - I don`t know what do do with this pot.

The answer tends to be different for different amps. Here's a brain dump.

1. Not all amps have a bias pot. The old Peavey PA amps had no way of adjusting bias. Some amps have a pot, but it's DC offset trim. Sometimes they have two pots, one is bias and the other is offset trim. Sometimes they have half a dozen and you need the manual.

2. BJT amps have emitter resistors and you can adjust the pot for about 10mV across each resistor. That's a good ballpark figure that should work for most output stages, erring on the cool side. Do this with the speaker unplugged. There's one resistor per transistor, so you can check them all to see what current each transistor is pulling, and find any unmatched ones. The 10mV figure applies to the hottest one.

3. Amps with lateral MOSFETs like the ECX... ones don't have emitter resistors. You have to break into the supply rail and adjust the pot for about 50-100mA per parallel MOSFET. 50 is probably better for MI/PA use.

4. Some BJT amps have bad thermal compensation, and if you bias them too hot they'll go into thermal runaway and blow up. There should be a thermal sensing transistor attached to the heatsink, or clamped to one of the main power devices, and you should make sure it goes back where it was if you replace the outputs. Lateral MOSFET amps don't have this thermal compensation, but ones made with vertical MOSFETs (like the Ampegs with their IRFP250s) should have it.

5. Most SS amps have so much negative feedback that they compensate the crossover notch. You'll never see it on a scope, you need a THD analyzer.

6. The only things that need matched are output devices that run in parallel. If you have 4 in parallel and one blows, you should really replace all 4 with matched ones, or at least ones from the same maker and batch.

7. Amps blow up for other reasons than bad bias.

[J M Fahey]

SS and tube amps are quite different in this regard.
Tubes are quite inconsistent in their parameters , specially modern ones, and often require bias adjustments, although it's not such a big deal as many claim .
Bipolar transistors are very consistent, and almost never require a new bias adjustment when replaced, even more so because the typical power transistor is used as an emitter follower or common collector, the least critical amplification stage, gain=1.
That's why you do not see a crossover notch, you don't have one !!
MosFets, might require a minor adjustment, although the golden rule still applies: if it works, don't fix it.
If you have no specific instructions, measure voltage across emitter resistors (typically from .33 to .1 ohms) and set current to generic 10 to 20 mA, which will do for most.
Example: 3 to 6 mV (millivolts) across a .33 ohm resistor.
Do the adjustment on a cold amp, play some music and check again after 10 minutes (heatsinks should now be warm to the touch).
If you have a reasonable value there, that's it.
If your idle current got much higher (say 100 mA ) your amplifier is thermically unstable and can runaway.
Not very likely on any decent amp, simply something to check.
As you see, I favor somewhat cold biased amplifiers, which I find much safer.
Peavey agrees with me, so I'm not alone.
As of your mysteriously blowing amps, you have repaired the blown parts, but not addressed the original problem which made them die, so look around.
*Maybe* they are oscillating or unstable, the client plugged everything he had on the output, covered heat sinks or ventilation holes with his sweater, whatever.
Good luck.

[Tilman]

Thank you guys for your kind help!

I am happy to now have at least some idea about that and a range of idle current I can measure and adjust.

For the BJT I think I totally understand the concept. I had a look into the Gallien Krüger and as you said all off the four output BJTs have emitter resisters (brown,black,silver) 0,1ohms so I should measure 1-2mV across each resister. Great!
Two of the 4 are blown (one on the P and one on the N side) so I will replace all four with the same brand.

Regarding the MOSFETs I am afraid I have to ask one more question:

The power mixer I mentioned is a Solton ACS 10/500. I have no schematic but from the PCB traces I made a sketch from the PA design. The picture shows one of the two channels. I am quite sure this channel blow up because on speaker cable was shorted.
As you can see there are two MOSFETs dead. One smaller "driver" and on big boy.
The strange thing is that almost ALL driver Mosfets are from different brands....
However, I will replace both N-Drivers with the same brand.

Back to the idle current : I could measure the voltage across R1 and R2 and adjust the
pot to a 100-200mA reading.
After that I need to measure the current for the whole PA side. The current should be 100-200mA + (50-100mA) for the big boy ? Right ?

Thanks!
Tilman

[Jazz P Bass]

Blown BJT output stage

If you do not have the schematic for a certain amp, I have found it helpful to look at the connections to the unknown trim pot.
The bias pot will be forward of the differential pair input. Between the driver transistors.
If a trim pot is attached to the diff pair, then that is typically your DC offset pot.
Enzo has some nifty advise on setting SS amp idle current.
Turn the trim pot full off.
Monitor the line current at idle.
Turn up the trim pot until the current starts to climb.
Turn the pot back a notch.
Depending on the amp, it should be 50-60 watts or less.

[J M Fahey]

What a *strange* set of output transistors.
Vertical and lateral MosFets practically in parallel, on the same amplifier?
Weird, and a biasing nightmare.
Maybe somebody repaired it earlier and thought "a mosfet is a mosfet is a mosfet".
Well, not in this case.
Can't believe the factory did that.
There's no "drivers" there , they are all sharing the load, at least as drawn.
Try to get some info from the factory, or, (hardest work), try to "reverse engineer" (aka copy) the schematic.
EDIT: check the Solton 600 schematic posted here, it should resemble yours. It uses IRF630/9640, which makes me think the Exicon is the "foreigner" there (might be the other way round, though) and also that that particular amp has probably passed through a few hands before.
http://music-electronics-forum.com/t21240/

[Tilman]

I did order the BUZ 906 already.....for about 20$....
The schematic for the Solton 600 looks VERY different though. From the way this unit is build I can tell that there are supposed to be 4 small and 2 large MOSFETs per channel.
Mosfets of the same size would just not physically fit in.
Regards!
Tilman

[jrfrond]

.....If you have 4 in parallel and one blows, you should really replace all 4 with matched ones, or at least ones from the same maker and batch.....

You wouldn't believe how much sh*t I have taken for this concept! Most techs replace the bad one and check current-sharing on the rest. What you DON'T know is that the transient generated when the first transistor shorted could blow a microscopic hole in the junction of it's SS brothers. I worked in military electronics for a brief period back in the 80's, and we were building servo amps for defense. The design was much like a power amp. If a single transistor blew on one side, ALL were extracted, tops were cut off (TO-3 cases) and examined under an electron microscope. There, you could clearly see damage that occured to the supposedly good transistors' junctions, which would not manifest itself until placed under heavy load, causing a hot spot and subsequent junction meltdown.

The moral of the story is: replace them all if in parallel and match them the best you can.

Good advice Steve.

[Enzo]

And when you are looking for the crossover notch, are you getting the entire waveform on the screen? We TEND to want the whole waveform on the scope. But working without a load, run the amp up with a sine wave to just before clipping. Now scope the output, but turn the scope gain WAY up, so you are looking mainly just at the zero crossing area. SInce the crossover point is the point of interest, what the peaks are doing is irrlelvant. You may find it a lot easier to see the notch that way.

Imagine you are looking for cracks in the base of the Washington Monument, or maybe the Empire State building. You don't look at a picture of the whole thing, you get a closeup of the base.

[Tilman]

Hi,

just a short update:
I replaced all TIP36 + TIP35 at the Gallien Krüger. Each transistor had an emitter resistor of 0,1ohms.

Steve suggested to set the idle current for about 10mV across each emitter resistor. This would correspond to 100mA.
J M Fahey suggested 10-20mA for each transistor. I choose this number (10mA/ 1mV across the resistor).
The wave form looks good with both values. I can see no crossover distortion even if I zoom the wave form as ENZO suggested.

I think I also found the reason why the thing blow up again. If I moved the master volume pot the thing started to oscillate. Tightening the pots nut and re solder the pots lugs fixed this.

Thanks for your help!
Tilman

P.S. : I am still fighting with the SOLTON.....there are no parts or informations available from the manufacturer. Bad.

[Steve Conner]

Steve suggested to set the idle current for about 10mV across each emitter resistor. This would correspond to 100mA.
J M Fahey suggested 10-20mA for each transistor.

Sorry, my original recommendation was a bit high. It might be OK for a hi-fi amp where you want minimum THD and don't mind it running a bit hot. But for a MI or PA amp you want it running as cool as possible.

[Alex R]

Enzo's idle current method has worked a treat for me since he told me about it. I don't worry about mystery trimmers - put a voltmeter on the output and an ammeter in line with the supply, twiddle the trimmers and see what does what.

[Tilman]

So you adjust for 0 Volt at the output with the pot that changes voltage at the output (if there is one) ?
With the pot which alters supply current - you do what ? Trim it until current rises and back off a little ?

[Alex R]

That's it, yes.

[Bernardduur]

revive this old thread!

I'm fixing my old Marlboro / Hohner G-50R that sounds super sweet with the Quadra sound blender. My only issue now is the BIAS

I've been reading the above and used the suggested 10 - 20mA for each transistor. Both transistors use a 0,51R resistor. I set the voltage over the resistors to approx 10mV. No heat is being made by the transistors on this setting.

When I first opened her up the BIAS was set to 62mV across the resistors; with this setting the transistors get a little bit warm.

Am I missing something? I don't want to blow up my amp but it feels that 10mV is a tad low.........

What do I gain with a higher BIAS setting? More output? More power used? A cleaner output?

Schematic: