bias a trace elliot ta100r

[lowell]

Just got finished doing a repair on a trace elliot ta100r and am wondering if anyone has any information on biasing these amps.

Thanks!

[lowell]

FYI this amp has 2 power mosfets driven by BJT's. I'm wondering what should be the DC voltage across the bias pot, which is b/t the GATES of the mosfets. Here's a schematic.

[lowell]

Enzo, JM, Steve Conner, J pbass, any of you guys able to give me basic instruction on biasing power mosfets? I'm under the assumption that they're similar to tubes in that the if Vgs is 0v then it's ON. Therefore the gate needs negative bias to limit quiescent current. ?

[Jazz P Bass]

Never had one of these on the bench.
But, it is a push pull output stage.
Therefore, I would monitor the mains currect that the amp is consuming at idle.
I would think 50 watts should do.
I never heard of 'negative bias' on a mosfet gate.

[mozwell]

Its a standard mosfet power amp stage.
Set the trimpot to MINIMUM resistance.
Monitor the power supply current with an ammeter (multimeter set to amps, use 10A range). You will need to disconnect either the positive or negative lead & put the ammeter in series.
Adjust the bias pot for approx 20 to 50mA on the ammeter

Alternatively, add a 0.33 ohm 5W resistor in series to each mosfet's "source" terminal.
You can then measure the voltage across this resistor to set the bias, 0.33 ohm 50mA = 16.5mV. Anywhere from 5mV to 16mV should be fine
This source resistor also will compensate slightly as mosfet temperature increases, so bias current will remain more stable with temperature.

Im not a big fan of power amps without emitter or source resistors, i just dont think they are as reliable. I am willing to sacrifice a few watts output, to get a more reliable output stage. Any paralled output devices, i think should always have emitter or source resistors, for current sharing & to minimise bias current shifts with temperature.

[lowell]

Mozwell so are you saying the the mains current reading will only be reading the mosfet current? Won't it also be measuring current for all other devices that are drawing current? Or is it so inconsequential that it shouldn't matter? Finally why 50ma?

[Steve Conner]

Enzo, JM, Steve Conner, J pbass, any of you guys able to give me basic instruction on biasing power mosfets? I'm under the assumption that they're similar to tubes in that the if Vgs is 0v then it's ON. Therefore the gate needs negative bias to limit quiescent current. ?

No, all power MOSFETs are enhancement mode. A Vgs of 0V is off, and positive voltage turns it on. (For N-channel devices that is. The P-channel ones are off for Vgs = 0, and turned on by negative gate voltage.)

With the BUZ901/906, a reasonable voltage across the bias pot would be about 1V. These are lateral MOSFETs with a low threshold voltage, not that different to the 0.6V of a BJT.

Mozwell's recommendation makes the most sense so far. At least, if you assume he meant to splice the meter into the DC supply to the output stage. Not the AC line. I'd also disconnect the speaker so that DC offset wouldn't skew the result.

[J M Fahey]

Agree: Mozwell's suggestion is the way to go.
And add, since you seem to be confused: nobody mentioned *mains* ; what he means is you put your current meter in series with one of the MosFets, so you are reading its current only, not the rest of the circuit.

[lowell]

"Monitor the power supply current with an ammeter" sounds an awful lot like putting the meter in series with the mains and the power supply. It's all good though. I get it now. Thanks yall. The thing is up and running with no crossover distortion.

[Jazz P Bass]

So how did you bias the output stage.
I would be curious to know what the mains draw was when you were finished.

[lowell]

haha you busted me. I actually put it as cold as it would go and then scoped the output w/ no speaker connected. I warmed the bias until crossover distortion was gone.

[Jazz P Bass]

Sounds like a good game plan to me.
By the book.
Thanks.

[lowell]

So this thing is back. It has a low volume fizziness on the notes. Very subtle. I monitored the current per Mozwell's suggestion and it's at 24ma. However, turning the bias pot does nothing. I have measured the resistance across the pot and it is indeed changing when it's turned. Not sure what's up here. All those diodes checkout with the meter.

TR4 C: -375mv
TR6 C: -1v

This seems odd to me. I'd expect to see these symmetrical.

[MarkusBass]

I monitored the current per Mozwell's suggestion and it's at 24ma. However, turning the bias pot does nothing.

I'm not sure what this means - previously you wrote that:

I warmed the bias until crossover distortion was gone.

Both statements cannot be true.
I'm not sure which Mozwell's suggestion you followed; he had two suggestions: to measure the power supply current, or to add a ballast resistor and measure voltage drop on the resistor. But he also suggested (maybe not directly) to verify that you can change the current flowing through the mosfets. Did you verify this?
There are several important notes regarding amp that use lateral MOSFETs (like 2SK135 and 2SJ50 in this amp):
- 2SK135 and 2SJ50 are very difficult to be purchased and you should be very careful in order not to short them.
- P1 trimmer changes the voltage between the gates of output transistors (and changes the bias). If this does not happen in your amp, it still needs to be fixed.
- the method of measuring the power supply current of the whole amp is highly inaccurate because you also measure the current of the Zener diodes (at IC1) and TR4 and TR6. Especially if you measure such low current like 24 mA. You have to measure the current of the MOSFETs only.
- the other method (with 0R33 ballast resistors) is also not optimal. It's because lateral MOSFETs have high resistance and usually such amps do not use ballast resistors. Remember that lateral MOSFETs have very special characteristics - check it on Wikipedia.
- the best method is to measure the current of one MOSFET only. But this may require cutting a track or desoldering the transistor - you choose.
- if you have problem finding what is wrong with the amp, you may desolder the MOSFETs and solder in their place two 1k5/5W resistors (from drain to source). Then you can work with the amp witout destroing the MOSFETs (and add them at the very end).
- I'm not sure whether the voltages on TR4 and TR6 should be symmetrical - this may depend of Vgs(off) voltages of both MOSFETs but I would expect at least one to be positive and the other negative. If you measure the (total) current of the amp, it is still possible that the MOSFETs are not properly biased.
- does the voltage on TR4 and TR6 collectors change when you turn the pot?
- do you have 0V on the speaker output?
- did you measure the amp with a dummy load (and full power)?
Sometimes simillar problems are caused by failed Zener diodes but first I would clarify the problem with bias.

Mark

[lowell]

Markus,
Thanks for the reply. Yes I DID warm the bias until the crossover notch was gone. Now, however, there seems to be no effect on my scope, or when reading current draw when the bias pot is turned.

I measured the current of 1 mosfet, not the entire PA.

- does the voltage on TR4 and TR6 collectors change when you turn the pot?
I will check this

- did you measure the amp with a dummy load (and full power)?
Not sure what you mean here. What am I measuring?

[MarkusBass]

I measured the current of 1 mosfet, not the entire PA.

I don't know what is the correct value of the current but I would set it to at least 50 mA and verify how hot gets the amp when you provide a signal to the input. If it gets hot, decrease the current slightly. Just calculate: power supply voltage (is it 60Vor less?) multiplied by 50 mA (0.05A) and you get 3W at idle - for me, it seems resonable.

- did you measure the amp with a dummy load (and full power)?
Not sure what you mean here. What am I measuring?

Signal to the input (1V or more), oscilloscope to the dummy load (4 Ohms) and you observe the signal on the output. You can measure the power of the amp before it starts distorting the signal.
You can also check whether the idle current decreases when the amp gets warmer.

Mark

[lowell]

Ok thanks Markus,

Well for now since I have the speakers disconnected, I've monitored the voltages on TR4 and TR6 collectors:

TR4 ranges from -1v to -.375v
TR6 does NOT change when the pot is turned

Does this clue you into something?

[MarkusBass]

Hard to say without some further measurements. Are you sure that the voltage on T4 (C) is negative? It should be positive. And the other should change in similar way. Would it be difficult to desolder the MOSFETs? It would be benefitial to check the amp without the output transistors - without speaker but please note that there is a DC negative feedback loop from the output back to TR3 (R14) so I would put some resistor (e.g. 220R) as a load. Do you have 0V on the output?
If you cannot easily desolder the MOSFETs, I would check voltages on TR6 (compare them with TR5). Maybe TR6 is shorted - this is the easiest one to be replaced (but first do some measurements). In similar amp I had problems with the diodes in the output stage (they testes OK but they were not OK). I hope it's not 2SJ50.

Mark

[lowell]

Markus,

I'll take some measurments first before removing mosfets.

Yes T4 collector is negative.
There is -70mv on the output with NO speakers connected.

TR6
C: -1v
E: -53.5v
B: -52.9

TR5
C: +18v
E: -53.5v
B: -52.9

So if I were alone on this repair I'd probably replace TR6. Voltages seem good except TR6 collector right?

[MarkusBass]

Yes, T4 collector is negative.
There is -70mv on the output with NO speakers connected.

TR6
C: -1v
E: -53.5v
B: -52.9

TR5
C: +18v
E: -53.5v
B: -52.9

So if I were alone on this repair I'd probably replace TR6. Voltages seem good except TR6 collector right?

-1V on TR6(C) seems to be a correct value. I would expect the voltage there to be within -1.5 to -0.5V. Can look at TR4 (I mean measure voltages on it)? This is a current source and it seems to me that it is not working correctly. While being there you can measure voltage drop on D6 (this is also a part of the current source) although it's probably OK because you have +18V on TR5(C). Have to checked other diodes in the output stage - including Zeners?

Mark

[lowell]

Mark,

TR4
E: 53v
C: -400mv
B: 52.7mv

Voltage across D6 is 660mv

Also of note:

590mv across R21.
540mv across R22
~1v across R23

Yes I've checked all the diodes in the PA for shorts/opens.

[MarkusBass]

Mark,

TR4
E: 53v
C: -400mv
B: 52.7mv

Voltage across D6 is 660mv

Also of note:

590mv across R21.
540mv across R22
~1v across R23

Let me quickly do the math:
The V+ rail voltage is: 53V (TR4-E) + 590mV (R21) = 53.6V
The voltage on TR4(B) should be:
53.6V(+Vrail) - (540mV (R22) + 660mV (D6)) = 52.4V so 52.7mV is rather not possible. In this case, instead of measuring the voltage on base, you should measure the voltage between the V+ rail and the base. This will tell you whether TR4 is correctly biased (in order to provide current). And this is most probably: 540mV (R22) + 660mV (D6) = 1.2V, which is OK.
The current provided through the current source:
590mV(R21)/220 Ohms = 2.7mA
The voltage drop on P1 set to 50% (110 Ohms): 2.7mA * 110 = 300mV
The voltage drop on P1 set to 100% (220 Ohms): 2.7mA * 220 = 600mV
I would expect slightly higher voltage - up to at least 1.0V but maybe this is due to some failed component.
Start with replacing TR6 and see whether this changes anything. If not, replace TR4. Measure both transistors when you desolder them.

Mark

[lowell]

Mark,

Yes there is always a discrepancy when I measure these. It seems everytime I'd take a measurement it'd be slightly different in regards to millivolts. Sometimes a couple hundred millivolts off. I figured this was just my meter being imperfect. So I'm sure your math is correct, and also measuring between the rail and the base as well. I'll report back. I'll need to order these transistors though so it'll be a few days if not a week.

[lowell]

Dang I got the wrong transistor and replaced TR6. Didn't help... ordering TR4 now.

[J M Fahey]

Don't worry too much.
It might ave been worse.
Imagine a Doctor pulling the wrong kidney

[lowell]

Okay, well, replacing TR4 didn't fix it. Still same results on both ends of the bias pot. I tested the original TR4 out of circuit and it tested good. There is 1.1v across R22 and D6 string therefore there is ~52.9v on TR4's base. The voltage rail is 54v.

[lowell]

So Markus I'm stuck here. I'm thinking one of those zeners in the feedback path and/or between the mosfets gate/source junction are causing this problem. How would you suggest troubleshooting this?

Can I lift D8 and D9? And D7/D10? It seems to me that these diodes are here to limit signal and DC biasing will not be affected if they're aren't in circuit.

[lowell]

Okay so I also realized my current measurement was wrong. I found some traces that I had not seen. SO... measuring current of JUST the mosfets I have 0ma on both of them, no matter where the bias pot is. Maybe these "new" mosfets have gone bad or were damaged during or after installation? I get OL on the resistance measurement of the mosfets, except I get 7.5meg from drain/source on the p-channel mosfet, and only 500k from drain/source on the n-channel mosfet. I feel that that seems off.

What do you all think of replacing with these?

IRFP-9240
http://www.mouser.com/Search/Refine....word=IRFP-9240

IRFP-240
http://www.mouser.com/Search/Refine....yword=IRFP-240

[lowell]

Okay well I guess I lost the assistance in this repair. I got these transistors and replaced the ones in the amp again. Now I have these readings:

TR6 C:-3v (previously -1v)
TR4 C: range is -2.3v to -3v on extreme bias pot settings (previously -400mv to -1v)

At this point I'm just throwing shit at the wall and seeing if it sticks. Not the best approach to fixing an amp. I don't really understand why this voltage has changed though. I certainly wouldn't think that changing the MOSFETS would affect the quiescent bias voltage. But again, maybe I don't fully understand how MOSFETS work. These are enhancement mode MOSFETS right? So the gate NEEDS a turn-on voltage. ?

[MarkusBass]

I got these transistors and replaced the ones in the amp again.

Which exactly transistors do you mean? Did you replaced 2SK135 with IRFP? This is not allowed - they are completely different type and one cannot replace the other. The first one is lateral, and the other is vertical. You may read about differences between these two here: http://www12.fairchildsemi.com/an/AN/AN-558.pdf - check that the gate needs a turn on voltage, otherwise MOSFET is not conducting. So you need eg. +/-1.5V on gates (on P1 trimmer).

Now I have these readings:

TR6 C:-3v (previously -1v)
TR4 C: range is -2.3v to -3v on extreme bias pot settings (previously -400mv to -1v)

I'm not sure how to read these measurements. There is a current source and turning the P1 trimmer should vary the voltage between 0 and let's say +/-2V. Does it work in this way, or not?

At this point I'm just throwing shit at the wall and seeing if it sticks. Not the best approach to fixing an amp.

During these few days you could already replace all semiconductors in the amp (without testing them). Alternatively, you can read about differential amp, current source and lateral MOSFETs. My advise is to remove the MOSFETs (whatever MOSFETs you have) and fix the amp without them. You can put 1.5k/10W resistors in place of the MOSFETS (drain-source) - just to be sure that you have 0V on the output. Instead of the speaker connect a resistor - someting like 470 Ohms/5W. Then, using a signal generator, an oscilloscope and a DMM make sure that the amp works. It means that you have nice sinusoidal signal on both ends of the P1 trimmer. And that you can vary the voltage on P1 up to +/-2V (in this case without the input signal).
You can lift the Zener diodes on the output - the are to protect the MOSFETs against to high voltage on the gates. Check each of the and solder them back. Make sure that the amp still works as expected.
Then (or maybe start with this) measure the gate resistors whether they are not open. This is a common failure in such amps.
Separately, you can check each (old) MOSFET with an external power supply and e.g. 16 Ohms resistor. Search for "MOSFET matching circuit". If they are fine, solder them back.

Mark

[lowell]

Which exactly transistors do you mean? Did you replaced 2SK135 with IRFP? This is not allowed - they are completely different type and one cannot replace the other. The first one is lateral, and the other is vertical. You may read about differences between these two here: http://www12.fairchildsemi.com/an/AN/AN-558.pdf - check that the gate needs a turn on voltage, otherwise MOSFET is not conducting. So you need eg. +/-1.5V on gates (on P1 trimmer).

Mark in this article it says that lateral MOSFETS are enhancement mode but it doesn't clearly state that vertical MOSFETS are depletion mode. Is that what you're saying? That vertical MOSFETS are depletion mode? Finally, I don't see the "mode" listed in the datasheets for the IRF. I DO however see that the G/S threshold voltage is +-2v. So that tells us that they're enhancement mode right? The 2SK135 is also enhancement mode so not sure, maybe I'm missing something.

I will add the 1.5k resistors and dummy load and see what results I get. Am I looking for TR4 and TR6 collectors to a positive and negative voltage? That makes sense to me, this way the gate/source voltages will be ON.

[MarkusBass]

Mark in this article it says that lateral MOSFETS are enhancement mode but it doesn't clearly state that vertical MOSFETS are depletion mode. Is that what you're saying? That vertical MOSFETS are depletion mode? Finally, I don't see the "mode" listed in the datasheets for the IRF. I DO however see that the G/S threshold voltage is +-2v. So that tells us that they're enhancement mode right? The 2SK135 is also enhancement mode so not sure, maybe I'm missing something.

All I'm saying is that 2SK.. is a lateral MOSFET and IRFP is vertical. You cannot replace one with other. If you use IRFP in this amp, you will get a thermal runaway and the amp will not work to long. If you are looking for general information about MOSFETs, you can look here: MOSFET - Wikipedia, the free encyclopedia , or just type "lateral versus vertical MOSFETs" in google. You cannot expect all information to be in the datasheet.

Am I looking for TR4 and TR6 collectors to a positive and negative voltage? That makes sense to me, this way the gate/source voltages will be ON.

On top P1 trimmer pin (from V+ rail side) you need positive voltage and on the other (from V- rail side) you need negative voltage.
Have you check the gate resistors? You could check also all the diodes - desolder them and check with external power supply and a resistor.

Mark

[g-one]

Maybe these "new" mosfets have gone bad or were damaged during or after installation? ... What do you all think of replacing with these?

IRFP-9240
IRFP-9240 MOSFET | Mouser

IRFP-240
IRFP-240 MOSFET | Mouser

Lowell: It would seem you have now replaced the output devices twice? Have you used the stock type J50 & K135 either time?
It would be helpful if you would say which type devices you have tried for replacements.
It is possible incorrect subs have been the whole problem since the original output devices were replaced.

[lowell]

Hi g-one. I have not been able to find the J50/K135 transistors. On various sites I've found that these should work: 2sj201/2sk1530.

Mark thanks for the info on that, I guess I'd need to redesign a bit, add a heatsinked Vbe multiplier so as to limit runaway with vertical mosfets. And I'm not going to do that.

Again, back to just using 1.5k resistors for now. I'll report back!

[MarkusBass]

Hi g-one. I have not been able to find the J50/K135 transistors. On various sites I've found that these should work: 2sj201/2sk1530.

They will work but they are in a completely different package, which could cause some problems when trying to use them. They also require slightly higher voltage on gates to work properly (so same small changes to the amp may be required). If you are looking for better equivalents, you may take a look at BUZ901/906. I also see (quite often) 2SK135 on ebay. In Poland I can even buy them in several shops - hard to believe that you cannot buy them in the States.

Mark thanks for the info on that, I guess I'd need to redesign a bit, add a heatsinked Vbe multiplier so as to limit runaway with vertical mosfets. And I'm not going to do that.

Yes, you would need to rediesgn the amp. In my opinion this is not needed since you can buy either original transistors, or eqivalents. What about the transistors that were originally in the amp, are they shorted?

Mark