SVT 4 PRO

[lowell]

I just finished repairing a blown power amp B. Power amp A was fine and working. I tested every component in the PA as well as in the bias circuit (Vbe multiplier?).

On pg 1 of the PA schematic, Q203/204 were toast, and this was visible the moment I opened this unit up. I'm not totally clear on SS terminology but I believe this section of the circuit is the splitter, and provides the 2 out-of-phase signals to drive the PA. There were shorted mosfets AND many shorted gate resistors in the PA. I replaced all mosfets and all resistors (gate, ballast, feedback) in the entire PA. I also replaced Q202-Q207 on the previous page in the splitter/bias circuit. I also replaced R218, R219, and R222 which were all burnt.

Wow, these amps are a serious undertaking as far as assembly and disassembly. So that being the case I figured I'd ask for some pro advice at this point. I've finished the repair and reassembled the amp. I always use my bulb limiter when firing up a new repair like this, but I'm not sure that is sufficient here. This was a TON of work, and I don't wanna fire it up only to watch my work burn up and go down the drain. Is there a more precise way of testing such a job that is, more sensitive?

Thanks in advance.

[Jazz P Bass]

I feel your pain, Lowell.
If all the static, unpowered checks prove out, then it is time to plug it in.
The lamp limiter is a good start.
A variac would also help.
I like to monitor:
#1: mains draw. (KillaWatt)
#2: B+ & B- Vdc.( 2 Fluke VOM's)
#3: The base voltage of the output stage.(2 Fluke VOM's)
#4: test the lower voltage power rails.
If all of that passes the power on test, I'll ditch the lamp & send in a signal to the unloaded output.
If that passes, load it up.

[lowell]

JPB,
Thanks for tips. So I'm guessing monitoring mains draw, I don't wanna see anymore than a couple amps at the most with no load or signal? So use the limiter AND variac together? What variac setting do you suggest?

[Jazz P Bass]

I use the variac to initially bring up the amp.
(SS amps scare the s_it out of me)
A solid state amp generally will not pull more than 50-60 watts at idle.

[lowell]

Ok so around 500ma at idle then. 60/120= .5

[Enzo]

Lowell, just an aside from a few posts back

There is no phase splitter, well there kinda is, I;ll explain in a minute. The amp output stages are driven by the op amp. Those HI and LO feeds to page 2 are in phase, they are just separated by a DC voltage. The op amp yanks the whole thing up and down to the music. Throughout it all, Q106 107 - the bias circuit - they maintain a steady voltage "space" between the HI and LO drives. The drives move in unison other than their DC offset.

So where is the phase splitter? It is the power supply. Instead of using inverted circuits using the same supply, we simply feed the same signal to the positive and negative sides, but since the power rails ARE opposite polarity, the same signal has opposite effect on each of those sides. SO really, the split power supply is your phase splitter. I have never seen it written that way, so I apologize if I stole someones lesson.

[lowell]

Good call Enzo. After reading your posts I said "duh, I knew that." That circuit just provides the upper and lower drive signals but I see that NO they are not out of phase, they are indeed in phase. Thanks for clarifying. Gonna try powering up w/ variac and meter and see what's up here. I'm very nervous, as these matched mosfets were about $25.

[lowell]

Ok so I've been comparing PA A and PA B. They look about identical. I have the variac and bulb in place as well. With the Variac at ~60VAC my bulb starts to glow. The fan also turns on at this point. Not sure if that's ok or not. Here are voltages.

G +/- 3v
D +/- 39v
S +/- 8mv

The G and D measurements are within a couple hundred millivolts from PA A to PA B.

When I crank the variac to 120VAC the bulb glows fairly bright. I'm a bit unsure at this point what to do. Everything seems good, however the bright bulb with the variac at 120v and the fan being on are making me hesitant.

[Enzo]

You might adjust the bias a bit colder and see if your bulb gets less bright.

[lowell]

Ok will do!

[lowell]

Ok so yeah something's not right w/ the bias circuit. The pot AP201 does nothing to the gate voltages. And I'm now realizing that the gate voltages of the upper half of the PA are -2.5v... and should be positive, not negative.

[lowell]

Sheesh, feel like I've had this happen too many times. I've checked and rechecked things and can't figure out how I'm getting these results:
My variac is set to 80VAC and the bulb is still in the circuit.

Q202 Vbe: -500mv
There is 128mv across R219 so it seems Q202 is on
There is 500mv across D206 and D209 so it seems that that string is on and current is flowing thru it. There is also ~-300mv between R220/221. (This is the same result on the working PA)

Q203 Vb: 2.5v
Q204 Vb: -3.2v

With the bias pot at zero ohms there is -2.2v on it. On the working PA there is -500mv at this point. All resistors in this string measure correct and are indeed connected. I also get identical ohm measurements across C210 from the bad PA to the good PA: 1.3k (with bias pot at zero)

I thought maybe something on the following page could be skewing the results so I checked to see if "output B" was 0v and it is. I'd think if something were off in the limiter circuitry that this point would be skewed up or down... maybe I'm wrong. I have checked the limiting transistors and surrounding components and they seem fine.

I replaced Q206/207, but it's possible there's something wrong with them.

I'm at a loss. I am still trying to fix a Trace Elliot amp that has bias problems as well and I'm at this same point with repairing it... stuck.

[lowell]

Can anyone suggest a safe way of firing this up with the bias circuit disonnected from the PA? Can I cut the traces and then connect the MOSFET gates to ground or something? I'd like to separate the circuits for more accuracy in diagnosis.

eck, I suppose I could take all those damn MOSFETS out again and install resistors in place, as I think Mark suggested in a previous repair, but damn... lotta work.

[lowell]

Took the dang thing apart again... not as hard the second time around. Kickin' myself... forgot to solder one Gate lead on one of the MOSFETS. whoops. I still don't see how that would affect my biasing circuit though so I'm not concluding that that is my problem. Gonna keep tryin some stuff and am double checking all other connections at this point while I have it out.

[lowell]

Enzo, Mark, JPB,

Correct me if I'm wrong, but turning the bias pot fully CCW is COLD right? It seems to me that this increases the Vbe on Q206, thus turning it on more, thus reducing the difference between +/- bias voltage rails, THUS reducing quiescent current in the MOSFETS.