Ampeg B100R drawing too much current (output section shorting?)

[velo-hobo]

This is a continuation of the zombie thread I revived here: Ampeg B-100R Melting Bridge Recto

(apologies for the resurrection, if a moderator would like to move the more recent/relevant posts from that thread to this one, that might be helpful)

Quoting 52 Bill:

Did you actually remove all 4 outputs transistors? Did you test them all out of circuit before you put them back in? I can't say for sure, but as Enzo hinted, you may have caused some additional damage by running the amp that way. I try and remove as little as possible to test and repair anything that is in front of me.

It is very easy to test the outputs for shorts while in circuit, finding which one of two or more paralleled devices is bad takes a little more work, but in only pulling the bad devices, it allows you to test the rest of the circuitry before ordering parts. If you had just found the short and pulled Q5 you could then power up the amp through your limiter to see if the amp would have powered up normally. If it still drew too much current then you could look at the drivers, etc.

Check all of the transistors and diodes in the power amp in circuit. If anything doesn't check right, then pull something out to test it. If the power transistors have already been pulled then test them all out of circuit and then put them back in until you complete diagnosing the other problems. Try and limit the variables that you add to the mix until you know what is wrong with the amp.

Yes, I pulled all 4 output transistors to check them out of circuit. Q5 is shorted from emitter to collector. I pulled Q5-8 and the amp powered up fully with no excess current draw. Q6 and Q7 seemed ok so I put them back in to see if the amp would stabilize, but something still draws too much current. I pulled Q8 because I am under the impression they more or less work in pairs - is that correct? (btw I am running this all on a light bulb limiter and variac, with no load on the output)

So either something is wrong with Q6/Q7, or there is another component that will only draw too much current with the output transistors in place.

I will double check Q3-Q9, and the 6 smaller 1n914 diodes in the power amp. I gave the smaller parts a cursory inspection but I need to take a closer look to see what else might be faulty. I should mention again that Q3 looked like it had a cold solder joint on the emitter.

Finding the short in Q5 kind of distracted me from the other stuff for a moment, and I didn't have time last night to investigate further.

Like I said, I ordered 2 whole new complements of semiconductors back through IC2, for about $20 shipped from mouser. I'll check the smaller SS parts I mentioned above for problems before installing the new stuff. Besides those and the bias and emitter resistors (already checked those Rs), are the any particular caps or resistors I should take a closer look at?

[52 Bill]

(btw I am running this all on a light bulb limiter and variac, with no load on the output)

Ooh, double protection. How familiar are you with using this setup with both variac and LBL? I use one or the other, but I have never used both at the same time.

Maybe one of the other guys here has done this before and can offer some advice. In my experience, the LBL does a great job of saving fuses and burned components, but it does have a few drawbacks. When you plug a working amplifier into a LBL, at turn on there usually is a brief flash of light as the amp draws its' turn on current and then settles down to idle. Some amps flash more than others and some don't seem to flash at all. This flash is caused by a short high current imbalance as the power supplies and power amp circuits come up to their balanced operating point.

When you slowly raise the voltage going into the LBL with the amp under test, you may be seeing what is normally a msec of high current draw as a major circuit problem. I know that some amps will not fully power on through a LBL and can fool you into thinking that there is a some sort of short circuit. Early SWR amps will do this, must be something about their bias circuit that causes this.

After checking and rechecking everything in the power amp, I'd suggest reinstalling the output transistors that are still ok and try the amp with just the LBL and see if the lamp lights up and stays lit. If it does, then there is still a problem to be found. If it doesn't then you will know that the indication of a short was caused by the combination of Variac and LBL.

[velo-hobo]

Yep, I do know about the inrush current as the caps charge up etc (the reason for slo-blow fuses, yes?) - I also use the LBL when I am working on tube amps and am familiar with the flash that settles to a dim glow. I guess I just hooked up the variac too as a force of habit, since often I am powering up old amps that have been sitting around for a long time and may have capacitor issues.

I will try using just the LBL for the next round of testing, since the variac isn't really adding much value to what I am doing at this point. Although I did notice that before I pulled the output transistors, the bulb would light up with the variac putting just 20V on the mains. Putting Q6 and Q7 back in, the bulb lit up at around 80V on the mains. I didn't go further as it got brighter with higher voltage and didn't dim down, as I would expect from caps charging up.

I'll have some more time to take another look at this amp tonight or tomorrow night and report any new findings. Otherwise, just waiting for the new parts to arrive.

SS amps are sort of newish territory for me, although I wouldn't say I am a tube expert either. Some of the electrical engineering aspects of this stump me as I don't have formal education in the field - I just absorb information easily and learn quickly from experience. Gear repair and modification is a sort of hobby for me, although in this case it is potentially going to save me hundreds in repair work or replacing the amp entirely. Plus the satisfaction of solving the puzzle myself (albeit with a good amount of help from folks like you!)

[Enzo]

Q5,6 are the V+ side and Q7,8 are teh V- side. For testing you really only need one of each installed - IF ALL ELSE IS OK.

But there are other reasons for large current draw. A shorted transistor sure will, but if the bias circuit is open, then both sides will turn on hard - a direct shunt across both power supplies. Q9 and its resistor friends are the bias circuit. If it is open or just can;t turn on enough, then you have large current draw. For test you can short across Q9 - short the opposing output transistor bases together. This results in an increase in crossover distortion, but also the amp runs maximally cold. If that stops excess draw, then your bias is the problem.



For me, the bulb is a go/no-go thing. I use the variac myself. The variac is not very useful without monitoring the current it delivers. People tend to focus on the voltage, but the current is what matters. When I turn up the variac, I am not looking for how many volts I can "get up to." I am watching the current, and I want to see if I can bring the unit up withoout having current start to ramp up on me. If that happens, I usually donlt care if it is at 90v or 60v. Oh I might derive some information from that, but it is not the point. If your current is already starting to soar at 20vAC, then you still have a hard short somewhere. And that includes my "both sides running" scenario.

[velo-hobo]

Checked the drivers and bias circuit, seemed fine so I put in the new output transistors. They were smaller than the originals so I had to fly the leads from the board to the Qs, which was a little bit of a pain and time consuming. The amp powered up fine with no excess current draw. I got ahead of myself and forgot to check the voltages in there before I hooked it up to the speaker and tested it with a bass. Within about 5 minutes of playing at moderate volume the fuse blew again, with the familiar burning smell coming off the board.

I checked it over and all 4 of the output transistors had shorted between the emitter and collector! Obviously I screwed up by not checking the circuit more thoroughly before testing with a load, but what could have caused this?

I checked the emitter resistors after this event and they still seemed fine. Pretty sure the bias network is not open either. I have another new set of Qs to put in there but of course I haven't solved the root issue yet.

As an aside, I have another one of these amps that is working fine. I am hoping to use two of these for an upcoming tour, both for the purpose of slaving one of them in bigger venues, but also to have a backup amp should one of them die on the road. Since this one has been giving me a lot of trouble, I just bought another used one as insurance. I would still like to try to fix this amp, but now it's not a high-priority project for me, and I don't see it being trustworthy on the road anymore anyway. It'll probably just exist as a practice amp from now on if I actually succeed in repairing it.

[52 Bill]

Checked the drivers and bias circuit, seemed fine so I put in the new output transistors. They were smaller than the originals so I had to fly the leads from the board to the Qs, which was a little bit of a pain and time consuming.

Please explain. How did you mount the new transistors to the heat sink?

I got ahead of myself and forgot to check the voltages in there before I hooked it up to the speaker and tested it with a bass. Within about 5 minutes of playing at moderate volume the fuse blew again, with the familiar burning smell coming off the board.

I checked it over and all 4 of the output transistors had shorted between the emitter and collector! Obviously I screwed up by not checking the circuit more thoroughly before testing with a load, but what could have caused this?

Not adjusting the bias setting after the transistors were replaced could have cause this. Or it could be the smaller size transistors couldn't handle the power that the amp was drawing.

...I don't see it being trustworthy on the road anymore anyway. It'll probably just exist as a practice amp from now on if I actually succeed in repairing it.

While you may have had a few problems with the amp, there's no reason that it can't be repaired and put back into perfect working order.

[velo-hobo]

The new Qs had mounting holes that fit the old insulating bushings and screws. Because of the smaller form factor, the leads wouldn't reach all the way to the solder pads on the pcb so I soldered short lengths of insulated 20ga solid Cu wire to bridge the gap. I've seen this done on old Wurlitzer 200 amps too, although I assume in that case it was also to accommodate thermal expansion since the transistor leads basically pointed straight at the board with no bends.

I do suspect that the new parts couldn't handle the power requirements unless something else is wrong with the amp. I'll have to try to dig up a datasheet for the old Motorola transistors and see how they compare to the new ones (I bought some made by Fairchild Semiconductor). I thought transistors of the same part number would be interchangeable but I guess I'll have to look at the specs a little more closely.

I'm terribly unfamiliar with adjusting the bias on SS amps. I suppose since the new Qs were a different brand this would need to be done, although the bias resistors in these amps are specifically called out, at least in my version of the schematic. Later versions had a trim pot for adjusting bias, I think.

I am pretty confident it can be fixed! But I worry about having enough time right now to develop a deep enough understanding of the problem to feel confident about my repairs.

[velo-hobo]

Just looked at the Motorola spec sheet and compared it to the Fairchild sheet. The specs look pretty much the same, except the Fairchilds have a Collector Dissipation rating of 80W, while the Motorolas have a Total Device Dissipation of 125W.

So that would appear to be the problem!

Looked on Mouser and ON Semiconductor makes a 125W TIP142/147. The form factor is similar to the larger Motorolas, so these would seem to be the ticket.

[g1]

Differences between Tip142 (or Tip147) packages are a function of the part number suffix, not the manufacturer. It is the suffix that determines the case style.
It is possible that the incorrect case style did not mate up with the heatsink properly, leading to overheating and failure.
But it is also possible there is another fault which caused the originals to blow, as well as the replacements. Be sure to go over everything very thoroughly this time. It would be helpful to check all voltages against your other amp which is functioning properly.
From the other post, Enzo mentioned there are 2 versions of this amp, one from around 1997, the other from around 2005. Make sure you have the right schematic and bias info.

[velo-hobo]

Sure, I would guess that too - essentially the same internals, but a smaller form factor would likely affect rated dissipation. I wasn't trying to say a difference in manufacturers was relevant, it's just that what is readily available at the moment comes from different suppliers. But thanks for clarifying, I was unaware of the different variants of these parts until I took a closer look.

By examining the spec sheets it seems pretty obvious to me that I overlooked the difference in dissipation ratings. It seems a likely explanation for blowing all four output transistors at once when the original problem was at least superficially traceable to just one output transistor failing.

I have new parts on the way but I will be double checking all the related components in the output section, as well as replacing the emitter resistors with new ones.

And yes, I am aware that there are two versions of the circuit - both my amps are the older variant, so I am working off the correct schematic here.

[Enzo]

You do have to be careful sourcing those types. Our amps do not use the TO220 style. This transistor pair is used extensively in Crate and Fender SS amps, and as a sub for the BDV64/65 types in MArshalls, to name a few. They are usually the TO218 size or the nearly identical TO3P. TO247 is great, but you need to make sure they have enough room, as they are a little larger yet than the others.

[velo-hobo]

Yep, the ones that I cooked are TO220. The ones I have on the way are TO218. I still have an extra set of the smaller ones though, so I'll hang onto them in case I need them for something else down the road.

What is better/different about the TO247? Or were you simply saying that they would also work fine?

I considered putting in additional heatsinks for each Q, but I think the amp would also need a small fan to make the most of that modification. I've had this amp since it was new in 1998 without any other trouble til now, so I'm not yet inclined to go there with it.

[Enzo]

The existing heat sink is OK.

The TO247 package is larger still than the TO218. It would tend to cool a trifle easier. If it fits the space, it is a viable alternative, that's all. I;'d stick with the TO218s.

[Steve Conner]

TO3P and TO264 are the largest of the plastic packages. TO218 is a bit smaller, and TO247 is smaller again. TO220 has no place in the output stage of a bass amp.

[velo-hobo]

I got the new TO218 transistors and put them in with new thermopads. While I was at it, I also replaced all the big power/emitter resistors (and bumped up the 620ohm Rs from 3W to 5W), Q3-4, Q9, and IC2. I also reflowed all the solder joints for the knobs and jacks, just cause it seemed like a smart thing to do while I had the board out of the chassis.

Everything seems to be operating stably and within tolerance. I played with the amp for about a half hour with no problems. I'm going to give it a longer workout this weekend at a rehearsal, so we'll see how it holds up.

Thanks for all the help everyone!