Ampeg B2 - Stopped working.

[frankthompson@37qs.freeserve]

Hi, I'm hoping that there are some knowledgeable guys around who can help me in my predicament! On stage I use an Ampeg B2RE coupled up to a SVT 210E and an SVT 15E. This set up has worked faultlessly for two years until Fridays gig. When it was switched on for the sound check it initially worked but there seemed to be a lot more 'popping' than usual when other kit was powered up. After aprox 2 minutes it stopped working. The power was still on and the limit light was very active. If I listened very closely to the 210 cab I could hear distorted sound but it was very quiet, also there was a lot of heat at the rear of the amp amd the fan was on. On investigation on saturday I found that the 15E cab was open circuit. Also if I took an output from the 'Preamp Out' socket into another amp it worked so I deduced that the fault must be in the power amp section. Another thread on this site said that there could be problems with the 'Anti thump' relay. I checked the relay and found the three contacts were shorted together inside the relay so shorting out the output signal to ground. Not being too sure of what I was doing I checked out a few components around the output area and found the six 5 Watt (0.47 ohm) resisters in the MOSFET circuits were open circuit and also three resisters were obviously burnt, R35, R45 & R21!
Can anybody please give me some advise as to what has caused the problem and how I should proceed.

There is a cicuit diagram of the amp at the following link.

http://www.euthymia.org/image/42704h0.pdf


Thanks in anticipation
frank

[Ray Ivers]

Frank,

Until Enzo can get here and tell you what really happened , my guess would be that your amp was the victim of a dead-shorted output. The first place I'd look would be inside the 15E cab, as the open-circuit reading could be caused by a short burning a wire or blowing an internal fuse (just a guess). I'd also check all your speaker cables and the inside of the 210E cab as well (I'm assuming you'll be using another head to gig with, and it would be a real bummer for an unresolved fault to toast that one too).

If you can verify the cause of the failure as being external to the amp, you could check each PS/PA component and replace as necessary, then cautiously power-up using a light-bulb limiter or Variac; if you can't verify the cause, then the fault is probably inside the amp, and you might be better off taking it to a tech unless you're 100% confident of your SS troubleshooting skills.

Ray

[frankthompson@37qs.freeserve]

Hi Ray
Thanks for the quick reply,
The speaker in the 15E was open cicuit and has now been replaced and works perfectly with another head. All the cables have been checked and are serviceable. I haven't checked inside the 210, I'll do it tomorrow. I was very surprised to find that the Ampeg doesn't s/c protection on the output, I would have thought that quality gear had all the bells and whistles.
Wrt taking it to a tech, I am confident and capable of doing the repairs and if I can get a few pointers as to possible causes I'll proceed. If not, then it's off to a tech. I know it's probably false economy but I don't like spending money unless I have to.
Frank

[Steve Conner]

If your anti-thump relay had failed and shorted the power amp output to ground, and then you kept on playing, that would explain the burnt-out power section. Here's my guess as to what happened:

Power amp goes a bit dodgy and starts putting out excessive DC: enough to burn out the 15" but not the 2x10".

Eventually the protection relay notices the DC and kicks in. Unfortunately since the contacts are shorted, it shorts the power amp output to ground and burns it out completely.

I would replace all the burnt resistors as well as Q3, Q8, Q4 and Q13. I'd also check the power MOSFETs and if any were dead I'd replace them all. However, those MOSFETs are not really meant for audio use, so Ampeg might have specially matched them at the factory. If you just bought any old devices from Mouser, they might not share the current evenly. (Notice the bias adjustment instructions saying they have to match within 5mV.)

[frankthompson@37qs.freeserve]

Hi Steve
Your thoughts about how the sequence of event make a lot of sense. I'll have to make some enquiries about matching the MOSFETs.
Thanks
Frank

[Enzo]

How it got that way is acedemic at this point. No protection circuit is bulletproof. Your relay contacts were welded together, that is a pretty dead short. Open loads don't matter to solid state outputs.

However, if you bought the thing new, isn't there a five year warranty? You should have three years left. Let them sort it out for free. Better hurry though since they are closing the field service. Last claim date is 11-15-6.

Those open .47 ohm ballast resistors is a good indicator the MOSFETs are shorted. R21 burnt means so is Q4 in all likelihood. When outputs are shorted, it is a good idea to replace drivers anyway.

Resistors are just conduits for current, they don't usually burn up on their own. So when one burns, you need to look at the path of that excess current. R45? I'd bet when the MOSFET shorted and opened its ballast, then R45 carried the load through Q3. Q3 is a limiter, and I bet it is shot now. Might as well check the S26, R46 pair too. Same story with R35 and Q8, and their friends. I am gonna also bet all those 47 ohm gate resistors are stressed if not open - change them.

A word to the wise, you only need one output xstr per side for the amp to function. I suggest of the MOSFETs are shot, remove them and replace only Q14 and Q10 until you are sure the rest of the circuit is not going to blow the new ones right up. Once it runs on the pair, fill in the rest of the blanks.

Q3,8 are the limiters, and you can disable them by lifting D25,28, but best to fix them. But I chose output Q14,10 specifically because they were the ones the limiters sample from.

Check Q13, R24, and check the whole bias string between them. Also chek the diodes and zeners between the + and - side gate drives, D15-18.

Left up to me, I would just check every part on the page. Certainly verify the DC voltages on the print.

[frankthompson@37qs.freeserve]

Thanks Enzo
The amp was purchased used so is out of warranty.
You've given me a lot of info there to think about and a lot of checking to do. I wont be able to check till next week when we are back of the road. I'll post when I've got some results.
Thanks to all so far for the help.
Frank

[Sock Puppet]

Those open .47 ohm ballast resistors is a good indicator the MOSFETs are shorted. R21 burnt means so is Q4 in all likelihood. When outputs are shorted, it is a good idea to replace drivers anyway.

Hi Enzo,
Could you clarify something for me?
I built a MOSFET amplifier a while ago which would protect itself from operating into a very low impedance by conducting less as they got hot (negative temperature coefficient).
If one transistor had shorted in the output stage then couldn't the others be OK? If any did happen to be OK, is it good practice to replace them anyway?

Thanks,

S.

[Steve Conner]

Hi sock puppet,

I don't know what Enzo will say, but I couldn't resist putting in my $.02 worth on the MOSFET thing:

There are two kinds of MOSFETs, lateral and vertical. Lateral are the old-fashioned kind, and vertical are the newer ones developed for use in switching power supplies.

Only the lateral ones have the magic properties that make them supposedly "Short circuit proof" (assuming that zener diode on the gate) and able to share current without ballast resistors when paralleled. The classic MOSFET amp application circuits were designed around these devices, such as the 2SK135 and 2SJ50.

VMOS were never meant to be used as linear amplifiers, but if you're careful, they can be made to work. There is a big incentive for manufacturers to do this as they're like 1/4 the price of lateral MOSFETs and can handle more current. They tend to have a positive tempco at low currents, and a poor spread of threshold voltages, which is why you see the relatively large 0.47 ohm ballast resistors in that Ampeg schematic. Also, they aren't self-limiting, so a current limit circuit is needed.

The reason for replacing them all if one burns out is to try and ensure that you get them all from the same batch, so they share better.

Lateral MOSFETs are still made by Exicon, and are still (IMO) a good way to build a simple, tough, but not terribly cheap power amp. I have a hi-fi amp with about $200 worth of them in it

[Sock Puppet]

Thanks Steve.

I'd no idea thay were so expensive now, the 2SK133 and 2SJ48 ones I used were remarkably cheap at the time.

S.

[Enzo]

Well try shopping for a pair of K134/J50. Now go shop for a second note on your house to pay for them. I have a dead four channel power amp here with a ton of them in it. I am using the amp as a MOSFET supply rather than repairing it because each transistor in it represents a repair I wouldn't otherwise get. The amp is worth a lot more parted out than repaired and whole.


On bipolars, a E-C short on one usually then takes ALL the current away from the parallel others. So they are usually then OK. Unfortunately if there is a B-C short it generally takes out the whole row.

In MOSFETs, a short from D to S could act the same, sure, and you might need replace only a pair. But in his case, the extra clue is the burnt open ballast resistors. The open resistor means the FET associated with it has probably failed. or at least was stressed hard. And just like bipolars, what if the rail gets onto the gate buss? Poof. perhaps the amp died from a failed driver, which puts the rail on the gates and turns the FEts on hard. Relay drops out and arcs from the current, welding itself shut. It could happen.

I expect the lot of his xstrs are shot in this case.

[Steve Conner]

Hi Enzo

Those Exicon FETs come in TO-3, and I think you ought to be able to use them as replacements for the 2SK135 and 2SJ50. The ECF10N16 and ECF10P16 seem about right. Exicon also do higher voltages than the 160V those are rated at, which could be useful. I've seen C-Audio amps that ran the 2SK135 and 2SJ50 off of +/-100V rails, putting up to 200V across a 160V device. (C-Audio went bust, go figure...)

They're not too cheap though, about double the price of a MJ15024 or whatever, so a repair might not be economical anyways. It's a shame as those C-Audio things were nicely built apart from the crappy FETs