Blowing ouput transistors when clipping, why?

I'm brand new here so I thought I would give a little background info.

Some of you may remember the company Mark V electronics that sold amplifier kits, the guy who ran the company closed up shop, but before he did he let everyone know who his supplier was. That company was Sound Master International out of Hong Kong, web address:

http://www.hkiol.org/template/0001/template.htm

A while ago I ordered some of the 300W Amp kits and built them. They worked great if I used a low power supply voltage such as +- 40V DC, if I hooked up a low impedence load such as 4 ohms and got clipping, everything was ok physically (except the produced sound, yuk). However, if I upped the voltage to the recomended +-75V, when I would get clipping I would lose half of the power transistors usually, as well as the op amp in the pre amp stage. For the life of me, I can't come up with a solution since my electronics background is limited to hobbiest activities. I can troubleshoot and repair the amps, but I would prefer to be able to drive them closer to their rated outputs to get more power from them. I tried replacing the output transistors with MJ15003's but still had the same problem.

I can post the schematic for these amps if anyone can help me troubleshoot them.

Here are the schematics and the instruction manual as well.


I'm glad that I found this forum, what a great way to get some more electronics experience.

http://www.lvcm.com/zombiess/amp/circuit.gif
http://www.lvcm.com/zombiess/amp/parts_list.gif
http://www.lvcm.com/zombiess/amp/powersupply.gif
http://www.lvcm.com/zombiess/amp/page1.gif
http://www.lvcm.com/zombiess/amp/page2.gif
http://www.lvcm.com/zombiess/amp/page4.gif
http://www.lvcm.com/zombiess/amp/page5.gif
http://www.lvcm.com/zombiess/amp/page6.gif
 
BLMN, if I remember corretly, that is what would happen, but I have blown the other ones a few times, but I believe most of the time it was Q10 andQ12. I would also lose Q7 and Q8 and Q5, Q6 and Q13, but not always. It just wasn't very consistant when blowing up! I would loose IC1 as well, but not always.

If this amp would always blow the same things repeatedly, then I could probably figure it out. But it seams almost random. I checked the voltage ratings on all of the transistors and even replaced some of them with higher ratings for more headroom, but I had no luck with it. My only success has been with lowering the supply rails down. I am trying to get a new transformer to bring my rails up from +-40 to +-55- +-60 or so to see if that doesn't give me more power without blowing things up. The only other thing I can think of is to run a limiter on it. Does anyone have plans for a simple limiter to limit the max input to something like 1.2 volts?
 
I have a suspicion about the way they designed the protection circuit Since the bases of Q3 and Q4 are referenced to ground instead of the output, as the power supply voltage is increased, the protection is progressively disabled. Maybe try changing the 3.9k's at the bases of Q3 and Q4 to 3 ohms while at the same time connecting the junction between the 1N4148's to the output instead of ground. I am not sure if 3 ohms is the right value for you, but if the voltage across the .5 ohm balast resistors goes up to 10 (20 amps of current), then about 50mA goes though the 180 ohm resistors on the bases, raising the voltage on the 3 ohm ones by over 100mV, added to the bias on the 1N4148, the transistor should turn on.

The diodes at the bases of Q3 and Q4 temperature compensate the protection circuit. But that may not really be needed. You may decide to just remove those diodes and use 15 ohm resistors instead of 3 ohm.
 

Nelson Pass

The one and only
Paid Member
2001-03-29 12:38 am
Strong possibility that the output devices are being
allowed to saturate at clipping, which results in
a slow recovery time of the saturated transistor.
This results in "cross conduction current" in which
high current flows from the positive rail through the
output stage to the negative rail, and is particularly
a problem if you clip the amp continuously at higher
frequencies. You can tell this is the case if the current
draw of the amp goes up significantly when you clip
it without a load.

If this is the case, then probably the bank of devices
which does not fail is the saturating bank.

You can make a simple limiter for about 1.3 volts with
4 signal diodes at the input, 2 pair in series for 1.3 volts
in each direction, and using an input resistor to limit the
current.
 
Nelson, I have tried the diode thing before, but when I was checking them with my scope and inputting a sinewave, i noticed that the output was being shaped by the diodes above about .8 volts with two of them in series. I have tried multiple types of diodes and all give a similar response. Is there a better way to do it with an op amp where it will just limit the overal signal. I guess a compressor would sort of do this, but I want the full dynamic range, just not to go above a pre-set level.
 

blmn

Member
2001-02-01 2:43 am
.
Jeremy,

At a first glance, it seems to me the current sample is feeded from Q9 and Q11 only, and they are, for this voltage (75V) being working out of their safe area. heatsink size and thermal stability must be considered too. I think you have to put the voltage down and modify the circuit allowing the current sample being feeded from Q10 and Q12 too. Maybe the circuit was designed to work at +/-75V with 8ohms loads only;

subwo1,
The current in .5ohm resistor at the colector of Q11 shares the colector current of Q6 too, so, it might be right and the absence of the .5ohms resistor at Q12 seems to me not "elegant". I´d prefer to put one there, but it seems to me the circuit could work without it.
 
Oh yes, Nelson, on second thought Jeremy should leave in the diodes in the base circuits of Q3 and Q4 , but hook them to the output instead of ground. They not only temperature compensate, but they also allow the protection to turn on softly and more gradually.

The resistors between the base and those diodes should be set to higher values to start and then test the amp and, like Nelson said, check the outputs for cross-conduction (shoot-through). They should not get too hot when driven to clipping with no load. A rule of thumb is that they should not get hotter clipping no load than they are just sitting at idle. Start with those resistors at say 20 ohms and with the power supply at +/- 75V, and decrease them until the protection circuit is no longer over active. I hope you can work it out without blowing any more outputs. Been there.

Also, If your power supply is stiff, it would be better not to run the amp at +/- 75V with a four ohm load, like blmn said. That may actually play a big part in the problem.
 
Oh wait, sorry, forgot that if there is no load, then the protection circuit cannot kick in. You will not be able to test it like that. If you have a variac, gradually raise the voltage maybe. Problem is that if the SOA is exceeded even for a fraction of a second, BJT's can blow pretty easily.

Here is my latest thought. Hook up the load you plan to run, be it 4 or 8 ohm. Make the protection resistors a rather high value. Use the variac to test it at gradually increasing voltages, all the wihile monitoring how hot the output transistors get. Watch yourself, there is 150V between the cases of the PNPs and NPNs.
 
Thanks for all of the advice guys, I think what I am going to do with the amps is just find some transformers and run them at a lower voltage so that I don't ever have to worry about blowing things up. I'll also use a limiter as well just to be on the very safe side. It's not as though I really nead the full amount of power from them, it would just be nice. I have two crowns in the house for my big power needs, a ce1000 and a ce2000.
 

djk

R.I.P
2001-02-04 4:23 am
USA
Several points.Coming out of HK the outputs are probably counterfeit.Real MJ15003 won't run on +/- 75V, nor will the 2N3773.Crown used ten 2N3773 at 67V in the DC300A.You say your design only has four? Throw them away and buy MJ21194 from a US industrial distributor.The TIP29C/30C will only handle +/- 50V, throw them away and use MJE15030/15031.Add a Baker clamp to prevent common mode conduction.Tom Holman, of THX fame, used this on the APT Model 1.It works.Anybody that can post schematics? I can send a drawing via snail mail.The TL07X series will invert the output phase and latch up when overdriven.Put in something else, like an OPA604.
 
If you have a good power supply (+/-75v , 1500VA ) your amplifier can reach 450w/4ohm rms and 600w with sever clipping.
I don't think that 4 MJ21193/94 will not blow after a short time if you run in 4 ohm mode. If you run in 8 ohm you have a chance with this good power devices.

If you want to run in 4 ohm , replace with Sanken 2SC3264/2SA1295 and blow cold air on heatsik with one PC fan.

Don't forget to match properly your power devices, becouse if you do not this, you will have problems again !

If 2N3773 is branded " Toshiba " it is counterfit .

Good luck !

PS. Tell us if everything is ok
 
I purchased the outputs myself, the original outputs were not MJ15003's. The original's were Tesla KU606's. I thought that the 15003's would be OK since they are rated for 140V 20A and 250W. I thought the voltage rating was OK since they are only seeing either +75 or -75 because each one is tied to the + rail and ground or the - rail and ground. Not the + rail to the - rail. Was I wrong to think this? When I found the spec's on the KU606's they were much lower than the MJ15003's. I think they may have had a slightly higher voltage though.
 
well isn't the +-75V power supply my real issue then? If I am exceeding the maximum rating then things will have a tendancy to blow. This could be my real issue. I didn't realize that they would see the whole 150V. So if I drop the voltages down to say +-60V, things should be OK. I need to get a variable transformer and do some more testing.