Reasons for transistor failure

[Bigdrums]

I am trying to repair a Krell kma 100 mkii mono amp. I have a second amp that is working properly. These are class A amps with huge power supply's and draw near 8 amps at idle.
Each house 16 metal cased transistors in the power section-8 PNP and 8 NPN. Originals are marked Krell A and Krell AA. Replaced with MJ15024/MJ15025
My question is Is it possible for an open emitter resistor to cause it's transistor to fail (short)? Or is it more likely that a shorted transistor would cause the emitter resistor to become open?
After replacing all 16 output transistors and some drivers I brought up the power on the variac to 25% and without seeing much current draw (less that 1 amp) blew out 1/2 the transistors.
Trying to understand the sequence of events and cause.
Also it would help to have a schematic for this particular unit but so far are unavailable so if anyone might have one, it sure would help.
I'm almost ready to give up on this effort after many hours and much expense.
Thanks!

[Jon Snell]

Open circuit emiter resistor means no current will flow. Short circuit transistor means high current will flow. Everything else is down the the amount of power available to destroy whatever is in the transistor's path.
I have this in my files ;
krell_ksa-100_mkii_2.zip

[vintagekiki]

Krell schematics
On some portals you must log in or sign up

http://www.hifi-manuals.com/Krell

https://elektrotanya.com/?q=showresult&what=krell

https://elektrotanya.com/krell_100-mk2_sch.pdf/download.html

http://www.audiocircuit.com/Home-Audio/Krell

https://audio-circuit.dk/krell/

https://www.diyaudio.com/forums/solid-state/45352-krell-schematics.html

https://www.diyaudio.com/forums/solid-state/42357-reverse-engineering-krell-kma-160-a.html

https://www.diyaudio.com/forums/solid-state/312263-krell-kma-100-a.html

https://www.diyaudio.com/forums/solid-state/308366-kma-100mkii-ksa-50mkii-rebuild-sanity-check-requested.html

www.amplimos.it/images/krell_ksa-80_200_kma-160_400_technical_info.pdf

https://www.youtube.com/watch?v=6jaov-4d0og
^{Krell KMA 100 Repairs and Restoration April 2016}

[Mick Bailey]

What was the original fault before you replaced the transistors? You mention open emitter resistors, but were these open beforehand, which side were they on, and how many were affected?

What did you use to measure current draw? If the amp normally idles at 8A and blew half the transistors at less than 1A @ 25% voltage, then something doesn't seem right there because the power transistor dissipation is less than 12.75W (0.25x51x1). This doesn't strike me as something that would cause half the output transistors to blow. If you had a fault that turns one side on fully and swings toward one rail, then the output transistors are sharing at most (ignoring all other loses) 12.7W/4 = 3.175W. Not enough to blow a power transistor or open an emitter resistor. I wondered if you had a domino-effect failure but this doesn't seem likely given the conditions.

I have had catastrophic failures when unintentionally installing fake replacement power transistors.

[loudthud]

An old Phase Linear trick: Usually one transistor shorts and saves the others. If you carefully measure resistance Collector to Emitter across each (for example) of the NPNs, the shorted one will be the lowest and the others will be higher by twice the Emitter Resistor value. This assumes that none of the Emitter Resistors have gone open.

[Mick Bailey]

Good advice - often because they're in parallel this causes someone to think they're all shorted when only one is, especially as emitter resistors are fractional Ohm in value.

[Bigdrums]

What was the original fault before you replaced the transistors? You mention open emitter resistors, but were these open beforehand, which side were they on, and how many were affected?

What did you use to measure current draw? If the amp normally idles at 8A and blew half the transistors at less than 1A @ 25% voltage, then something doesn't seem right there because the power transistor dissipation is less than 12.75W (0.25x51x1). This doesn't strike me as something that would cause half the output transistors to blow. If you had a fault that turns one side on fully and swings toward one rail, then the output transistors are sharing at most (ignoring all other loses) 12.7W/4 = 3.175W. Not enough to blow a power transistor or open an emitter resistor. I wondered if you had a domino-effect failure but this doesn't seem likely given the conditions.

I have had catastrophic failures when unintentionally installing fake replacement power transistors.

What is the best way to determine if the transistors are counterfeit? Should I break one of the originals and one of the new ones open to see?

I have an amp meter on the variac and I thought that if I kept it low like that I would be safe. The swing to one rail seems possible because the original reported problem was 100 ohm smoked resistors on one side of the pre amp. The same thing happened after I replaced the bad transistors (and bad resistors) The feed to these resistors was from the amp section, which I thought was through the TO-3's (one from each side of the amp sourced back to the P/S. This would make no sense because of the huge voltage difference. But there is a jumper tying each side together also that I can't understand. I checked the emitter resistors when I changed out the TO-3's and after the 25% test some of them were open. I monitored the voltage drop across the pre amp resistors that wanted to burn and at 25% got a 20 volt difference. Obviously it needs to be more or less nil.

[Jazz P Bass]

Did you have a load connected for the ramp up test?
Best to not have one at that time.
If you are totally unsure of the repair, power up the unit with a current limiting lamp in series with the mains.

[Enzo]

Where did you buy the transistors? If you got them from Mouser or Digikey, they are likely fine. If you bought them from ebay, well, I never buy there.

[Bigdrums]

Did you have a load connected for the ramp up test?
Best to not have one at that time.
If you are totally unsure of the repair, power up the unit with a current limiting lamp in series with the mains.

NO load on power up
What wattage lamp would be appropriate for a device pulling 8 amps?

[Bigdrums]

Where did you buy the transistors? If you got them from Mouser or Digikey, they are likely fine. If you bought them from ebay, well, I never buy there.

Ya, I got them from ebay and am challenging their authenticity with the seller now.

[Mick Bailey]

I use a 100W lamp for most jobs, sometimes a 60W for smaller amps.

If you get multiple transistor/emitter resistor failures on one side this is usually because the transistors are all turned on hard on that side due to a fault further back, or one transistor has a C-B short, which is less likely. The approach I always take when the output transistors have shorted is to leave them removed and once I am reasonably happy that all of the faults have been identified downstream, power up the amp slowly and monitor the output transistor base voltage. Then I install just one pair of transistors (one either side) and bring the amp up slowly on the variac whilst monitoring the output for any DC offset, which would mean that the output transistors are pulling towards one rail due to a rising base voltage.

When buying transistors you really need them to be matched, or from the same production batch. If they're a random assortment you can get problems with uneven load sharing. This is evened out to some extent by the emitter resistors but you can still have problems with mismatches.

The first check with possible fakes is to research the component type and see if they've been reported as fake. Sometimes the batch codes are all the same, the printing is different, there is print where they should be laser etched, the casing is rough or has other distinguishing marks.

[g1]

The first check with possible fakes is to research the component type and see if they've been reported as fake. Sometimes the batch codes are all the same, the printing is different, there is print where they should be laser etched, the casing is rough or has other distinguishing marks.

One very rough rule of thumb I use is this: if they are an obsolete part that are hard to find or can't be found here, yet they are popping up from China, they are probably fake.
I ask myself why anyone in China would ever have had them, or why would they be getting them now. They have to be NOS, no one is building obsolete parts. So why would they have them?

[Enzo]

You are quite right^^^^

But I don't even go that far, I just ONLY buy parts from real suppliers like Mouser or Digikey. WHatever small savings I might get from CHina, will totally disappear the first time one blows up on me.

[Mick Bailey]

https://www.diyaudio.com/forums/part...24-beware.html