Output Transistor replacement Crown Amp xs1200

[BDB]

Hello all of you beautiful people,
i have a crown xs1200, i bought it knowing it was blown up. Tore it all apart and quickly identified two failed Output transistors on channel two. Amp is two channels total. Each channel has 8 OP devices. 4 npn and 4 pnp.
Removing the two duds (one shorted, one open) allows the amp to power up and run like normal. No faults. Sound comes out. No dc offset. (before i removed the two failed devices there was about 12v of offset on the channel) ive only tested these two devices, reason being: it works fine without them. So my question is, should i just replace these two? all four? i ordered two replacements but have not put them in yet. I have not checked the emitter resistors and i have not checked any drivers. Figured the drivers were good because the amp still powers up and works. i know when the OP fail that its probably caused by something else in the circuit, however i am assuming that since only two of the four failed that it is just the devices and not something else in the circuit causing this?

 

[elnaldo]

When you have output transistors in parallel, you need to match them for Ic, so they share the load. If not matched, one transistor will take most of the current and fail (electricity always go to the easiest way )

In your case, having one npn and one pnp to replace, you should replace all 8 output transistors with a matched set. Crown used to number the transistors measured at the factory so you could buy just one of the same type. Or buy a matched set from them.

If not, you need to measure them, biasing them in a test circuit and measuring the voltage across one emitter resistor.

I can post the procedure extracted from an old Crown service manual. It explains how they match the output at the factory.

 

[BDB]

Hello, i might of misled you, its two pnp transistors. or two npn i forget which but they are the same type.

 

[BDB]

PNP mj21193

 

[elnaldo]

OK, so if you blew 2 of the same half, you should replace at least all 4 in parallel with a matched set (match Ic, the current they let flow under the same bias voltage). Check if they have some number written on them, with a marker. That would be the factory group.

If not, perhaps they were replaced already.

I had very good results using On Semi MJ15024 as output transistors. Same Voltage, Amperage and Wattage than your actual devices ( a bit lower gain, min 15, max 60) (MJ15024 are original transistors in many Crown and Peavey large amplifiers). Also, very close manufacture tolerances. From the batch I bought, all were very close in specs. I could use them without measuring and matching, since their performance was very similar.

 

[ZOOM]

Matching is nonsens, there is no need for it in this time and age. Tis a practice of dinosaurs long dead, sadly their legacy tries to live on due to willful ignorance.

 

[BDB]

Alright that sounds good thanks

 

[BDB]

uh oh two conflicting opinions :O

 

[Hyperion]

No, it is not nonsense, our esteemed colleague is referring to the fact that transistors these days are manufactured to a much tighter spread of characteristics. He completely forgets that you won't necessarily need to replace ALL the transistors, IF you can match some you buy (from a reputable source!) to those already installed. Otherwise you might need to buy a full set for one side and at least check/match accordingly.

If you buy from auction sites and particularly from sources in China, you may find that you have bought a bunch of counterfeits - beware.

 

[SoundOfSound]

When you have output transistors in parallel, you need to match them for Ic, so they share the load. If not matched, one transistor will take most of the current and fail (electricity always go to the easiest way )

In your case, having one npn and one pnp to replace, you should replace all 8 output transistors with a matched set. Crown used to number the transistors measured at the factory so you could buy just one of the same type. Or buy a matched set from them.

If not, you need to measure them, biasing them in a test circuit and measuring the voltage across one emitter resistor.

I can post the procedure extracted from an old Crown service manual. It explains how they match the output at the factory.

I would like to see this procedure and how close the parallel transistors need to be matched and how complicated it is. I have a Denon with a blown channel that uses 2 sets of PNP/NPN pairs per side. In my case I was thinking of using NJW3281G/NJW1302G pairs that are supposed to be matched to within 10% from the factory.

http://www.onsemi.com/pub_link/Collateral/NJW3281-D.PDF

I'm still looking around for reliable information about matching parallel pairs so seeing the crown procedure would be great.

Would it happen to be in any of these manuals?

http://www.schematicsunlimited.com/c/crown

 

[BDB]

Hi sound of sound, it is my understanding that matching npn to pnp is unnecessary as they're not exactly in parallel (they operate on two different rails) however since you said PAIRS of pnp and npn that means that there's 4 devices per channel? if so i am informed that you should match the two pnp's together and the two npn's together.

 

[elnaldo]

you need to match the transistors in parallel, not the complementary. It's true that nowadays manufacture tolerances are much tight and better, and I'm not an expert on this (for sure there are people working at a transistors factory or very close to the manufacture procedures , here at AK, and of course their opinions are much better informed) , but anyway I wouldn't assemble a large amp (with 4 transistors in parallel we are talking about hundreds of Watts) without matching them.

At least, after installed, I'd play a test tone and measure the voltage across the emitter resistors, to see if they match. If one transistor is taking all the current, it will fail, perhaps starting a cascade event.

I would like to see this procedure and how close the parallel transistors need to be matched and how complicated it is. I have a Denon with a blown channel that uses 2 sets of PNP/NPN pairs per side. In my case I was thinking of using NJW3281G/NJW1302G pairs that are supposed to be matched to within 10% from the factory.

http://www.onsemi.com/pub_link/Collateral/NJW3281-D.PDF

I'm still looking around for reliable information about matching parallel pairs so seeing the crown procedure would be great.

Would it happen to be in any of these manuals?

http://www.schematicsunlimited.com/c/crown

This is how Crown did it in the '70s. From the DC-300 service manual at the Crown site.

 

[SoundOfSound]

Thanks BDB and Elnaldo,

Looks like the preferred group ranges from about 24% to 14% on Vbe.

In my case, Denon PMA 900V, I also have 4.7 Ohm 1W MOX +/- 5% (NBF) resistors hooked up to each of the bases on all 4 transistors per channel. There are also .22 Ohm 1W MOX +/- 5% (NBF) emitter resistors.

I think I will hand match these being they are only about 10 cents each. Perhaps that will help to balance the load a little bit.

Thanks for the info.

 

[Dackard]

Hi there SoundOfSound,
I have received an old Rotel Ra-931 to fix. I want to install that pair NJW3281G/NJW1302G for my client. Are you satisfied with the sound they produce?

Dackard.

 

[Hyperion]

Hi there SoundOfSound,
I have received an old Rotel Ra-931 to fix. I want to install that pair NJW3281G/NJW1302G for my client. Are you satisfied with the sound they produce?

Dackard.

Welcome to AK Dackard!

Please note that you should not post a query for a Rotel on the end of a thread about a Crown xs1200 - these are totally different and advice given for the Crown may be wrong or even dangerously wrong for the Rotel.

What you should do is start a new thread titled "Rotel Ra-931 problem" - or similar, and ask your specific question there.

Sorry to jump on you for your first post, but getting posting protocol right can lead to excellent and helpful results, I hope you appreciate that.

Enjoy your stay, there are a lot of tremendously helpful people here.