Krell KAV 250a deader than a dodo

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Hi everyone this is the first time i've posted hope to start an interesting project the kav250a i've recently got has 1 channel working and the other dead. 6 of the krell output transistors have gone! (wasn't me lol) and i've been on google all night looking up this amp and the bjt's which are motorola 9710/9711 (npn/pnp) there are 7 of each type, a darlington driving 5 output devices tied with 5 x 0.33Ω emitter resistors for each half.

Basically this amp didn't cost anything - and i don't have the spare £££ either to do a krell repair so how about using different output devices?

Any info on this would be greatly appreciated since i would rather find equivalents that will "work" plug+play than having to modify the driver board which tested ok (substituted in the other channel) and i'm not too worried either about trying to meet the original spec - the krell devices are high bandwidth is that their only feature?

Look forward to chatting to you knowledgable people i've got a fair amount of experience building high power amps for PA/disco and making speakers but this is probably the nicest bit of kit i'm likely to own! :scared:
 
Ex-Moderator
Joined 2002
Have you checked with a scope? Does the driver stage give a nice clean sine if you run it with a signal generator? If so, you should be good to swap the outputs. I'd swap the output resistors as well whilst you're at it, the short could have damaged them.
 
Hi,
MJ21193/94 are almost identical to the 95/96 types.

These are not widebandwidth with Ft=4MHz, but are good for the power they can pass reliably.

Did Krell use a Darlington transistor in the KAV?
or did you mean a darlington connected transistor pair?

When you swapped the board over, did the drivers go with the PCB? and only the output devices and Re were not swapped?
 
update on krell service....

Hi, thanks for the replies and info guys I've been working on the amp this afternoon and had some success...I stripped it right down and cleaned the chassis first then did some minor work on the psu board (new mains fuseholder - someone had mashed it !) and then tested the psu on and off load (88.6 / 91.2VDC) it's ok.

I did some tests on the working channel, DC offset is fine 25mV but there is +/-1.2VDC on the driver outputs is this ok? It is working perfectly tho - had it running for a few hours into a 4Ω bass bin earlier.

The next stage is testing the dead channel, first the driver board like you suggest with a scope, I'm not sure how to go about this I use my PC for measurements normally don't own a scope....can I set up an attenuator from the +/- drive outputs and safely leave the output stage disconnected? I don't want to fry the amp (or my PC!) by overloading the line input......

Yes the drivers are the same as the outputs (9710/9711) and the drivers are in a darlington pair feeding the 5 outputs on each side. I've sketched the output board, the terminals marked "drive +/-" are the test points. According to the datasheet the max gate voltage is 5VDC but what sort of voltage can I expect here for attenuation purposes (i.e. connection to computer for measurement?)

Thanks a lot for your responses here is the sketch....
 

Attachments

  • krell_250a_op.jpg
    krell_250a_op.jpg
    63 KB · Views: 1,643
Re: update on krell service....

digitallake said:
Hi, thanks for the replies and info guys I've been working on the amp this afternoon and had some success...I stripped it right down and cleaned the chassis first then did some minor work on the psu board (new mains fuseholder - someone had mashed it !) and then tested the psu on and off load (88.6 / 91.2VDC) it's ok.

I did some tests on the working channel, DC offset is fine 25mV but there is +/-1.2VDC on the driver outputs is this ok? It is working perfectly tho - had it running for a few hours into a 4© bass bin earlier.

The next stage is testing the dead channel, first the driver board like you suggest with a scope, I'm not sure how to go about this I use my PC for measurements normally don't own a scope....can I set up an attenuator from the +/- drive outputs and safely leave the output stage disconnected? I don't want to fry the amp (or my PC!) by overloading the line input......

Yes the drivers are the same as the outputs (9710/9711) and the drivers are in a darlington pair feeding the 5 outputs on each side. I've sketched the output board, the terminals marked "drive +/-" are the test points. According to the datasheet the max gate voltage is 5VDC but what sort of voltage can I expect here for attenuation purposes (i.e. connection to computer for measurement?)

Thanks a lot for your responses here is the sketch....

I have noticed that some analogue scopes go for peanuts nowdays on ebay. Perhaps it would be worth getting one? Look for one near you so you can pick it up, as they are heavy and can be damaged. Also, you may want to see it in operation before you hand over your money.

Oh, and make sure it comes with probes!

If you are near Oxford you can borrow my spare Philips 25 MHz scope instead.

P.S: Just noticed: Cornwall. Not near Oxford!
 
scope testing

Hi, thanks for the advice on getting a scope shame I'm miles from you thanks for the offer of borrowing one though. I've managed however to do some tests today and come to the conclusion that both the driver boards are fine. I've had one running but wanted to make sure it wasn't a fault with the driver section that caused the output devices to go short but all looks good it produces the same results with either board.

:)
 

Attachments

  • scope_tests.jpg
    scope_tests.jpg
    92.5 KB · Views: 1,036
I have just learnt that Krell also used some 484xxx prt numbers that are basically MJ1302a's. It's possible that the 97xx's are also MJ1302a's/pairs. The MJ1302's are 30Mhz parts.

ON does not sell these anymore but sells the TO-264 version MJL1302a/Pairs.

I am very sure that the 21193/4/5/6 will work, though I would keep the driver/ pre-drivers as is 97xx because of their better gain.

I have not seen a pre-driver, driving the next stage via an Inductor (L1/2 in your diagram)... is this because of high FT devices?

Yes I do remember Krell literature reading drivers are OP devices in this amp...

PS: What number Hex key did u use (or was it torx) to open up the heatsink to expose the TO-3's?

PPS: Krell had to buy 1 million OP devices from MOT to get exclusivity.... with those many devices.... one can see them being used a lot. ;)

HTH...
 
pinkmouse said:
Excellent. Looks like you're good to go with a new output stage then. Don't forget to change the resistors as well, and get some new isolation pads for the TO3s. Will you be redoing the other side to match?

Hi, thanks for the reply pinkmouse! I did take note of your earlier tip about replacing the emitter resistors - do you mean replace the other power resistors on the board (between the driver and output devices) as well? Also, there is a total of 12 x 47uF / 100V elco's (1 next to each TO3) on the +/- supply rails to ground - replace these? The only components left are the 2 inductors mentioned below and 2 diodes which are tested ok.

In reply to your question about doing the other channel, I am in 2 minds whether to or not since it is working....if it ain't broke don't fix it comes to mind! However (on the other hand ;)) while dismantling the output stage which is in need of tlc I noticed that most of the TO3 insulators came apart very easily (brittle due to heat?) no matter how careful I was removing the transistors. Now the question: as long as the other output stage remains intact (heatsink not undone, etc.) is it going to be ok assuming the insulators are in a similar state??!
 
K-amps said:
I have just learnt that Krell also used some 484xxx prt numbers that are basically MJ1302a's. It's possible that the 97xx's are also MJ1302a's/pairs. The MJ1302's are 30Mhz parts.

ON does not sell these anymore but sells the TO-264 version MJL1302a/Pairs.

I am very sure that the 21193/4/5/6 will work, though I would keep the driver/ pre-drivers as is 97xx because of their better gain.

I have not seen a pre-driver, driving the next stage via an Inductor (L1/2 in your diagram)... is this because of high FT devices?

Yes I do remember Krell literature reading drivers are OP devices in this amp...

PS: What number Hex key did u use (or was it torx) to open up the heatsink to expose the TO-3's?

PPS: Krell had to buy 1 million OP devices from MOT to get exclusivity.... with those many devices.... one can see them being used a lot. ;)

HTH...


Hi K-Amps I came across your earlier post about the equivalent transistors while searching on google for info before posting on this site - kept ending up here! So...I've been comparing transistors and I appear to be able to get hold of the MJ21193/4 but they are out of stock...the MJ21195/6 arn't available where I've looked (my normal supplier, Farnell) but there are lots of others listed for that power and voltage I'm not really sure what I'm comparing are any of the following any good and safe to use?

MJ15003/4 (20A 140V 250W)
http://www.farnell.com/datasheets/15068.pdf

MJ15023/4/5/6 (16A 200/250V 250W)
http://www.farnell.com/datasheets/67535.pdf

I'm not sure on the use of the inductors in parallel with the resistors between the driver sections but I first thought they were resistors look identical not sure how to read the colour code...are they acting as a low pass filter?

There are 4 driver transistors, all which need to be kept as Krell? I need to properly test the transistors to get 2 good NPN's and 2 good PNP's but I've got a suspicion that there is only 1 good NPN of the original set, the other 6 NPN's are dead and all the PNP's are OK - fault caused the output to fail catastophically on one side :hot:

The hex key is 3.5mm and there are 3 other sizes of hex inside, all the internal fixings are hex and all the external bolts for cover, etc. are torx drive. I'm shortly hoping to get a better schematic done of the output stage and possibly also the driver and psu if time allows and post them here with photos etc. of the amp but my digicam is currently OOC.

Yes I think I read the same thing as you there when originally searching for the krell transistors to try and find equivalents - that's a lot of transistors wonder if they've got any for sale ;) ha doubt it but trying to think of a use for all those PNP's though....?

Thanks for the info both .....hope to do some more to the amp tomorrow if time allows I'll post an update then.

:)
 
Jim:

Do not use the 15003/4's they are 140vce devices and will fail. They have a much smaller die size than a 21195/96's. Don't use the 15025/5's either, they don't sound as nice as the newer devices... though they will work... to me they sound steely... thats why i preferred the sound of the KAV-250 to the KSA-250 which uses 150024/5.

My guess is the inductor filters out US frequencies and stops the OP stage from oscillating...

Tell you what, you post the schematics (or email to me) and I send you 5 pairs of 21195/6's for free (you just pay postage).

:)
 
schematic update

Hi I have had time today to do a better diagram thanks for your further tips and the offer of the transistors how much would it cost to send them to me in England? However I'm not sure on my ability to accurately draw out the driver circuit it's complicated! I'll try though...might have to get the MJ15024/5 instead if you think they'll work - difference in sound quality.....? I will still be using the Krell driver stage and I think that they must be half decent otherwise they wouldn't of been used at all in the KSA-250 but I haven't heard it!

PS Had to scale the image to fit - max width is 1000px or you can get the full size version here http://www.digitallake.co.uk/KAV_250A_OUTPUT.gif
 

Attachments

  • kav_250a_output_2.gif
    kav_250a_output_2.gif
    65.9 KB · Views: 1,012
How do I read the inductor? It's buggin me! It's resistor sized and shaped, green with colour bands is there a chart somewhere like resistor codes (which I always remember using a very racist rhyme that I learnt at college!!!) I darn't post it here....also I'm not sure on the type of capacitor I'm looking at pictures of caps on Farnell to find out but there are 100,000's taking a while.....:eek:
 
Hi Digi,
that schematic is very legible.
Don't apologise, it's better than many we see posted, be proud.

I can't understand why the pre-driver is so big.
Why would Krell have adopted a low gain, low fT, high Cob, less linear device, operating decades below it's optimum range, when a decent 200V pre-driver device could apparently do the job better. Any thoughts?
 
Hi, K-Amps I am trying to get my digicam going for pictures this project is not going to be done soon I've like 10 others in progress (not Krell amps lol) but hope to keep this thread going in the meantime with progress updates. If you are sure that you want to do this deal on the transistors I should be able to get the schematics done within the next 3-5 days I am extremely committed with work at the moment and this (unfortunately) isn't work....

Andrew, the first sketch I did was terrible that's what I was referring to it was done very quickly on the bench - I then used it to do a better one using cad. The drivers used 9710/1 I'm not sure but I've heard that Krell had a lot of these transistors they were a special order from Motorola/ON so they're probably trying to get rid of them!

Thanks for the input check back on this thead soon for more updates!

Take it e z .....:cool:
 
Take your time Jim !!!


Andrew, having owned the KSA-100mk.II, the KSA-250 and the KAV-250a, I can say this without hesitation... the KAV-250a sounds better by miles compared to the other two.

Case in point... a KAV-250a still runs in my HT... where as the other two have long gone the way of ebay.

As far as why they used em... as Jim says they had tons of them and they had to use them somewhere... plus they sound darn good... so I am not sure I will question their use as they are implemented.
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.