POOGED Hafler transistor blowup
I have a POOGE modified (a la Walt Jung) Hafler DH-200 that was working fine, but having reviewed the old article from Audio Amateur, I was going over it, refreshing my mind about the layout and checking connections. Based on a comment in the article, and thinking it might be a good idea, I loosened all the screws on the output transistors, dabbed a little contact cleaner on the screw thread only, and tightened them down again.
When I fired it up, only to check it (without any input or speaker loads), I heard an unusual amount of hum, then noticed that very quickly the heat sink on one side started getting very hot. I shut it off immediately, and hooked up inputs and speakers to verify. I ran it for another 15 seconds, with music, and clearly the left channel was now out (right played okay). The left had an intermittent light popping sound. Around this point the protection circuit popped in.
I shut it off and found the heat to be coming apparently from just one upper transistor on one side. Everything else seemed to stay cool (except the heatsink taking up the heat). Maybe the screw tightening thing was a foolish thing to do with 25 year old transistors? Has anyone heard of this before? (Actually, this side always ran a little hotter than the other one.)
I have built kits and changed out caps, but otherwise I am not a very experienced DIY'er. Never did transitors. I have a meter which is supposed to test transistors. I've never used it, but I thought of pulling the two MOSFETS on that side of the heat sink and checking them with the meter. The operation might kill them both, but I presume I would have to replace them as a matched pair.
--1) Okay, so first of all, does this seem like a likely scenario--that I set up a vulnerable 25 year old transistor for failure?
--2) How can I verify that this is the problem?
--3) If it is, advice about a good place in the US to get these MOSFETS for the DH-200. The numbers on these are: 571134, and under that 8044 (I noticed the numbers on all the output transistors are not exactly the same, but each pair are--maybe having to do with matched pairs? Anyway, these are the ones of concern.)
Thanks if you can help.
As I am not getting any responses, I wonder if I should have posted this to an existing thread. I'll try posting my questions to an existing Hafler DH-200 thread.
I cannot help much with your amplifier issues, but I will suggest patience. 12 hours is not long enough to wait for a reply before asking again. Remember that this is a holiday weekend, so responses will be slower than usual.
It sounds like you may have knocked something else loose or moved something into contact with something it shouldn't touch. My DH500's gate stopper resistors have long exposed leads. Check in the area of the hot transistor for a lead touching something it is not soldered to. Reheat the solder joints in the area of your efforts in case you cracked some.
Unscrewing the devices after 25 years will probably have damaged the micas.
The standard DH-200(both factory and kit model) had 2sets of Hitachi 2SJ49/2SK134 Mosfets with room for 3 sets. 2SJ50/2SK135 can be used, old Siemens BUZ900/901 or the new Magnatec BUZ900/901
New ones cost, scout another DH-200 for parts.
Do a search,must be plenty of info on DH200s here.
I agree that you first should check the mica insulator under the apparently overheating MOSFET output device.
If you had a Variac and a couple of meters you could test this channel using MOSFETs from the other channel by ramping up the voltage very slowly while monitoring the DC offset and bias current draw. But, if not done properly this could be dangerous.
I use a bona fide Hafler MOSFET checker here and can tell in a matter of seconds if a MOSFET is still good.
There are a variety of spots to check voltages and troubleshooting strategies to follow and I am sure some othes here will chime in with some of them.
Do you have the schematic and a good VOM and perhaps a Variac?
I have some extra MOSFETs here if you need them.
Thanks for the help. If the mica insulator is cracked, will I be able to see it easily by pulling that MOSFET off the heat sink? (I am trying not to do anything rash until I get a good idea of what I am doing.) It gets hot so fast on that one, I think it must be shorting.
I don't have a Variac, but I have thought of rigging a series of cords with various resistors (properly insulated) to achieve a similar effect of cutting the line voltage and raising it in stages, if necessary. I have a new auto-ranging digital multi-meter and an old tiny analog mini-meter. The digital unit also has a transistor test section. I have to know whether the transistor is NPN or PNP type, and which leads are the emitter, base and collector. These are not things I know off hand, or whether this can even test a MOSFET, but I sure enjoy learning about this stuff. (The meter has testing holes for the leads, and says: "LCD display will show the approximate hFE value at the test condition of base current 10uA and Vce 3.2V.")
I built kits years ago as a kid (mostly Dynaco, not Haflers) and have just started to work on these old amps. I own both a modded DH-200, and an unmodded one that I would like to POOGE as well. Since I can see what I am doing by direct comparison, as well as following the instructions, I thought this would be a good way for me to get going again, and then maybe eventually be able to build an amp from a design on this forum. But the transistor blowout threw me for a loop!
At this point I am concerned about turning the amp on again without removing that MOSFET. I've considered changing out the complete heat sink with one from the other amp--but of course that's a risk without checking first.
There is one other thing I did that I am wondering about, when I checked over the amp. I replaced the old twin-lead line cord with a grounded one, and I attached the green ground lead of the new line cord to the amp chassis. This was probably coincidental, and I doubted this would cause a problem on one channel (and I actually disconnected it again before turning the amp on the second time--MOSFET still got hot).
Again, I really appreciate the help, and am particularly glad to hear from Dick West, who I see from reading the forums, has a lot of experience with these old Haflers. Yes, I might need a couple of those MOSFETS! I do have the original kit building instructions printed out, which has a schematic, such as it is. But I am kind of in second grade still on reading schematics--willing to learn, though. I also have a copy of the Audio Amateur POOGE instructions.
Given my limitations, as well as what I have on hand, what would you suggest?
There are so many possible variables it is difficult to give too much advice. But, here is a list of steps I might follow.
1. Get a Variac. You can pick up one with a 5 amp capacity off an eBay auction for around $50. It is an excellent investment for a POOGER.
2. Study the schematic for the DH-200 and the DH-220. Learn the difference between NPN and PNP transistors. Also, learn the difference between N-channel and P-channel FETs (output MOSFETs).
3. Study the voltage chart in the DH-220 manual and get a feel for which points should have the same voltages in the DH-200 circuit.
4. MOSFETs are very rugged but I believe the 2SJ49s are more prone to failure than the 2SK134 devices. Did you perchance switch them around when you were messing with that channel?
5. Capacitor failure is frequent given the age of your amp. The two filter caps on the PCB probably should be replaced or at least the power supply rails voltages checked carefully.
6. You can measure many voltages on the bad channel AFTER removing the MOSFETs. This way there is less chance of a problem creating more problems. This is a circuit somewhat similar to an op-amp and feedback from the MOSFETs is required to keep some voltages settled and in spec, but at least you can get a ball park feel for how the circuit is operating.
7. You can check the viability of the suspect mica insulator with your VOM set to measure continuity. Obviously, there should be no continuity between the MOSFET case and the underlying heat sink.
8. What is the serial number of your amp? Which version of PCB does it have? Personally, I would not mess with the older phenolic PCBs. Their traces were glued on so too much soldering heat will dislodge them. Also, the driver transistors on the PCB (the ones with the small metal tops) may have cold soldered joints because the phenolic does not allow a hot enough solder bath in manufacture. If you have phenolic you should reflow all solder joints.
9. There are many voltage points to check but you can't do this if a MOSFET is overheating so you have a catch 22 except to remove the MOSFETs and do some sleuthing with the PCB by itself, with a Variac to carefully control voltage applied.
Good luck. . . BTW, I gave up POOGing 15 years ago. My favorite amp here has a pair of PCBs from Musical Concepts applied to a DH-200 chassis and transformer. The more recent PCB upgrades from Musical Concepts make for a stellar performer far superior to the the DH-200 or DH-220! If you have good soldering skills you can install the kit from MC.
My $0.02 ;)
A couple things may have happened here. The hot mosfet was probably okay, the cold one(s) on the other polarity would be the suspects here.
You may have over tightened the device and cracked the die inside and / or cut the insulator causing a short. It's possible the insulator for the screw is missing or damaged, so the screw may short against the heatsink.
The capacitors are suspect as mentioned.
Whenever you disturb a power transistor, it's a good idea to clean everything up and use fresh grease and insulators. Take care not to over tighten the parts when you reinstall them.
Now, if you said plastic TO247/264 casings, i'm great at cracking those.
|All times are GMT. The time now is 07:19 AM.|
vBulletin Optimisation provided by vB Optimise (Pro) - vBulletin Mods & Addons Copyright © 2014 DragonByte Technologies Ltd.
Copyright ©1999-2014 diyAudio