About 13 years ago, my kid had a rave party and sent square-wave like audio through my 2325. One of the right output transistors shorted and sent 60 VDC through my right JBL and blew out the woofer coil. The protection circuit - well.... didn't. The unit absolutely needs a full recap, but that is phase 2. Right now I am just trying to get the right amp module working. 13 years ago, I had ordered a bunch (a dozen each) of the output xstrs and the 747's HFE's were within a reasonable range and grouped closely. The1116A's were in two groups >1000 and <150. The also had a smaller "dome" (Cheap Chinese knockoffs?). I shelved the unit.. Now - I have replaced the TO3 2SA747A's (PNP) and 2SC1116A's (NPN) with matched pairs of these questionable (Chinese?) devices. I have replaced the three lytics and checked every smaller onboard semiconductor out of circuit. Also the power resistors are OK. I have powered the board with my current-limited bench supply that does + and - 35 VDC. Lower than the required 60 VDC but probably OK for testing. The speaker output shows 5 V DC offset which is far beyond the adjustment range, also. The voltage indication for the idle current is very high. Could these crappy output xstrs be the culprit? They test OK... Any thoughts?
Thanks, Russ KD4JO
Thanks, Russ KD4JO
Update: Even though the schematic shows the input power ground pin tied to signal ground -> it is not (ostensibly to achieve a star ground I guess), and when operating the module outside of the receiver for test, you have to jumper the power ground pin to the RCA input connector shield. Suddenly, the 5 VDC offset at the speaker pin disappeared and I was able to adjust R713 for a 0 VDC offset. I couldn't get much above a few mV with R740 using the J708/702 indication of idle current, but am not putting much stock it it since I am only powered the module with +/- 35 VDC from my bench supply.
How big is your bench supply? Perhaps you could SOA-test one of those transistors to see if they’re up to snuff?
Or measure base-emitter and base-collector capacitances from the ones in the good channel and compare your “replacements” to them?
Or measure base-emitter and base-collector capacitances from the ones in the good channel and compare your “replacements” to them?
Thanks - but I don't trust power BJT's that have such a crazy differential in HFE from sample to sample. They certainly don't meet max/min spec!!
And God knows how good the internal heat-sink substrate is!! You will never know until they get HOT!!
I also don't want to pull the other side because - you know - if it ain't broke.....
I have ordered 2 sets of Motorola MJ21193G/MJ21194G's - hoping their HFE's match up!!
THD and IMD tests under load will tell...
Thanks for your help! At least I solved the DC offset question (see my intermediate post)
Russ KD4JO
And God knows how good the internal heat-sink substrate is!! You will never know until they get HOT!!
I also don't want to pull the other side because - you know - if it ain't broke.....
I have ordered 2 sets of Motorola MJ21193G/MJ21194G's - hoping their HFE's match up!!
THD and IMD tests under load will tell...
Thanks for your help! At least I solved the DC offset question (see my intermediate post)
Russ KD4JO
Older parts would have hFE’s all over the place. Jap parts were better about this, but still they weren’t as well controlled as they are now. New ON’s won’t be as tight as new Sanken’s, but at least they will work and not blow up.
Even if I wasn’t going to use those “suspect” parts in the amp I’d still be curious enough to test them. Or pop the lid on one and have a look-see.
Even if I wasn’t going to use those “suspect” parts in the amp I’d still be curious enough to test them. Or pop the lid on one and have a look-see.
Yep!! Mr. Dremel and I did some 2SC1116 can opening earlier - will post pics tomorrow!!
They are labelled as Sankens (NOT!!). How the heck do you know you are not getting a knock-off anything nowadays?
As far as testing them - my bench supply only does +/- 32V @ 5A so will have to give them the acid test in the amp later.
The new xstrs are NOS (non-"G") Motorolas (I know I said they were G's - oops) before Freescale and On-Semi - they have LEAD in them - OMG!.
Russ
They are labelled as Sankens (NOT!!). How the heck do you know you are not getting a knock-off anything nowadays?
As far as testing them - my bench supply only does +/- 32V @ 5A so will have to give them the acid test in the amp later.
The new xstrs are NOS (non-"G") Motorolas (I know I said they were G's - oops) before Freescale and On-Semi - they have LEAD in them - OMG!.
Russ
Imo just use the onsemi parts from a reputable supplier and not be so concerned with matching hfe. Give up in trusting Asian ebay sellers of such items. Once installed and biased up check the bias current balance.
A 64V 5 amp supply is plenty enough to test those transistors. My test is to rig one up as a 50-60V 4 amp current source (on a big heat sink, or with the fins in an ice bath) and see if it holds up for 10 seconds. 12 volt supply feeding the base, trim the emitter resistors to the desired current, and put a big toroidal supply on a variac until I get enough Vce. If it blows it just takes out the fuse in the variac. If you use a bench supply, it should current limit if the device goes poof.
13 years ago you might have gotten lucky. Last C1116’s I bought were back in 85 or 86, from MCM. Those were real. The amp they’re in was still running, last time it was used.
13 years ago you might have gotten lucky. Last C1116’s I bought were back in 85 or 86, from MCM. Those were real. The amp they’re in was still running, last time it was used.
Using the +/- 32 VDC current limited bench supply, each rail draws about 300 mA of current. Using it as a voltage amp with no load, I get 30 dB of gain and low THD of around .02% on my Tektronix SG505/AA501 test set. With this PS, it clips at 20 V RMS - 28 VP-P / about 50 watts RMS into 8 ohms (if I had a load connected)- sounds about right. This is with the 2SA707A/2SC1116A "questionable Chinese" BJT's. See attached teardown of 2 of the 2SC1116A's
Attachments
Those actually don’t look fake. like I said, you might have gotten lucky. But, just because the NPNs are real doesn’t guarantee the PNPs are, either. Both would need to be tested.
Testing the whole amp board unloaded won’t necessarily tell you anything. You would have gotten the same result with “2N3055-inside” fakes. Full voltage, under load, poof. Why I suggested an SOA test on just the devices themselves. A junker will blow out if subjected to 50 volts at 4 amps - pretty much instantly. Real ones will take it as long as the case can be kept cool.
Testing the whole amp board unloaded won’t necessarily tell you anything. You would have gotten the same result with “2N3055-inside” fakes. Full voltage, under load, poof. Why I suggested an SOA test on just the devices themselves. A junker will blow out if subjected to 50 volts at 4 amps - pretty much instantly. Real ones will take it as long as the case can be kept cool.
I didn't expect this no load test to tell me anything except that all the devices and circuit are operating in their linear ranges, which it does. I am just not set up right now to do a separate load test on the devices themselves - so I am going to go for it and run the amp into a non-inductive 8 or 4 ohm load (which I have), while monitoring things closely. I have replaced the 2 big 15mF 70V filter caps and all the caps on the PS Board too, BTW all the caps I have removed and replaced test good with low ESR's.
Thanks for your help - Russ
Thanks for your help - Russ