Thanks for the kind words, Ken! For whatever reason I've come across more of these than any other item, by far. At least I'm quite familiar with them at this point!
Oh man... this is one of , if not THE best video I have ever viewed on troubleshooting an amp. You really have helped the learning curve here for me. Thank you Sir !! I'll be looking at more of your content ! :- ) I'll be using the old Simpson 260 I have here for double checking too !
Just had my sui 717 re capped (the dreaded corrosive sansui glue) caps alone cost me £200. All in all it cost me £550. Money well spent. She's back to her normal brilliant self. Fantastic amplifier
Yeah, the supply chain issues we're facing around the world hit the parts world too. Interestingly, I just ran a spreadsheet for the 717 and came up with close to the same cost here, about $210 US (£173).
@@raygianelli3612 worth it though. I use 4 leak sandwich 600 speakers with my sui. Beautiful beautiful sound. Garrard 401 turntable (SME arm) ortofon concord cartridge
Your last bit of this video is exactly why I have numerous transistor testers, capacitor testers and multimeters. They all test a bit differently. If the component tests good on all of them, then it can be assumed (Ass-u-me-d?) that the component is good. If it fails the test on any of them and that tester is in good working order, then the component is almost certainly bad. One tester I have that I think might have caught this is my Sencore TF46 "Super Cricket" transistor tester. It would have show weak on the display and the tone it emits would have been weak or muffled with a leaky transistor. I have caught bad transistors with it before when others have shown them to be good. Also use my Simpson 260 sometimes in the exact method as you did. That somewhat higher test voltage will sometimes show the leakage just as you demonstrated.
100% in agreement. Is the Sencore your go to tester? The vast majority of the bad transistors I find are with the Fluke in diode test mode. The Peak Atlas is used mostly for hfe matching, the curve tracer for output transistor matching. I believe that low levels of leakage are difficult for transistor testers to find, and in many circuits isn't an issue.
@@raygianelli3612 Yes, the Sencore is what I usually use as the "gold standard". If it tests good on it, then the transistor always works. Another very handy piece of equipment that I use for leakage is an Electronic Design Systems Semianalyzer model 59C. It will test, non-destructively, for leakage up to 175 volts. That device would definitively find that E-C leakage and show exactly what voltage the breakdown occurs at. That Semianalyzer is getting very hard to find and is usually very expensive, but definitely worth it in a situation such as the one you demonstrated.
I just had the exact same issue on a Sansui AU-417 bias transistor was 2sc945 . Seem to be replacing lots of 2sc945 s recently . Very strange that your curve tracer didn’t catch it , my atlas pro 75 did . Leakage C to E ! Thanks for the video 👍
Joseph, I guess it really depends on the severity of the leakage. Without using an oscilloscope one would never know there was a problem. I seem to spend an inordinate amount of time fixing troubles that no one would know existed!
Thanks Wilfredo, you made my day! I really do believe in giving back, and I've learned so much on RU-vid and other web sites that I fell compelled to try and return the favor.
Thanks. It's kind of crazy that with all the great new test equipment available, that ancient VOM was the only thing that saw the leakage. That's why I'll always have one on my bench.
Agreed on both counts! Gotta get up to see you guys, but I burned off all my vacation time at the beginning of the year in preparation for retirement. Then the economy started circling the drain and I thought better of it!
Thanks Ray. This inspires me to get my dad's old 1960's Simpson meter from his basement. It has one of those odd (27volt?) battery's. Hopefully he removed the battery and has not leaked.
I remember there were 22 volt batteries, kind of looked like a long 9 volt, with a terminal at each end, IIRC. Don't recall Simpson using them, but maybe Triplett did. But absolutely check it out. An analog meter is occasionally just the right tool for the job!
@@raygianelli3612 I will grab the meter next time I visit. It might be a Triplett, both similar looking. Has 2 or 3 he found in the dumpster when he worked at Bell Labs. Also has a radio shack analog that he bought because the 22V batteries were hard to get in the 80's. On a side note, had to check my notes from when I recapped my 717. Replaced both TR06's with KSC945C to future proof it.
@@raygianelli3612 Yeah, he has several patents to his credit, but they are property of BL. Also wrote several chapters in tech books. I remember him showing me glass negatives of IC's, this was mid 60's. These were how they designed the circuit layout using photolithography. He his almost 90, and sadly his memory going.
@@scrappy7571 That must have been an amazing place to work when he was there. Sorry to hear about his memory. That's the ultimate betrayal, when your own body/mind goes, IMHO.
Thanks for the video been wanting to go threw my Au717 love the amp but right channel at low volume had distrotion nobody in Kentucky I know of works on these vintage integrated amps 😥
Good find to check with the Simpson, but, I wonder if you used the ohms function instead of a diode test on a 9 Volt battery powered DMM, as that would have tested the junction with a comparable voltage to the Simpson ( Simpson manual says it uses 7.5V on the R X 10K range) .It would also be interesting to check it on "Ye Olde Eico 443" compared to your new replacement.
Well, after reading this I had to go back out to the garage and check this out further. My Fluke meter is a 177, and in diode mode puts out 7.3V. In resistance mode it puts out 5.5V. The Simpson puts out 9.5V in RX10K. The 443 shows the bad transistor to be indistinguishable from a good one. The Simpson was the only one to show any leakage. I don't think the amount of leakage would be an issue in other circuits, but it was in this one. And TBH, without test equipment no one would have thought there was a problem with this amp in the first place!
@@raygianelli3612 Thanks for checking! It seems the diode test mode on just about every DMM will not indicate a high resistance leakage issue, no matter what voltage it uses to test with. One thing that is funny, I looked at the manual for a Simpson 270-3, which uses 5 x 1.5V batteries, but you must have a newer (poss series 5) as that uses one 1.5V and one 9V battery. I know the Eico 443 doesn't use a high voltage when using the transistor portion of the unit, but the diode test side can go as high as 2000 volts. It definitely paid off to use a tester, even the Simpson, that pushed the transistor a bit harder to find that fault.
@@poormanselectronicsbench2021 You're making me get my steps in today... another trip to the garage showed the Simpson is a 270 Series 4. I used to have a Tektronix 575 curve tracer. Huge old boat anchor, but you had great control of all the parameters pertaining to transistor testing.
very interesting video, I was just wondering would it be possible to see that leaky transistor in circuit with an oscilloscope? My thoughts are, if we turn the amplifier up to where it clips at the output, then probe backwards through the circuit checking each transistor until we find no distortion in the signal, would that be an alternative method of diagnostics to locate the faulty leaky transistor?
Dave, one of the things that makes troubleshooting direct coupled amps difficult is that the negative feedback will reflect the issue back into the input differential pair, and you'll see it throughout the stage. This is where understanding how the amp works comes into play. A little dumb luck never hurts either. 😁
Hi Ray. I have had same transistor problem. Checks good but is bad in circuit. I have found that my curve tracer will find the bad transitor that my meter says is ok. Have you tried a curve tracer? Bob
I did. I mentioned at the end of the video that I tried the Peak Atlas DCA75, Fluke diode test on DMM and my curve tracker. I really thought the curve tracker was going to show it, but it didn't. Perhaps if the leakage was worse it would have. Thanks for bringing it up!