advertisement


Cursed Amp Rebuild (or NAP/NCC200 Troubleshooting)

flashgo

flashgo

DIY Practitioner
Happy Boxing Day to All!

I have been attempting on and off to rebuild a NAP140, and now I have 2 NAPA6/3 and 2 NCC200 boards that are screwed up. I admit to some errors along the way, but I don't understand how anything I've done could've fried anything on the boards. At this point, I'm at a loss. Is there a troubleshooting guide or even strategy that I should follow? For example, can I leave out parts of the circuit and test out one bit at a time? If so, how do I go about this?

I should say that the HackerCAP is putting out a solid +/– 37 VDC that is super clean. Also, I can find no spurious connections to ground or to the case.

For the NCC200 cards, I misconnected the rails at first, so one card got two minus rails and the other got two plus rails. :rolleyes: With 2 minus rails, the emitter resistors overheated. I don't understand why anything would've fried since I don't see any current paths in the circuit with 2 minus rails, but I replaced the emitter resistors and the output transistors. Now with correct rails connected, I get no bias current. Does that mean that every transistor on the board is open circuit?

The NAPA cards have outputs pinned to one rail, despite replacing the OP transistors. The only idea I've come up with is that residual voltage on the PSU caps may have fried the OP transistors when I connected the rails. Before firing up the replacement OP transistors, I tried attaching the PSU ground first and then draining the charge in the PSU with a 100R power resistor prior to attaching the rails. The voltage would drift back up after removing the resistor, so could 150mV have killed the transistors? Is there a way to drain the PSU more quickly and effectively when the power is switched off? I'm working in an old 140 case, so there's not much room for additional circuitry.

Thanks for any help!
 
There is one thing that you can do to "test out one bit at a time" and that is to check the voltage across TR5 or the Vbe capacitor BEFORE installing the power transistors. You should be able to set this to 1.6 - 1.7 V (not mV) using the trimmer on both NAPA and NCC boards.

This is also a useful check that you've got the rails correct and is also pretty much non-destructive.

I would strongly advise removing the power transistors from all four boards and doing the above test as a start point to diagnosing any problems. Whilst you have the power transistors off you can check with a meter to see if they are live or dead (plenty of guides on the Interweb). If you can set the 1.6 - 1.7 volts without them and they are OK and the rails are OK you should get sound rather than frying.

malcolm
 
hi flashgo,

The emitter resistors overheating proves there was current flowing between the rails and ground.

I would only connect one amp up at the time and use 3.15A or 5A fast blow fuses in the rail wiring. To date, I have found the fuses to prevent the transistors blowing up. For discharging the PSU caps I use a resistor (a couple of k) and a LED on flying leads with alligator clips. It is slower than a straight resistor but I like to have the indication. I connect between ground a positive, then after the LED goes out between negative to positive.

Ensure the bias pot is adjusted correctly before powering it back up. If you can get one amp powered up without blowing the fuses measure Vbe of all the transistors. They should all measure 0.5 to 0.7 volts. A bad reading indicates a problem.

I am not aware of the "correct" procedure for the NC200 amps having never seen the documentation.

regards
 
Cheers Malcolm & Greg, those are great starting points. I'll have another go and report back.
 
misterc6 said:
There is one thing that you can do to "test out one bit at a time" and that is to check the voltage across TR5 or the Vbe capacitor BEFORE installing the power transistors. You should be able to set this to 1.6 - 1.7 V (not mV) using the trimmer on both NAPA and NCC boards.
Click to expand...

OK, I removed the power transistors from the NAPA board, connected the rails, and measured volts on either side of TR5/Vbe cap. It's -35.5V on both sides and 0 mV across TR5 and the Vbe cap. With power removed, and a diode tester across TR5, I get 0.75V. Does this mean that TR4 is non-conducting, possibly because TR1 is OC as well? Any help appreciated.

Meanwhile, the hfe measurement adapter for my DVM fell apart. Like I said, "cursed"....
 
With the transistors it's not necessary to check hfe (except to match TR1 & TR2). You just need to check that they still work as transistors. Here are two guides:

http://www.hobbyprojects.com/test_and_measurement/testing_transistors.html

http://www.kpsec.freeuk.com/components/tran.htm

Assuming the rails you've attached to be +/- 37V, then it looks like some of the transistors have gone open or short circuit. Start by checking/replacing TRs 4,5,6,7, & 8.

The power resistors in the output stage are difficult to measure as they have very low resistance. Given the problems you've had and their low cost I'd advise replacing them as a matter of course.

When wiring up using the HackerCAP you should ignore the way the 140 was wired and follow this scheme from the Avondale web site:

http://picasaweb.google.com/lh/photo/tetrIz7QKvzI7rKIopc9_Q?feat=directlink

Important points to note are that, with both NAPA & NCC, both of the grounds (input and output) should be connected to the dc end of the HackerCAP and that the negative speaker terminal should go to the ac end of the HackerCAP.

Finally, here's a link to my NCC200 build notes as published on the Avondale web site:

http://avondaleaudio.com/naim-amplifiers/power-amplifiers/ncc200/

Good luck :)

malcolm
 
Thanks, Malcolm!

misterc6 said:
When wiring up using the HackerCAP you should ignore the way the 140 was wired and follow this scheme from the Avondale web site:

http://picasaweb.google.com/lh/photo/tetrIz7QKvzI7rKIopc9_Q?feat=directlink

Important points to note are that, with both NAPA & NCC, both of the grounds (input and output) should be connected to the dc end of the HackerCAP and that the negative speaker terminal should go to the ac end of the HackerCAP.
Click to expand...

I have the grounds wired as per your reference, but with one exception. I re-used the NAP140's "0V 1" line that connects adjacent to the "signal shield". This line is connected to the middle of the HackerCAP (so closer to AC end), but it has 140pf inline. I did this because I ran out of ground pins on the DC end of the HackerCAP.

Also, I have 2 x 470uF caps across the rail connections with grounds connected to the DC end of the HackerCAP.

misterc6 said:
Finally, here's a link to my NCC200 build notes as published on the Avondale web site:

http://avondaleaudio.com/naim-amplifiers/power-amplifiers/ncc200/
Click to expand...

Thanks, these have been updated since the copy I've been working from. ;)

So now I'll breadboard a little transistor tester and get to work pulling transistors. I've got spares, but I still worry that I don't understand what caused all of these boards to go south. Maybe some extra LEDs across the HackerCAP will help to drain any "potential" for future damage. :rolleyes:
 
FlashGo

Whereabouts are you based? If you get stuck and somewhere near the south coast I would be happy to help out.

Cheers


Luke
 
Thanks, Luke, but I'm across the pond and then some. ;)

Actually, it's a good educational experience, and I need to be able to troubleshoot power amps. With some HackerNAPs on the horizon, this 140 rebuild with various boards is providing some much needed perspective and respect for the dangers and pitfalls in power amps.
 
Hi Flash...Happy Christmas...if you get really stuck send them over to California and I'll rebuild them for you...in the meantime that fact that you have full rail voltage on the output could also suggest that the 1N4148 diodes in the constant current supply to the front end (D1 and D2) are shorted...don't forget to check them too.

Neil
 
Cheers, Neil, and happy hols to you too! At least on the NAPA board I'm starting with, the 1N4148s are still good. Once all the creatures stop stirring tonight, I'll start pulling transistors. TR 4, 5, & 6 are first on the block.

EDIT: TR 4, 5, & 6 all good. I'll pull TR 7 & 8 tomorrow and check them.
 
This is a nice interesting thread for us novices, I wish you luck with your endevours flash.
My only advice is that I have found limiting potential currents to be a great help with my builds / bodges.
Either limiting current with a bench top psu fitted with a current clamp or temporary series R or even series light bulbs, heater ellements etc. I uses all sorts of things to limit current to non destructive levels while bodging away, I still have to find the fault of course but limit collateral damage.
Tony
 
Well, all of TR4 through TR8 plus all of the output transistors that I've pulled are good. E.g.,



Now what?

BTW, the bias pot is about midway. In terms of tools, I have a scope and sig generator on long term loan from a fellow PFMer as well as a bench supply and a Variac. I'm not proficient, but I can make them all work. Any and all help appreciated! If worse comes to worst, I can ship the amp off to Neil, but I'd really like to sort this myself if I can.
 
As I have identical symptoms on my hackernap (no voltage across tr5, rail voltage on output) I am watching with interest ! From quick in-circuit tests I think I may have dodgy MJE15030 driver, dodgy positive rail output power transistor, suspect in4004 on positive rail, tr5 suspect. It's a bit mind-boggling to me that all your transistors check out OK, but it is clearly grossly faulty. Your IN4004 diodes OK ?

In my case my plan of attack is to take out driver and power transistors and get to 2v (what I have on my working channel, hackernap will be different from NAP/NCC I guess) across tr5 before replacing driver/power trannies. Will have some hefty resistor in the power line this time. I am no expert in power amp circuit debugging, though, so happy to find out any better ways.
 
If the board is not working but not smoking then I suggest leaving it powered up, getting a schematic and writing on it the voltage at each pin of each transistor. Take a photo of this and post it.

Paul
 
simonbrown said:
In my case my plan of attack is to take out driver and power transistors and get to 2v (what I have on my working channel, hackernap will be different from NAP/NCC I guess) across tr5 before replacing driver/power trannies.
Click to expand...

2 V on a working board is fine; setting to 1.6 to 1.7 V before installing the power devices ensures that the bias at first switch is below the desired setting giving the opportunity to gradually increase it rather than rushing to decrease it in panic. :eek:

malcolm
 
Here are the V readings:



Nothing is connected to either the input or output. The case is floating around -100mV. There are 470uF caps connected underneath to the + & – rails. The grounds from those caps are the 2 green lines coming out from under the board near the – rail input and going round to the bank of ground tabs on the HackerCAP at the top.

Current draw is around 2W.

I'm going to put the voltages on the circuit diagram and see if they make any sense.
 
Curiouser and curiouser. The 1N4004 diode between the – rail input and ground was acting flakey. Sometimes it would conduct both ways and sometimes it would behave and only forward conduct. So I replaced it, and the signal input went to zero (instead of –36.4), but the voltage across the Vbe cap was moving all over the place.

After a few minutes, the signal input went to the minus rail, and the voltage across the Vbe cap measured zero. So back to the same old problem.

[EDITed out: The one thing that's in common with all 4 of my flakey amp boards is the Vbe cap. They're all from the same lot of mil spec 47uF caps bought on eBay. They measure 47uF, but I wonder if there's an intermittent short in these caps. I replaced with a completely different cap, and I've still got the same problem.]
 
Maybe worth having a look at the resistors, some will act like an anchor preventing parts of the circuit floating up to the rails.
 
The green ground near the input should be connected to 0v. It cannot be at the negative rail by accident. I think from looking at the board that without this at 0v the current sources won't fire up and what you're seeing might be expected.

Worth investigating.

Paul
 
You must log in or register to reply here.


advertisement


Back
Top