NAD 218 THX amplifier Repair.

[SantaClaw]

1 - Those results from the transistor tests seem OK to me.
2 - What about the transistors that are driving them?  Take those out and test them.
3 - How do the transistors on the good channel test by comparison?
4 - Do any of the components on the bad channel get more hot/warm than on the good channel?

1. Well, that's what I thought too.

2. This will be the small transistors type C4793, (Component Q333, Q329, Q321) type A1837 (Q331 and Q335) type A1837 ?
I haven't de-soldered them yet, it's past midnight.

3. I haven't de-soldered them either, though I plan on changing them too if I have to replace any of the others.

4. Yes, the resistor that blew... strangely enough. I never got the chance to see if the transistors got any warmer, as it failed before I could do that.

I am pretty clueless. The closest to an electronics education I have was when I did a year's school to become a car mechanic..

 What I can say that it was powering a set of Dali concept 10 speakers at the time it failed (6 ohm's) this amp is rated 4 ohms min. It produced low/muted/distorted sound on the left channel when it failed, so there was some sound, but not very good sound, no bass/punch and distorted high tones. Could it be the speaker relay somehow ?

There is a diagram posted further up in the thread, I posted a link to the service manual as a pdf

[Jwillis]

B(Beta) HFE or Gain are generally the same thing when describing BJTs so the 2SD1718 transistor might have a current gain anywhere between 60 and 200. The gain of the 2SD1718-Q will be in the range from 60 to 120, 2SD1718-S ranges from 80 to 160, 2SD1718-P ranges from 100 to 200.
The 2SB1163 transistor might have a current gain anywhere between 60 and 200. The gain of the 2SB1163-Q will be in the range from 60 to 120, 2SB1163-S ranges from 80 to 160, 2SB1163-P ranges from 100 to 200.
Your test is fine .If your  tester shows the pins 1-2-3 as resistors or open then the transistor is toast.
Download the data sheets for each transistor your putting to test and compare your results.

From experience ,Ebay is not bad for electronic components from China.But here's a tip.If the seller is only selling electronic components or items that are in that general area including modules ,test equipment ,ect.Your probably Ok .
If the vendor is also selling Lingerie and other weird stuff then you probably want to avoid them.
I have purchased probably 90% of my components from Ebay have have found no problems with them.
Not to say that i havn't received fake stuff ,but believe it or not it came from western Europe.I was indeed shocked.
Distributors like  Digikey is good but finding older components or obsolete products is tough. 

[Fire Doger]

The replacement resistor was 0.5 watts as schematic says?

[SantaClaw]

The replacement resistor was 0.5 watts as schematic says?

No, it was a 5w one, i didn't have any 0.5w 82K-ohm resistors on hand... Stupid as I am, I asked a radio amateur friend, and he said it would be fine.

I did order some, and I replaced it now when I de-soldered the transistors with a 0.5 watt one.

[Fire Doger]

The replacement resistor was 0.5 watts as schematic says?

No, it was a 5w one, i didn't have any 0.5w 82K-ohm resistors on hand... Stupid as I am, I asked a radio amateur friend, and he said it would be fine.

I did order some, and I replaced it now when I de-soldered the transistors with a 0.5 watt one.

5W is more than enough, if everything was ok it would be fine as your friend said.
I thought that you maybe used a 0.25W...

*For feature reference, replacing parts with higher maximum ratings is not always "the bigger the better", for example if you replace a 100mA fuse with a 10A fuse, a working circuit will work fine but in case of fail fuse won't blow up.
Accordingly a 5W resistor on a faulty circuit will put more stress before blowing up than a 0.5W one and may damage additional components or pcb tracks etc...

[james_s]

Ebay is a crapshoot for components, I've gotten some that seem fine but quite a few others are questionable if not obviously counterfeit. Sometimes a transistor that may be close enough to work in some situations is relabeled as some other considerably harder to find/more expensive transistor. The thing is, if you lack the expertise the evaluate the performance of the part, you never really know what you're getting and it's not worth the frustration. If you're just hacking around building hobby circuits it may not matter if you get a fake component but other times it does. I saw some tests somewhere on a potentially fake linear regulator, turned out it was indeed a regulator, but it did not come close to performing as well as the one it claimed to be.

[exe]

And even the $10 "POS tester" can do a competent job of simple hFE measurement for matching.

Shouldn't matching done at least at quiescent current of the amplifier? Power transistors optimized for higher current, measuring hFE at a few milliamps is not representative, imho. I'd say take 100mA (ideally several bias points, if curve tracer is not available).

Another thing is, I noticed that transistors from the same batch don't have much spread. So, buying "millions" may not help with binning . (but, of course, I never bought myself such quantities to draw conclusions)

And how much is a good match? . Is it even THAT needed? Do these expensive amplifier have a negative feedback to compensate for mismatch?

[Richard Crowley]

Shouldn't matching done at least at quiescent current of the amplifier? Power transistors optimized for higher current, measuring hFE at a few milliamps is not representative, imho. I'd say take 100mA (ideally several bias points, if curve tracer is not available).

Another thing is, I noticed that transistors from the same batch don't have much spread. So, buying "millions" may not help with binning . (but, of course, I never bought myself such quantities to draw conclusions)

And how much is a good match? . Is it even THAT needed? Do these expensive amplifier have a negative feedback to compensate for mismatch?

We can see for ourselves how closely the  transistors are matched by the measured values. The only "outlier" in my view appears to be that #6.

The purpose of matching is to equalize the "effort" by the transistors in parallel. So that one transistor isn't passing a much different current (more current or less current) than the other three transistors.  @SantaClaw hasn't yet measured the transistors on the good channel to see what a "normal" spread should be.

[SantaClaw]

Right, would anyone want to answer my question on weather I identified the transistors that drives the main fets correctly?


The top and bottom plate of this amp is held in place by no less than 36 m3 screws... lol

also I find the schematic confusing..... It says "Right channel" and "Left channel" but on the boards itself it says the opposite, even in the schematics... Where they drunk? I've dealt with the board that is marked Left channel on the PCB

[exe]

The purpose of matching is to equalize the "effort" by the transistors in parallel. So that one transistor isn't passing a much different current (more current or less current) than the other three transistors.

Wait, that's not the matching I meant . This matching looks to be easy to me, just buy enough bjts from a single batch. Usually this is the case when ordering from large distributors.

What I meant was matching between npn and pnp transistors. If this is not required, then task is easy (assuming no adjustment of bias current is required).

[Richard Crowley]

Right, would anyone want to answer my question on weather I identified the transistors that drives the main fets correctly?

Not sure what you mean by "main fets"?  The output devices in your amplifier are ordinary BJT (bipolar junction transistors). They are not FET (Field-Effect Transistors)

On the positive (upper) half of the circuit, the driver for Q354 and Q350 is Q334.  The driver for Q338 and Q342 is Q330  And working "upstream", the driver for Q330 and Q334 is Q325. 

And you can apply the same logic to the negative (lower) half of the circuit where Q326 drives Q332 and Q336 which in turn drive Q340, Q344, Q348, and Q352.

also I find the schematic confusing..... It says "Right channel" and "Left channel" but on the boards itself it says the opposite, even in the schematics... Where they drunk? I've dealt with the board that is marked Left channel on the PCB

Don't sweat it. That is a common error in the published documentation.  Things often get re-named in the process of creating the service documentation and proof-reading is not a high priority.

[Richard Crowley]

Wait, that's not the matching I meant . This matching looks to be easy to me, just buy enough bjts from a single batch. Usually this is the case when ordering from large distributors.

What I meant was matching between npn and pnp transistors. If this is not required, then task is easy (assuming no adjustment of bias current is required).

Right. Matching the PNP and NPN are more commonly found in smaller circuits where there is just one pair of output devices. 

But note that this circuit is DC-coupled throughout the power amplifier circuit. There is a special "sense tap" at the output which goes through R430, a 2M2 resistor for the "DC Servo" circuit which does the positive (PNP) against negative (NPN) balancing.

[exe]

There is a special "sense tap" at the output which goes through R430, a 2M2 resistor for the "DC Servo" circuit which does the positive (PNP) against negative (NPN) balancing.

Oh, didn't know there is schematic:). Anyway, that's servo thing is beyond my understanding (I wasn't even able to find that dc circuit). But I also see balancing resistors (sorry if that was mentioned before). So, I think, just buying a bunch of transistors from a single supplier should be enough. May be it's worth buying a few extra just in case.

I'd also replace all transistors connected in parallel at once for better matching, even if just one failed. Unless measurements show new parts are a good match to existing.

[SantaClaw]

I will be replacing all transistors on both channels. This amp has bridge mode, and I intend to use it.. I figure mixing parts would be a bad thing?

[Richard Crowley]

I will be replacing all transistors on both channels.

Are you including the small transistors as well as the large output transistors?

With all the transistors (large and small) removed from the circuit, I would absolutely do a thorough comparison between the good channel and the broken channel.  I would print out one or two copies of the schematic diagram and go through all the resistors and measure each one (in-circuit, but without transistors).  With power removed, unplugged, and filter capacitors discharged, of course)
I would write on the schematic the actual measurement of each resistor. 

I would compare each reading on the good channel to the same resistor on the broken channel and write down the actual measurements on the schematic diagram page.  You could write the good channel values on one copy, and the broken channel values on the other copy.  Or use two different colors, etc.  There is a possibility that a resistor failed (open, shorted, or changed value) which caused the failure in the bad channel.

And with all the transistors removed, I would check each capacitor (in-circuit) with the "POS tester" to at least get a "sanity-check" good-channel vs. broken-channel comparison.  You may see things like high series or parallel resistance (or "leakage") by testing in-circuit.  But if you get the same reading on the broken channel as you measure on the good channel, that is enough for a cursory comparison.  Of course if you see a significantly different reading on the broken channel (compared to the good channel), then it is time to take that component out of circuit and test it isolated.

This is a high-power, direct-coupled circuit.  It is very easy for a small, very inexpensive component (like a small resistor or capacitor) to fail and cause the transistors around it (and all they way downstream) to fail along with it.  Or, conversely, it is possible that a transistor failed and damaged resistor(s) and capacitor(s) around it.

This amp has bridge mode, and I intend to use it.. I figure mixing parts would be a bad thing?

I would feel much more comfortable with all new transistors, are at least with all the transistors tested out of circuit and shown to be good.  For that matter, I would test the new transistors as well before installing them.

When you remove the transistors (especially on the "good channel" be sure to label each one so you know where it came from.  The Q number will identify which resistor in the circuit, and "L" or "R" to indicate which channel it came out of.

Mr. Carlson has an interesting video on the subject of troubleshooting (BrEnglish: "fault-finding") audio circuits and using some of his inexpensive DIY test gear. 

https://youtu.be/jKr3rdX9oXQ

[SantaClaw]

I already labeled all the transistors I removed and the spot it came from

I'll have the chance to do more work on this tomorrow.

The uni-t 81c is actually a decent meter for normal measuring, by that I mean when it comes to functions, the oscilloscope part sucks though..  I also have an aneng an8009.

My component tester: https://www.banggood.com/DIY-M12864-Graphics-Version-Transistor-Tester-Kit-LCR-ESR-PWM-p-986954.html
I have two, I built a kit for this case as well https://www.banggood.com/Supporting-Shell-For-DIY-M12864-Graphics-Version-Transistor-Tester-Kit-p-989949.html
As with my meter, I can double check any value with the other one.

Though with regards to the component testers, they are calibrated using a capacitor... and since both where calibrated using the same capacitor, I would think the results would be damn near identical/with the same potential faulty readings.

The amp has been without power for 6 months or more, so I assume all caps are dead, though I'll measure the main filter caps just to be sure. I am a fan of mr carlson, though most of it goes way above my head

[Jwillis]

Arbitrarily replacing transistors is a bit extreme and expensive.

On an earlier post you stated that a "Protection Resistor" keeps blowing .Which one on the schematic is that .What is the reference number?

Bad regulation can cause components to burn up . Bad rectification can cause components to burn up. Shorts can cause components to burn up.
Did you confirm that the output voltages from Main power supply are correct? The schematic shows what voltage should be present at certain points.
I find that printing out the whole schematic helps reduce confusion.

Simply checking the resistance between transistor pins can point you in the right direction .If you get zero resistance between two pins on a transistor, reverse the probes and check again .If it is still  zero ohms that transistor is Shorted . If you get several Kohms between pins and Open or several ohms  when probes are reversed that transistor is probably OK.Move on to next transistor do the same .This can be done with components on board. You do not need to remove the components for this test..

Why do we do it this way? Removing and replacing components over and over Damages the epoxy gluing the traces to the board. So you want to avoid removing components unless you absolutely need to .

We can worry about Pairing transistors or where they come from after we find out which, if any, are at fault.

[SantaClaw]

Arbitrarily replacing transistors is a bit extreme and expensive.

On an earlier post you stated that a "Protection Resistor" keeps blowing .Which one on the schematic is that .What is the reference number?

Bad regulation can cause components to burn up . Bad rectification can cause components to burn up. Shorts can cause components to burn up.
Did you confirm that the output voltages from Main power supply are correct? The schematic shows what voltage should be present at certain points.
I find that printing out the whole schematic helps reduce confusion.

Simply checking the resistance between transistor pins can point you in the right direction .If you get zero resistance between two pins on a transistor, reverse the probes and check again .If it is still  zero ohms that transistor is Shorted . If you get several Kohms between pins and Open or several ohms  when probes are reversed that transistor is probably OK.Move on to next transistor do the same .This can be done with components on board. You do not need to remove the components for this test..

Why do we do it this way? Removing and replacing components over and over Damages the epoxy gluing the traces to the board. So you want to avoid removing components unless you absolutely need to .

We can worry about Pairing transistors or where they come from after we find out which, if any, are at fault.

I was told when this resistor blew, it was proof the transistors where shorted/broken, R345 it's a 82kohm resistor

[Jwillis]

Ok . Work your way from the obvious blown resistor outward .
That resistor ties the bases of Q317 and Q319 .That's a lot of current running out of  the Base of a transistor to fry a 1/2 watt resistor . Check Q317 and Q319 for short between emitters and base and collectors and base.

Looking at the transistors 2SA1478 / 2SC3788 label reading left to right Pin 1 is Emitter Pin 2 (center pin) is Collector and right most Pin 3 is Base .
What are your readings.

Check series Diodes D301,D303,D305,D307 and series diodes D308,D311,D313,D315 with Diode Setting of DMM. Forward voltage of 1N4148 should typically between 550mV and 650mV.

[SantaClaw]

Right, all set up, I'll print out the schematics and start work tomorrow.

[SantaClaw]

Right, time to get to work.

Cleaned it a bit

[SantaClaw]

No connectivity between any pins on Q317 and Q319 other than a short blip. I made sure I had good connections

D301 .636v  - Right Channel: D302 .637v
D303 .640v  - Right Channel: D304 .637v
D305 .637v  - Right Channel: D306 .638v
D307 .640v  - Right Channel: D308 .639v

D309 .631v
D311 .634v
D312 .537v
D313 .638v
D317 .658v
D319 .661v
D321 .594v
D323 .535v
D325 .669v

So the diodes I could find, looks good ?

The small transistors:
C4793:
1 Q333 B=143 Vf = 614mV
2 Q329 B=143 Vf = 614mV
3 Q321 B=173 Vf = 605mV

A1837:
1 Q331 B=160 Vf = 603mV
2 Q335 B=188 Vf = 571mV

So far no fault,  I suppose starting to check resistors is next ?

[SantaClaw]

Any suggestions ?

[Richard Crowley]

No connectivity between any pins on Q317 and Q319 other than a short blip. I made sure I had good connections

Not sure what this means?  What does "no connectivity" mean? 
Tested how?  Out of circuit?
With what? POS tester? with a DMM? with a curve-tracer?
What does "short blip" mean?

So far no fault,  I suppose starting to check resistors is next ?

I would say that "no connectivity" and "short blip" seem to indicate that Q317 and Q319 are blown.
Unless we are misunderstanding your testing methods and results?

What do you find when the corresponding transistors (Q316, Q318) in the other (good) channel measure by the same method?
Out of circuit in all cases.

[SantaClaw]

I put my DMM meter in diode check mode, and it briefly beeped when I measured between the legs.

So I need to de-solder q317 and q319 ?