NAD C325BEE Will not power.

[paul_g_787]

So I have replaced C23 and C24 now as I was not happy with their ESR being so high and so different from one another.
And I replaced R254 the other day.
And I also had previously replaced R45.
So these are all brand new parts now.

I have ordered some resistors to replace R219 and R230.

I also previously replaced C517, C518, C523, C524 and C444 as they all had very high ESR and C517, C518 had bulged and were reading waaay low (13.2mF and 12.7mF).


I have probed over and over again all the parts on the SMD-L module which I have removed and I cannot find any shorts on any of the transistors or the diodes.
All the other resistors on the module seem to measure just fine, as do the two electrolytic caps C203 and C204.

I am struggling to find why R254, R219 and R230 and R45 had all burned out. Has anyone got any other suggestions?
Maybe to do with the power supply capacitors I previously replaced? I am seriously guessing here. 

Something must have drawn too much current as mentioned, but what?

Or is it possible the resistors all just failed? It doesn't seem likely to me. 

With the SMD-L module removed, I just took some measurements in the 'power on' state.

I just checked all my voltage rails and they all appear OK:
+46 = 50V
-46 = -50V
+37 = 39.4V
-37 = -39.3V
+18 = +18.3V
-18 = -18.3V
+5 = 4.99V


I just measured the voltage across R45 and R46 again.
I get 2.75V (84mA / 230mW)  across R46 consistently.

However with R45 I get 3.54V (379mW), then the voltage quickly jumps to 4.0V (485mW) then the protect kicks in right away (probably because the SMD-L module is missing) and the voltage returns to 3.54V
Does this current seem high? or is this good? It is under the 500mW limit. 

[SilverSolder]

Is it possible to disconnect power to one of the two channels, and see if one or the other side will come up on its own?

[Kim Christensen]

I just measured the voltage across R45 and R46 again.
I get 2.75V (84mA / 230mW)  across R46 consistently.
However with R45 I get 3.54V (379mW), then the voltage quickly jumps to 4.0V (485mW) then the protect kicks in right away (probably because the SMD-L module is missing) and the voltage returns to 3.54V
Does this current seem high? or is this good? It is under the 500mW limit.

I would say that it's OK. It's only drawing 22mA more than the -18v rail (3.54V vs 2.75v across 33R) and there a few more circuits powered from the +18V vs the -18V rail. The protect relay being one which draws 43mA when the speakers are enabled and protect is off. This would explain the temporary bump in current after power up.

Looking at the circuit, I would estimate that the current draw on the +18V line via R254 would normally be very low (1-2mA) since it only provides base current to Q201 and some current via R29 which is 39K.
What voltage does R154 (Right channel) have across it?
ie: The only high current path to burn out R254 is C23 (which you checked and changed due to ESR) and Q201. You mentioned before that Q201 had some questionable readings on diode check. (0.2V +Base -Collector which seems faulty to me)
This is where a small bench supply would be handy. Especially dual supply one with tracking, even if it only went up to +- 35v or so. Then with a pair of +- regs on a breadboard for the +-18V, you could run the module up and check currents, and general operation without letting any more magic smoke out. Otherwise, it's a bit hit and miss with the DMM diode checks without removing components.

Or is it possible the resistors all just failed? It doesn't seem likely to me.

More likely some transistor or capacitor died and created a cascade of failures.

[paul_g_787]

I think it is safe to say that whatever took out R254 also likely took out R45 too.
And also whatever took out R219 likely also took out R230.

Would you agree with this?

Q203 is also a  bit discoloured but it is right next to R219 so it may have just been that which caused it.

Q201 reads as follows:
B+ to E- = 0.639
B+ to C- = 0.632
E+ to B- = OL
E+ to C- = 2.3V then rises to 3.0V then OL
C+ to B-= OL
C+ to E- = 2.3V then rises to 3.0V then OL

[paul_g_787]

What voltage does R154 (Right channel) have across it?

0.002V in standby and 0.008V in power on state.
so 0.2 to 0.8mA?? (over 10Ω)

This is with no input signal and the volume turned down to zero (as were all my previous measurements).

This is where a small bench supply would be handy. Especially dual supply one with tracking, even if it only went up to +- 35v or so. Then with a pair of +- regs on a breadboard for the +-18V, you could run the module up and check currents, and general operation without letting any more magic smoke out. Otherwise, it's a bit hit and miss with the DMM diode checks without removing components.

Yes it would be handy! If I manage to make any money on this amplifier I will definitely be putting it towards a bench power supply. I have looked at a few on CPC and eBay but the reviews say they don't do good current limiting below 200mA. Which is a bit naff.

More likely some transistor or capacitor died and created a cascade of failures.

I hope it is something this simple but it is proving very tricky to find.

[paul_g_787]

Just measured R146 and R246.
R246 is 1.77K (1.8K) however R146 is only 316R??
Confirmed taking the measurements with probes both ways round but the measurements are taken in circuit.

Edit: Just popped them both out of circuit and they are measuring fine on their own.

[paul_g_787]

Just took a step back and coloured all the components I have tested so far. (Don't think I have missed any).
See attached image.

I popped out R230 and it is indeed a complete open circuit!

[Kim Christensen]

Just measured R146 and R246.
R246 is 1.77K (1.8K) however R146 is only 316R??
Confirmed taking the measurements with probes both ways round but the measurements are taken in circuit.

The in-circuit measurement of R146 makes sense because of R453, R454 (headphone area) and R149 being in series/parallel with it.
ie: R453, R454 and R149 are effectively in series (220 + 100 + 68 = 388 ohms) so when you measure R146 it is like there is a 388 ohm resistor in parallel with the 1.8K (319 ohms) (R454 and R149 are connected together via ground path)

So a reading of 1.77K on R246 in-circuit points to a problem with either R452, R455, R249, or even connector/wire CB45/CZ45 or L21.

[Kim Christensen]

What voltage does R154 (Right channel) have across it?

0.002V in standby and 0.008V in power on state.
so 0.2 to 0.8mA?? (over 10Ω)
This is with no input signal and the volume turned down to zero (as were all my previous measurements).

Yea, I expected it to be quite low. The main power source for the daughter board is the +-46V rail with the +-18V just being bias and power for the very 1st stage.

I think it is safe to say that whatever took out R254 also likely took out R45 too.
And also whatever took out R219 likely also took out R230.
Would you agree with this?

That appears to be a correct line of thought. 

Q201 reads as follows:
B+ to E- = 0.639
B+ to C- = 0.632
E+ to B- = OL
E+ to C- = 2.3V then rises to 3.0V then OL
C+ to B-= OL
C+ to E- = 2.3V then rises to 3.0V then OL

That looks OK. It looks like it survived the high current though it's base that burned up R254. Hard to say 100% if it's really OK with just a diode check, but it'll have to do for now.

One other thing. Since you are powering it up with the SMD-L module removed, how does the rest of the left side amplifier behave? What is the DC voltage on L21? (The output)
I had a thought, that if the rest of the left channel circuitry is OK, then you should be able to make the left output settle at near zero volts by putting a resistor (Try 10K) between pins 3 and 6 of CZ21. This would make the protect circuit happy and kind of verify the output circuitry. Essentially, the voltage L21 should follow the voltage at CZ21-3.
note: CZ21-6 is ground and CZ21-3 is connected to the bases of Q210 & Q211.

[paul_g_787]

The in-circuit measurement of R146 makes sense because of R453, R454 (headphone area) and R149 being in series/parallel with it.
ie: R453, R454 and R149 are effectively in series (220 + 100 + 68 = 388 ohms) so when you measure R146 it is like there is a 388 ohm resistor in parallel with the 1.8K (319 ohms) (R454 and R149 are connected together via ground path)

So a reading of 1.77K on R246 in-circuit points to a problem with either R452, R455, R249, or even connector/wire CB45/CZ45 or L21.

Just checked those resistors and they are all looking good.
R452 = 217.86Ω 
R455 = 98.19Ω
R249 = 68.3Ω

Also checked the same on the Right channel which are measuring a little low.
R453 = 195.58Ω
R454 = 94.0Ω
R149 = 66.3Ω
Edit: I disconnected the headphone board (CB45/CZ45) and these then measured the same as the Left channel. I think this is due to the fact the Left channel module is removed from the main board.

And while I was there I checked:
R150 = 0.03Ω
R250 = 0.04Ω
These are measuring low I believe due to the inductors L11 and L21 being in parallel with them but they are close. I can always remove them and test out of circuit if necessary but they are not discoloured in anyway that suggests they got too hot.

R151 = 9.78Ω
R251 = 9.95Ω
These seem good to me!

[paul_g_787]

I think it is safe to say that whatever took out R254 also likely took out R45 too.
And also whatever took out R219 likely also took out R230.
Would you agree with this?

That appears to be a correct line of thought. 

OK that is good! I am learning (albeit slowly)! 

Q201 reads as follows:
B+ to E- = 0.639
B+ to C- = 0.632
E+ to B- = OL
E+ to C- = 2.3V then rises to 3.0V then OL
C+ to B-= OL
C+ to E- = 2.3V then rises to 3.0V then OL

That looks OK. It looks like it survived the high current though it's base that burned up R254. Hard to say 100% if it's really OK with just a diode check, but it'll have to do for now.

It is the best testing I can do at the moment but there is no obvious direct short circuit on Q201 which is a good sign!!

One other thing. Since you are powering it up with the SMD-L module removed, how does the rest of the left side amplifier behave? What is the DC voltage on L21? (The output)
I had a thought, that if the rest of the left channel circuitry is OK, then you should be able to make the left output settle at near zero volts by putting a resistor (Try 10K) between pins 3 and 6 of CZ21. This would make the protect circuit happy and kind of verify the output circuitry. Essentially, the voltage L21 should follow the voltage at CZ21-3.
note: CZ21-6 is ground and CZ21-3 is connected to the bases of Q210 & Q211.

That is a good idea to check! I assume I would need to wait for my replacement for R230 to be installed first before testing these things? R230 appears to feed the base of Q216 via R269.
I have 10K 0.25W and 1W resistors in my parts bin. Would 0.25W be sufficient?

[Kim Christensen]

I disconnected the headphone board (CB45/CZ45) and these then measured the same as the Left channel. I think this is due to the fact the Left channel module is removed from the main board.

Ah yes of course. Brain fart on my part. R249 wouldn't be in circuit with the Left channel module removed. 

I have 10K 0.25W and 1W resistors in my parts bin. Would 0.25W be sufficient?

0.25W will be good enough. Just kind of tack it in there with full lead length so you can easily tack another lower value in parallel if needed. I'm not sure exactly how much base current Q210/Q211 will require, but starting at 10K ensures nothing really bad can happen with this experiment. Q210/Q211's base current cancel each other out, so the resistor will only needs to pass the mismatch between the two.

[paul_g_787]

So my 82Ω 1W resistors arrived! I have installed a replacement for R230 now.

I had a thought, that if the rest of the left channel circuitry is OK, then you should be able to make the left output settle at near zero volts by putting a resistor (Try 10K) between pins 3 and 6 of CZ21. This would make the protect circuit happy and kind of verify the output circuitry. Essentially, the voltage L21 should follow the voltage at CZ21-3.
note: CZ21-6 is ground and CZ21-3 is connected to the bases of Q210 & Q211.

It did indeed. I installed a 10k resistor as you instructed and the amplifier does indeed power on normally! The protect turns off when it should and I get a green LED!
I measured the voltage across this 10k resistor and it is 0.16V.

One other thing. Since you are powering it up with the SMD-L module removed, how does the rest of the left side amplifier behave? What is the DC voltage on L21? (The output)

Now the amp is switching on normally I measured the DC voltage on both channels output. (With the 10k resistor in place of the left module).
These voltages were stable and not flickering about (after about 1 min power on time).
Left Channel = -0.1678V
Right Channel = -0.0042V

(Yes this is a negative voltage with the + probe connected to + and the - probe connected to GND)

Interesting the left reading is higher than the right but this may be due to the module being removed perhaps?

[Kim Christensen]

Interesting the left reading is higher than the right but this may be due to the module being removed perhaps?

Yes. The negative feedback path, which zeros the DC output voltage and reduces distortion, is part of the module that was removed. So it's not surprising that there is a bit of offset voltage on the output of the left channel. It's pretty good, so the output stage must be fairly well balanced.

Looking back at your picture of the SMD-L module, what caused the goo-stain that appears between Q28 and Q29? I wonder if it was electrolyte, which is conductive, and it might be from C204. Have you checked those two caps for ESR (C204 & C203) or was it just for shorts?

When you do reinstall the repaired module, and before power on, setup your meter(s) to measure the voltage drop across R230 and R254 so you can quickly turn it off again if needed.

EDIT: changed link. The photo Hansson posted was the one I was looking at originally.

[Per Hansson]

Looking back at your picture of the SMD-L module, what caused the goo-stain that appears between Q28 and Q29? I wonder if it was electrolyte, which is conductive, and it might be from C204. Have you checked those two caps for ESR (C204 & C203) or was it just for shorts?

I agree, I think it is even more visible in this picture: https://www.eevblog.com/forum/repair/nad-c325bee-will-not-power/msg3957044/#msg3957044
SMD caps exposed to this much heat and of this vintage are notorious for failing, I would just replace them now that you have the module out Paul since it is not a job for someone without proper tools (that might come after you I mean).

[paul_g_787]

Interesting the left reading is higher than the right but this may be due to the module being removed perhaps?

Yes. The negative feedback path, which zeros the DC output voltage and reduces distortion, is part of the module that was removed. So it's not surprising that there is a bit of offset voltage on the output of the left channel. It's pretty good, so the output stage must be fairly well balanced.

Oh I see! Yes that makes sense to me now. 

Looking back at your picture of the SMD-L module, what caused the goo-stain that appears between Q28 and Q29? I wonder if it was electrolyte, which is conductive, and it might be from C204. Have you checked those two caps for ESR (C204 & C203) or was it just for shorts?

I only measured them in circuit with my capacitance LCR meter and both of them were 22µF exactly.
C203 appears to have some goo too looking closer now you mention it. Perhaps this is what took out R219 and R230!?

I will remove them and check their ESR then report back.

When you do reinstall the repaired module, and before power on, setup your meter(s) to measure the voltage drop across R230 and R254 so you can quickly turn it off again if needed.

Good thinking!

[Kim Christensen]

C203 appears to have some goo too looking closer now you mention it. Perhaps this is what took out R219 and R230!?

Could have. Especially if it flowed under one of the transistors like Q29, bridging CB, and causing it to turn on more than it should. Then the circuit, by the very nature of it's design, would attempt to compensate by driving complementary pairs (Q26 & Q28) harder which would cause excessive current to flow through R219, etc.

[paul_g_787]

OK removed C204 and C204. I think they are pretty toast.

C203 measures 21.3µF 63.2Ω ESR
C204 Measures 22.2µF 45.3Ω ESR

I find it strange that they still have their capacitance despite the ESR being so high.

Now, I originally though the goo on the PCB was just flux, as the whole amplifier has crusty flux all over where they never cleaned it off during manufacturing. 

So perhaps the goo is indeed the electrolyte from these capacitors. It would make sense as the owner said that he was playing music and then it started to sound all distorted.
He described it as if the distortion slowly faded in over about 15 seconds then the red LED came on and he heard a click (the protect circuit). So maybe he was hearing the transistors driving harder as you described due to the short.

This is all starting to add up very slowly!

As for replacement capacitors, I have two options in stock:
22µF 35V NP (Non-polarised) but these are quite hefty.
or 47µF 35V polarised ones which are pretty tiny.

Would either of these types suffice for C203 and C204?
What exactly is the function of C203 and C204 in this case?

Edit: Now that C203 and C204 are removed I was able to get a good reading on R202 and R204, which I am please to say are reading just fine!

[paul_g_787]

Just removed R219 in preparation for a replacement. The bottom pad is absolutely obliterated and there is a large black smoky crater where it used to be!! Poor R219! Must have got pretty toasty before R30 blew.

I have checked for shorts from the damaged trace but it appears to be good. I should be able to join a through hole resistor between D23 and R201 around the edge of the board to make the connection.

Interestingly, NAD replied to me (after a bit of a wait) with a copy of the service manual! How helpful! Top marks on their part, only Marshall and Behringer have ever been this helpful in my experience!!
Their engineer also said that these modules are always failing on this model amplifier and they used to just swap them for a new module, unfortunately the modules are now discontinued.

[SilverSolder]

Can the components that fail be beefed up a little?

[paul_g_787]

Can the components that fail be beefed up a little?

Yes I believe so!

The two SMD resistors I have/intend to replace with 0.25W through hole ones. (Because I have these already). My theory is the larger size of these compared to the SMD ones will allow for better heat dissipation, especially as they are mounted off the edge of the board away from the other parts. This should avoid future pad destruction.

The SMD capacitors have been/will be replaced with through hole types with the legs bent in an upside down T shape to attach to the pads. This will allow them to stand away from the board and away from the direct heat of the other components.

The R45 fusible resistor and R46 also were mounted flat on the PCB, this resulted in the destruction of R45's pads. I have mounted the replacements up off the board with 1cm gap to allow the heat to dissipate better.

The fusible resistors of course have been replaced with the exact same power rating.

I am hoping this will all make this module more reliable in future.

[Kim Christensen]

Would either of these types suffice for C203 and C204?

While it's a bit of a waste of a NP capacitor, I'd go with the 22uF ones just to keep things as close to the original design. The 47uF ones will double the time that it takes on power up for the bias currents through Q26-Q29 to stabilize.

What exactly is the function of C203 and C204 in this case?

They decouple the bias circuits for the output stages on the module. ie: Q203, R201, R202, R203, & C203 set the DC bias voltage for Q26. There's flipped version of that for Q27.

Edit: Now that C203 and C204 are removed I was able to get a good reading on R202 and R204, which I am please to say are reading just fine!

Hmm... Sounds almost like C203 and C204 were leaky in more ways than one. Do you have a capacitor leakage tester? If not, you can use a small power supply or 9V battery to try and charge up C203/C204 via 100K or 1M resistor. You can then calc the leakage current by measuring the voltage remaining across the resistor after a few seconds. If you were curious.
Because a very leaky C203/C204 would increase the bias current and burn up R219, R230, etc.

[paul_g_787]

While it's a bit of a waste of a NP capacitor, I'd go with the 22uF ones just to keep things as close to the original design. The 47uF ones will double the time that it takes on power up for the bias currents through Q26-Q29 to stabilize.

I mean, I have 48 of them in stock as it was cheaper to buy 50 than two for another repair last year. I never can work this pricing out sometimes. I may as well use these rather than paying out for new parts and costing more money.

Hmm... Sounds almost like C203 and C204 were leaky in more ways than one. Do you have a capacitor leakage tester? If not, you can use a small power supply or 9V battery to try and charge up C203/C204 via 100K or 1M resistor. You can then calc the leakage current by measuring the voltage remaining across the resistor after a few seconds. If you were curious.
Because a very leaky C203/C204 would increase the bias current and burn up R219, R230, etc.

I do not have one but that is a good idea! I will certainly try that later on just to see.

In the meantime I have attached a new R219 and the two new capacitors C203 and C204.

[paul_g_787]

View from the other end.

[paul_g_787]

OK, I re-installed the module and the voltage readings across the resistors were as follows:
R254 10.56Ω 0.005V
R230 81.70Ω 0.751V

So all being good here I connected some speakers and it is working great!!!!!!

Kim Christensen thank you for all your helpful advice!!! I certainly would not have managed to repair this if it wasn't for your excellent help! 
I have certainly learned a fair bit on this project.

Please see attached an image showing the two SMD modules. On the left is, well... the left channel. You can work out the rest.

I have given the right channel SMD module a very good clean with IPA to make sure it is spotless, just in case of future right module oozing.

I followed the calibration procedures in the service manual. Both channels were way off! But this is all measuring good now.