SX980 with PA3004 chip PCB substitute

[Ray66]

An SX980 that I’m restoring has a problem with the protection board where the relay does not energize. I believe the problem is with the PA3004 IC so went ahead and replaced the it with the ‘substitute PCB’ designed by merlynski and alexhastings (see PA3004 chip PCB substitute). A big thanks to Alex for sending a populated board for me to try! The good news is that the protection relay now engages when the unit is powered up. Unfortunately, it engages almost immediately. There is no time delay. Below are the voltage readings I was getting initially with the defective PA3004 chip, and after replacement with the PA3004 substitute PCB:

Pin# Expected Initial with PA3004
Substitute
pin 1 +13V +12.3V +12.3
pin 2 N/C
pin 3 +0.7V -0.26V +0.79V
pin 4 ~0V ~0V +0.004V
pin 5 GND (0V) 0V 0V
pin 6 -7V -0.95V -13.5V
pin 7 0.75VAC 0.72VAC 0.53VAC
pin 8 +7.2V -0.29V +0.013V

Initially Pins 3, 6, and 8 were all way off the expected voltages which is why I believed the PA3004 was faulty. With the PA3004 substitute PCB in place pin 3 is now OK as it can feed the base of Q2 and engage the relay coil. The voltage on Pin 6 is high compared to the expected -7V (taken from the service manual schematic), but Alex suggested in his description that -7V to -13V is OK. Pin 8 continues to be off, although now different. Finally, the AC on Pin 7 dropped by 200mV with the PA3004 substitute PCB installed.

A few other comments:
- I am not the first one to work on this unit as some components appear to have been replaced.
- Q2 identified in the SM as 2SC1438 had been replaced with a ECG194. I had planned on replacing this with KSC2383 but tested the ECG194 and confirmed it was functional NPN with gain ~ 120.
- As you can see from the photos the relay cover is missing. The solder looks untouched so I’m assuming this was the original but I do plan to replace it.
- I’ve replaced all capacitors
- I checked the voltages on binding posts 1-10 of the AWM-124 protection board and all appear to be as expected.

I inspected the PA3004 substitute PCB when I received it and found that one of the SMD resistors had fallen off in transit so replaced it, but it’s possible there’s other damage that I didn’t catch. I’ll keep scratching my head but am open to any troubleshooting suggestions.

 

[petehall347]

are the capacitors original values ?

 

[Ray66]

are the capacitors original values ?

Yep. It's been noted in other posts that there are two 47uF caps shown on the schematic in parallel and connected to Pin 8 of the PA3004, but they find only one capacitor. That is the case for the one I'm working on. Probably the way it came from the factory.

 

[petehall347]

try populate the missing one . it should slow things down some .

 

[Ray66]

try populate the missing one . it should slow things down some .

I thought about that but when I say the relay clicks on immediately I mean it's essentially instantaneous. Doubling the capacitance likely won't do much. The fact that Pin8 is hovering around 0V instead of +7V suggests a bigger issue.

 

[alexhastings]

hmm... on the board, is there any way you can check continuity through R9? I believe that's the one that disappeared in transit right? It provides a path to ground for the zener diode D5, which pulls down Q9b and breaks the voltage connection to pin 3, which is what pulls pin 3 low when there's a fault... Based on our conversations so far, that is where I would start! using your multimeter in ohms mode, check the resistance from D5 (anode) and Q9b to pin 5 - if it's open instead of 150kohm, there's no electrical connection where there should be one there, and you should resolder R9. Hope this helps!

 

[Ray66]

hmm... on the board, is there any way you can check continuity through R9? I believe that's the one that disappeared in transit right? It provides a path to ground for the zener diode D5, which pulls down Q9b and breaks the voltage connection to pin 3, which is what pulls pin 3 low when there's a fault... Based on our conversations so far, that is where I would start! using your multimeter in ohms mode, check the resistance from D5 (anode) and Q9b to pin 5 - if it's open instead of 150kohm, there's no electrical connection where there should be one there, and you should resolder R9. Hope this helps!

I confirmed the connection from 150K resistor R9 to GND along with its connection to D5. While I was at it I beeped out all other connections to the GND plane. I resoldered R9 to be sure. Voltage readings remained the same, and the relay engages immediately. Well, to be exact the first time I power it up after a long break there is a minor delay (<1sec), maybe b/c all caps on the protection board have had time to discharge. The relay engages immediately on subsequent power ups. One thing I did notice is that R7 on the PA3004 substitute PCB I have is a 68K, while the schematic specifies 75K. This is the path for charging the capacitor tied to Pin 8 I believe. I'd be surprised if a 10% lower resistance would have such a drastic effect on the charge time but I don't fully understand the circuit.

 

[alexhastings]

My biggest worry is that the board is pulling pin 3 high even though pin 8 is low... that should not be happening... I wrote out the expected voltages for the circuit. Can you check the voltages around Q9 and D5 to see how they match up to these? Another thought - let's make sure we don't have current leaking through Q9 when it's not supposed to - with the system off, can you measure the resistance from Q9c to Q9e? It should be super high... anything else means the voltages is leaking through when it's not supposed to!



This is teaching me that I should probably be testing these on the bench before they go out. I didn't want to solder the pins on just in case changes need to be made to fit it, but desoldering pins is much easier than troubleshooting a board!

 

[Ray66]

Thanks for the tips Alex. I'll try and work on this today. Will likely have to attach your PA3004 substitute PCB to the protection board with a ribbon cable to allow for easier probing of the voltage nodes. Stay tuned.

 

[alexhastings]

Cool - I just got another board from this batch on a test jig here, supplied by +-12V, and 14.5VAC on Pin 7. Looks like my Q9b is a little higher - around 2.8V

 

[Ray66]

The attached image has the voltage measurements superimposed on the schematic with the expected values. As you can see several on the right hand side are way off. Something is pulling the node connected to Q7b low. I confirmed that the components with terminals connected to V+ show 12.3V, and that all with terminals connected to GND show connectivity to Pin 5.

 

[alexhastings]

hmm something crazy is going on with Q9... If there is no voltage at Q8c, I don't see any way you should be a positive voltage at Q9b... Why don't I send you a new one... I don't trust the board... would love to get it back at some point and see why it's doing what it's doing... I'll send you out another one (a known good this time)

 

[Ray66]

hmm something crazy is going on with Q9... If there is no voltage at Q8c, I don't see any way you should be a positive voltage at Q9b... Why don't I send you a new one... I don't trust the board... would love to get it back at some point and see why it's doing what it's doing... I'll send you out another one (a known good this time)

Thanks Alex. I forgot to mention that I confirmed there was no leakage E-C leakage on Q9; those terminals show as open on the DMM. That doesn't necessarily mean that device is functioning properly however. I suppose I could pull it off the board and test. BTW, Q9 is missing from your BOM Excel file. I think it's supposed to be included in the list on row 9 'Q1-2, Q4-6, Q9: NPN BJT'.

 

[alexhastings]

Good catch - you are correct - I'm changing it now.

 

[alexhastings]

I'm curious - is the zener diode in backwards on the board perhaps? The stripe on the package is HARD to see - I need a bright flashlight, and I wouldn't put it past myself to have gotten it wrong...

 

[Ray66]

I'm curious - is the zener diode in backwards on the board perhaps? The stripe on the package is HARD to see - I need a bright flashlight, and I wouldn't put it past myself to have gotten it wrong...

It looks like D5 is oriented correctly. If you look at the pic I had PM'd you earlier showing the missing R8 you can see that the cathode is connected to Q8 as per the schematic. D5 was soldered in a bit crooked but I'm pretty sure I confirmed there wasn't a bridging short hiding under the pads. I've been travelling since Friday and not home til Tuesday. I'll re-check when I get back.

 

[Ray66]

I'm back in town and spent a bit more time looking at this. I removed D5 and used an inexpensive TC1 component tester and confirmed that it was a functioning zener diode. Confirmed no bridging underneath and soldered it back in. No change in operation, relay kicks in immediately. I had tried to probe Q8 and Q9 on board but I think the tester got confused. I might try pulling each off and testing but not 100% confident in my SMD soldering to get them back on.

 

[Ray66]

A new PA3004 substitute PCB was provided by @alexhastings yesterday and installed. Good news! The protection relay now engages as expected (an approximate 5s delay after SX-980 is powered on). The attached photos show the protection board before and after the PA3004 substitute PCB was attached. Note that I had to bend the leads of C2 and C4 when I replaced those to make room for it. The 'after' photo also has the relay replaced.

 

[Ray66]

A super big thanks to Alex for his generous help with this. Guys like him are an asset to this community.

 

[Ray66]

Before I connect good speakers I figure I should check that the protect function is working as intended. I figured that the best way to test is was to push the DC offset on one of the channels using the trimmer resistor on one the power amp boards. I tried that swinging the left channel from +1V to -4V, but the relay did not turn off. My protection board had the original 2SA733 Q1 that is part of the circuit checking the offset voltage of the amp outputs. I swapped in a new KSA992F that I've read is a suitable replacement. Still can't get the relay to trip when I push the offset well away from 0V on either channel. Perhaps I am misunderstanding how the protect circuit should work. Open to suggestions on how to confirm the protect circuit is in fact protecting.