Issues with SMP (NE5561) from Oberheim Matrix-12 synthesizer (stupid me)

[bartsx]

 Hello,
I could bite myself in the behind over and over again  by accident I shorted two connection +55V and 9V AC (D6/D8) in the power supply of my vintage Oberheim Matrix 12 synthesizer.
Of course the synth stopped working and the SMP is acting strange.
It looks like the NE5561 tries to start up, I can see the 30Khz duty cycle come up on the FET (R11/Q1) but collapses again after 0.5sec and starts over, not sure what is causing this behavior and if the NE5561 is defect due the Vcc connection coming from the +55V line (which I shortened by accident to the 9V AC line).
I’m able to measure the ~30Khz frequency at Pin 5 of the NE5561
I found out that the 12V Zener diode D7  is/was shorten.
I’m also not sure why they use the middle connector 4 from the transducer to connect it back to +55 (C6/D6) and what role D7 plays in this setup and why they bring the +55V line back to Vcc of the NE5561.
Of course I’m worrying what else might have been destroyed in the synth itself due to my mistake.
Any help is highly appreciated.

[bartsx]

As far as I understood the NE5561 the internal referenz voltage is 3.75V times the gain factor determines the regulation output voltage in my case 5V.
The gain stage is configured by R13 (1.24K) and R16 (220K) which would be already a high gain factor times the referenz voltage of 3.75V would by far exceed the 5V the regulation is based upon
What do I overlook?

[rastro]

Don't bite yourself too hard - this kind of shit happens.  I'm not an expert but here are my observations and suggestions.

Pin4 Gain:
I think the voltage divider is actually R13 (1.24K) and R15 (3.5K); notice these are both 1% resistors and form a circuit from 5V to Ground. R16 looks like part of an intergrater - possibly for start up or stability. Using ohms law, the voltage at the R13/R16 junction should be around 3.5VDC with 5 volts applied.

The 55VDC:
This appears have several functions.  It appears to act as a reference between the high and low side of the SMPS.  It also provide 12VDC VCC to the NE5561 (PIN-1) since D13 zener diode shunts it to 12VDC.  I also looks like the 9VAC is referenced/riding on 12VDC again due to D7 zener diode shunting the 55VDC to ground at the center tap.

General:
An excessive load on the secondary side can cause the SMPS to start up for 1/2 second then shut down (over-current protection).  Since you noted problems with the 9VAC circuit you may want to disconnect the secondly terminals on T1 (3, 4, 10) and see if this unloads the circuit.

Questions:
1. Did you replace the shorted D7 zener diode?
2. Why where you originally troubleshooting the supply?

Regards
-rastro

[bartsx]

@rastro

Thank you very much for the Feedback and your kind words.

Pin4 Gain, I uploaded the block schema for the NE5561 and as far as I understand it Pin4 Gain is the output of an OP-Amp which is feedback using R12 to the negative input of it Pin 3 (Fb) therefore I assumed it would affect the gain?!

Concerning the +55V I was wondering why they need it to begin with, the voltage where the regulation of the NE5561 is based on comes from the +5V side (A8,B8,C2,D2,A4,B4) and Vcc should be provided by R7 and D13 12V zener diode so why bring it back from the +55V side?

I upload a screen cap from my scope showing the duty cycle on R11 and the over current voltage on R19, besides the spikes at the beginning of the cycle I would expect somehow a steady increase of the voltage until it hits the 0.5V trigger for current protection which shuts off the duty cycle on pin7, not the spikes as seen on the scope capture.
Now the voltage on R19 drops but the duty cycle is not kicking in again, which it should based on the datasheet of the NE5561.
When I look at the scope picture I assume that the “fuzzyness” of the duty cycle shortly prior to shutting of complete is the due to the fact that the over current protection is already kick in and the subsequent drop of voltage on R19 under the 0.5V threshold immediately activates the cycle again.
Why the duty cycle stops completely for the remainder until it repeats this behavior is beyond me. Having said that I also monitored the Voltage on Pin 3 (Feedback) which stays at ~5V for the whole time which would be correct as far as the regulation goes.

Yes I replaced D7 with a new 12V Zener Diode and I could measure 12V on pin4 of the transducer.

I was working on the CPU board to fix the noise issues for VCO2 when I noticed the solder blob on the connector back to the power supply. I grabbed my desoldering Station to clean it up. The whole PCB was on angle where I couldn't see that by desoldering the solder blob I accidentatly created a short between the two line and when I powered it back on it didn't boot and just one LED lit up 

[rastro]

The primary side of the SMPS for the most part appears to be functioning.  It's starting up and putting out a PWM signal and the external transistors are switching the voltage to the primary side of T1. 

SMPS typically have over current an/or over voltage protection.  Usually these are overridden/muted for a brief period at start up while the power stabilizes.  However after the timeout period it will activate a shutdown if a power error persists.  It appears this is what is happening on your system with the 0.5 second timeout.

An overload on the secondary of the transformer will get coupled through to the primary side as excessive current draw.  The secondary can be overloaded due to: Shorted coils in the transformer, failure in one of the regulators/bridge rectifiers, or one of the supplied circuits shorting. 

You can sequentially disconnect the various power supply branches on the secondary side to see if it unloads or clears the shutdown fault.  You will not be able to apply this to the +5V or 55V branch since it is required for feed back and power for the controller.  This method will not eliminate the possibility of a short in the transformer.  Since 55V was sent through one leg of the secondary and in the path of the shorted zener diode; that leg/coil carried the same current which burnt up the diode.  So a damaged transformer coil is a good candidate based on the history of the system.

-rastro

[bartsx]

You are correct that an short/overload on the secondary side would trigger the overload protection of the NE5561.
What I don't understand is the need for the +55V back to Vcc of the NE5561, shouldn't it be sufficient to operate the NE5561 using D13/R7.
Concerning the short of the secondary side if the current was to high due to the short between 9V AC and the 55V wouldn't it be more likely for the copper windings to burn through (like a fuse) instead of getting so hot that the isolation melts and causing a short?

[rastro]

What I don't understand is the need for the +55V back to Vcc of the NE5561, shouldn't it be sufficient to operate the NE5561 using D13/R7.

In my experience "most" SMPS have the primary side isolated from the secondary.  There's typically hundreds of volts on the primary and it usually does not share the same common as the secondary.  Even the reference feedback voltage is typically isolated through an opto-coupler or transformer.  This is one reason that CAUTION is recommended with the o-scope ground when probing SMPS signals, particularly on the primary (AKA high side). The primary common may be at a different level from you household ground and therefore your o-sope ground (smoke may ensue).  You may want to view eevblog video "how not to blow up you scope". 

Your power supply seems different in regards to typical SMPS I've seen.  It uses relatively low voltage (~13V) on the primary side and the sense voltage is directly fed back to the controller (no isolation).  I'm not sure what else the 55V is used for on the system but it may have something to do with ground referencing it back to the primary side???. 

Concerning the short of the secondary side if the current was to high due to the short between 9V AC and the 55V wouldn't it be more likely for the copper windings to burn through (like a fuse) instead of getting so hot that the isolation melts and causing a short?

It could go either way short/open - both can cause an imbalance.  It's also possible that it may not be significantly damaged.  That's why I suggest that you unhook all three related taps of that secondary leg to see if you can clear the apparent overload condition.

Also I would not leave the power supply run in its error cycle for extended periods - it's probably straining the system. 

-rastro

[bartsx]

I was aware of the ground issue with the scope but I watched the very informative video anyway.
You are absolutely right that this particular SMP is rather special with his low primary voltage.
Having said that I found the issue and the Matrix12 is working again.
During the whole time I had no load on the SMP which caused the regulation not to work correctly due to the huge capacitor on the 5V line (sense line for the regulation).
Then I decided to put a load on that line and guess what the whole thing came to life and is working fine no over current protection and all secondary voltages are fine even the 9V AC.
The idea with the load came after I unhooked the secondary side and did some testing with only reconnecting certain parts of it, as you suggested.
I’m still trying to make my way around this rather unusual design and keep you guys posted.
Again thank you very much for your advice and help during the last couple of days.

[rastro]

Hello bartsx,

That's really good news.  Glad to hear there is no damage past the shorted zener diode.  As you discovered most SMPS will not operated correctly without a load - I thought you had it hooked up to the instrument.  Although I suspect you could run it with just the 5V loaded since this is the voltage monitored by the controller. 

I was trying to find this site earlier to send you: http://www.repairfaq.org/sam/smpsfaq.htm
It's has some general guidelines but also discusses loading and grounding/common voltage dangers.

Out of curiosity if you go back to pin 3, what are you measuring when the system is functioning normally?  My understanding from the schematics is you should read somewhere around 3.5 volts since the 5V reference goes through the R13/R15 voltage divider. 

Regards
rastro

[bartsx]

@rastro,
Looks like you are correct as far as I remember On Pin3 I could measure 3,7V.

[rastro]

Hi bartsx,

Yes this would make sense.  In normal operation you could expect to read around 3.5V-3.7V based on the R13/R15 voltage divider and internal 3.7V internal reference on the controller chip.  The 5V reading on pin-3 (mentioned in response #3) was high probably because there was no load on the 5V supply causing it to drift a few volts higher.  If 7V is applied from the 5v power supply to the R13/R15 voltage divider it would come out to around 5V at the junction pin-3.

-rastro