Help me repair this HP8560E power supply

[DC1MC]

Hello everybody, I've finally got myself a SA with TG at an affordable price. The catch is that it has a "kicking" power supply and a 10.7MHz IF (if one knows what is different on 10.7 vs 10MHz IF, please kindly let me know).
The power supply when started goes on for 1.5s and off again and the cycle repeats  , during the period when it was on, the fan was starting to rotate and some control LEDs were turning on.
Even on the table, with all loads removed, including the external HI-voltage block, the PS behaves the same.
I've measured all the capacitors, including the strange 3-legged tantalums, and all are remarkably good.  Same with the semiconductors that I could measure.
Any hint from more experienced forum members, where to look, what fails often on this PS model and other things that may help, are extremely appreciated. The power supply model number is 08564-60031, and it's used in in a wide range of HP/Agilent SA.
I've attached for convenience the schematics of the PS and the service manual, maybe it will also help someone else in my situation.

 Thanks for your help,
 DC1MC

[DC1MC]

Strange thing, I had a reply and now it's gone, what had happen  ?
The advice was not bad, the transistors are OK (removing them is another pain), and when my separation trafo will arrive I'll do the measurements under power.
In the mean time all other advices are still welcome.

 Best regards,
 DC1MC
 

[SaabFAN]

You should watch the videos about similar units on "The Signal Path". There's at least one complete overhaul of a PSU in there.
The parts that fail in these PSU are almost always the power-transistors.
If they are okay, check if the gates/bases of each Transistor get rectangular waves and also check the bypass-output of the internal voltage regulators of the PSU-Controllers (if there's no datasheet available, those pins are most of the time only connected to a capacitor very close to the chip). If no PWM-Signal arrives at the MOSFETs, or the voltages at the Capacitor-Pin of the internal regulator is "strange", there's something wrong with the controller-chip. You CAN do these checks without a isolation transformer, but you have to be extra careful when doing so. If you're lucky, a mistake adds a Scope to the list of devices to repair, if unlucky...

Getting transistors out of the PSU can be a real pain. If possible, somehow unscrew them from the heatsink. That makes thing MUCH easier (the iron doesn't have to heat the heatsink too). If the Transistor is clearly dead, just clip of the leads and desolder the remains
Also if you have hot air available, a nice trick is to heat everything up to about 120°C - You can do that with a hair dryer if necessary. Makes it much easier to desolder stuff. Just don't overdo it with the hot air. Otherwise the board will delaminate and probably rip traces in the process.


About the 10.7 MHz IF: 10.7 MHz is a common IF in Superhet-Receivers. I'm guessing HP went with a common IF to use cheaper parts, or it made construction of that stage simpler.

[jfphp]

1. Some HP (US or Taiwan brands with HP PN) power supplies don't run alone if the fan is not rotating (837XXA/B). Current drain detection on the fan supply ?
2. Almost all the power supplies, like in computers, don't run without load.
First : test without power supply all the supplies lines (5V, +-12, 15...) with an ohmmeter and look for shorts : it is typical for a switching power supply to be on and off with a short on a supply line.

[DC1MC]

1. Some HP (US or Taiwan brands with HP PN) power supplies don't run alone if the fan is not rotating (837XXA/B). Current drain detection on the fan supply ?
2. Almost all the power supplies, like in computers, don't run without load.
First : test without power supply all the supplies lines (5V, +-12, 15...) with an ohmmeter and look for shorts : it is typical for a switching power supply to be on and off with a short on a supply line.

1) The full schematic it's in the first message, there is no fan check that I can see.

2) The output circuit always had some load, I'll do detailed checks in the WE, I've washed the board of a very fine pitch black dust (diesel fumes, air pollution  ), and I'll power-on the low-voltage part of the high-voltage part    to try to verify the SMPS controller and the auxiliary circuits and  if I get pulses and stuff, I'll gently add some power in the hi-voltage part, I have a PS that can output 60V and I hope to see if the whole assembly works.

 Cheers,
 DC1MC

[DC1MC]

OK, I'm baffled, after exchanging the semiconductors (transistors, diodes and all the IC except the NE5560N that didn't arrive yet) the behavior is EXACTLY the same   
That is, powered on from the mains, the kick-start circuit, tries to  well, kick-start the PS, the voltages are coming, and then are going down, no abnormal power consumption, nothing.
If I power the 14V from an outside PS, as recommended in the service manual, everything is there but no power is delivered, just some pathetically small output voltages , I'm at the end of my wits  .

So now I'm offering a BUG HUNTING BOUNTY - I'll PayPal/GitCard/Amazon WishList/whatever on the tune of max. 10 EUR for information that will lead to the bug killing, that of course does not includes the eternal gratitude .

The NE5660N controller from Phillips is way too "analogiky" to me, and in the actual schematic I have very few cliúes how the "C - Buck Regulator" block is even biased, only that the voltage on the 5V1 Zener is around 3 volts. It does send a lot of pulses on the new MOSFET but this doesn't produce too much output .

I can make measurements and post oscilloscope screen shots and also do sensible modifications tomorrow at request. Really this design is pissing me off and I've never encountered until now this controller.

Any takers are welcome, I'm on verge of throwing the junky board in the electro-garbage and implementing a new PS, if not for the CRT high voltage,.

 Waiting for your participation,
 DC1MC

P.S> If the winner is not interested in the 10EUR bounty, he can redirect it to a charity of his choice or to a forum member with less means.
Also, if anybody around Mannheim that wants to get some beer/würstchem/cola, has experience with SMPS using this chip and want to give it a try I can travel to him on this Sunday.

[DC1MC]

HEEEELP, the bloody power supply is not working, don't know what else to do  , I'll change the NE5560N but I don't understand what stops it, if someone knows what to measure please let me know, I'm willing to test anything, what a fu--ed design  .

 Depressed,
 DC1MC

[rf+tech]

Hello DC1MC,

I have very few cliúes how the "C - Buck Regulator" block is even biased, only that the voltage on the 5V1 Zener is around 3 volts. It does send a lot of pulses on the new MOSFET but this doesn't produce too much output

From the attached schematic clip, the buck stage is supplied from +14 VP. VR131 provides 5 Vdc to U131. The 6N137 datasheet recommended Vcc is 4.5 to 5.5 Volts.

With U131 Vcc of 3 Volts, correct operation cannot be guaranteed. The presence of only 3 Vdc indicates excessive voltage (I*R) drop across R132 and/or R134.

These details provide a very broad hint as to what needs to be measured, with 14 Volts from an external power supply.

RF+ Tech

[DC1MC]

And this is the problem on my side, the resistors measure OK, the semiconductors are ALL changed, including the optocoupler and Zener, and I have no clue where is this thing in contact with ground, via which strange current path.
Also when powered by the "kick-starter" the MOSFET opens and there are some pulses in the drain of something like 140V amplitude  .
Now what should I do ?
 Many thanks,
 DC1MC

[rf+tech]

Hello DC1MC,

The broad hint is to actually measure the voltage drop across both R132 and R134. Apply Ohm's Law to determine the current flow though each and post the results.

I have no clue where is this thing in contact with ground, via which strange current path.

Look at the schematic again - observe that the gate drive "ground" is floating, it is referenced to Q102 source, not chassis ground. Measurements within the buck regulator gate driver are referenced to VR131 anode/U131 pin 5/U132 pin 3/TP102/etc. Please confirm the voltage directly across VR131. If this is not 5 Volts, then the above I*R test results should point to which path to follow.

The high-current path out of the buck regulator (Q102 source) follows the heavy lines, it is quite clearly shown. On the primary (line/mains) side of T101, -Vrail (TP101) is the return - there is no chassis ground connection. Bog standard mains-powered SMPS practice.

With this knowledge, and as a test of your understanding, please tell me where the primary side current feedback signal is derived?

RF+ Tech

[DC1MC]

Hi RF+Tech, I have measured the supply voltage over the 5.1V Zenner some while ago. The buck circuit with the optically isolated pulse excitation is new to me. Naturally to have some current passing trough R132 the drain of the MOSFET has to go under 14V. In any case I will try to put the scope ground on TP102 that seem to be the reference point for this little circuit, and measure the TP103 and power for the driver chip and optocoupler.
I have made a couple of scope screenshots, first is gate-drain voltage (blue) vs the 5V1 and second is also gate-drain voltage vs the power supply of the gate driver, practically 14V after the CR131.
The thing that baffles me is that during the brief period when the PS is powered by the kick-start circuit, everything seem to be OK, if I put an external 14V it will not behave the same (tried with everything from 12 to 14V).
The current feedback seem to be in the drains of the other MOSFETS where there are those 5 x 1ohm resistors.  No strange values there and also nothing after the LM339 amplifier/comparator.
If you can make any sense of these screenshots is great, if you need more I'll gladly do them.

 Many thanks,
 DC1MC

[capt bullshot]

Your scope traces look quity unhealthy for this circuit. You wrote gate-drain voltage, I suppose you measured gate-source?
I'd say the voltage across VR131 is supposed 5V, not 3V - this might be a pointer to either excessive current consumption of U131, defective C132 or R134. Voltage across C132 is expected at ca. 13V, with some ripple due to discharge of this capacitor while the MOSFET is on.

In case you'd use TP102 as the reference point, you'd either have to float your scope or use a proper isolated scope like the Tek THS or TPS series, or a decent diff probe. Floating the DUT input with an isolation transformer and earthing TP102 through the scope may lead to various kinds of meaningless measurements, since TP102 is a high dU/dt switching node. Earthing this node applies the high dU/dt to otherwise quiet circuit parts, they might not like this treatment.

In general this circuit works by charging C132 through CR131 and R132 while the MOSFET is off (and TP102 is at -VRAIL) and having C132 supply the driver while the MOSFET is switched on (and TP102 is at +VRAIL). So did you check R132, C132, CR131, R134, C131, VR131 ?

Next thing may be: Remove the MOSFET (or at least disconnect its gate and short gate to source), so TP102 stays quiet and you can observe the waveforms at U131, Pin 6 and so on ...

[capt bullshot]

Did you check the driver waveforms at TP207 / TP203? These should drive alternating Q104 and Q103, synchronized to the buck regulator at TP202

What is happening at TP208? (Over_current)

The behaviour looks like some protection circuitry is kicking in, but I cannot locate such a circuit this in the schematic

Does the regular 14V supply (R101, CR201, C201) work?

U202A appears to do under voltage lockout, what happens at TP205?

[jfphp]

As explained in a previous post, there are two basic rules in testing a switched power supply :
1. Isolation transformer for the mains
2. Isolator for the scope for ex Tek A6902A/B. Floating the scope is dangerous and can exceed the isolation specs for the scope.

[rf+tech]

As capt bullshot points out, the scope captures show that the 5 volt supply is dropping to 2 Volts when Q102 gate is pulled high. The greater the current drawn when U131 switches, the greater the droop on the 5 Volt supply. This is why a request for measurement of the I*R drops - the scope waveforms suggest excessive current flow from the 5 Volt supply.

Naturally to have some current passing trough R132 the drain of the MOSFET has to go under 14V.

 

This statement is false. Why? Q102 is not conducting, therefore the source (that you seem to refer to as the drain) is always less than 14 Volts.

Get the datasheets for U131, VR131 and U132. Study them. All of these components must draw some current to function. There will be I*R drops created, even with Q102 disconnected. 

Moving on, current feedback derived from the voltage drop across R112,R113,R115,R116,R117 to -Vrail, is fed to comparator U202C via R212. This current return path to -Vrail depends on Q103 and Q104 alternately switching the current flow through the respective halves of T103 primary.

Loss of drive to Q103/Q104 would result in no I*R drop across the paralleled source resistors and no output from comparator U202C. With neither Q103 or Q104 conducting, there is no return path to -Vrail.

Presumably, TP207 and TP203 will not show the presence of gate drive. U207 gate driver is driven from TP211, the 80 kHz clock from comparator U202B ouput. TP204 should have a sawtooth waveform, that comparator U202B converts into a square wave.

RF+ Tech

[DC1MC]

Replaced the NE5560N, no change in behavior. 

Some more measurements:
 - TP207 & TP203, DC-DC block MOSFETS gates.
1.1 Overall signal block shape.
1.2 Zoomed in, to observe the phase.

 - TP105 & TP106, DC-DC block MOSFETs drains.
2.1 Overall signal block shape.
2.2 Zoomed in, to observe the phase.

 - TP211 & TP204, 80 KHz Buck clock and sawtooth wave.
3.1 Overall signal block shape.
3.2 Zoomed in, to observe the shape.

 - TP202 & TP102, Buck signal on the NE5560 going to the opto isolator  and Buck MOSFET source.
4.1 Signal shape.

Now what  , I'm out of ideas, why is this POS not working, please help before I chop it into small pieces.

 Depressed,
 DC1MC

P.S> sorry for the drain <-> source SNAFU !!!

[capt bullshot]

So this looks like the PWM circuitry is working.
At some point it just turns off ...
Check if 14V power is delivered from the DC/DC transformer (via R101, CR201)
Check the output voltages at the rectifier block (-6.8, +6.8, +17, -17, +34), are they ramping up like the DC/DC drain waveform? Is one of them missing or too low?

[SaabFAN]

Did you take a look at this video:?
It seems like you have a similar problem.

Also the NE5560 has a Pin called "DEMAG: OVERVOLTAGE", which is compared to a 0,6V internal reference. If the voltage on that pin is too high, it disables the outputs.
I haven't analyzed the Buck Control Circuit in detail, but I somehow got the feeling you're chasing ghosts in that part of the PSU and the actual problem is somewhere else.

[DC1MC]

It could be an issue somewhere else, I will disconnect all the secondary windings, to see if in the low-voltage part isn't some gremlin  .

Thanks for the video,
DC1MC
 
EDIT: What the rotating frak I've just seen in this video  , some really useless crap and he avoided to actually offer any kind of useful information, "...so uhmmm... yup... I have changed some components.. [puts an elco backwards]... I find some other issues as well... don't really remember what... but now it should work..."
The most useless repair video EVER, this is so typical, I've seen this in my day work, cover your incompetence in lots of words, in the end not saying anything useful or giving away any of your "precious" crap.
A web page with 3 rows of text to the point would have been 100x times more useful.


 

[DC1MC]

So this looks like the PWM circuitry is working.
At some point it just turns off ...
Check if 14V power is delivered from the DC/DC transformer (via R101, CR201)
Check the output voltages at the rectifier block (-6.8, +6.8, +17, -17, +34), are they ramping up like the DC/DC drain waveform? Is one of them missing or too low?

Hi capt b, the output voltages at the rectifiers output are lovely, the 14V power is delivered and if I try to power the 14V rail from an external power supply the buck pulses are ceasing after the trafo makes a strange high pitched "tzik" sound, I'm kind of scared to not be the trafo, then it's done.
I will try to measure more around, but I think the if you're bored tomorrow I will not refuse a visit invitation in either direction  , it seems that with whatever skills and TE I have here I'm not able to repair it .

 Cheers,
 DC1MC

[SaabFAN]

Have you had a look at the troubleshooting-guide in the Service Manual?

http://www.siliconinvestigations.com/hprep/hp8560/08560-90081-A-srv.pdf
It starts on Page 434 and describes what kind of Voltages or Waveforms should be present at the testpoints and also points at the component most likely to be defective.

[Rog520]

Have you tried isolating the functionality of the supply by disconnecting its outputs from the SA and, for example, gradually loading down the +5v output with an active load? 

[DC1MC]

The PS is staying on the table, and has no loads, I will fabricate some load, I will kick myself hard if suddenly it will start under load  .

[SaabFAN]

With no load, I would deem it quite probable that a Overvoltage-Protection kicks in.

Have you measured TP206 and TP208?
Service-manual: "If the signal at TP208 goes high before the signal at TP206 goes low, an overcurrent condition has been detected. Suspect a short in the secondary (Output Rectifier, Voltage Regulators, or another assembly)."

Edit: What just popped into my head: Were the Transistors isolated from the Heatsinks? If they were, check if that's still the case with the new components.

[SaabFAN]

I just looked at the Schematic for a bit.

I've got two more areas you should check:
- RT201 (Temperature-Sensor) - Should be 100 Ohms when cold
- U205 (5V Regulator) - Supplies a bunch of components and comparator-voltages.

There seems to be no Overvoltage-Protection, except for some Zener-Diodes. If there's something wrong with the regulation, they conduct and most likely create a overcurrent-condition. It might also be that one of these zeners is toast and starts pulling a rail low before the specified voltage is reached, also creating a overcurrent-condition.

Btw. I don't think you need to load any of the rails because of Section F (Linear Regulators). But just to be sure, connect 4 or 5 LEDs to the 5V-Rail to draw about 100mA. That should be enough load to get the regulation in line. It certainly is enough to load a PC-PSU