CubeRevo HD PSU repair

Thanks stroker once again.
Yes, those two wires are J7 and J9.
I’ve done the following several times just to make sure I had a consistent story. I unplug the output cables from the main board, and check the reference voltage on V5S and get 5.1 as before. I then check from the chassis to J7 and get 5.1 again. I then plug the cables back to the main board, thereby linking the PSU and front panel boards. No voltage between the chassis and J7. I unplug the cables from the main board, and now I still have no voltage. After a while, though I do – perhaps after I measure across one of the GND terminals rather than the chassis.
After three or four goes, I disconnect the front panel board but connect the PSU to the main board. Now I do get the usual 5.1v between the chassis and J7.
Resistance between P/S and GND is about 0.8MegaOhm.
Resistance across C33 is about 1.8 kOhms. (I measured across the two blobs on the underside.)
Shorting J7 and J9 had no effect and I did make sure I had voltage in J7, and did pause several seconds..
Yes, I do have a mutimeter with a diode checker. And yes I do have a soldering iron but have only done crude jobs, such as loudspeaker terminals where there is no chance of the flux spreading to anything important! And I do not have any desoldering kit.
I live in ******, but I’m from Pembrokeshire so go through Newport several times a year – next time probably late January. But really stroker, I’ve put you to far too much trouble already.

In case you were wondering, I didn't type "I live in ******", but gave the name of a county west of London. Surprised it fell foul of the ****ography police or whoever!

OK, and neither did I type ****ography!

Quote:

---

Originally Posted by jphilweybr

OK, and neither did I type ****ography!

---

Please read the forum rules, especially this one which refers to the board auto censor:

4) no *** out partial names of other sat retailers or @uction sites etc , ie:- do not try to cir***vent the board censor , including posting emails,
those names are starred out due to the fact we have a board sponsor here.
these censored words and emails are also censored in any pm,s so that is why they dont appear there either
( its a board auto-censor )

I'm not worried about the time it's taking I get a buzz from fixing things and like a challenge.
Your location in you message was blanked out presumably to protect your privacy.
I only asked to make sure we did not live around the corner from each other.

When you get a chance I want you to check some diodes I am hoping you can do this without having to unsolder them.
Using the diode check on the meter you should get a reading usually 0.6 and infinity with the leads swapped.

Near the board part number there are 2 diodes, a black one with a silver stripe at one end and a glass one with a stripe D91 I think it is.
So use the diode check then swap leads and make a note of the readings then do the same for the next diode.

The next check is inside PC41 which is in between the transformers.
If you look on the bottom side of the board one of the pins is marked with a dot, this is pin1.
Check for a diode between pin1 and the next pin(above) which is pin2 again reverse leads and repeat the check.
You can repeat the same check on PC91 so we can compare the results.

PC41 is my main suspect it a photo-coupler or opto-isolator which consists of an LED and a photo-transistor.
It is used to transfer signals between 2 circuits while maintaining electrical isolation.
PC41 is part of the on/off control circuit.

That's enough diode checking for now.
Now try soldering a wire link between pin3 and pin4 of PC41 these are the pins next to where it says PC41 on the board.
Don't use too much solder as the link will need to be removed later.

I am hoping the link will turn on the main part of the power supply bypassing PC41 and the previous circuitry.
Put everything back together and power on.
If it powers up the standby switch will not work but this will prove the power supply itself works and the problem is PC41 or something earlier in the circuit.
If it does not power up properly it suggests the problem is on the hot(mains) side of PC41 probably the switching device on heatsink.

It may be useful if you can check the details on PC41 and the switching device IC91 on the heatsink.
PC41 can probably be replaced by any bog-standard opto-isolator.

Thanks Also. I did read the rules when I joined and now re-read them.

I didn't actually type any asterisks to start with, but I did make a light-hearted to what I called the, and here I inserted a word meaning indecent videos etc, "police". It was the auto-sensor (I assume) that substituted asterisks for the first four letters of that word.

Anyway, thanks again - great site you help to run.

And apologies, I typed Also instead of Aldo and didn't check. Typing errors are not unknown frome, and I should have checked. Sorry.

Thanks stroker. Probably you and I have one thing in common - don't give up easily - always want to try something else.

Will do the diode checks tomorrow.

I think you are right there.
I really want to breath new life into this box.
For me replacing old with new is taking the easy way out.

OK, brief progress report, stroker.
First diodes D91 and D93 near the back.
These show 1 in one direction, which the multimeter instruction says means infinity and 483 and 502 respectively in the other, which I take to mean about 0.5v.
As regards PC41, all my pins seem to have a black background. So, looking at Picture 6 of the resource you found, I took Pin 1 to be the one at the top, near the letters PC, and Pin 2 to be the one just below it. Here I get infinity in one direction (red probe to Pin 2) and in the other direction unstable readings – they hover in the 1300 to 1500 area. Reading across Pin 3 and 4 gave similar results.
Using the same pin numbering for PC91. The top pair gives 1534 in one direction and hovering in the 1200 to 1500 area in the other. Pins 3 and 4 gave infinity in one direction and again unstable in the 1200 to 1500 area in the other.
I’ll have to break off now, probably for the rest of the day.

Sorrry stroker. Just re-read, more carefully, your previous post, where you said that on PC41 Pins 3 and 4 were next to the letters PC. As it happens, my diode check readings were the same for both pairs.

But to double check, when I look at the underside of the board, with the mains input to my right and the DC outputs to my left, the top pin is Pin 4. And it's that pin I'll connect via soldered wire to the one below it.

From your test both diodes are OK.

The pins on PC41 I want you to link are on the right as you look at it in your previous post.
To put it another way imagine a line running from top to bottom through the middle of both transformers and PC41, on the left is the low voltage "cold side" and on the right is the mains "hot" side. We want to link the 2 pins which are both on the hot side. These pins connect to the photo-transistor inside PC41 and we are simulating the transistor being turned on irrespective of the input signal to PC41. This will hopefully fire up the second switch mode circuit but all depends on how the switching device works - we may need to look up a data sheet for the device for which we will need the device type number.
Of course fitting the link may have no effect but I really think it is worth while doing.

The diode tests I wanted on PC41 were on the cold side pins. Best not to worry about pin numbers in this case. Don't worry about the diode test for now.
I now see PC41 is in a very strange package which I haven't seen for a very long time there is no pin1 dot and no semi-circular cutout at the pin1 end but there is a ridge along the one side. Looking with the ridge on your left pin 1 is at the top with the other pins numbered anticlockwise from pin1.

I hope this not too confusing - good luck!

Well yes I had managed to confuse myself. Because PC41 was written on the hot side and PC91 was on the cold side, I had assumed that all four pins on the hot side were J41 – while being baffled by what was the purpose of the links on the upper side of the board!
Anyway the diode check across the two pins of PC41 on the cold side gives infinity and 1308.
IC91 gives:
STR
G6 351
76 20

PC41 and PC91 difficult to read, and impossible to photograph. Best effort:

PC-1 7X1 not too sure about the 1, could be 3 or just a vertical line
81 7v
?? CB where ?? is some kind of symbol like an inverted U below an upright U.

Shorting the two pins on the hot side of PC41 had no effect, with or without the front panel attached. Small confession here though. I didn't solder a link. I was able to turn the board upside down and power on from there. (It’s quite stable as there’s a sizeable flat plate attached to the heatsinks on the cold side of the transformers, and I’d made a jumper to by-pass the switch at the back – keeping the wall socket at my fingertips.) I then tried linking with thin-nosed pliers with no effect. Perhaps that’s not good enough?

Friends arriving soon and staying for the weekend, so you won’t hear from me for a few days now.

Happy New Year, stroker, and hope you’ll still be ready with your excellent advice in 2017!

OK great info.

I have found a data sheet for g6351 so may be able to make more sense of things.
I probably won't post again til Sunday.

Thanks and happy new year to you too.
Remember Bob the builder - can he fix it? - yes he can!

I hope you had a good night - just had a quiet one here.
Anyway as I said I downloaded a datasheet and a sample circuit diagram.
In our case the standby on/off circuit has been added, any component references will be different but still useful info.
I tried to upload the files here but it did not work - may try again later.

Here's a few things I would like you to do.

First I want to check the resistance of one of the shorts you reported earlier which I think may be due to low value resistors in the circuit.
Measure between the heatsink of IC91 and J1 (between R95 an R96) I would expect to see about 68 or 680ohms depending on the value of R94.
If R94 is blue-grey-black it's 68 ohms if the black stripe is brown it's 680.

I now want to do some voltage measurements on the hot side, it would be helpful to have a spare hand so try to attach your black lead to to IC91 heatsink.
We will be using this point as a common but beware it is not at chassis potential (more like 300V) so should not be touched when powered up.
Now measure voltage on the end of R98 nearest the maims input, expect something up to about 35V dc.

Next measure voltage on the end of R95 nearest the mains input, expect 0V or the same as on R98.
Keep the meter on R95 and press the standby button on front-panel, does it change?
I would expect it to toggle between 0V and the same as on R98 each time you press.

Next measure the voltage on the end of R98 furthest from the mains input and repeat the standby button check.
I would expect 0V changing to about 10V but it may be difficult to measure this.

Finally but only if you have a reasonable voltage on R98 in the first measurement.
Remove mains power and use a piece of solid wire to link R95 and R98 each at the end nearest the mains input.
Just do this on the component site of the board with a couple of blobs of solder - will remove again at some stage.
This should connect power directly to IC91 so when you now power up IC91 will kick in and provide 5V and 3.3V
If it does power up the transistor next to R98 I think it's Q91 is probably faulty - see if you can ready details from it.

If it does not power on IC91 is probably faulty. I found some STRG6351's on line for about £6 inc. postage.

Thanks stroker.

I probably won't be able to get back to you today - other duties.

Well, I’ve been a bit naughty and abandoned another job I should be doing to have a look at this.

First, the resistance between the heatsink and J1 is 680 ohms, and yes R94 is blue/grey/black. I did find it difficult to get a stable reading, but when I did it was 680.

Secondly, the voltage between heatsink and the hot side of R98 is 21.1. Actually, it’s also the voltage to the cold side. This was also the voltage through R96.

Between, the heatsink and R95, the readings danced around a bit but almost settled at around 1.3. Again, the same to the other end.

Holding the meter probes steady, pressing the standby button had no effect anywhere, and nothing powered up.

The resistance of R98 is 9.94 kO, and of R95 4.62 kO.

By the way, I checked these with two different multi-meters in case one was bust.
Thought I’d better stop there to let you cogitate.

It may be no longer of interest, but Q91 reads:
A733
TCJ1D

I'm sorry I only just noticed your post.

R98 is 10K and R95 is 4K7 so your readings are ok.

From your other measurements all looks as expected apart from the pesky standby button having no effect.
Could it be the fault is actually on the front panel?
My only concern is the 21V is it high enough? But I suspect it is.
We need to get this 21V to IC91 so if you can tack a wire between the hot ends of R95 and R98 I am reasonably confident it will burst into life.
As I said you can tack this link on the top side of the board since it is only meant to be temporary.

If it does power up you will only be able to turn it off with the main switch only.

By the way if you want to have a break from this for a few days it's not a problem for me.

Thanks Stroker.
So I linked the hot end of R98 and the hot end of R95 and no sign of life.
With the link in I checked the voltages and now they were:
To each end of R98 11.7v as opposed to 21.1 previously.
To the hot end of R95 11.7 as opposed to about 1.4v previously (so makes sense?)
To the cold end of R95 4.6v as opposed to 1.4v previously.
I suppose the only sign of life I’d notice would be the fan on the main board, and that does require 12v.
No, I don’t need a break, but I may be slow to respond now and then. And don’t you feel you need to respond quickly either.

Well the results were very disappointing but hey that's life!
I had a longer look at the datasheet for IC91 and it's more complicated or clever (depending on your point of view) than I thought.
It needs a minimum of 15.8V to turn on preferably 20V and may turn off below about 11.1V so we are struggling with what we have.
It also has several shutdown modes any of which maybe the cause of our problems, eg. over-current, overvoltage.....
All of these shutdown conditions are somehow sensed by the same pin on the device - hard to get your head around!

The part of the power supply we are concerned with produces the main 5V, 3.3V and 12V hard drive supply.
I have had a concern from the first pic I saw but I dismissed it perhaps too lightly.
So I want to do a couple of checks on this to eliminate it as a cause.

Your previously measured 90ohms between 5V and ground but I cannot see a check on the 3V3.
Please measure between J15 near the long output connector and GND or chassis this should be 3V3.
Also recheck 5V to ground, I think I may have said this is on J11 previously but that is not correct.
The easiest place to check 5V is probably on the end pin of the hard drive connector - I think it is marked on the board.
If 5V and 3V3 both measure the same they may be shorted - if so check with the meter.
If they are shorted (a long shot I admit) try removing the large heatsink/plate held in place with 2 self tapping screws.
Then check is the short is gone. If it has try powering up it will be ok without the plate for a while if it does come on.

I don't suppose we could be that lucky so perhaps you can poke round IC41 which regulates the hard drive 12V supply.
It looks like it's got 4 connections one of which is ground, check for shorts to ground on the other pins but there are capacitors present which may show as a sort until they charge up.

Just looking for clues at the moment. A problem on the main board could easily cause the problem so may be worth checking again with it disconnected.