Omega MML200 Coutant Lambda PSU repair

[Crystaleyes]

Hi all.

Recently I picked up a vintage HP 37717B communications analyzer which uses one of these MML200 Coutant power supplies. Needless to say the PSU is not working, and surprisingly there is apparently nothing available on the web regarding repairing one of these. No stories, no schematics, zilch!
There are however many companies offering repair 'services', however I don't have the luxury of sending it off, so am intending to do the repair myself.

I admit right now that I am no expert on smps repair, although I do have a grasp on the principle of the general design:- Filter, rectification, chopper, feedback etc.

The PSU was totally dead on all the outputs, so, when I first opened this PSU up I quickly located two RIFA caps and an BTA12-700B thyristor (crowbar protection?) which were bad. After replacing these there was somewhere around +6vDC on one of the outputs. I assumed that the unit was still bad (although perhaps this was perhaps a standby voltage?). The output comes from three separate modules which are labeled:

6A - +12v, 0v and +5.2v, 0v.
12A - +12v, 0v, +S and -S
25A - +5v, +S, -S

God only knows what the +S and -S are supposed to be?

Anyway I thought it would help to see if we were getting any chopper signal on the gate of the single main 10A 900v 2SK1539 N-Channel FET.

Here, I made a schoolboy error.

Although the information is in my minds data banks, for whatever reason I didn't remember that some SMPS heat sinks are LIVE (in this case +362vdc). To compound matters, I used a damaged Chinese DMM's frequency option to check any Hz on the FET gate, whilst using the LIVE heatsink as the return.

Doh!

After a clear pop, all output voltage was lost and the FET was shorted across all three pins.

The FET is now replaced with something similar but the PSU pwm controller is not switching the FET gate.

Looking at the datasheet for the UC3844D current mode PWM controller it would appear that the 3844 output should be connected directly to the FET gate. Obviously sticking a high voltage onto the gate (as I did) is also sticking it onto the pwm output signal path, which is where I am now. Checking which components are on this trace in order to find out what has been damaged.
It doesn't help that I have no means to measure frequency here, other that a scope (which I've never used for this)...

Therefore I suppose my question here would be, "has anyone here ever repaired one of these, or even how to test one of these pwm controllers with only limited equipment?"


Many thanks in advance and loads more photos if requested

The last photo has the PWM controller circuit board and pin 3 is the output

[Crystaleyes]

So anyways, after a quick look, there is an sot-23 with the markings A6p which has shorted.

Looking online hasn't really helped as some say its a BAS16 type diode whilst I have seen say it as a PNP and another a zener.

[CapLeaker]

Take the sot23 off the board and see what the traces do. And compare them to the data sheet. For example pad #2 doesn't do anything or is shorted via a trace to pad 1, it's a diode.

If you are just short of this part and it is a diode, just put temporary a 1N4148 in there instead of the BAS16

[R_J]

If you shorted all 3 pins of the fet, you also damaged the pwm ic.
The +S and -S are the remote voltage sense lines

[Crystaleyes]

Thanks for the replies.

CapLeaker
I shall be pulling this sot-23 off as I'm determined to do my best to fix this PSU.

Pin 3 of this particular component is connected directly to the pwm controller output pin, in fact there's not actually a lot between the controller and the FET Gate. There is only one other sot-23 (markings CP) and then a 1Ω ¼W resistor connecting the Gate.
As for the first component, after looking through various datasheets I am now certain it is a BAS16 diode manufactured by Philips in Hong Kong.

R_J
Thanks for the +S and -S info..

I've actually already ordered a a few UC3844D pwm chips and K1539 Fets so let's see what happens when the BAS 16 sot-23 is removed? With a bit of luck there will still be an output on the pwm controller.

Wish me luck

[R_J]

Check if there is a source resistor between the mosfet source and ground, it will be a low value around 1 ohm or less, It may also be open.

[Crystaleyes]

There is a 10k actually a resistor from the source pin to ground although due to it being in parallel with a torroidal inductor it measures around 1 ohm across the resistor.

Back to the sot-23 question, I removed the suspect component and it is indeed a diode and is not shorted. Upon closer inspection there is a 15 ohm resistor, again connected in parallel, which confused my readings. As for testing the output pin of the pwm controller, that won be so straightforward due to it being hidden by one of the main filter caps.

On a different tip..


When I collected the unit, the PSU was in two parts, as in the mains wires weren't connected. There are two red and two black, connecting to four pins. The only reference I could find was a photo online (see attachment), and I had taken it for granted that the two blacks and two lives were at the same potentials.
After actually 'looking' at the wiring in the mains section today, it is clear that this is not the case, and it seems more than likely that some other damage was done to the input and filter section, but that is not such a major concern.

It is really late here but I'll post some more pictures of the PSU mains input wiring tomorrow

[R_J]

The 110v and the Link terminals should connect to the 110/v220v slide switch, the ac line will connect to the L~N terminals The slide switch would be closed for 110vac operation and open for 220vac operation

[Crystaleyes]

Thank you R_J. This is info I could have done with last week.


So, after taking a closer look, I can now see how it is supposed to be wired. Something I couldn't readily find when first searching for info, so hopefully this thread can serve as a reference to future techs.

It is as you mention above only there is no option for 220vac with this particular unit.

Pin 1. Earth
Pin 2. Neutral
Pin 3. Live
Pin 4. Link (connected directly to Pin 2 and wired to 115v voltage selector)
Pin 5. Link (return from voltage selector)
Pin 6. MLU - (I don't know what MLU is...? however neither are connected)
Pin 7. MLU +

I can also now see why I had trouble when I put it onto the bench as I didn't bother to bring the mains input section, only the smps section, which I connected it up incorrectly. Even though there are two red wires, one of them (Pin 5) is actually part of the Neutral line.
Without the AC input section, and with only a vague photo for reference, last week I assumed both reds in the photo were AC live and linked pin 3 to pin 5. This would explain why when I went back to the unit yesterday, there is no longer AC at the bridge rectifier.

But hey, we all live and learn, and with the unit only having been connected through a dimbulb, the damage should be minimal.

Below are some clearer photos of the AC input section, and I know I am no-one to tell Hewlett Packard how to make power supplies, although why they didn't use two wires of the same colour for the voltage selector 'Link' is beyond me...


Anyway, there is a week or more for the Fet to arrive so I'll get the SMPS input section back up and running and see where we are.

[R_J]

It is as you mention above only there is no option for 220vac with this particular unit.

Yes there is, without the LINK connected it operates on 220v With the link connected it operates on 110v That is the poupous of that slide switch in picture 5. The two wires from the 115/230 slide switch connect directly to the LINK and 110v terminals In the 115v position it shorts 110v to link, in the 230v position the 110v and link terminals are not connected together

[Crystaleyes]

Yes there is, without the LINK connected it operates on 220v

They say that every day is a school day, and indeed you have just taught me something which I didn't know.



Thank you

[Crystaleyes]

A quick update.

The parts are due to arrive in another two weeks.

In the meantime, I got the input section sorted and once again have voltage on the Fet drain, but no trigger from the PWM onto the Gate.

As mentioned the two large filter caps don't allow access to the pwm controller so I'm thinking of when I next have a moment, connecting them off of the pcb via some long wires. Seems like the best option as I write this, and I do wanna see where the circuit is failing.


C

[Crystaleyes]

Well, there has been some progress, and some hinderance.

Once again the Brazilian customs have shown themselves to be 0% for 'progress' and 1000% for keeping Brazil well and truly a third world non-entity.

The good news is the Fets arrived from abroad (as NO-ONE sells them actually 'in' Brazil).

The boring part is that the pwm controllers have been given a 100% import tax, which purely on principle shall not be getting paid.

Don't get me started on Brazilian backwardness and corruption as I'll probably get deported, however, a $1 (for 5) pwm product that CAN NOT be bought inside the cuntry for any money, receives a 100% import fine. As does the $6 postage..
Plain and really fucking dimwit simple theft.

Anyway, I'm not paying, nor shall I be informing the thieving cunts of my decision, as I hope that it helps clog up their warehose storage space.


Back to the PSU... Let's hope the problem is not in the chopper circuit, eh?


What a fully fucked-up, retarded country Brazil is...

[CapLeaker]

Well, Canada isn't much better either in this regard. Shipping costs and taxes and other regulation made a single coin cell battery worth CAD$78 (that was 5 years ago). We cannot order any Lithium batteries from Mouser, DigiKey Newark, Allied or any other outfit like that.

[Crystaleyes]

This is what happens when you have communist globalists running the show.

I feel for you in Canada. I forgot that you lot have Justin Castro fucking you over in his attemt to wipe you all out

[Crystaleyes]

Regardless of dodgy agendas, the 1539 FET is now installed.

Considering that it is somewhat of a mission to reassemble all the PSU even before mounting it into the 37717B, does anyone have any suggestions as to how the PSU can be tested beforehand?

It's late here now but I'm thinking that there could/should be a standby circuit which will only activate the pwm chip with the correct signal.


Currently doing 12 hour days so only get to the bench late at night...

[CapLeaker]

Incandescent light bulb in series trick?

[Crystaleyes]

Good idea, athough unfortunately that didn't help in this case. No voltages are present on any of the ouput pins.


A quick half-disassemble reveals that the Drain pin has voltage however nothing is triggering the Gate.

This is half what I expected due to the unit needing to be 'switched on' although perhaps it is permanently on once the mains switch is closed?

Someone with experience of these PSU's or some kind of schematic would be priceless here...


Posting some pics just coz it's possible and may help in some way

[CapLeaker]

I.read a few documentation bits and found that A: they want a minimum load and B: the PSU can be had with a “remote” turn on/off and I think yours has it.

[Crystaleyes]

Hi again.

Yeah mine has the remote turn on and off..

I've not actually had time to get back to it hese last fews days but am planning on connecting the filter caps via some long wires in order that the pwm controller chip can be accessed. At least that way, it should be possible to confirm (or not) that the chip is being powered...