ebay power supply

[John S]

Will do and send a PM when I get a reply on times and dates.

[ecat]

Well, they have arrived. Impressive beasts, especially the big hole in the middle to facilitate passive cooling.

irving2008, could you possibly link me to the UCC2985 application note or spec sheet, google is not helping :(

[irving2008]

No, I cant find it again either... but if memory serves (and I'll have to look when I get home) I didnt find it before but a near match - enough to give pinout and reference circuit clues... but I dont recall offhand what it was or where I found it...


*edit* might have been a typo.... UCC2895 datasheet

[ecat]

thank you anyway.

An older part perhaps? There is certainly enough info on the, what I suspect to me similar, ucc2884, 3884.

If I can track down a comparator feeding the opto-isolators then the switching chip may be irrelevant. You mention a second opto, could this be for over-voltage?

Output -> voltage divider -> comparator -> opto... if nothing else the feedback circuit could be replaced.... heh... maybe :)

Guessing again, the comparator - in custom chip or not - is running from +5v or +12v, so there must (-ish) be a voltage divider external to the custom chip/comparator etc.

Anyway, we know these are good for 42v.
Drivers that can cope with the full 56v are not so expensive (?)
On a personal note, if the worst comes to the worst, I now have the makings of +/-56v psu @ 500w which will make for a stonking sub-base amp :)

[ecat]

Some initial investigation:

PWM controller : http://www.digchip.com/datasheets/pa...C2895N-pdf.php

The interesting parts on the surface mount board are, guessing here as I've not buzzed things out or even applied power yet!!! :
Op-Amps : lm224 equivalents
Comparators: lm239 equivalents

Now I guess the fun can begin :)

[irving2008]

well I got as far as setting up my test bed again yesterday to restart the investigation but got sidetracked (watching TV while probing 400V isnt a good idea)... This link gets you to datasheet and applications notes.

[ecat]

probing 400v at any time is not a good idea imo!

Anyway, that daughter board is a nightmare, everything is way to small and the board would be trouble to remove: I think I have the comparator outputs to the 2 optos, via 2 transistors so looks good :) The bad news is all the fun gubbins for this is behind the capacitor bank making it almost totally inaccessible :(

Undeterred, lol, I moved on to the inputs to the daughter board and here I may have hit the jackpot... maybe.. The op voltage is divided down by a factor of ten and fed to the daughter board. Yay. These dividers are on the main board and not surface mount. Yay. The bad news is there are two voltage dividers and I have no idea what the second one is for, then again I have no idea which of the 2 we need anyway.

Divider 1 feeds op div 10 (measured) to daughter board pin 12
Divider 2 feeds op div 10 (measured) to daughter board pin 8

You should be able to trace the resistors I'm talking about from these pin numbers.

If I'm right about the lm224 equivalents then pin 12 feeds to a unity gain op amp, pin 8 feeds to a different opamp that actually 'does stuff'. I've no idea what it does but it's not unity gain.

So, why two dividers and two opamps ? 1 to catch the default 42.9v and the other for the variable feature? I don't know. Do I need to change both? I don't know.

These boards were not made for easy disassembly so I've measured the divider resistors in circuit, of course the values measured this way do not match the values required for a div 10 series pair. Ho-hum. One resistor in each pair is big and easy to see - they are not quite under the heatsink and just behind the output coils, I can't see a ref number now the thing is back together :( -, so if you have reasonable colour vision perhaps you can let me know the value?

Anyway, before I risk life and limb by snatching the soldering iron from the coals (should be hot enough by now) I'd like your opinion on which of the two possible modifications would be most suitable...

1) Configure the psu so the variable range works over, say 43v to 47v (ish)
2) Configure the psu so the 'non-regulated' voltage can be set to XXv and the variable input can be ignored.


I'm off to try and find a zener...

[irving2008]

Where did you think the optos are?

Ideally would want to be able to force the voltage down further, tho that might not be viable...

Where were you going to put the zener?


Since I've fried this board anyway I think I'll pull it to bits, get easier access to the parts....

[ecat]

A couple of hi-res images if needed ...

http://www.i2net.me.uk/files/cnc/min...oard-hires.JPG
http://www.i2net.me.uk/files/cnc/min...oard-hires.JPG

Back of the board showing the 2 optos (2 x 4 pins), the track feeding pin 20 (Error Amp) on the PWM and the 4 res making up the 2 voltage dividers. (ignore the pencil marks, they may or may not be correct!)




Top of board, only 1 of the optos is visible.




Close-up of output coils, one of the two large voltage divider resistors marked (R34). The large resistor from the second divider (R11) is right next to it.



The zener is just so I can set up your test circuit... still not found one :(

My idea is to modify one or both of the voltage dividers by fitting parallel resistors, this is a nice easy modification and so long as we don't ask for the impossible the whole of the control circuit should play along with us :)

Example:
In fixed mode the op voltage is 42.9v and the voltage at the centre of each divider pair is 4.29 volts. This 4.29 is the critical voltage in fixed mode, the control circuit will do every thing it can to maintain 4.29 volts at the centre of the divider pair. If we reduce the value of the lower resistor in the divider then this resistor will drop fewer volts and the divider ratio will effectively increase. If we select a value that gives us an 11:1 divider for example, in order to maintain 4.29v the output voltage would have to be 47.19 volts. Yay, in theory :)

Reducing the value of the upper resistor in the pair reduces the divider ratio which should allow us to modify the output voltage range in variable mode.

What I don't know for certain is the value of the resistors in the pair, my colour vision is not so good so I can't read the value of R34 or R11.Once we know the value of one of the resistors we can calculate the rest :)

[irving2008]

OK, going to whip the heatsink and daughter board off this dead one so i can get a good look... be back later...