Any electronics whizzes recognise an axial diode slightly smaller than a
1N4001 bearing just the marking "4L" along with a blue cathode band? It's used in a switch mode power supply of 1993 vintage running at around 25kHz. No other clues I'm afraid.AWEM
It *might* be an PR1004L-T 400v 1A fast recovery, but I wouldn't count on it...
Is it doing anything "special"?
If you can say where it's used (input rectifier, output rectifier, etc) and the rating spec of the SMPSU, we can almost certainly give you a current part number to use as a replacement.
No. I'd look at the circuit to see what it needs.
NT
OK the full story - it's a switch mode supply that is integral with the internal low power RG221 drive card of a Mitsubishi Meldas MR-S12-100B-Z33
2kW & 1kW twin AC servo control unit from my Traub TND350G CNC lathe. The diodes in question form a full wave rectifier hung on one centre tapped winding of the smps transformer and generate two lines, a +ve and a -ve, that having gone though a choke end up as the inputs to a 7815 (+15v reg) and 7915 (-15v reg). I do not know the power consumption of these supplies but it's obviously less than the 1 amp the regs are rated at. It'll be more than a few 10's of mA as they are mounted on small upright heatsinks. These supply lines are two out of about twenty independent supplies some of which are floating relative to each other +/- 300 ish volts as they supply the high and low sides of the six arms of the two three phase H bridges in the unit. The unit failed with a 'pop' like a fuse blowing and I traced it to a blown track on the -15v supply.The fault: loading either supply with a very minimal resistive load (22k) drops the supply to less than 12v - presumably on load it needs to be 15v plus a margin of perhaps 5v across the reg. I had thought that the switching regulator (which is a Meldas special thick film hybrid DK-466) had failed, but I can load the 5v rail (which is another winding on the same transformer) quite heavily at a couple of amps and it stays within 10mV of
5v. Both 15v regulators have been removed from circuit thus removing any loading, and in any case a temporary lab supply has been substituted in their place on a low (20mA) current limit, and can raise the rail to 15v with no problems. So in my book that leaves a faulty transformer winding having one high resistance (but it measures ok at DC) or faulty rectifier diodes. I initially substituted 1N4007's but they obviously weren't fast enough as the off load voltage went down to 10 ish, I then tried the only schottky devices I had to hand - 1N5819) and again the output was pitiful. The "4L" diodes are now back in and the the off load volts being +35 and -39.I could buy a replacement drive but don't fancy the price - there is one similar here:
The diodes may be a red herring but worth a replacement I think
AWEM
For what they cost, replacement is certainly worth a try. Almost certainly there will be caps associated with these diodes, though they may be further along the path they are still worth a check as well.
There are a series of diodes like the 1n4001-7 range but with a UF prefix which are faster and suitable for SMPS eg UF4007
Bob
Yes, one of them had popped it's cork with the original fault. They've been changed.
AWEM
that might have buggered the diode.
is there any sign of overheating of the transformer..might be a shorted turn...but there are specialist LV fast diodes for SMPS available. IIRC its a stock place to put a Schottky.
So a 100V shottky might be a good starting place. get em off ebay or somewhere.,
The transformer looks fine and smells ok - I've had it out of the board, tested it's pin out using the trusty Avo so I could more easily trace the circuit, and I've Meggered for leakage between windings and it all looks ok - obviously the windings are very low resistance at DC. A shorted turn I would expect to cook up when it's running, and also load down the other windings that are working
AWEM
So 1A 40v fast diodes will do fine
NT
Thanks Lee that sounds promising, though it seems they were discontinued in
2008 ! I'm trying to find an equivalent on line but no luck so far. As to their application see my other post for the nitty gritty details.AWEM
Beware - loads of fake and substandard components on ebay. You really want somewhere like Farnell (or CPC), Digikey, etc.
As for reverse voltage, you lose half the point of a schottky diode of you go significantly higher than needed. If you have a scope, I would put it on the open-circuit secondary windings to see what the peak voltage is, and select a reverse voltage rating 20-50% higher than that.
yeah only way to test that is an L bridge and look at Q...
I think these may be UF4004.
No harm in getting a pack and replacing the lot really
Diodes are all about peak forward current peak reverse current forward voltage and recovery time. That's it really. As long as recoivery time is good, and they will handle the forward current and reverse voltage thats enough really.
Won't that give you the RMS voltage? And you want it to be able to handle the peak? In which case 20% is pushing it... but I'm software, what do I know?
Andy
A scope is one of those things that shows a wiggly line on a screen. You can see waveform and measure what the peaks come to, so the 20-50% higher that that is fine.
Not a lot about hardware by the looks of it. B-)
But hey there are loads of hardware bods who couldn't write a Hello World programe in any langauge.
With a scope you can see the peak.
Exactly.
More than enough.
MBQ
Looks like the diodes were a Red Herring anyway!!!!
Turns out (using my 'scope) that the feedback loop for the chopper chip uses the 5v rail as it's reference for it's control loop. With the 5v unloaded there was insufficient current passing though the main V-Fet switching device so the supply for the + & - 15v rails, although there, was at a high impedance. Load up the 5v rail and all of a sudden the supply to the 7815 &
7915 regulators becomes sensible. Just load the + & - 15 supply and the V-Fet drive drive stays unaltered.Should have thought of that as a lot of PC supplies are the same
AWEM
A diode needs to handle twice the peak Vout of the winding, plus a bit of headroom in case circuit conditions shift a bit at some time.
NT
I've bought loads of components off Ebay without problems. Only thing vaguely like that I did have problems with were Ni-Cads for a power tool. They were rubbish - I should have guessed by the low price. Although not so low as they should have been given the poor spec.
HomeOwnersHub website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.