Help!!! Some good days... some... not so good..

Linear and non linear

Help!!! Some good days... some... not so good..

Postby chrismb » Sun Oct 02, 2011 1:24 pm

Not sure what happened....

...but with HV applied to the chamber, all of a sudden the cooling fan to the turbo wound itself up to jet-turbine speed, just before the power supply to it packed up.

I figured that meant the power supply had gone mental. From the same output, though, were connected both of my high vacuum gauges, my PKR251 and AIM-S. U-oh..I feared the worst - if these babies got a full wack of excess voltage then did they survive?...

PKR251 - looks dead.

AIM-S - still looking normal.

So did the PKR251 die from an HV transient in the chamber, then killed my PSU, or the other way?

I have a Farnell LT30-2. It is a classic analogue bench supply with two outputs, each with 4 x 2N3055H to regulate each supply. The other half of the PSU was supplying a CCFL feeding some big V's. That side is fine.

The duff side blew its 2A fuse. After refitting a working fuse (and removing the loads!!) there is a 48V output on the terminals that cannot be regulated by the control knobs on the front.

Well, no point crying over it. Best get on with a fix...

Not sure how to approach the PKR251. It looks quite dead, draws what looks like a near short circuit worth of current at just a couple of volts. Any thoughts welcome. I may just have to manage without it, which I can just about do.

My PSU is where I'd like to get some quick responses from. This is a basic thing, transformer cap and regulating transistors. Not one of these delicate SMPS things, it has seen biiiig transients before and never had a problem. I have dismantled it, thinking it is probably one of the transistors. Do you think I am on the right path? I can detect a short to ground, in-circuit, on the transistors of the failed side that is not present on the working side. Maybe a transistor, maybe something in the circuit further back? I have desoldered one of the transistors, that one is working. Looks to me like the 4 transistors are in full parallel, so is the best first-step to desolder all of them and check them out one by one? Or are there better ways to start the investigation?

thanks in advance...
chrismb
 
Posts: 620
Joined: Thu Aug 05, 2010 6:32 pm

Re: Help!!! Some good days... some... not so good..

Postby johnf » Sun Oct 02, 2011 3:48 pm

Chris
time to put some big transils on everything
I use 15 volt ones to protect 12V circuits they are fully on @ around 21volts
Your output transistors have probably survived. I find that it is usually the drive transistor that feeds the 2N3055's that ends its life. The 2n3055's have so much junction capacity that they let the nasty spike do its thing to the driver.
johnf
 
Posts: 433
Joined: Sun Aug 08, 2010 3:51 pm
Location: Wellington New Zealand

Re: Help!!! Some good days... some... not so good..

Postby chrismb » Sun Oct 02, 2011 5:25 pm

D@mn.

It would've made it all a whole lot easier just to have to replace the 2n3055's.... not sure where to keep digging into it now, then.
chrismb
 
Posts: 620
Joined: Thu Aug 05, 2010 6:32 pm

Re: Help!!! Some good days... some... not so good..

Postby chrismb » Mon Oct 03, 2011 1:07 pm

OK, looks like I gave some duff info - I disconnected the main control board and still had the low resistance fault. But it was low resistance between collector and emitter (if I interpret the data sheet right - none of the makes are particularly clear which leg is which...)

I desoldered them all, and one shows a short, c to e.

The others seem to be a bit like units-10s of megs, c to e, also. Should I be changing the set? Why this happened I am still puzzled over. I guess the c/e just broke down with over-voltage fed back to it from the PKR experiencing some plasma HV?

I guess that might've weakened/damaged any one of the transistors, so I suppose I should change them all?

So, what's the difference between 2n3055, 2n3055E, 2n3055G and 2n3055H?

..I'm debating whether I should/could replace with MJ15024 as an upgrade (higher CE volts, higher power handling, better frequency response)? Any other suggestions for upgrading the 2n3055? Alternatively, would using MJ13333 be OK? There is an offer on at Farnell at the moment, for them.
chrismb
 
Posts: 620
Joined: Thu Aug 05, 2010 6:32 pm

Re: Help!!! Some good days... some... not so good..

Postby johnf » Mon Oct 03, 2011 4:51 pm

Chris
Careful about putting higher ft devices in --it may oscillate

I just tested 20 2N3055's two were in the 4meg range +to collector minus to emitter 1 short circuit the rest better than 60meg

the shorted one was used so now is in the circular filing cabinet the two leaky ones were RCA brand ie 20 years old NOS
I suspect junction poisoning through dopant migration ie time + temp causes dopant migration

looks like you should replace all 4
johnf
 
Posts: 433
Joined: Sun Aug 08, 2010 3:51 pm
Location: Wellington New Zealand

Re: Help!!! Some good days... some... not so good..

Postby chrismb » Mon Oct 03, 2011 5:04 pm

Thanks for giving me a point of comparison, John. So if temp and age causes a degradation, it sounds like I should replace all 16!! (I have 2 of these power supplies, 2 channels of 4 paralleled in each! And they are all RCA brand too, and all 'UOS'! I can't afford to let the transistors give up voltage regulation and destroy what they are powering.)

John, the MJ13333 have a very slightly lower (like ~0.1V for 1A collector current) turn-on voltage than the 2n3055. They have a flatter, more linear response. Would I be right in saying that is the right way around to protect against inducing oscillations, or is it a more complex function of the behaviour of the driver circuit? Sorry if that is a 'beginners' question.
chrismb
 
Posts: 620
Joined: Thu Aug 05, 2010 6:32 pm

Re: Help!!! Some good days... some... not so good..

Postby Doug Coulter » Mon Oct 03, 2011 7:07 pm

I have pounds of old 3055's in stock someplace, I think. They were widely used in power supplies, for two reasons -- too slow to oscillate easily, and pretty robust safe operating area, which isn't the same as the volts and amps taken separately. I'd use the same type if at all possible, as many of the engineers of these old supplies kind of slipped one by, and as John said, putting in a purportedly "better" transistor there can lead to headaches very quickly. I think you can still even get them new, if not let me know. Mine are from the days I worked at DEC -- they're pretty old, but they never were low leakage types, that might just have been normal measurements for them. A lot of newer types didn't have the same safe operating area (combined volts and amps) and actually aren't as good in this job. Those things were really tough for power transistors "in the day" and many more modern ones with higher volts and amps ratings still don't have the SOA ratings these have. I note digi now lists these with two different Ft ratings - you almost certainly want the slower -- they were harder to melt with transients.

Screenshot-28.png
SOA for 2n3055


Any feedback loop has to be compensated for gain and phase to close the loop without oscillation. The turn on voltage and linearity don't matter (much at all). It's the speed of the transistors. The slow speed of the 2n3055 made it easy to close a loop without oscillation by the beginner engineers they usually put to power supply design, so putting in a faster, gain-ier one may make it oscillate -- more gain means the loop gain has to hit zero more quickly before the phase error builds up, and changing that would require other changes in the control loop. Also, the faster the transistor, the thinner and less robust the junctions against 2nd breakdown (oversimplification, but generally true) -- which is almost surely what happened.

Those high voltage arcs can really deliver some spikes to things in the room -- everything looks like an antenna when you have that many peak watts (or the other turn of a transformer). I've had to take "heroic" precautions in lead dress, routing, and shielding to prevent losing gear here. Even a badly routed mains power cord can pick up enough stuff to fry the contents of things, and ground loops, don't get me started.
Posting as just me, not as the forum owner. Everything I say is "in my opinion" and YMMV -- which should go for everyone without saying.
User avatar
Doug Coulter
 
Posts: 3515
Joined: Wed Jul 14, 2010 7:05 pm
Location: Floyd county, VA, USA

Re: Help!!! Some good days... some... not so good..

Postby chrismb » Tue Oct 04, 2011 1:19 am

The MJ13333 has a larger SOA:

mj13333_soa.jpg


...and is listed (variously) as 2MHz, while 2n3055 (variously) as 3MHz.

(I say variously because I have seen variations on that. I guess it is either not too critical or not well controlled).

So, slower, plus a bigger SOA too?

I will add (for information - if anyone comes looking here for LT30-2 Farnell power supply) it may be useful to know that mine is currently running 48V unregulated. So I guess that is the CE voltage to look up on the SOA diagram. As there are 4 parallel transistors, so I suppose we need a point at 48V/0.5A within the SOA.

Looks to me like any transient current draw above the regulated current setting, with that 48V on the 2n3055, is marginal.

At 48V, 2n3055 is out of its SOA at 2A drawn for 1ms, whereas 12A can be drawn through a MJ13333 for a ms.
chrismb
 
Posts: 620
Joined: Thu Aug 05, 2010 6:32 pm

Re: Help!!! Some good days... some... not so good..

Postby Doug Coulter » Tue Oct 04, 2011 9:12 am

I see 3a/40v on the 2n3055, but only 2a/40v for the MJ here. That's not bigger - it's only 2/3 as big, just to pick one point on the curve. You'd have to look at it where it actually operates, of course. Most PSU's don't have that much drop across the output transistors except in fault conditions (uh oh, that's what you had, right?). SOA limitations are what drove the concept of "foldback current limiting" in earlier designs.
Posting as just me, not as the forum owner. Everything I say is "in my opinion" and YMMV -- which should go for everyone without saying.
User avatar
Doug Coulter
 
Posts: 3515
Joined: Wed Jul 14, 2010 7:05 pm
Location: Floyd county, VA, USA

Re: Help!!! Some good days... some... not so good..

Postby chrismb » Sun Oct 16, 2011 7:21 am

I have now successfully installed the MJ13333s and they appear to be working perfectly. I have loaded them up with full volts, then low volts and high constant current. They are running slightly cooler than the 2n3055s by a couple of degress, but that might simple be for better thermal compound.

I am very happy with these. They were on offer at 97p each, so I bought 50 of them and, as these LT30-2 farnell psu's have 4 parallel power transistors as voltage followers, I went through all 50 marking each with the hFE at 0.5A collector current to pick a matched set. The three remaining 2n3055 in the failed channel has hFE of 77, 82 and 120. Not exactly matched!! The hFE of the parts I used are 105, 106, 106 and 106! This being the advantage of buying a big batch and having the time to sort them!!

PA160826.JPG


PA160825.JPG


So they have 7 times the voltage withstand capacity of the 2n3055, higher current rating and higher SOA for pulsed currents. Best of all, they were less than half the price of the 2n3055 they replaced. Good result!

I did test a small batch of these first, to check they had the right characteristics, before buying a batch. So I still have 54 of them, which I will be using for an assortment of future tasks, not least of which is to swap out all the old 2n3055s from the other channels on this, and also my second LT30-2 bench supply. I am still anxious that my PKR251 blew up because of a failure on this psu, rather than caused the psu failure. I shall now move on to reparing the PKR251.... with your help, of course!!...
chrismb
 
Posts: 620
Joined: Thu Aug 05, 2010 6:32 pm

Next

Return to Analog

Who is online

Users browsing this forum: No registered users and 4 guests

cron