Sorensen DCS33-33 repair. (ZOMBIE =!!!!!!)

[coppercone2]

So this is one of my last old 'gambling debts' that I want to try to fix, its getting a little old to only have one of these functioning.

So the problem is the device is 33V, and it was getting stuck at 16.5V. However during probing I managed to short the LDO out some how (5V behind the heat sink), I think one of the channels to ground, not input to output, given how the chip pinout is..I am not sure why it went bananas after shorting a LDO to ground, it is possible I some how shorted input to output (3 to 1) but I don't think so, maybe a control loop went bananas.

But this caused it to make a horrible squealing noise while it was on and then to blow a fuse. I have left it alone for the longest time.

I have a sorensen XKW schematic, but I am not sure how its different from DCS.

I took the PCB out and I have no idea how to approach it. I was  first going to replace the 3 LDO's, since I have them (lm337 and 2x ti 5V reg), and I know something bad happened to the rail with my multimeter probe, then see if it at least works without triggering a fuse.

Why a 5V LDO caused squeeling/whine from what I guess is magnetics i don't know, cascade failure comes to mind because it is only a 1 amp reg, I measured it in circuit and I don't see evidence of a power down low impedance condition,


Is it possible there was some kind of 'hiccup' caused by the interruption of the power rails when I abruptly changed voltages on one of the regulators?, and that it was some kind of soft fault? I thought to replace the fuse and try it again.

[THDplusN_bad]

Good Day,

the limited advice I can give:
* A friendly link to a loooong repair thread for the DCS series from Ametek/Sorensen. https://endless-sphere.com/forums/viewtopic.php?t=95184
* And then there is the user's manual, which I suppose you already have. It is available at https://www.powerandtest.com/power/dc-power-supplies/dcs-series in the Manuals area.

Section 3.1 Power Circuit (A2 Assembly) in the data sheet provides a high-level block diagram, and reasonable circuit description. The squaling noise is a good indication that something in the PWM control loop is defect. You might have fried a supply for this control loop circuit named "auxiliary supply voltages for the PWM circuit" in the manual (page 3-3). I would start looking there.

You are in the realm of high-power DC/DC conversion, so I trust you know what you are doing regarding safety at work etc. 
I have been there (partially at my job, partially for my hobby, i.e. with Tektronix SMPS power supplies for their "classic" 7000-series of oscilloscopes), done that and I have decided that it's just not my cup of tea...

I have the XG600-1.4 model and it is a "beast" in a positive sense.... 

Good luck for your repair.

Cheers,

THDplusN_bad

[coppercone2]

I do have a working unit but I would want to not work with the bare PCB if possible, and going back and forth between the chassis and the PCB is a serious time sink and something will likely get damaged after the 10th install remove operation.

This one is a little better to work on then most SMPSU, since its flat and not compact. I have some more advanced technology now so I was planning on

1) using tektronix minigrabbers preplaced during power off to measure difficult circuit junctions
2) wearing goggles
3) making sure not to lean over the board (this is how you usually get shocked, working too long and deciding to put your hand some where to reduce back strain)

Since I got a tip on where the bad guy might be, maybe I will desolder some devices and test them on a bread board (discretes near the region where you mentioned) and replace cheap IC as a preventive measure (i.e. quad cheap op amp). My op amps were  bad on the linear sorensen repair, but the job here is much more tricky because its all based on pwm.

linear sorensens also scream unless you replace all the caps btw.. mine got 4x as more quiet when I replaced every cap on the board. I suppose I should test all of them with the meter also in circuit.

I also do have a working one, and I never compared the two, but I need to get it up and running so I can cross probe them, right now I expect its too damaged to turn on until i find the fault I caused. At least running good enough not to blow a fuse.  I find it weird that a 5A fuse blew without damaging a part visible either... really hope something sank the current and stayed alive and it will power up but I will at least replace the LDOs first since I have em. First time I did this on the kitchen table, this time I setup another work bench with all the hard core soldering tools with another ESD mat on another table so I can repair the big boys without clogging my lab.

 I am banking on this job being made easier and cheaper by the pace desoldering iron... I thought to bread board test alot of the transistors, since I can actually do it now with a reasonable work load.

mmm shit that guys finger is fucked up, maybe I should buy some multimeter probe shrowds also to reduce risk of short circuit with the probes. For mains stuff I do have a scope meter also.

[coppercone2]

I think this job will be delayed until I get this new insulated probe set, nothing I have quite fits the bill for this job and those fingers look really messed up, it looks like he got burned with fireworks. This piece of shit is not worth that kind of damage, I would attack it with a hammer.

And I bought new multimeter hook tips, I should not be using stuff fixed with heat shrink (my budget was non existent).

Good call on additional safety instead of going at it like a psycho

UT-C01 (unlikely to use, but should replace)
and fluke TP38

[THDplusN_bad]

I think this job will be delayed until I get this new insulated probe set, nothing I have quite fits the bill for this job and those fingers look really messed up, it looks like he got burned with fireworks. This piece of shit is not worth that kind of damage, I would attack it with a hammer.

And I bought new multimeter hook tips, I should not be using stuff fixed with heat shrink (my budget was non existent).

Good call on additional safety instead of going at it like a psycho

UT-C01 (unlikely to use, but should replace)
and fluke TP38

Yep; better safe than sorry. You might want to return to this one when you have some leisure time and you are in the right mood for it.
I agree those photos from the burned fingers are disturbing..  

I have also learned to "get disengaged", say to stop working on a repair when the thing is simply beyond repair.
But it took me several years to understand that a good decision to stop is not quite the same as "giving up".

Cheers,

THDplusN_bad

[coppercone2]

there is always a board swap or brute force fix by replacing a whole bunch of parts, it just increases costs.

chances are you might get a component thats on the verge of being bad also if you just start replacing stuff randomly.

beyond economical repair often means putting confidence that ebay won't screw you on another one that you are not quite ready to buy when the price is right and then it costs 2x when you want it, so often a bigger random repair is actually worth while so long its not prohibitively expenisve

[cdev]

Whoa, this is a big one..



20v 125A

[coppercone2]

IIRC those are the SCR ones, they should be very robust and strong and hard to break and easy to repair without a complex PCB.

Usually most lab supplies, the old ones, are SCR at high power, or if linear.. have a SCR front end to reduce LDO drop ( tracker hybrids). That might be just a SCR followed by LC filter and thats it.

They are high noise, if you look at the power supply offering tables, I recall that in the older sorensen literature , and lambda as well, you start to get bad noise levels as you go up in current.

I think the gigantic HP ones are quiet however, but they are like 2x that thing.

Personally, I stay away from anything involving power SCR. I almost feel like thats putting a weed whacker in the kitchen.. not in my lab. I do not like thyristor power supplies. They make me think of lawn care equipment engines.


This might be quiet, https://www.ebay.com/itm/HP-6483C-600V-DC-15A-Power-Supply-TESTED-400VDC-25A-10000Watt-10kW-3-Phase-AC/303135257941?hash=item469444f555:g:vwgAAOSw1mdeqiHv

I think they have the same style but low voltage higher current. I don't even wanna think about them, I would not feel comfortable around them unless it was setup like a welder in a weld shop and I had protective gear. It is also SCR but I think they do something better then the other brands because a long time ago I recall looking into what a low noise big ass supply looks like, and I got around that far.

No. thats not it, it has 600mV rms

[coppercone2]

Ah, this is a quiet one, but half the current

https://www.ebay.com/itm/HP-6259B-0-10-VDC-0-50-AMPS-Huge-Analog-Power-Supply-Newly-Calibrated-CheckedOut/313176620713?epid=1220319319&hash=item48eac7fea9:g:dVYAAOSwQBNfLsBj

[coppercone2]

taking another shot at it.

The other one broke, because the mains input connector was not actually soldered in , or soldered in very poorly. When I took it apart I regreased all the heat sinks and cleaned the contacts etc. Also had a bad socket.

So I decided to take another try with the one this thread was about. I decided to replace all the IC's and transistors, because I had quite a few of them from other stuff, but I did have to buy some parts. Currently waiting for the remaining parts to come in. Some were a little pricey, but I can't help myself, those bus bars look too nice

having two working ones would be pretty awesome. If I get them both working I can setup a little electrochemistry "line" without having to swap leads and readjust settings all the time.

[coppercone2]

well I replaced every part on the main board that was a IC.

One of the transistors does not have orientation labeled.

Anyway, I assembled it correctly the first time, same problem as years ago, the 500mA slo blo fuse glows. The resistor connected to the transistor that was poorly labeled got hot.

Before doing anything, I noticed that earth isolation was rather poor on the unit I am repairing, compared to the other one.

Anyway, then I thought MAYBE i put the transistor in backwards, so I actually put the replacement transistor in backwards, and it arced something and shows 60V, so its really destroyed now.

I am thinking there is maybe something wrong with the transformer. Anyway since I think I fried a bunch of power electronics I put it in the trash now. But I suspect this mishap was actually a blessing, given the poor impedance reading to earth from the terminals. I will keep the GPIB card or whatever is in there,and the heat sinks, blower fan. The unit is not very valuable and the transformer is some tape wound shit, so I am not even gonna think about it. 

[coppercone2]

Whoa, when I was taking it apart for break down I found something ! The OVP front panel pot is replaced with a bourns 10 turn pot and its installed so you can't even turn it , like totally upside down!

what the hell is this?! its supposed to be a 5k . does not work. I started experimenting on it before final dissasembly, I cut the offending resistor that was connected to the OVP transistor. stopped making noise but output is 55v. but anyway with the difference in isolation to ground and like the rattling noise it makes (like a rattle snake) i am gonna dispose of this. starting to give me the hebee jibies . i am not gonna get wasted by some rattling power supply doing weird shit

[coppercone2]

I suppose that its stable with the R93 cut, its possible to acutally do trouble shooting on the PWM with an oscilloscope. I wonder if I should do it for the sake of experience. i threw it out but its like jumanji I heard buzzing and clickin from the trash can

[coppercone2]

I FOUND A SHORTED DIODE. CR44. NOT SURE HOW I MISSED IT. I THOUGHT I TESTED EVERY SINGLE DIODE, INCLUDING LIFTING LEGS. MUST HAVE MISSED THIS ONE

PUT IN CHEAP ASS 6.8v DIODE. WORKS FINALLY. JUST NEED TO REPLACE R93 (3.3 OHM 5 W) BECAUSE I LET THAT BASTARD COOK WHILE i WAS FINDING THE PROBLEM, ITS PROBOBLY COMPROMISED

I thought it was a regular diode so I think the first time i measure it maybe it was reading some wrong voltage, then maybe after I bypass the fuse and let it run overload for a while it went open circuit. not sure. don't renember.

but regardless I might have broken the diode shorting the IC next to it, the supply previously was stuck at 16.5V max. When I replaced everything and found the diode it goes to 33 now. so there was some other fault I cleared by carpet bombing.

[coppercone2]

actually during reassembly I found another problem. The GPIB/LAN control card thing was pinning it when it got to like 2/3 max voltage (18V or so).

I removed the cards and used a section of the card shield, cut with a sheet metal auto shear, to cover the opening in the back. works great.