PDA

View Full Version : OT - Mastech HY3020D Problem (Power Supply)



CCWKen
09-11-2010, 06:13 PM
I've got a power supply that went ape on me. It still has output but the problem is that the output is not controllable. Neither the current or voltage controls affect the reading on the meters. When powered on, the volt meter reads 53.0 with the CV light on. Adjusting the voltage dials has no affect.

I set up a test solution with anode and cathode connected and the current reading jumped to 32A right off the bat and the current dials had no affect on the meters. The PS seems to be pegged at full output. And yes, I know it's only supposed to be 30v/20A but evidently it's design is capable of higher output.

I suspect the problem is in the feedback/control circuit but I can't figure out the only schematic that's available on these power supplies. I've looked for obvious problems like burned traces, swelled caps, etc. but can't see anything out of the ordinary.

Has anyone experienced the problem and found a fix? Anyone an experienced electronics tech that might offer suggestions?

Thanks in advance, Ken

J Tiers
09-11-2010, 09:39 PM
There is fixing, and there is finding out.....

The very first place to look for fixing, is solder joints, which if cracked, can lead to an open, with obvious issues of a feedback circuit not being connected, etc.


Failing that, you may have to do some "finding out" which will then help you "fix".

The power supply no doubt has a higher voltage source, which it "regulates" with circuitry, to limit voltage and perhaps current.

If it is unable to regulate voltage, then either the series pass element is "on" all the time because it is bad, OR because the circuit which drives it is not responding to the voltage control input.

The pickoff is normally near the output terminals. Look for a wire or trace that is connected, or is supposed to be connected there. Follow it back looking for bad connections.

Some supplies have both current and voltage settings. However, they may likely use the same element for control, and if there is a bad connection in the section that turns the series pass OFF, then it won't limit anything.

if it is an SMPS type, the elements won't be as obvious, but they are all still in there.....

If you can find the control for the series pass device, you may be able to find out if the drive circuit is "trying" to turn it off...... if it is, the feedback is OK and something in the series pass is bad.

otherwise the feedback has the problem.

if you have a PDF of the schematic, I'll give you an email address and maybe I can help you figure it out.

CCWKen
09-11-2010, 10:04 PM
Thanks Jay. The pdf can be downloaded here:
http://fileshare.eshop.bg/download.php?fileid=50114
You'll need to "zoom" in to see much. If the direct link doesn't work, let me know.

When it's operating correctly, the PS is variable allowing you to set constant current or constant voltage. (up to 30 volts or 20 amps.) You set one to max then set the other to the desired output.

Forgot to mention:
It's a "linear" PS. Also, I've disconnected and reconnected all the inter-board connectors thinking there may be corrosion but that didn't change anything.

J Tiers
09-11-2010, 10:53 PM
OK, yes, an older linear supply.

A little confusing, because it regulates the positive side vs ground, to control the negative output...... same result, but different from regulating the positive output directly, allows a much lower drive voltage. Simpler in some ways, all controls are near ground.

The point of drive is at the base of Q1 over near the transformer, the 2SC1008.

As I suspected, it operates by turning ON all the outputs via R10 from the +12V, and regulating by "diverting" drive from the base.

So teh problem is that either an output is shorted (one of the 2n3055 devices in teh central group), or that the drive is not able to be "diverted".

The main regulating device seems to be the lower op-amp under the outputs. The other one appears to control current.

So, at the lower LED, LED2, there should be a voltage which is trying to reduce the drive to the output devices......meaning that voltage should be "reasonably low" vs circuit ground (the circuit is NOT referenced to chassis ground). Circuit ground is the + output.

First, check the +12V at the regulator at top of schematic.

Also check the -6V at DW1 shown below it.

There are some apparent test points......

if 12V and -6V are OK, check test point 6, at the join of the two LEDs. That should be close to circuit ground, but somewhat "+", maybe about 3V + at no load.

Also, TP11, at other side of lower LED should be about 1.2 V below TP6.

Pins 2 and 3 of U3, the voltage regulating op-amp that is connected to TP 11, should be equal and at approximate circuit ground.

TP12 at the top of teh capacitor C11 should be also close to ground.

One of these will not be correct........ report back which.

Since the current controls don't work, I am suspecting that the problem is not voltage feedback. The current limit uses the same ultimate control point, at the base of Q1, but is separate up to there.

I suspect either the -6V is bad, or a problem exists in the outputs area (Q3, Q3B, Q7, Q7B, Q8, Q8B) or drivers area (Q1, Q2, and the 2N3055 that as I see the PDF, is labeled "Q?". might not be a bad transistor, there are resistors etc there also.

CCWKen
09-11-2010, 11:02 PM
I'm printing your post as we speak. I'll chew on it a while and make the tests tomorrow. (Meters are out in the shop.)

Thanks Jay!

J Tiers
09-11-2010, 11:07 PM
get the whole thing..... I apparently edited after you posted.

CCWKen
09-11-2010, 11:24 PM
Got it! Thanks.

rdfeil
09-12-2010, 01:46 AM
Ken,

Jerry's procedure is right on the mark. Now possibly a quicker way to narrow the search, IF the output and driver transistors are socket mounted, would be to pop out the driver transistor and see if the output goes away or stays at full output. The driver is labeled Q? on the schematic. It is the same as the output transistors, so it should be easy to spot. With it removed is the output is still at max you have one or more shorted output transistors. If it goes away then either the driver you removed is shorted or there is a problem earlier in the control stage. Just a suggestion to possibly save a little time.

Robin

PS: Mouser.com has the transistors for a good price and they have no minimum order.

whitis
09-12-2010, 03:50 AM
Here is a little more detailed circuit description.
The four bottom secondary windings of the transformer supply the output power with relays selecting appropriate taps to minimize the voltage which needs to be dropped across the pass transistors which reduces heat and power waste. The next winding up is used for a little boost winding that runs on top of the raw filtered DC to provide a little extra voltage to drive the base of Q1/Q2/Q?; this minimizes the dropout voltage so it can operate with the output close to the raw DC. There is extra circuitry (most of the right hand side of the schematic) to determine when to operate the relays which you can ignore. The top winding makes the 12V logic supply and a -6V and -15V to give the op amps a little room to operate. The curious thing about the logic supply is that the "ground" for those is not the "-" output but the "+" output of the power supply. This is because many of the components can't handle the high voltage and almost all the action takes place around the positive rail, so they float the logic supply on top of the output voltage.

Transistors Q3, Q3B, Q7, Q7B, Q8, and Q8B are paralleled to make a bigger pass transistor. If any of those transistors is shorted, it could cause your problem but when you drew 30Amps you would have seriously overloaded the emitter resistor (which would be taking about 9 times as much current as normal) which might be expected to explode which you probably would have noticed. Resistors R103, R101, R99, R10B, R101B, and R99B are the emitter resistors on the transistors; these are required for the transistors to share current evenly. They also double as the current sense. R98, R100, R102, R98B,R100B, R102B sum the voltages across the emitter resistors.

Q1, Q2, and G? boost the base drive for the 6 main power transistors, which have low gain. R10 operating thrugh DW2 and D11 which drop the voltage slightly, turns these on by default, then the current and voltage limit circuitry pulls this voltage down as needed to shut off the main transisotrs. They do this through two diodes, which also happen to be light emitting diodes. This is a rather dubious practice because LEDs do not like reverse voltage so one of those could have failed shorted . Make sure you try to turn the output off by turning both the voltage and current knobs to their minimum settings simultaneously, not one at a time, with a little load. The two op amps connected to LED1 and LED2 are the voltage and current limit circuit (the top one being the current limit circuit.

Note they could have used the logic supply as the boost voltage for Q? but since the main pass transistors have low gain, a lot of current flows into those bases and would overload the logic supply. So, that is why there are two supplies floating on top of your output voltage, rather than one.

Your failure is likely somewhere between the two LEDs and the 6 pass transistors.

First do the test of running both current and voltage knobs (coarse and fine) simultaneously with a little bit of load that is enough that the current limit would kick in. Submerge just a tiny amount of electrode in your electroplating tank or use another container with a weaker electrolyte. You do not want to draw 32A for any length of time. Adjust the tank so you are drawing within the rating of the supply. It is safest not to exceed 1/6 of the rating just in case one of the pass transistors is shorted and the emitter resistor is overloaded.

Check the +12V, -6V, and -15V supply voltages for being somewhere in the right neighborhood.

Connect a cliplead to the positive output of the power supply, call this "logic ground". Short the junction of LED1 and LED2 to logic ground and note what happens to your output voltage. Repeat with base of Q2. Then the base of Q?. Shorting the bases of the 6 pass transistors is not advisable, you are likely to damage Q? or the boost supply. This will quickly narrow the problem.

Also measure the voltage about 1.2V expected for a 20A load, proportionately less at lower current) across each of the emitter resistors ( on the 6 pass transistors with a significant load connected to the output of the supply. If one of those is high and the others are low, you have a shorted pass transistor. But if that was the case, the emitter resistor would probably be glowing/burning/shrapnel.

Look for bad solder joints, hairline fracture where the traces meet the pads, and bad components in the area of the circuit that is not behaving properly.

Black_Moons
09-12-2010, 04:17 AM
Wow nice detailed explinations. Im just gonna vote ++ on the 'shorted output transistor' theory as the first thing you should check, as that fits perfect with the symptoms you discribe, and its a common failure if the power supply has been abused and does not have the worlds best internal protection circuity.

CCWKen
09-12-2010, 07:15 PM
... and its a common failure if the power supply has been abused and does not have the worlds best internal protection circuity.
LOL... Abused? Hardly. It runs in an air conditioned shed and I don't recall ever going over about 10-12 amps for plating. My tanks aren't that large. Don't know about the "worlds best internal protection circuity". What does that mean? I don't own the World's best anything. Do you? :cool:

I'm disassembling the PS to locate the test points. The component layout doesn't match the schematic but I'm able to find the parts. And I think most of the test points are at the inter-board connectors. The components are located on five major boards and in two huge fan cooled heat sinks that are half in and half out of the cabinet, mounted together. I don't know if I can get to the power transistors and still have a running PS. They're buried in the heat sinks.

Here's the Boards as I can figure them out.
1. Display and display driver board. (Amp and volt)
2. Pot and LED board. (Course/Fine-Amp, Course/Fine-Volt, CV, CC indicators.)
3. Relay board.
4. Input Filter board. (Large caps and diode array.) There's 28 1N5408 diodes instead of the 8 shown in the schematic--Four groups of seven.
5. Control and component power board.

Other small boards:
1. Output filter board w/output terminals
2. Input voltage selection
3. Power switch mount

And two boards inside the heat sinks that connect the power transistors. (Four on one and three on the other.) Hmmmmm. It looks like that's where the driver transistor is located. Dang it!

It looks like I'm going to have the PS spread out on the desk and no chassis to speak of. I gotta run back out to the shop and get another load made up. Can I just use a 60W, 120v light bulb for the load?

J Tiers
09-12-2010, 08:37 PM
Sure.

Actually, you don't NEED a load, the thing should regulate with the load open.

h12721
09-12-2010, 10:54 PM
Hi Fellows, After U1 and R 1 there is shown an SCR, what is with this Thing?
Look how the gate is connected.do I miss some thing ??
H.

CCWKen
09-12-2010, 11:49 PM
Hi Fellows, After U1 and R 1 there is shown an SCR, what is with this Thing?
Look how the gate is connected.do I miss some thing ??
H.
Yeah, it took me a while too but if you look to the right, above and below C4, you'll see that it's labeled as WL431 (DW6). It's really a TL431 or a shunt regulator.

I had a dickens of a time getting the heat sink apart but I may have found the/a cause of the initial problem. This would explain why the PS was working one day then about a week later of being idle, went ape on the next power up. Apparently, some plating solution entered the outside heat sink and dried to a sticky coating over the Collector, Base and Emitter traces of one of the power transistors. I'm guessing this electrolyte caused a short or at least killed something. I hope this didn't burn them up but just created a resistive short since the power transistors are ganged together. I'm going to clean up the board and check the traces. The electrolyte would have been a strong alkaline solution, around pH 13. :rolleyes:

I remember bumping the tank and causing a splash but didn't realize it reached the PS some 3+ feet away. I'm tired and will report back tomorrow after I clean the sticky goo off the board.

What do think? Could the power transistors still be good?

rdfeil
09-13-2010, 02:47 AM
Ken,

Clean it really good and give it a try. With a short like you describe there is a very good chance that the transistors are still OK. The power transistors act as a simple voltage regulator where the emitter voltage is approximately .6 to .7 volts lower than the base voltage. What has happened is the base circuits are getting voltage from the collector circuit which is very likely 50+ volts. This will cause the transistors to effectively go into full conduction and woohoo you get lots o'volts out :eek: . Anyway, clean and test. If it is still bad just continue the troubleshooting. Let us know what you find out.

Robin

whitis
09-13-2010, 11:23 AM
Hi Fellows, After U1 and R 1 there is shown an SCR, what is with this Thing?
Look how the gate is connected.do I miss some thing ??
H.

If it was an SCR, it would be Overvoltage Protection (OVP). If the voltage goes too high on the 12V supply, the SCR triggers and crowbars (shorts out) the supply protecting other circuitry in the event of failure; of course, the OVP often triggers falsely and causes a failure of its own but that type of failure doesn't blow up other circuitry.

However, the TL431 programmable zener diode is a very handy shunt voltage reference, and that is what you are probably looking at here. I have used them and they are common in power supplies.

"Abused? Hardly. [... I use it for electro-] plating. "
That is a form of abuse, right there. Not that you shouldn't do it but you need to be aware that you are putting your supply at risk merely by operating it in a corrosive atmosphere. Put some distance between it and the plating tanks.

Clean that board real well to get rid of the residue and try not to spread it either (yeah, not easy). Hose it down with water, then rinse with distilled water. Then bake it dry at less than 158degreesF (70C). If you use the kitchen oven, check temperature carefully and take precautions against lead on the circuit board contaminating your oven racks or anything else which might come in contact with food. Beware of radiant heat from the heating elements causing localized overheating. Just the pilot light may be enough on a gas oven. Boards are normally cleaned in water, alcohol, or certain other solvents. The trim pots are the components most likely to be affected by water/solvents and may have a plastic cover or kapton tape on them when circuit boards are cleaned in production. If you can find the datasheet for the trim pots, check what it has to say about cleaning. Give the board time to dry before powering up. Tape may cause ESD, so be careful with it, too - ground it after unrolling and cutting.

J Tiers
09-14-2010, 12:28 AM
On teh schematic it's completely poled wrong for an SCR, and would not work with gate shorted anyway.

As a TL431 variant it makes sense...... 2.5V temp compensated source, used as a reference for the output voltage. On my version of the schem, the designator appears over to the right of the adjacent capacitor...... I actually thought the missing font error I got had "eaten" the designator for it.

So, is it fixed yet?

If it's seriously splashed, I'd wash the thing in deionized water, and dry with a hairdryer held at a judicious distance...... just don't soak the pots (unless they need cleaned) and there will be no trouble drying them.

There is a goodly amount of current gain, so a splash across the 10K and Q1 could really cause trouble.

CCWKen
09-14-2010, 03:20 PM
Haven't had a chance to fully clean the board yet. After I pulled the heat sinks apart it set for about 36 hours and the goo turned a little hard. I sprayed with distilled water and kept wiping with an acid brush but it's taking a while. I think I'll try an old tooth brush next. The power transistor that got the worst looks to be one of the second group of three. It's the one that had the goo over all it's connections. There's no sign of any burning or high heat and I still have my fingers crossed.

I'm neck deep in the body shop and running late on a project so I'm having to work this in at night.

I really appreciate the help and the following. I'm spread pretty thin at the moment so it may be a couple of days before I try to fire the PS back up.

Forgot to mention:
The pots are at the other end of the unit and were spared (front). The heat sink(s) is/are at the rear. That's where the goo is.

Black_Moons
09-14-2010, 03:57 PM
As for abuse, maybe I should of said (ab)use. Sometimes just use near capacity is enough.

Also, if you ever had any shorts, that counts as abuse. Some supplys you can short all day, others might fry the first time, while others might fry after a few shorts, or an extended short.

rdfeil
09-23-2010, 12:48 AM
Ken,

Did you get it repaired? If so, did cleaning it work or were there bad parts?

Robin (just wondering :confused: )