Hitachi four pole capacitor

[jefti]

Hi guys,

I've got a laser spot welding system which is tripping the mains fuses. It is a 3x220v system. I've meggered out all transformers, which seems to have no faults. Now I've looked at the input on the main PCB where some rectifying is taking place. There are two large capacitors located there, of which one is a little bit bulged up. I've took em out and one is indeed bad, losing voltage straight away, while the other one seems OK.

I want to replace those caps, since this is obviously one point of failure. The caps are 4 pole ones, and I'm having difficulty finding a replacement, they dont seem readily available. I've added two pictures showing them. If anyone can point me in the right direction on where to find these caps, that would be great.

Also, can a faulty cap like this cause a short in such a way that the fuses will trip?

Thanks!

[Chalcogenide]

They are 4-pin snap-in capacitors, available at most electronic distributors. They are standard electrolytic capacitors with a more robust mechanical mounting.
It's unlikely that just a capacitor failure could cause the fuse to trip, unless the capacitor itself presents a very low DC resistance (in which case you wouldn't be able to charge it up at all).

[Dave]

Hitachi HP3 capacitors. Apparently the 4-pin version is available upon request and two of those pins are there merely for mechanical strength. You could just replace it with a suitable 2-pin capacitor and add some RTV silicone around them for mechanical strength.

Catalog for HP3 series
Capacitor pinouts (scroll down)

[jefti]

Thanks for the quick replies! The two other pins indeed seem to be "dead". I way trying to charge these caps with just a 12VDC supply. If there is a short across the cap, they would probably draw alot of amps, right? Any other way to test this?

[Dave]

Put an ammeter in series and apply the voltage to the capacitor. Allow some time for the current measurement to settle. The leakage should be well below 1mA in a good capacitor.

But considering that one of them has bulged, you can be pretty sure that they need to be replaced.

[PChi]

Yes the capacitor could be shorting out internally and blowing the fuse. The sneeky part is that it can then heal and appear to be OK for a while. Many years ago I had some Philips snap in capacitors do exactly that. You hear a click from the capacitor then find that the fuse is blown. Changing the manufacturer to BHC fixed the issue. Both suppliers now gone or been bought up.

[jefti]

Thanks for all the help! I've ordered some new capacitors and will replace both (there were 2) just to be sure. Ill keep you guys posted. I find it quite difficult to troubleshoot what appears to be a power supply failure if it trips the breaker right away. There is not much what can be measured then. I thought it was a good idea to power up the circuit with both mentioned caps removed, but that violently blew a varistor, so that will be replaced now, too 

[wraper]

Yes the capacitor could be shorting out internally and blowing the fuse.

With a chance smaller than one in 1000 of similar failures. Shorted electrolytic capacitors are extremely rare.

[The Soulman]

I thought it was a good idea to power up the circuit with both mentioned caps removed, but that violently blew a varistor, so that will be replaced now, too 

Better check the rectifier diodes, they're suspicious.

[jefti]

Thanks for the tip. I will check the rectifier bridge tomorrow. It's a three fase unit, with hardly any components in front of it if I recall correctly, so basically just running on 3x220vac. I'm a bit cautious of just hooking it up to the three fases, is there anything I can do to test it in a bit safer way? Can it be tested open circuit?

[LA7SJA]

I use a multi-purpose meter with insulation measurement function so that I can check the capacitor and the reverse of the diodes with 250v-500v-750v or 1Kv. If you don't have one try borrowing from an electric installation firm or a trade school.

Uni-T have a few http://www.uni-trend.com/productscatelist_96_96_96_96.html

Johan-Fredrik 

[SeanB]

3 100W lamps in series with the power leads will work, giving current limiting into the fault but not too much drop in normal use.

[jefti]

Those are handy tips. I've got the rectifier tested, the megger test was OK and also on 120v the rectifier seems OK on all phases. I've gotten into some more thinking and am puzzled on why the varistor blew. The PCB has an AC circuit and a DC circuit. The caps that I mentioned earlier are on the DC side, while the varistors seem to be in place for surge protection on the AC side, they are mounted in between fases. Testing the circuit without the caps on the DC side cannot have been the issue I believe, Since they are behind the rectifier.

The connector marked CN1 is the mains in. Z1 is where the blown varistor was. If I megger out the traces of Z1 I get a megaohm resistance, so no short circuit. The varistors are TNR 23g431k, so nominal voltage of 430 volts. How come one can blow so violently on 3x220v (thus 380v in between fases)? My guess is only when they are faulty. But before all this, the breaker just tripped instead of blowing anything up, so I think there was another fault causing that. That may still have been on the DC side, so no way of testing that until I have the new parts.

For now I am planning to wait until I get the new parts, and then testing the board without the rectifier first, so effectively only testing the AC side.

What do you guys reckon?

http://imgur.com/a/dNgTD

http://imgur.com/a/f9r79

[jefti]

Just to clarify, with megging  out the traces I mean between the connections where the varistor used to be.

[LA7SJA]

Sounds like a plan  .
Did you use the "Megger" on the cap's and rectifiers? And as SeanB proposed, a 40-100W lamp in series with the power leads will prevent your AC test from going up in smoke in case of a short.

Johan-Fredrik

[jefti]

Since all the components in this part of the circuit are mains voltage, I figured it would be OK to use the megger on the low (250V) setting. So yes, I've used that to test the rectifier.

[SeanB]

430V varistor is not intended for connection to 380VAC, it is only for 230VAC across it, and if connected across phases your unit is a 110VAC 3 phase one, where there is 208VAC between phases, not a 3 phase 380V one.  Just connect a single phase and neutral, not all 3, as a 3 phase full wave rectified power supply will be over 600VDC.

You have a very common single phase unit with integrated bridge rectifier, power device block, and while it is called 3 phase 220VAC, it is really only a single phase device, when used in a non 115VAC US network. Only way that a 400V capacitor will work there is if they are connected in series, but if they are parallel then it is a single phase version.

Pictures of the board top, board bottom and a little more info and pics of the whole system will help as well. We thrive on images, preferably reasonably clear and within forum size limits.

[jefti]

I think you are wrong SeanB, the unit really is a 3 phase 220 volt one. Also, the rectifier is on the DC side, while the varistor is on the AC side. My phase to phase voltage is round about 400 volts, which should be OK with a 430V rated varistor, right?

I've got some more pictures of the boards, i've uploaded them online to keep their full size and detail.

Just to clarify the pictures, CN1 is the mains input. The AC circuit is on the side of the CN1 connector opposite of the transistor rail with heatsink. After the CN1 connector, there are a few varistors connecting the different phases, after that a few capacitors, and after that straight to the connectors for different parts of the machine.

The DC circuit is on the other side of the CN1 connector, where there now are just the holes where the rectifier used to be, marked REC1. From the rectifier, the DC passes through a inductor (presumably for filtering), after which is goes to the two big caps mentioned earlier. From there it goes to a unknown component for me (https://www.google.nl/search?q=sr-25ap&espv=2&source=lnms&tbm=isch&sa=X&ved=0ahUKEwj2i7-3sanTAhVIa1AKHXdEDK4Q_AUICCgB&biw=1455&bih=665#imgrc=e1Fx-OukMFLQXM:), after which is goes through some power resistors and to the mosfets/transistors.


http://imgur.com/a/9fprz front of the board

http://imgur.com/a/W7hJS the rear of the board

http://imgur.com/a/eCPMm

[SeanB]

Yes, the plate says it is 220VAC 3 phase, which is not the standard EU voltage of 400VAC 3 phase that you will get out of the 3 phase socket in Netherlands. Made in Japan where you have either 100/110VAC 60Hz, or 115VAC 50Hz mains in Tokyo, and have 220VAC 3 phase supplies in both areas, depending on which power network you are on. You need to connect one phase only to there, and one other phase to neutral, not all 3 phases, unless you have a 3 phase transformer that will give you 220VAC to the board.  As you have a 3 phase plus neutral connection there, you will probably have to get a transformer, 3 phase, that will provide 110VAC 3 phase to you. Probably best to get 3 site transformers, rated at 1kVA, and connect the primaries in star to the incoming 3 phase supply, and take the secondaries, remove the centre earth connection, and do them as a star connection.

You could simply connect L1and L2 to line, and L3 and N together, and have the power supply side work, but I think those other outputs it has ( J1 CN4 area) will not work properly, as they also are 115VAC systems, controlled by the SCR and optoisolators there. not sure what they drive, likely heaters, fans or something like that, and those are all going to be 115VAC devices.

The green block you are puzzled about is a current transformer, using a hall effect sensor to give a voltage proportional to the DC current flowing in the primary single turn loop, and it is a common part used to give feedback of the power output of the laser supply.

Z2 and Z4 are varistors, the voltage rating is the maximum peak DC voltage it will not break down at, if you connect ot 400VAC there is a very good chance it will blow, as it has around 560V peak across it every mains cycle. no wonder they blew, I am just amazed the survivor did not, it either was at the high end of the range, or the other 2 died to protect it.

[jefti]

Allright, so you are saying that i probably screwed up some components when connecting it to my 3x220 VAC or 400VAC across phases supply? Those different ways of noting the supply voltage really are confusing 

You may be right though, the first time I switched the unit on i did hear the transformers hum for a second or so, before it blew the fuse. I think i got a couple separate 220-110V transformers laying around. Would I be able to manage the correct supply with three separate transformers? 

I dont exactly know how to draw this in a schematic, but if i understand correct, I could connect T1 between L1 and N, T2 between L2 and N, T3 between L3 and N, so basically have a star setup with three transformers on the primary with the N as shared point. Only the secondary is vague to me, will there be a connection on the shared N over there, or will this just be a "floating" point between to secondaries of the transformers? The device does not have a N connection, only a earth.

[LA7SJA]

What color is the wall socket? Red or blue? And the number of contacts in the plug?
 
Johan-Fredrik

[SeanB]

Yes 3 transformers will work, and wire them like you say, just measure after connecting ( and before connecting your board) so that phase to neutral is all 115VAC, and phase to phase is 220VAC, that way you check you have not wired one secondary backwards, and thus made an imbalanced supply. Here the actual phase rotation is not important, but please ensure the voltages are correct first. transformers must be able to handle 4A on the secondary side each as well, so they do not get overloaded.

when it is running remember to check as well when moving them.

[jefti]

It's a red socket, with L1 to L3, neutral and ground. Between neutral and L it is 220V (more like 230 actually), and between phases it is 380 (more like 400V).

[jefti]

All right. I will have a look around to ensure I have the right transformers, hook them up, and do some measurements. After I've got all the replacement components, I can go ahead and test the board.

The input from here has been a great help! Thank you all for the help so far!

[LA7SJA]

Red is used to indicate 400V 3 phase, your device is made for the blue one. I used to construct controller electronics on high-pressure hot-water washing equipment so this is not the first time i seen the result of 400V on a 230V machine.

Johan-Fredrik

[jefti]

To complicate it even further, the unit also came with the red 5 pole connector, which fits snuggly into my wall socket.

[jefti]

Could you guys help me think for a bit? I have some trouble finding the right transformers for my cause. I picked up a variac, but figured this would probably not work since it is not isolated, thus the N is not floating. Another small step down transformer that I have has the same issue. Will a isolation transformer in front of this solve this, or do I need to get a isolated step down transformer?