HELP, Electronics wizards needed!

[DICKEYBIRD]

I'm in the middle of a big job with my CNC lathe and the control quit this morning, blows the on-board fuse repeatedly. It's an old (1996) Denford Smartstep/3 control but has low time on it.

I examined the board and it looks like the device pictured below, a KBU806 rectifier may be fried. The only spot on the board that looks like it's been hot is where it's soldered on by its 4 pins. The solder around those pins has obviously gotten to the melting point and all other pins on the board look slick & shiny.

The question is whether something else caused it to fail or did it just die of old age. I was worried about the big capacitor next to it because the cap is bulged but I posted that question here a while back and some posters said it was probably OK. It's been working fine for almost a year in light use. When checked with an ohmeter, it measures low resistance for a moment, then high. I think that's how you test 'em without a proper tester? Would a fault with that capacitor overload the rectifier? If I could find a place open today (not likely these days) I'd replace both of them. Ya'll eleck-trickal experts have any thoughts?

[DFMiller]

The bridge does look like its been running hot. Ther could be a shorted diode in it.
I would be checking the bridge then look for other issues like a short on the motor drive to blow the fuse.

Take the bridge out and check the 4 diode using your ohmmeter.
Dave

[lakeside53]

If the bridge rectifier has a shorted diode, the capacitor will have been stressed.

If the cap is bad, the bridge can be damaged, but your test (at a couple of voltmeter volts) shows the cap isn't shorted.

As suggested above, the overload may be further down the line. Disconnect the load, and see if the board still blows fuses. Even then, it could still be a driver.

[MaxHeadRoom]

As mentioned it would pay to remove the bridge and test, an ohmmeter will not necessarily show an open or a correct reading on a diode or rectifier, If you have a range with a Diode symbol on your meter, this applies enough current to forward bias each diode for a correct check.
If it normally runs consistently hot I would bolt a piece of aluminum to it as a heat sink, making sure the aluminum does not contact anything else.
The cap shown often show a raised top, the underside is the place to check for leakage, although if may be best to replace it if you can find an equivalent.
If it still blow the fuse, it is most likely one or more of the driver IC's.
Max.

[MotorradMike]

It's very unlikely the bridge failed on its own.
If indeed there is a problem with it, it was likely caused by something else failing.

Can you see any signs of overheating elsewhere?
I'd look carefully around the drivers bolted to the heatsink.

If you do take out the bridge you're not going to have an easy time of it. Best way is to heat all 4 pins up simultaneously.

[John Stevenson]

Don't even bother testing, just replace it and the cap.
You are looking $15 total, how much is your time worth ?

[DICKEYBIRD]

Thanks all. I forgot to say that I unplugged all the stepper motors first and it still blew the fuse.

I looked closely at the rest of the pins & components and I don't see evidence of overheating.

I will replace both parts assuming I can get them Monday as all the old shops I used to go to on Saturdays with my Father when I was a kid don't exist anymore.

Can anybody tell me how to test the bridge once I get it off? I sorta remember the 4 diode bridge concept which brings up my other bad-freakin'-luck issue. I typed in the diode p/n into Google, clicked on a link to a Chinese site that had a "Data sheet including test procedures" and it went to a .pdf download page that immediately killed my shop computer. It's an oldie with W2000 and is weak in the malware protection area. When it rains, it pours.

If a stepper driver chip failed (shorted I assume?) would it show visible overheat signs or do you have to just replace them? Any way to test?

[Evan]

What John said. Replace the cap and the bridge. Put a heat sink on the bridge as suggested. Any of those green caps from that period are suspect.

Also, there is a good chance that one or more of the pins on the bridge actually have a bad connection due to a very hard to see crack in the solder. That will cause overheating of the part because of losses heating the pin and then the entire part. Same applies to the capacitor. That is a common cause of part failure when a poor connection at the board then causes the part to overheat. The root cause is poor wave soldering of large pins caused by poor board preheat when soldering.

[Evan]

I typed in the diode p/n into Google, clicked on a link to a Chinese site that had a "Data sheet including test procedures" and it went to a .pdf download page that immediately killed my shop computer. It's an oldie with W2000 and is weak in the malware protection area. When it rains, it pours.

Don't feel bad. My system was clobbered by a virus last night from a malignant web site. I was using Chrome configured to not allow anything but the basics. So much for Win 7 security. Microsoft Security Essentials trapped it but not before it had managed to install some scam virus checker software which also turned off the AV software and Task Manager. I had to go into safe mode and remove that manually. Grrrr.

[MaxHeadRoom]

I see you have L298's for drivers.
It is difficult to check them under power when they are blowing fuses, sometimes reducing the power supply with a variac can be done.
If you do manage to remove them, check with an ohm meter between each output and +v and then common, one or more may show very low resistance.

http://pdf1.alldatasheet.com/datasheet-pdf/view/22437/STMICROELECTRONICS/L298.html

The bridge should show the diode arrangement on the side, with a good diode and a regular high impedance ohmeter without the diode range, they may show high resistance across all pins.
Max.

[DICKEYBIRD]

Thanks Max; that site looks a lot like the one that got me into trouble when I tried to download a .pdf on the KBU806. Do you know of one that isn't in Chinglish and is safe?

"L298 : DUAL FULL-BRIDGE DRIVER
STMicroelectronics

Your require pages is cannot open by blow
Reason : Connect this pages through directly deep link"

[DICKEYBIRD]

Don't feel bad. My system was clobbered by a virus last night from a malignant web site. I was using Chrome configured to not allow anything but the basics. So much for Win 7 security. Microsoft Security Essentials trapped it but not before it had managed to install some scam virus checker software which also turned off the AV software and Task Manager. I had to go into safe mode and remove that manually. Grrrr.

Actually I do feel bad because I don't have the computer knowledge or the patience to fix the dang thing right now.

If I can get the CNC control working long enough to do a couple hundred more of the shaft groove repairs I'm doing, I'll have enough money to replace the shop PC.

[DICKEYBIRD]

Well [email protected], it looks like getting the rectifier off to test and see if it's tango-uniform is a no-go. I'm unable to get it to budge even with a big iron and a solder sucker. It's mounted too close to to clip the pins from the other side as well so I guess I'll have to crush it and remove the pins individually unless someone's got a better idea.

[Evan]

Solder a piece of 14 gauge wire across all 4 pins and then desolder it with a hot iron.

Come to think of it I was looking for a data sheet too when I ran into that virus.

[DFMiller]

Yeah,
Evans got it right. If one Iron will not do use two. The copper wire is a good idea to spreader the heat. Use the Irons and lots of solder to get it heated up. Then pull the device.

I never use Solder Suckers with parts in the holes. A great way to destroy cheap boards.

Dave