Announcement

Collapse
No announcement yet.

XMT304 repair

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

  • mwahl
    replied
    I seen the group of 304 XMTs up on ebay a few months ago and thought about it myself. I ended up finding a fairly new one up there for around $1200 that looked like new. Got lucky it worked fine and has little use on it. I have took the ebay gamble and came out ahead 8 out of 10 times. I got a multimatic 200 (the one in the portable pelican case) and paid near nothing for it and found out just the power cord push on terminal had came off. Slid it back on and it worked fine. Some times it takes a bit more though. A lot of the time you can make one good one out of two if you can get them cheep enough to make it worth while or just order a few parts. Its a gamble tho,

    Leave a comment:


  • jjohn76
    replied
    Originally posted by Bushytails View Post

    I found some of those too, but they had 1.25" terminal spacing instead of 1.125" (if I remember right) to go with the larger can, so you'd have to slot the holes in the pcb as well as deal with other mounting issues, like some way to support the back without the stud to mount it by. I finally decided that I needed the welder working now, and spent the money on the miller parts. A lot more money than I wanted to spend on a pair of capacitors.
    Yup, there is a direct replacement Nichicon I found too, but it's only saving $40 on the pair...

    Leave a comment:


  • Bushytails
    replied
    Originally posted by jjohn76 View Post

    Yeah, the only issue I had was finding the right size cans. The 550C's and 550CE's I could only find decently priced in a 3" diameter can. They may be able to fit, I just never tried them.
    I found some of those too, but they had 1.25" terminal spacing instead of 1.125" (if I remember right) to go with the larger can, so you'd have to slot the holes in the pcb as well as deal with other mounting issues, like some way to support the back without the stud to mount it by. I finally decided that I needed the welder working now, and spent the money on the miller parts. A lot more money than I wanted to spend on a pair of capacitors.

    Leave a comment:


  • jjohn76
    replied
    Originally posted by Bushytails View Post

    Heh, I was kinda hoping to see new capacitors when I opened this one up... but nope. one of the few parts they didn't throw at it.

    Unfortunately, I really don't have the money to drop $250 on replacing working parts... and I've already spent way too much on these welders. Now that I have two, I'll probably need to sell one. Somehow the bills keep going up while the paycheck keeps going down...

    The CDE 550C series seems to be low-ESR high-ripple inverter capacitors, with fairly good ratings. The actual miller part number is a special order with no spec sheet online, but you can read about the series at http://www.cde.com/resources/catalogs/550C.pdf . I didn't look up the UCC ones, since my unit had the CDE ones, but I'd imagine they're something similar.

    Yeah, the only issue I had was finding the right size cans. The 550C's and 550CE's I could only find decently priced in a 3" diameter can. They may be able to fit, I just never tried them.

    Leave a comment:


  • Bushytails
    replied
    Originally posted by jjohn76 View Post
    I replace the buss capacitors on any machine I keep because of ccawg's recommendation. It's better than replacing caps, the input bridge, and the switch at a later date. The older Sanrex Thermal Arcs (outside the crappy plastic case, they're awesome) use 8 each 470uF caps instead of the two big ones like the Millers. I haven't actually compared the ripple current capacities, but would guess the thermal arc's design can handle more ripple current.
    Heh, I was kinda hoping to see new capacitors when I opened this one up... but nope. one of the few parts they didn't throw at it.

    Unfortunately, I really don't have the money to drop $250 on replacing working parts... and I've already spent way too much on these welders. Now that I have two, I'll probably need to sell one. Somehow the bills keep going up while the paycheck keeps going down...

    The CDE 550C series seems to be low-ESR high-ripple inverter capacitors, with fairly good ratings. The actual miller part number is a special order with no spec sheet online, but you can read about the series at http://www.cde.com/resources/catalogs/550C.pdf . I didn't look up the UCC ones, since my unit had the CDE ones, but I'd imagine they're something similar.


    Leave a comment:


  • Bushytails
    replied
    It took me the entire Blackheart Revolution album and half of the Flesh Is The Law album to write that.

    Part two was the part about the second welder with the bad capacitors.

    Listening to industrial rock helps most repair projects!

    Leave a comment:


  • Noel
    replied
    My cable provider sucks, I want a revolution to. Yea!
    That's a five minute song and I just finished reading the post as it went into the second song on the album play list, Vampire Lover.
    Thanks BT for the lengthy reply. WOW! And theres a part two? Let me shake your hand.
    Truly, that's a WOW! I'm surprised they don't have a diagnostic port of some kind to plug a system function tool into.
    I'm saving this to my sticky file on welder repairs for future reference. Admitedly, I'm dazed in reading it. Listening to Industrial rock probably didn't help that but It was playing in the back ground. You have me thinging maybe I can. Once again, thanks.

    Leave a comment:


  • jjohn76
    replied
    I replace the buss capacitors on any machine I keep because of ccawg's recommendation. It's better than replacing caps, the input bridge, and the switch at a later date. The older Sanrex Thermal Arcs (outside the crappy plastic case, they're awesome) use 8 each 470uF caps instead of the two big ones like the Millers. I haven't actually compared the ripple current capacities, but would guess the thermal arc's design can handle more ripple current.

    Leave a comment:


  • jjohn76
    replied
    Originally posted by Bushytails View Post

    Here's the thread that I was thinking of: https://weldingweb.com/showthread.ph...MT-304-Problem

    I would imagine the boob shape indicates failure with internal arcing, and is a sure sign of a bad capacitor.
    I remember reading that thread and most everything ccawg posted. I have seen a few capacitors with an air blister on the top where gas has vented, never anything on the side of the case.

    Leave a comment:


  • Bushytails
    replied
    Originally posted by jjohn76 View Post
    Man, this is almost to a T my experience with a known bad eBay. Except yours was much better written. And... well, I didn't replace the switch until after testing the arc and finding I needed a switch... I haven't seen a boob on a capacitor yet, only signs of leakage and stinkage...
    Here's the thread that I was thinking of: https://weldingweb.com/showthread.ph...MT-304-Problem

    I would imagine the boob shape indicates failure with internal arcing, and is a sure sign of a bad capacitor.

    Leave a comment:


  • jjohn76
    replied
    Man, this is almost to a T my experience with a known bad eBay. Except yours was much better written. And... well, I didn't replace the switch until after testing the arc and finding I needed a switch... I haven't seen a boob on a capacitor yet, only signs of leakage and stinkage...

    Leave a comment:


  • Bushytails
    replied
    Previously, on Adventures In Welder Repair, we had just gotten welder #1 working... (continuation due to post length limit)



    Might as well go into the second xmt304's troubleshooting as well, since this post needs to be even longer... (and the stereo has now advanced to the next cd, flesh is the law)

    Got this one knowing it didn't work, and the guy I got it from told me it showed a help 7, but he got it that way and had no idea of its history. I never powered it up to test.

    Did the same initial process as the first one. No dust in this one, though - nice and clean! Went through the ohm/diode readings. Input rectifier module is entirely toasted in every way. Like, of the six diodes, two are good, and the scr has all three terminals shorted together. One of the IGBTs measures shorted. Hrmm, one of the capacitors does too! Visually inspect capacitor. Has a boob (sorry, someone in another thread called them that, and it's laughably appropriate) on the side and a puddle of electrolyte under it. Ah. Pull out the screws holding the pcb to the capacitors and slip a thin piece of plastic (I used two tags from potted plants, because they were handy in my driveway) between the back of the board and the terminals on the capacitors, so I can measure them separately. IGBTs now measure correctly. Woohoo! Not as many blown parts as it could be! The earlier bad reading was due to the shorted capacitor across them. Capacitor still measures shorted. Repeat the whole checklist, and everything else measures good, including the bleeder resistors often blamed for capacitor failures. Notice the plastic insulator around the power switch shows signs that the power switch likely let its magic smoke out, but the switch itself is clean. Likely failure sequence is the capacitor failed, this blew out the input rectifier module, this blew out the power switch, someone replaced the charred blob of a power switch and it didn't fix it, and up for sale it went. Ordered two new capacitors and a new rectifier module. Installed and put everything back together. I used the thermal compound that came with the new input module, but it kinda sucked - I would use Aavid Thermalcote (my favorite white goop) if I did it again.

    Now, time for the big test... Is it going to work, or are all my new parts going to instantly blow because of some other failure that I didn't detect? Used the old TV repairman's trick, and stuck a light bulb in the power line. A 500W halogen in this case. If something is shorted, rather than blowing all my expensive semiconductors, the light would just turn on. Flipped the switch, and it powers up perfectly, but switches to a help-6 a few seconds after completing the power-up sequence. Makes sense, there's a light bulb in series with the power. Re-wire for normal power, and it seems to work perfectly. Grab the leads, give it a test. Shuts down if I turn the current over 100A, but that turned out to be a legitimate issue with my power to it - rather than using my oscilloscope to monitor the various voltages while trying to weld, I just took it to someone else's house and verified it worked on their power correctly. Easy tests first!


    So, no magic... Just plodding through repairs one step at a time.


    Ok, this post is long enough. Stereo is now playing Terrorvision. Somehow it's dark out now, and I didn't get anything done I was going to do. Oh well. Hopefully this helps!

    Leave a comment:


  • Bushytails
    replied
    Originally posted by Noel View Post
    In my books your still a welder wizard. One day I'm going to find a dead one and try my luck, until then, I'm listening and paying close attention.
    Well, I'll give you some insight into my troubleshooting process (at least, what of it I can remember six months later), then, just in case it helps... *puts on some music, since this is going to be a looong post. random album: Genitorturers - Blackheart Revolution*

    First, I blew all the dust out, then went over the pcbs with a dry, soft-bristle paint brush to get the dust out from around components. Not a repair step, but it made it a lot nicer to work on. I also did a visual inspection at this time, looking for anything visibly smoked. Didn't find anything.

    I followed the pre-power-on checklist in the service manual, which checks the major power components. That's how I found the bad 200 ohm resistor on the sub power pcb - a resistance check was much higher than it should have been. I didn't have the right resistor to replace it, but I had two 100 ohm resistors of suitable power rating, which I put in series and dangled out the side of the unit using jumper wires. I didn't find any other issues with this step. The semiconductors all measured good. The most important thing to look for here is anything that measures 0 ohms, because nothing should. Use a diode check as well as ohms, as different measurements will be more useful with one or the other. Neither should ever read 0.

    Still wouldn't power up correctly. I noticed something smelled warm. Not burning, just warmer than it should be. With power disconnected and the big capacitors verified discharged, I started running my hand over the pcbs. I quickly found the heatsink for one of the voltage regulators was hot, indicating one of the power rails was being shorted. A check with a multimeter with the unit on confirmed one of the voltages was low, about 3 volts instead of the 15 volts it should be. 3 volts is much better than 0 volts, at least from a diagnostic perspective - because it means that whatever is pulling the rail down is dissipating some heat itself! A dead short, which would give a reading near 0 volts, doesn't dissipate heat - p = vi = 0 if v is 0, but if v is 3, and i is big enough to make the regulator protect itself, p is going to be noticeable. Disconnected power again, verified caps discharged again, went back to feeling for overly warm components. Only felt for about 30 seconds at a time, then reconnected power, let the unit warm up, disconnected power, verified caps discharged, went back to feeling parts. Didn't want to miss something because it cooled down. Found Q39 was hot. Did some quick reverse-engineering, since that section looked nothing like the section of the older version that I had the manual for. No real trace tracing or anything - just some figuring about how it must work. I knew it had to generate a voltage on the output, because it appeared to be the replacement for the lift-arc tig section of the schematic, and that's how the lift-arc function works. It only has one switch. That switch is capable of shorting a rail to ground. Flyback seemed the most likely topology. Any kind of push-pull would need two switches, and flyback topologies tend to have low-resistance primary windings that would short a rail if left on. I desoldered one end of Q39. Lazily, from the top side of the pcb. Lifted it up. Yay, no warm smell, +15 back up closer to normal. Q39 measures shorted even when removed from the circuit. I didn't check to see whether it failed on its own, or a failed driver circuit left it on too long and melted it.

    If you have a thermal camera, this process is a lot quicker, and you can even find less-obvious problems by seeing which traces are warmer than the others. I've actually since gotten one, for another project. A flir i3 reconfigured for i7 capabilities. But, I didn't have it at the time, and the good ol' finger worked again.

    So, now it has all its power rails, but it's still not working. Big capacitors aren't charging. Contactor is clunking. The power rails are all there, but all too low. Verified that the control transformer was being linked for 460, while I'm feeding it 230. But I hear the contactor clunking! huh? The manual has test voltages for the contactor coil. Verified it's being told to pull in. Funky. Tried shorting the current control to the contactor (pc1 has a circuit that applies a high current to the contactor to pull it in, then a low current to hold it in) to see if that would make something un-stick, only succeeding in making louder noises. Looks like I might have a bad contactor. I had the top pcb loose at this point, so lifted it up (after disconnecting power and re-re-re-re-re-verifying the big caps were discharged. see a trend here?), disconnected the coil leads from the contactor, and connected an external power source to the coil, in this case the 18V battery from my cordless impact driver that I'd been using to run the screws out. Nice clunk, but meter showed the normally open contacts still weren't closing. Time to pull the power PCB. Something like 32 screws. Found the mechanical test button on the contactor wouldn't push in. Pulled contactor, went inside and started pulling it apart. I found the interlock that prevents a pc1 failure from trying to pull in both the 230 and 460 halves at once (aka circuit breaker test mode) was jamming and not letting the 230 half pull in. It's an exceptionally crappy plastic design that lowered my opinion of miller for including it in a product. This contactor is of substandard quality. I adjusted it slightly with a file so it wouldn't hang up, checked that all the contacts were clean, and put it back together. yay, both test buttons work now. powering it with the 18v battery now correctly engaged it. put it all back together, reattached the pcbs, etc.

    Applied power, and... yay! it tries to do things! First thing I verified was that the capacitors were charging correctly. With a meter on each one, from a cold start, they'd rise over the span of a few seconds to near the final voltage, pc1's relay would click turning on the scr in the input module, they'd jump to the final voltage and hold solid. This is good! this means the power half of the unit is likely working correctly.

    The adjustments on the front don't quite work right, and the current display shows 280A even though there's nothing hooked up to the output. No output voltage unless you turn the current control way up. The hall effect sensor, that measures the output current, seemed the obvious place to start. Found it it was putting out 2.8V (and there's a 1V/100A ratio) at no current. Verified both of the power rails, at the connector to the sensor, were correct, so it wasn't giving bad readings because of bad power. OK, it needs a hall effect sensor. They were probably welding with it like this! Tried seeing if the calibration pots could at least make it pretend to work, but no response from them. With the hall effect sensor failed like this, it should still be possible to weld, but won't have current control. Also, the high readings with no actual current seemed to be pissing off the feedback circuitry, and the main inverter section was hunting, but should still be testable.

    So, borrowed the leads from my bobcat, grabbed a rod, and... Help 6. ****it. Did some checks to rule out easy, obvious input problems. Was dark, and as my workspace is the middle of the driveway, packed up for the night, and posted on here. And then someone helpfully pointed out it was under warranty, and off to the LWS it went. I re-soldered Q39 first so PC1 would show as failed and they'd fix or replace it, rather than not notice that lift-arc didn't work and ship it back to me still broken.

    So, now you know... when people say things like "determined the problem was Q39", they sometimes actually mean "I poked things until I found the one that was hot". Sounds a lot less wizard-y when you know the details, doesn't it?

    I'm pretty sure now that the help-6 was correct, and had I tried with a better power source to the welder, it probably would have welded. With no current control due to the failed hall effect sensor, I was overloading my power source. I got a second xmt304, and ran into the same problem if I turned the current up too high, while verifying the welder worked correctly at someone else's house.



    Blah! I just tried posting, and it won't let me post because it's too long. Cutting off the second half of this post and posting it as a second post. To be continued.
    Last edited by Bushytails; 10-04-2019, 10:10 PM.

    Leave a comment:


  • jjohn76
    replied
    Definitely Noel! I referenced Bushytails work to help me troubleshoot a lift arc circuit on an XMT. Relooking the Dynasty DX manual, it is the old design, but has all of the XMT PC1 subsections better parsed and helps segregate the CV control subsection, if that makes any sense.

    Leave a comment:


  • Noel
    replied
    In my books your still a welder wizard. One day I'm going to find a dead one and try my luck, until then, I'm listening and paying close attention.

    Leave a comment:

Working...
X