Problem with welder and Inverters

[DanB]

I used to run my place off a single Trace SW4048 inverter, and my 240V loads would run off an autotransformer.  It worked pretty well although the setup was a bit under powered for the welder.  Last summer I replaced the inverter and the transformer with two Trace SW5548's, now I've got all I could ever want.

But I have this annoying problem.  About 10% of the time, when we energize the welder the inverters shut down in error mode.  We don't have to strike an arc to make this happen, just pull the trigger(which energizes the transformer in the welder).  It's not a big deal to turn the inverters back on and try again ~ but it's annoying especially because it crashes my computer and my satellite modem.

Any thoughts how to solve this?

[Flux]

Dan, this is flux doubling on the transformer. If you switch on at the peak of the voltage all is well but if you switch on at voltage zero the first cycle tries to take the core to twice flux and it saturates causing a very large current surge.

The only thing I can suggest is that you include a bit of resistance in series with the transformer primary. Ideally you only need it during switch on and you could use a contactor to short it out when running. If the welder is designed to do this on every trigger pull then you will hit the problem occasionally. Some welders have the contactor on the low voltage side and once running it is no problem, but most have the contactor on the primary. Having another contactor timed half a second later to short the limit resistance  will cure it but is a bit of a tricky mod.

You may be able to use the very minimum resistor to stop the inverter tripping and still be able to leave it in series without affecting the welding too much, that would seem the simplest solution if you can get away with it. 

If it is a single phase mig welder with a capacitor bank you have the additional surge when charging the capacitors. When the capacitor surge and the core saturation come together it will trip.

I don't like advising people to fit capacitors across the input of devices working from inverters but if you are sure your inverter will tolerate this then that may also cure the problem but it's not quite the same as for motor starting where the issue is low power factor.

[just-doug]

if you really have more than ample power,maybe the inverter is just to slow ramping up the power.might try a bigger base load to cause the inverter to a higher power level before the welder is engaged.

[joestue]

i'd start with a time delay and a second relay.

wire up a relay to the trigger, to energize the transformer through a resistor (i'd start with 4 ohms) and a 100-250mS time delay energising the main relay/contactor/ssr whatever is in there now.
on the other hand if it already has an SSR in it, you might as well build a peak voltage turn on circuit.

[Flux]

I like that, you could then use a fairly small relay for the start. My way would need a fully rated contactor to bypass the resistor.

I doubt that it is sophisticated enough to have a ssr, if it had the makers would have probably gone for the peak voltage turn on.

Flux

[ChrisOlson]

But I have this annoying problem.  About 10% of the time, when we energize the welder the inverters shut down in error mode.  We don't have to strike an arc to make this happen, just pull the trigger(which energizes the transformer in the welder).  It's not a big deal to turn the inverters back on and try again ~ but it's annoying especially because it crashes my computer and my satellite modem.

Dan, I got dual 4024's and I weld with them all the time without a problem.  If I weld continuous the master will start the generator for Peak Load Management to help out.  But the inverters run the welder with no problem, otherwise.  But one difference is that I got a T240 load balancing transformer in step-down on the split-phase output of the two inverters to balance the 120 volt loads on the two inverters evenly.

I sort of wonder, before you modify your welder as per the suggestions, if the real problem isn't unbalanced load on the inverters?  My 4024's will deliver 14.4 kW for 9 seconds.  But if the master already has a decent load on it, and you apply a heavy split-phase load like from a welder, the master will kick out due to overload, and shut down the slave in the process because you exceeded the surge load on the master.

If you never re-configured your home wiring for the split phase system so the legs of the split phase service are balanced, I could see how the welder would very easily kick the inverters out if one is already carrying a 2 kW load, while the other is idle, and then you apply a big surge load to the 240 split phase.

The Xantrex XW has auto load balancing of the split phase service built in.  But with stacked SW's or SW Plus units, you have to use a load balancing autotransformer or the inverters will never be balanced.
--
Chris

[DanB]

Hi Chris - certainly the loads are never balanced, but my loads are small ~ so often times even the master inverter might be running 100-200 Watts of load when the other one is only searching (till the welder comes on), but I don't think this should be the problem.  And also - this only happens about 10% of the time, or less - usually everything works fine.  I prefer to run all my 'normal' loads off one inverter if possible because it's well more than up to the job with lots of room to spare.  I only have the 2nd inverter to run 240V loads - my only 240 loads are the well pump and the welder.

It's interesting you don't seem to be having similar problems though.

[RP]

when the other one is only searching (till the welder comes on)

I wonder if this is the problem.  An easy test would be to put 100Watts or so of light bulb load on the lazy inverter before welding

[DanB]

It doesnt help if I take the slave inverter off search so that it's always on.  But yes, putting a constant load on it during welding might help, easy enough to try.  Flux's comments make total sense ~ I prefer to find an easy solution.

[ChrisOlson]

Well, then another thing to look at is your battery bank and wiring to the inverters.  If the bank and wiring is not sized properly to deliver the full surge power to the inverters during surge loads, while keeping the voltage above the LBCO, the inverters will kick out on overload.

You'd have to look at the specs for those inverters, but I know mine pull 720 amps at full surge power.  If the battery bank and wiring cannot deliver those 720 amps to the inverters with the voltage above 22 VDC at the inverter input, you will not get full surge power from the inverters.

I would suppose that your 48 volt bank would have to be able to deliver half that, or about 360 amps, while keeping the input voltage above 44 to keep the inverters lit during surge at 14 kW load.  Might want to check that out, or lower the LBCO in the inverter menu if you have it set too high.  Adding dual inverters to a bank that was originally sized and wired for a single inverter might require a bank upgrade.  It took 24 Rolls T12's to keep my inverters lit at full surge power.  And that was the bare minimum.  I added 12 more, so I got 4,500 amp-hours feeding the inverters, and the bank will now deliver the required surge power with ease, even with the bank run down into the 23's on voltage.  I was not able to weld with mine until we spent $10 Grand on a new battery bank.

Our normal loads average about 1.2 kW 24 hours a day.  So the autotransformer to keep the inverters balanced was a necessity for us.  At only 200 watts load on one inverter, that shouldn't be an issue.  But I can't figure out how you can run your inverters very efficiently at that load.  The inverters are way oversized for your system then.  Look at the efficiency curve for your inverters, and they don't reach peak efficiency until you get about a 500 watt load on them.  They are least efficient at zero load, and the curve from zero to peak is very steep.  But 200 watts is less than 80% efficient.

This is the power efficiency curve for our 4024's, as an example:



You need to at least turn on the coffee maker and pull 1,200 watts because otherwise you got about the same situation as bolting a small block Chevy on a lawnmower    
--
Chris

Edit to add some information for folks who might read this:
With any given battery bank, whether it be wired 12, 24 or 48 volt, the amount of amps each battery in the bank has to deliver to meet a given load is the same.

Let's say you have four batteries hooked in parallel on a 12 volt inverter.  The draw on the battery bank is 1200 watts at 12 volt nominal.  This means the batteries have to deliver 100 amps to make the 1200 watts.  Each battery delivers 25 amps of the total load.

Hook the same four batteries in series/parallel for 24 volt.  Apply the same 1200 watt load.  Total amp draw is 50 amps @ 24 volt nominal.  Each parallel group delivers 25 amps of the total load.  Each battery in each series string delivers 25 amps to meet the load.

Hook the same four batteries in series for 48 volt.  Apply the same 1200 watt load. Total amp draw is 25 amps @ 48 volt nominal.   Each battery in the series string has to deliver the full 25 amps of the load.

So for folks thinking a 48 volt system will be superior to 12 or 24 when it comes to meeting surge loads because of less amp flow to the inverter, it is not.  The only difference between them is the size of wire required from the bus to the inverter feed studs.  Whether the bank be 12, 24 or 48 volt configured, every battery (and interconnect) in the bank still has to be able to handle the same amps without excessive voltage drop at the inverter input to get the full surge power.

I learned this the hard way - when we upgraded our old 12 volt system to 24 last year I couldn't get any more surge power from the new inverters than I got from the old 12's.  Until I put enough battery power on them.  And with dual inverters, it takes double the bank capacity that it requires with one.

[oztules]

try a long extension lead and see if that stops it.


.............oztules

[ChrisOlson]

Not to say any of these suggestions might not work.  But knowing what I know about those inverters, and living with almost the identical units here every day, there is just no way those inverters can't deliver enough power to start the transformer in that welder.  Those units are famous for being able to deliver 2.5x their rated power for 9-11 seconds, and 10% over their rated power for up to an hour.  I have measured my 4024's at 14.4 kW @ 240 volts for 10 seconds.  That's 60 amps and the breaker feeding the welder is only 40 amps.

Those inverters also have SCR's in them for soft start on things like well pump motors, air compressor motors, and so on, that can pull 600% of their rated power on inrush for starting.

If one inverter is not overloaded because of imbalance between them, then there's either something seriously wrong with the welder (and the breaker feeding it), or you're not getting enough power to the inverters.

If you have a single 48 volt string of like 1,000 amp-hour 2 volt cells feeding dual inverters, that's your problem right there.  Take note of when it happens.  If the batteries are up close to full capacity with good incoming power from wind and solar to the bus, and it seems to work OK - but when they're down a bit it kicks the inverters out - you just need more battery on them.

When you got dual stacked inverters, if you can't get enough power to them with the batteries you got, you're probably better off to just use one inverter with a step-up transformer on it so you can get the full surge capacity from it.
--
Chris

[rossw]

If you have a single 48 volt string of like 1,000 amp-hour 2 volt cells feeding dual inverters, that's your problem right there.  Take note of when it happens.  If the batteries are up close to full capacity with good incoming power from wind and solar to the bus, and it seems to work OK - but when they're down a bit it kicks the inverters out - you just need more battery on them.

Not sure I agree with you here, Chris.

I have 2T500/2 cells. They're only 500 amphour, 2V. The product spec sheet states internal resistance of 0.18 milliohms, and a short-circuit current of 11,667 amps.
Thats the ones with just 2 terminals per cell.

The same cell with 6 terminals (3pos, 3 neg), same capacity, states 16,154 amps.

I expect a 1000AH cell would be substantially higher. Cables and connectors likely make more difference than the cells themselves.
At 48V, I use a rough "rule of thumb" that it's 25 amps per kilowatt of load.
A 4kW load would be sucking 100A. 200A at 8kW. It's not even 10% of the short-circuit current of the batteries.

[ChrisOlson]

At 48V, I use a rough "rule of thumb" that it's 25 amps per kilowatt of load.
A 4kW load would be sucking 100A. 200A at 8kW. It's not even 10% of the short-circuit current of the batteries.

Well, it may not be.  But I know that it took 24 Rolls T12's to power my little 4024's to full capacity with 4/0 cables going to each inverter from the bus.

I just thought of something.  It was either last Friday or Saturday when I was doing some welding on just inverter power.  I pulled on the welder cord and it had developed a loose wire in the plug.  The two legs of the 240 shorted together in there and there was a huge POP! and sparks came out of the plug.

I was like, holy frickin' crap.  But the inverters did not kick out.  It just tripped the breaker to the welder.  And at the time, I would guess we had roughly 1-1.5 kW of other loads going when it did that.

So I don't know, but I think I'd be putting an ammeter on the line to the welder and measure what it actually draws when you energize the transformer in it.  If it's below 60 amps @ 240, which I know I can get out of my 4024's, then I'd start looking at whether or not you got enough power to those inverters.

I have 2T500/2 cells. They're only 500 amphour, 2V. The product spec sheet states internal resistance of 0.18 milliohms, and a short-circuit current of 11,667 amps.

Ross, I just went out and took a gander at one of my batteries.  It says CCA is 648 amps.  I think CCA is at 0 degrees F and the voltage has to stay above 7.2 (12 volt batteries).  In series, that's only 14.4 volts, and the inverter will immediately kick out due to overload if the voltage drops below the LBCO, which I have set at 22.0 in my inverters.  I've never loaded one to see how many amps it can deliver and keep the voltage above 11 when it's at 80% SOC.  But a rough guess would be around 100 amps.

Two of those are hooked series for 24 volt, and the series pairs feed the bus in parallel.  When you hook them in series and pull 100 amps, the series pair should still maintain the 22 volts required to keep the inverter alive during heavy surge.  The original 12 pairs should, in theory, be able to deliver 1,200 amps to the bus.  But when the inverters are drawing 720 amps during full surge, the voltage drop in cables at that temporary high amp rate is probably why it took 24 of them to get the full power from the inverters and still keep the voltage above 22.0 at the inverter lugs.  That voltage drop problem should be a lot less of an issue on 48 volt at only half the amp draw.  But DanB has bigger inverters too, and to get the full power from them will take serious batteries.

I know when I was leaving with my new inverters, when I bought them, the guy I got them from asked me just what sort of batteries I plan on putting on those.  I told him what I had for a bank at the time, and he sort of smiled and said, "you're going to need a lot more than that."  $10,000 worth of batteries later his prophecy came true  
--
Chris

[joestue]

Not to say any of these suggestions might not work.  But knowing what I know about those inverters, and living with almost the identical units here every day, there is just no way those inverters can't deliver enough power to start the transformer in that welder.  

My Lincoln 225's turn on transient worst case is as follows:

first current peak: ~200A
second: ~80 amps
third: ~60
fourth: ~50.
it takes about 10-15 cycles before the saturation is gone; 160-200ms.

At max current with the output shorted (225amps) it pulls 70-75 amps.
See attachment for an image of the full load current photo shopped onto the turn on current transient.

Edit: forgot to convert from rms to peak. 300 amps is the turn on transient, when behind at least .15 ohms of supply resistance. meaning the 240vac line drops to 200 volts during the first 8.3ms.

[rossw]

Ross, I just went out and took a gander at one of my batteries.  It says CCA is 648 amps.  I think CCA is at 0 degrees F and the voltage has to stay above 7.2 (12 volt batteries).  In series, that's only 14.4 volts, and the inverter will immediately kick out due to overload if the voltage drops below the LBCO, which I have set at 22.0 in my inverters.  I've never loaded one to see how many amps it can deliver and keep the voltage above 11 when it's at 80% SOC.  But a rough guess would be around 100 amps.

Ahh. We're not comparing apples with apples though.

You said a 1000 amp-hour single-string. That implies each cell must be 1000 amphours, be it at 2V or 6V or 12V.
I can't imagine a 1000 amp-hour 12V battery having a CCA rating as low as 648.
648 CCA would be (my WAG) about an 80-100 amphour vehicle battery. 

That voltage drop problem should be a lot less of an issue on 48 volt at only half the amp draw.  But DanB has bigger inverters too, and to get the full power from them will take serious batteries.

I didn't consider my 500 amphour bank to be "serious" batteries either, but they would easily deliver way over 5,000 amps from a single string. I sure as hell wouldn't want to be there when it happened though!

I know when I was leaving with my new inverters, when I bought them, the guy I got them from asked me just what sort of batteries I plan on putting on those.  I told him what I had for a bank at the time, and he sort of smiled and said, "you're going to need a lot more than that."  $10,000 worth of batteries later his prophecy came true  

Yup, absolutely!   

[ChrisOlson]

My Lincoln 225's turn on transient worst case is as follows:

That's the same welder that I power up here every time I use it, with not a single problem with dual 4024's.  Mine is an older one with the copper windings in it, but that shouldn't make any difference.

DanB, have you looked at these settings in your inverters?







22.0 is where I keep mine for LBCO.  With a 15 minute delay.  But I'm pretty sure I read in the manual a long time ago that the reason you can set it as low as 16.0 volts is for surge power.  Under the full power rating, the LBCO delay is active.  During surge the inverter will kick out immediately if voltage drops below LBCO.

16.0 on a 24 volt system = batteries wrecked.  So I've never set it below 22.0.  My bank is at 24.8 right now and just for giggles, when I just snapped these pics with my cell phone, I set both of them at 24.8, then rushed out the shop and turned on my welder to see what happens.  You might know it started the welder right up, so I guess that didn't really prove anything.  But I got SW Plus inverters, where yours are just straight SW's.  So I have to push the "SAVE" to save the settings, and your inverters don't have that.  I didn't push "SAVE" when I did that so maybe the setting wasn't active until I "SAVE" it.  I don't know.

At any rate, it's something to try.
--
Chris

[ChrisOlson]

I can't imagine a 1000 amp-hour 12V battery having a CCA rating as low as 648.  

No, no, no.....

I was just using the 1,000 amp-hour 2 volters as an example.  I think DanB told me once he has some old telecom batteries, so I assumed they're 2 volters.

My batteries are T12-250's - you should be able to find them down aways in the chart there:
http://www.rollsbattery.com/content/specifications-renewable

I got the 12's because that's what the guy at the RE place where I got my inverters from recommended for dual inverters on 24 volt.  He had a bunch 2 volters there but he said I'd have to buy two strings for dual inverters, and then I would never be able to pull a battery off the bank without the power going out.  So he ordered the T12's for me and it took a month to get them.
--
Chris

[rossw]

I can't imagine a 1000 amp-hour 12V battery having a CCA rating as low as 648.  

No, no, no.....

I was just using the 1,000 amp-hour 2 volters as an example.  I think DanB told me once he has some old telecom batteries, so I assumed they're 2 volters.

I think we're banging on about the same thing, and on the same side here

You said "If you have a single 48 volt string of like 1,000 amp-hour 2 volt cells feeding dual inverters, that's your problem right there."
That's what I was doubting.

If he has an old bank of ex-telecom batteries that are a bit old and tired, probably pretty soft, then yeah. I'd agree with your assessment.
But I don't think I'd call them a "string of 1000 amp-hour cells" if they were in that condition.

[ChrisOlson]

If he has an old bank of ex-telecom batteries that are a bit old and tired, probably pretty soft, then yeah. I'd agree with your assessment.
But I don't think I'd call them a "string of 1000 amp-hour cells" if they were in that condition.

Well, yeah.  I think we're on the same page   

I guess I was going by what I thought DanB told me he had for batteries.  And that was quite some time back.  Assuming I remember correctly, and I think he told me he got those inverters at the Energy Fair in Custer, I sort of wondered at the time if he had enough battery power to run them   

So that's where I was coming from on that deal.

But those old telecom batteries are tougher than hell, usually.  If they are a little soft, just give 'em a good boil at 64 volts for a couple days and they'll probably come back to just like new again (if you can stand the stench of acid in the air)   
--
Chris

[thirteen]

I'm sure you have looked but is it a possible connection inside the welder becoming loose where it heats up and then works fine for a long period of time. Could there maybe aluminum wire feeding the welder. It has a tendency to get loose because it will heat up and swell up  then cool down and after several times it might become loose. Just a passing thought.

[SparWeb]

You guys might be right, but I have a completely different theory.



Dan has wired the DC from separate strings of batteries to two busbars (pos and neg)  The big wires run up from there to the inverter.  Not a lot of effort was taken to keep the wire runs close to eachother.

I have a client at work that needs lots of DC power but also has sensitive magnetic survey equipment.  They have a lot of practices to keep the DC from creating magnetic fields, which it can do if the current "loops".  Not only does a current loop create a magnetic field (that deviates their magnetics) but it also is an inductance loop, which prevents their inverters from rapidly ramping up the DC.  Their solution is to run all DC wires in single complete out-and-return circuits, ground their circuits close to the breaker panel, and twist the runs of wire along the whole length, too.  Their inverters were used for some high-power equipment, too, and when it switched on the inverter would trip its breaker, no matter how high the breaker was rated or how heavy-duty their inverter was.  That was how they discovered that their wire runs had to be bundled.  They had been ignoring the magnetic spike on the instruments when the equipment started (or tried to start), but then they put 2+2 together and realized that the inductance causing the magnetic pulse was also causing an impedance on the inverter's DC supply. 

Dan might be experiencing a similar effect.  The welder turning on is a rapid spike in load, much faster than 60Hz.  The inverter needs to get the corresponding DC from the batteries, but the inductance loop poses an impedance on the battery wires during the ramp up.

I seem to recall reading a product bulletin on the subject from Xantrex, but I haven't been able to find it.  If you read the Xantrex SW5548 installation manual (note: the pre-2001 manuals do not include this warning) you will find a caution about the wires being taped together.  This could be the reason for that caution:  some users may have encountered this problem in the past and called Xantrex about it.

If you don't happen to have your 5548's user's manual any more, good luck to ya finding it on the "Schneider" website.  I don't know if anything is even there anymore, or if they have re-written everything all over again.  I have a copy on my home computer so I could e-mail it to you, if you need it.

The solution, if this is the cause, if to move the busbars closer to the batteries, and re-arrange the wires so that there is as little space as possible between the heavy cables.  Then twist together or tape together the runs of wire where they are parallel to each other.

[ChrisOlson]

If you don't happen to have your 5548's user's manual any more, good luck to ya finding it on the "Schneider" website.  I don't know if anything is even there anymore, or if they have re-written everything all over again.  I have a copy on my home computer so I could e-mail it to you, if you need it.

If you don't have a manual for a Trace Engineering or Xantrex inverter, or any other equipment (GSM, ALM, Autotransformers, etc) you contact Schneider Electric's RE division Customer Support and tell them what model equipment you have and what manuals you need for it.  They will email you a link to download the manual for it in PDF format.  The link stays active for one week from the time the email is sent.

I needed the pinout on the jack that goes to the GSM on my master inverter, which is usually authorized dealer-only information, and they even sent me that.
--
Chris

[DanB]

Interesting discussion so far and thank to all who replied.
My battery is 'tired' - its a 72V Forklift battery.  That said though, the voltage hangs pretty well and they it is behaving nicely - it's not low voltage cutoff that is the problem.  The inverter that shuts down says 'overcurrent' ~ I find that hard to believe though.  The wire hooking it all us is probably not quite up to par for teh 'full' load of the inverters (especially peak load) however, it's a short run and I never ask for much from these inverters (they're overkill).

You talked about efficiency Chris and I know that's an issue with these inverters, it's why I only run 1 inverter and leave the slave on search.  My 'phantom loads' (stuff that is always on) adds up to about 200W.  Usually we use a fair bit more I guess (if we're awake).  In any case, I rarely have issues with low batteries - I have a lot more coming in than I can use especially this winter.

I'll try some of the easy ideas here first.

[ChrisOlson]

My battery is 'tired' - its a 72V Forklift battery.

How do you get a 72V battery to work?  Isn't the max input voltage 68 on those inverters?
--
Chris

[DanB]

sorry Chris should've been more clear.  It's a 78Volt battery for some sort of fork lift, or mining equipment....  made up of 2 volt cells.  I 'tap' 48Volts worth, the other 24 are either dead, or being 'preserved'.  In the last year I've had to remove some cells and tap other ones, but it's being quite good lately.

[ChrisOlson]

OH.  Below is the chart for those inverters.



Full surge power on two is 360 amps.  The 5548 does not have any more surge capacity than my 4024's - stacked configuration of either one is 78 amps @ 240 volt, or 18.7 kW.  I know my inverters will deliver 14.4 kW (that I've measured) for up to 9-10 seconds.  But they would not do it before I replaced my batteries.  If Kristin was running hot water in the bathroom sink so the water heater is on, then fires up her Super Turbo Pro Megawatt by Revlon® hair dryer, then the 'fridge decides it's now time for Defrost Mode, the freezer comes on, plus a bunch of other normal stuff going - and now the pressure in the pressure tank drops and the well pump fires up.  That was the formula for the dreaded 05A Error - Over Current.

24 brand new Surrette T12-250's and it laughs at it.

Still got your autotransformer?  Unplug the stacking cable, wire that transformer up and try it with just one inverter using the transformer.  I don't think you got enough power to feed two of 'em with a tired battery.  All it takes is one lazy cell.
--
Chris

[DanB]

You're right definitely, my battery is not up the task of providing these inverters what they could want.  But it's definitely up to what I need/use.  I don't think the battery is the problem here (could be wrong).  That said... my welder runs so much better off both inverters providing 240V than it did when I used the autotransformer - I don't care to go back to that.

[ChrisOlson]

I was suggesting it, not as a permanent solution, but merely as a test to see if it reliably fires up the welder every time.  If you can start the generator and run the welder fine, then you're barking up the wrong tree changing anything in the welder.  A 6 kW standby gen is not capable of the surge power of just one inverter.  But if you got a DC problem you can get more surge power from a single inverter than you can two of them.

That's just all the things I would've already tried the first time it did it.  I don't know enough about the fine points to stick capacitors or relays on the welder.  Where I come from when a breaker kicks out or something quits you either got a short or you ain't got enough power.

Edit: if the problem ends up being not enough power, then there's a simple solution - do what I did before I spent the money on new batteries - start the generator on Peak Load Management (so the internal chargers aren't operating).  When on Peak Load Management, you have the combined surge power of both the generator and inverters and the inverters will "blend" the power from the gen with their output.  Just remember that you have to set the max input amps in the menu for what the gen can deliver during surge to prevent the gen breaker from kicking out or having the master inverter spit the gen off because of gen overload (freq below 55 Hz) during surge.
--
Chris

[Flux]

I don't know anything about the inverter in question, but inverters usually have a defined current limit beyond which they trip. Generators just go flat.

Using one inverter with the auto transformer would probably avoid the start up shut down as the auto transformer will have enough resistance and leakage reactance to absorb the welder transformer current surge.

I think the trip issue is probably an instantaneous thing that the inverter objects to. The surge from a big highly fluxed transformer can be very large, depending on the type of core. The initial surge is only one half cycle and then it falls rapidly, but if that exceeds some absolute limit in the inverter it will shut down.

I suspect poor batteries or poor wiring may make things worse, but that depends on the capacitors on the input of the inverter.

Genara tors just lop the peak off any transformer surge. Even the mains usually dips a bit due to local wiring and takes the sting off.

Someone suggested a long extension, this is not a good thing to have with welders, but perhaps just enough to stop the tripping may not be enough to interfere with the welding.

I wouldn't expect a welder transformer to be the worst device, toroidal transformers tend to be the worst offenders. I have measured a surge of 200A when switching on a 30A variac on a not particularly stiff supply.

I don't know what current sensing there is in the inverter, it could possibly be influenced by stray fields from loops in the battery cables but it seems a bit unlikely.

Flux

[DanG]

I know swapping out 160-pound inverters is a PITA to see if the intermittent problem is repeatable with a different source - but I'm wondering if the inverters age and component value 'drift' on the watchdog circuit may have lowered the threshold...

[ChrisOlson]

I don't know anything about the inverter in question, but inverters usually have a defined current limit beyond which they trip. Generators just go flat.

The inverters in question will trip out on overload if the input voltage drops below LBCO during surge events.  The reason is because say the voltage drops to 35 volts and you need 14 kW surge.  At 35 volts you need 400 DC amps to meet surge, which is over the inverter's max input to the big primary transformer that steps the DC battery voltage up to high volt DC.  The minimum input voltage to meet surge has to stay above 39, and if the battery can maintain that @ 360 amps the inverters will put that full surge power for 9-10 seconds.  Few battery banks can do that.

You have a better chance of making it happen with just one inverter that you do with two of them, if you're short on raw amp capacity from your bank.

The guy I bought my almost identical inverters from has been selling and servicing Trace/Xantrex inverters since they first built them.  He told me that with dual inverters you must have dual battery strings or it won't work.  But they have to have a common ground and positive point (bus) that is close to the inverters to keep the cable run from the bus to the inverters under 6 feet with 4/0 cables feeding the inverters.  Without using a bus, the balancing (parallel) interconnects between the battery strings can be smaller than the cables feeding the inverters.  But they have to be common so the internal chargers work right when you have stacked inverter configuration because the master controls the slave for output current to battery.

Cobbling two inverters on to one battery string will not even come close to working.

Using one inverter with the auto transformer would probably avoid the start up shut down as the auto transformer will have enough resistance and leakage reactance to absorb the welder transformer current surge.

What the transformer does is take inverter output and make a mirror image of it with the sine wave 180 degrees from the input.  The two are combined, using the inverter's output as one "leg" of the split phase and the transformer's output as the other "leg", 180 degrees out of phase.  You get 240V between the inverter's output and the transformer's output, and 120 from either one to neutral.
--
Chris

[joestue]

Not sure where you're trying to go with this Chris.

390 amps is how much his welder draws from a 240vac line.. 1 out of 10 times.
Obviously neither the batteries nor the inverter can supply the power. Whether the voltage drops on the battery or it is dropped in the switches and the transformers in the inverter is immaterial, and the current won't get that high anyway because the impedance of the inverter and battery is too high. so the inverter trips out due to either low output, low battery, or too much current. its that simple. it is not possible to fix the problem with a stronger dc supply, and if you want to fix it with a bigger inverter then you need a 50kva inverter with a 100kva surge rating.
 
All the auto transformer does is provide enough impedance to keep the inverter from tripping out. the extra .5 ohms is probably all it takes to solve the problem.

if you are curious why it works for you, here's one reason.

The two legs of the 240 shorted together in there and there was a huge POP! and sparks came out of the plug.[...] It just tripped the breaker to the welder.

Another reason is that not all transformers are the same.