Trace SW5548 Problem - Which Board/What Controls the Inverting?

Hi Everyone,
I have been having troubles with my SW5548UPV inverter (Grid tie batteryless inverter outputting 240VAC) and looking to all you fine folks for some help. I have not seen anyone discuss the UPV versions, but I believe they operate the same and according to the manual, the only differences are software and minor hardware differences.
My question is whether anyone knows which part of the unit actually controls the inversion process? Just so you know, I am not an electrical engineer but know enough to fix some things but could also get myself into trouble here so please bear with me.

Short story: I have been trouble shooting this unit which comes online with DC power, locks onto the grid AC1 is fine (solid green light), however it never begins the inversion process (does not hum at all) and sits there with the line tie LED blinking orange. There have been a couple of times when the inverter begins, but then shortly afterwards the LCD display will start flickering and and the normal hum of the inverter sputters and stops and the line tie LED goes back to blinking orange. I think this has only happened maybe two times out of the 10+ times I have powered it on. Right now there is no other signs of any issues other than the unit just sitting there (mocking me). My thought is that whatever controls the inversion process is malfunctioning, however I have no clue what controls that. Is it done on the main control board, or the power MOSFET board?

Long story: I bought this unit about 2 years ago and it was online for a couple of months when it stopped working. The inverter was not outputting to the grid and when I checked the DC voltage, it appeared to be sinking massive amounts of power, enough to bring the panel voltage down below 24v on a 48v system without outputting anything. I took the unit offline knowing there was an issue and since then I have been running a smaller SW4048UPV without any issues. Until recently, the SW5548UPV has just been sitting in my basement as I had no real need to get it back online. I have since acquired more panels and would really like to get this beast back online so I pulled it apart and did some more testing. The positive and negative terminals appeared as shorts and in taking the MOSFET board apart, it made sense that there were likely bad MOSFETs, though there were no signs of burnt FETS (I think there are people on this forum that may have experience with replacing FETS and whether there could be other potential issues when these go bad that I don't know of or tested). I pulled a few of the MOSFETs off of the board and did a quick test on them and sure enough, some were complete shorts while others were OK. I replaced all of the power MOSFETS (52) which were originally part number RFP40N10 with IRF3710 MOSFETs. Specs are a bit different but seemed comparable (better RDS, higher current). I got everything back together, applied power and a row of FETs on Mama bear went up in smoke. I think there may have been a bad FET (they were from China) so I replaced the bad row, made sure there was no shorts, and now the unit powers on and everything displays correctly without sinking massive amounts of power, but like I said the inversion process does not even start. I know on my SW4048UPV, when you apply DC power it starts humming right after it powers to generate the sign wave before trying to sync with the grid even if it has not locked onto the grid, so I don't think it has anything to do with the grid is not even beginning the inversion process. I'm lost at this point. My thought is that the inversion is controlled by the main control board, however I was able to secure a control board (albeit a different model) and switched the 5 main chips and it still acts the same. Not sure whether the mishap with the mama bear MOSFETS could have fried something else, but a visual indicates no black marks anywhere else on the power MOSFET board. I think whatever controls the inversion process is not even attempting to (as if there was some fault it detects) because scrolling through the Tech Menu, it clearly shows "Inverting - NO". If it showed YES for inverting, but was not actually inverting then I would conclude there may be a hardware issue somewhere. None of the normal faults occur and no other LEDs light up and it says there are no errors.
One last question I had is whether anyone knows what each of the 5 main removable chips on the control board do? The extra control board I have is not for a SW5548UPV, but if I knew which chips contained the program specific to the SW5548UPV, I could try swapping out the generic chips that I could without doing more damage to see if maybe one of those chips is bad.
I'm sure there is something I didn't cover which may be useful in figuring this out so please ask away if you think there is something I missed that may be relevant (I'm sure there is something important that I have not thought of putting in this thread).

Any help or suggestions would be extremely helpful!


  • Cariboocoot
    Cariboocoot Banned Posts: 17,615 ✭✭✭
    Re: Trace SW5548 Problem - Which Board/What Controls the Inverting?

    Welcome to the forum.

    I'm going to suggest you shop for a new inverter. For one thing they don't really have a lot of user-serviceable parts inside (you may have noticed). For another it's old (it was old two years ago when you bought it) and getting parts to service these old SW units is problematic. A few places offer to service them, but how long before it fails again?

    Unlike the battery-based SW's where the #1 problem is a sticking AC relay when one of these GT units dies it's usually the FET's and drivers alike, if not the microprocessor control as well. In other words cascade failure (which you've seen) and you end up replacing everything.

    Which is pretty much the same as buying a new inverter, only without any warranty.
  • boB
    boB Solar Expert Posts: 1,030 ✭✭✭✭
    Re: Trace SW5548 Problem - Which Board/What Controls the Inverting?

    Isn't this the battery-less grid-tie version of the SW5548 that was used for the SMUD project (Sacramento Minicipal Utility District) in the mid 1990's ?

    I don't remember a lot about them but the battery terminals would connect directly to the PV, with a large external computer
    grade capacitor, and it would start selling back to the grid and try to track the input voltage to generate maximum power output.

    What I forgot was if it could operate as a regular inverter but you said it did before it broke so I guess so.

    3 of the 5 ICs are old Microchip (PIC) 16C57 microprocessors which are OTP (one time programmable).

    One handles the inverter waveforms (fast pic) and pulls those pre-programmed waveforms for varying DC input voltages from another one
    of the 5 IC's which is a window-less PROM... 2732 I think...

    Another micro (slow pic) handles the charging algorithm, gen start, A/D reading, voltage regulation etc.

    Third micro (Display Pic) handles the LCD and pulls out the menus from another plug-in PROM
    and sends that up to the LCD.

    That's what I remember from 15+ years ago anyway.

    If you get it to stop smoking and your extra PCB is for a 48V SW inverter, it should work if you swap it in.

  • teamhyd
    teamhyd Registered Users Posts: 2
    Re: Trace SW5548 Problem - Which Board/What Controls the Inverting?

    Hi boB,
    Thanks for the reply. Yes, this is one of the inverters from the SMUD project (mfg in 96') and because it's a bit unique I would really love to get this online especially because of my set up. The PV array does connect directly to the battery input and yes, there is a large capacitor on the input. So it is basically a 48VDC in, 240VAC (split phase) out, batteryless grid-tie inverter which is hard to come by. And I don't need to stack two to get the 240VAC to feed into the house/grid. The only other inverters I can find to replace this thing are those Chinese made extruded aluminum ones that will likely only output half of their rated power efficiently. Even the hydrid ones which take 48VDC in (outback, Xantrex XW) require at least 100AH battery capacity to function properly.

    So I got the power MOSFET board to the point where it keeps the magical smoke in and like I said before, everything powers up and seems to work fine except it does not start the inversion process. I did not understand what you meant by:

    "What I forgot was if it could operate as a regular inverter but you said it did before it broke so I guess so."

    but this IS grid tie only and the way it is suppose to operate (which I imagine is simialar to a regular SW unit) is that is powers up on DC input and immediatly it will start to generate the sine wave and will search for AC1 in. If it gets a good AC input, the green light comes one for AC1 good and it will begin the sync process. I see the AC1 voltage cycle up and down as it will tries to get a lock and once it does, the large output relay engages and connects the inverters sine wave withe grid and starts feeding back into the grid. The issue I have is when I power on DC input, the inversion process does not even start (no humming). If I scroll through the "meters" menu, it is verified that there is no AC Volt output from the inverter as it displays 6 to 10V which is likely just random noise on the display. I can connect the AC1 input and it will get a lock and will sync, but because there is no AC Volt out, the relay never engages. So that's where I'm at, trying to figure out what tells the drivers for the MOSFETs to start working. Is this inversion process controlled via one of the replaceable fast pic chip on the control board? Is this even controlled by the control board or is it done on the MOSFET board? Could the MOSFETs that fried have taken out the drivers on the MOSFET board or does this sound more like a waveform control issue?

    Are the 48V SW control board unique, or is it just the 5 swappable ICs unique? The control board looks like it only gets low voltage through the ribbon cables, does the MOSFET board step the DC voltage down to 12V, thus all of the control boards are the same other than the 5 ICs? The extra board I have is from a 2512 (so 12v version), however the part number on the board is the same and looked almost identical so I swapped out the 5 ICs and like I mentioned previously, it acted exactly the same so that board seemed to work which makes me wonder if the original board is also fine and whether one of the ICs may be bad and just not initiating the inverting process. This at least suggests to me that the control board are the same. The only difference I noticed is that instead of showing 52VDC in on the display, it showed like 76VDC which I'm guessing is due to a slight difference of the boards...but the fact is powered up and did the same thing as the original board leads me to believe the actual original board is fine, but maybe one of the ICs is bad.

    Attachment not found.
    Can you identify which of the ICs on the board may be compatible from board to board (i.e. if I used the chip from the 2512 board) to see if maybe the chips are bad? The stickers had come off of all but two of the chips. If I was to identify the chips from left to right, the lower left chip that still has a sticker would be #2, as #1 would be 3 chips to the left of #2 and is missing it's sticker. #3 is in the center of the picture and has a sticker, #4 is just to the right of #3, and #5 towards the right hand side has remnants of a sticker still, but no numbers. If I could figure out whether any of the ICs from the 2512 board are compatible with the 5548, then I could try those chips in it's place. My guess is that I would have to at least keep the chip that has the main program difference with the UPV version otherwise potentially more smoke. Anyone have a control board from an SW5548 or have any chips from an SW5548 control board? Maybe I am barking up the wrong tree, please let me know...maybe it is not the control board or the chips at all?

    I have to hand it to you, you have a great memory considering the details you are providing. Please keep posting, it is the only way us newbies will gain this information in hopes of keeping these things alive. As Cariboocoot said, replacing it is likely the most logical choice (which may happen), but I would really like to get this going considering my options and it would be nice if I can revive an 18 year old beast.
    Thanks for the information. If you can think of anything else or if anyone else reading this has suggestions, please post!
  • boB
    boB Solar Expert Posts: 1,030 ✭✭✭✭
    Re: Trace SW5548 Problem - Which Board/What Controls the Inverting?

    Yeah, I'm not sure exactly what I was trying to say in that one sentence ! I haven't been drinking, I swear !

    Well, any of the chips that are NOT in a socket are all compatible and I ~think~ maybe the auxiliary power supply on the
    left on that picture should be good also.

    I can't remember if the all of the processors need to be changed or not ?

    I wasn't sure if you had applied AC grid to this guy or not but it should work OK if you are trying to grid tie it.

    I thought maybe that you were trying to get it to work as an off-grid inverter which could be done if the right
    chips were replaced.

    I will see if I can dig something up to jog my memory.