TheBackShed.com

Forum Index : Solar : String Inverter service lifetime

Author

Message

GWatPE

Senior Member

Joined: 01/09/2006
Location: Australia
Posts: 2127

Posted: 05:02am 11 Mar 2009

Copy link to clipboard

Print this post

I have just repaired a Solar String Inverter. This is a SunPower 1200W model made by SMA.

The unit presented with a bang when connected to solar panels, 180VDC only. The unit had no burnt smell, and did not present any shorts. I powered the unit up on a power supply. All good at 160VDC. The unit showed 2 green ligts for ON and PV OK. There was a fault light active and the grid light was OFF. This was still OK as the AC was not connected yet. AC was then connected, and the fault light went OFF and the grid light started to flash. All OK still. As the connection established DC current was only 10mA at 160VDC. Waited the 2 or so minutes and then bang again. Quick disconnection. Assumed switching devices had failed. Major surgery to remove all filters and AC line transformer to see the main PCB. Loosened the main heatsink screws and was able to inspect all PCB components on both sides of the board and measured the switching devices. All 4 devices [not mosfets] measured the same, and no shorts. No visible signs of arcing anywhere. Was not looking good. Decided to remove the main DC bypass filter caps. 5pin plated holes, large thick copper lands. Not an easy feat to remove without damaging the PCB. First cap looked good from the outside, aluminium can with heatshrink intact and all writing clear. On closer inspection found that the mounting pins were deformed and the positive pin was loose. This would not have been good. Decided that the other cap had to come out as well. There was provision for 3 caps on the PCB. The other cap was a bit more stubborn. A bit like pulling teeth. Finally it was out as well. This cap had obviously failed. This was the source of the bang. There was a slight trace of electrolyte under the can, and this had over time accumulated under the cap. The daily cycling of the inverter had allowed the liquid to escape and build up and finally allowed the voltage to track from the positive terminal to the metal can.

here are some pics

The caps are not able to repaired as one pin on the can is loose on both caps.

Decided to replace the 2x 385VDC @ 1000uF with 3x 250VDC @ 680uF. I only intend to connect 180VDC of panels to this unit. I had to insulate the bottom of the caps and work the pins to get them to fit. Reassembly of the internals and time to fire it up again. The unit works a treat now. Very low DC standby current, but the unit consumes 11W from the grid at all times.

I suppose I am trying to work out approx service life of these type of units. This unit was replaced and deemed unservicable, so was to be binned. The problem only seemed to be caps. The manufacturer had decided to not fully populate the board, and this appears to be the weak point.

Have any readers had warranty covered, or other failure of string inverters. There is a bit of conjecture on service life expectancy. This unit is only 8 years old. I would not expect it to have been repaired normally, so I don't think this is an acceptable life for a grid inverter system where the panels are guaranteed for usually 20-25 years. I would be interested in reader comments on this.

I am not sure how you could influence manufacturers to design for a longer life.

Gordon.

become more energy aware

davef
Guru

Joined: 14/05/2006
Location: New Zealand
Posts: 499

Posted: 06:59am 11 Mar 2009

Copy link to clipboard

Print this post

Something I find suspect is the plastic sleeving appears to have crept a lot on one cap. These caps look like they or the unit has been running at elevated temperatures.

If electrolytic caps are run near their dissipation ratings (ie high ripple current) than 8 years is probably pretty good.

Most colour TV power supply faults in the mid 80s were due to dried out electros. The ESR rating is a key parameter for heat dissipation calculation when using these units.

3 caps could help if their combined ESR value is lower than the 2 original caps. I assume these 3 caps are all in parallel.

GWatPE

Senior Member

Joined: 01/09/2006
Location: Australia
Posts: 2127

Posted: 07:33am 11 Mar 2009

Copy link to clipboard

Print this post

Hi dave,

I cannot say what the service life was like. I only rescued the unit from certain fate at the tip. The caps were embedded deep within the unit, and there is only convective cooling. The caps could only really be inspected after removing everything from the cabinet. The heatsink is 12" square. I am planning on only loading to 800W initially. These caps had esr of 30 or so mohms at 100Hz. The unit is exceptionally well bypassed for the switching frequency. Pitty about the 100Hz area. The replacement caps are in parallel, and are of high temp rating [105 degree], with low esr as well. I have had the unit running today and there was no temp rise in the caps I put in. I will be monitoring my other inverters once warranty has expired. The non contact thermometers are a great tool for detecting hotspots.

Gordon.

become more energy aware

sPuDd

Senior Member

Joined: 10/07/2007
Location: Australia
Posts: 251

Posted: 07:54am 11 Mar 2009

Copy link to clipboard

Print this post

G'day Gordon,
I'd say 8 years for such a device is fairly good. As it was only caps that failed I'd give it a tick of approval. I do electronic security, and the Japanese cameras we use get 8-10 years in a sealed housing in tropical heat & rain. And by then they just need the caps replaced. We have a 12month swap out policy for anything else with an SMPSU in it.

The component access sounds tight, which is odd considering the unit is designed to occupy a wall and could be big enough to allow better access. But hey, it all depends on who made it and why.

I suppose if it was expected to last 5+ years before a failure, the manufacturer could have a service exchange agreement in place at time of sale that says they will send an exchange unit out each 5 years to swap with your unit... like a regular service on a car.
Also has the advantage of being able to upgrade retuned units to prevent field failures.

sPuDd..

It should work ...in theory

Print this page

To reply to this topic, you need to log in.