SmartUPS 3000 SU3000RM Won't "Cold Start" from battery

[Anonymous user]

A year ago I got a SU3000RM and didn't have any batteries to test it with. So I had hooked it up to 2 small wall wart SMPS that were in series with eachother. (30V + 18V for 48v total)  just to test to see if the unit would power up before purchasing batteries.  It cold started  doing this the first couple of times but it seems to have fried something. Possible a voltage spike from one of the SMPS wall warts. I didn't put any load on the unit at the time of testing with the wall warts because they were only rated for an amp or two.

Now the unit functions normally if you plug it in first, power it up and the power fails. (I now have batteries to test it properly)  but won't do a cold start.

Attempting to do a cold start causes the overload light to turn on an and the alarm beeps.

Any Idea where to start looking for problems? Bad MOVs?(are there any in the output circuit?) Bad diodes? Faulty FET? 

It seems to be drawing more current on start up than it should but my DMM isn't quick enough to give a reliable reading.

It is a SU3000RM J3U with a 640-0734K PCB in it.

Thanks for any advice

[Brad_C]

I'd start by querying the status registers to narrow down what the UPS thinks the fault is. Towards the bottom of this page gives you the bit definitions, and looking at the protocol will tell you how to get that data from the UPS : https://networkupstools.org/protocols/apcsmart.html

Start with registers 1, 2 & 3.

It'd also be interesting to check both the input and output voltage measurements and the power measurement to make sure they are about where they should be (Check voltages while plugged in and on mains, and the load with a known load of some kind).

None of the units I've seen have MOVs on the UPS side of the input isolation relay and that should be open for a cold start.

[Anonymous user]

Thank you for  your reply.  Sorry for the delay in writing back. Things have suddenly become really busy at work and I haven't had much free time to play around with projects. I will definitely get a serial cable built and test this out as soon as I can.  Hopefully in the next couple of weeks. I will get back to you with the results. Thanks again!

[Anonymous user]

Things ended up being more busy than I expected. I have finally gotten back to working on this UPS. Got a serial connection working with it and can see when it says there is a power failure but can't see any error in the terminal when it shows overload on start up.

Tested and works in normal power failure mode with no obvious problems.
Cold booting / starting it from battery fails every time with overload error, but with a light bulb connected to the output it flashes on for a split second. 

Tried disconnecting output from transformers before it goes through the Current Transformers on the output of the PCB. Still same overload error.

Start-up curren't from batteries doesn't seem excessive, won't blow the 4A fuse I had on hand and threw in to see.

I'm not sure how to get diagnostic info from the serial terminal on this to see what else could be the problem.

Next thing I want to try is to bypass the Current transformers (CT1 CT2) on the hot and neutral output and see if that makes a difference. Maybe just remove them and try?

Thanks in advance for any help!

[Anonymous user]

I have gotten it to cold start now. I'm not sure it is fixed yet but it seems to be. I replaced C48 a 6.3v 100uf electrolytic capacitor after studying the schematic for along time to see what might be the problem. It is in the circuit that the current transformer is part of. Just as a guess I thought it might be causing problems, especially since it is a CapXon cap. Hopefully that was the issue.

I imagine there are other bad capacitors on this board because they are mostly CapXon. So I plan to replace most of the electrolytics on it with some Nichicon caps so this unit will hopefully operate for years to come.

[Anonymous user]

The double conversion architecture provides tight voltage regulation, frequency regulation ,and zero transfer time to battery during power events.

[Brad_C · Accepted solution]

On 13/4/2021 10:52 AM, Mikey said:

Hopefully that was the issue.

Sounds like a good result. Nice work, and thanks for coming back to report. Many don't.