My install and review of the AIMS Prius 2kW Pure Sine Wave Inverter for Backup Power Generator

Yes, I do not want to connect mine again until the ground to the negative terminal is removed. I do not want to depend on having to keep the unit isolated from ground. As you said, there are too many ways to create an inadvertent ground.

With the B- grounded, if you were to get another ground on the B+ terminal you would have a short circuit path that could create some really high currents.

 

It really bothers me that the SGII did not display a DTC code. It would let me clear the yellow warning triangle but only displayed No Pending Codes.

Am I correct in thinking that even though it was the yellow warning triangle (and not the dreaded red triangle) symbol and the message Check Hybrid System that a DTC should have been triggered and displayed by SGII?

I suspected a ground fault like Bob but it sure would have been nice to look up a DTC and have some sort of troubleshooting path to follow.

If I am correct I need to send an email to LinearLogics and ask why the SGII did not display the code.

 

AC or DC?

If DC, it suggests they have a 'center tap' in the traction battery and this provides the Ground Fault signal.

Bob Wilson

 

Excellent!

Now this sounds like a serious, ground fault detector signal. It 'could' be noise but that would take an oscope.

Bob Wilson

 

I received permission from the supplier to remove the cover to the inverter and look for any obvious reason for the DC- connection to ground. I could not find anything obvious and without a wiring diagram it would be difficult to troubleshoot.

I sent an email asking for a wiring diagram. I also explained that I cannot use an inverter with a negative ground with the Prius.

 

I will try to bring an oscilloscope home this weekend and see what the AC signal looks like.

 

I will let you know what Jack Chen has to say in his response. I checked my disconnect plug for voltage balance to chassis ground but did not think about leakage current. I will try mine tomorrow to give us a comparison.

 

Walking the dogs, I figured out what may be going on:

The typical way to generate an AC signal is an "H" circuit. It operates in these states:

1. Positive cycle
   • S1, S2, S3, S4 OFF - this is no power applied to the load
   • S1, S4 ON, S2, S3 OFF - this applies a positive voltage to load
2. Negative cycle
   • S1, S2, S3, S4 OFF - this is no power applied to the load
   • S2, S3 ON, S1, S4 OFF - this applies a negative voltage to load

So the computer or logic circuit simply passes through these states changing the duration of the time OFF and ON_OFF, to create a set of steps that follow a sine wave. But then look at a standard AC power source:

Normally, "Neutral" and "Ground" are at the same potential. But in our case, the "H" is switching B+ and B- with the default likely to be B-. But it doesn't matter to the ground fault circuit which one is attached to the case and vehicle ground.

I suspect if we carefully look at the B+ and case, we will see it periodically show up on the case. Also the B- will periodically disappear. Both would occur at the 60Hz generated signal.

MITIGATIONS

Isolate inverter chassis, insulated 'box' with air-flow, and pass only "neutral" and "ground" to power plugs external to the insulated box. This is fairly simple but assumes no AC loads have an independent 'ground' that can reach back to the car to return the ground-fault signal. In effect the inverter and loads are kept isolated from car ground. Fortunately, the car is on four, air inflated insulators and wearing shoes in dry weather minimizes the 'via driver' return path.

Chokes on B+ and B- leads high enough inductance to block the ground fault signal yet low enough DC resistance to not add too much heat. The problem is finding the right size chokes for a poorly documented, ground-fault signal. The reported 5 Mhz signal from "New Car Features" needs to be better qualified.


At first the toroidal isolation transformer on the traction battery input side had me a little confused. But I suspect it now serves to pass the DC and block the ground fault detection signal plus any EMI from the buck inverter.

Bob Wilson

 

Thanks for the drawings. I was going to do that last night but the NCAA championship ball game was on

With the AIMS unit being a pure sine wave inverter you would need something after the inverter output to smooth out the waveform. There is a fair size transformer on the inverter board but I could not follow the tracings from the top side to see the connections. It appears to be on the DC input side of the board and not the output side. It may be like the toroidal isolation transformer that you indicate above. Without a wiring diagram it is hard to tell.

I would prefer to isolate the DC side from ground within the inverter but that will depend on AIMS to provide the isolation. I just do not like the idea of having to provide the isolation between the case and ground.

 

Having more time to read your post above during lunch, I did not read the DC- to ground while in operation. I have only read this with the inverter disconnected. I disconnected the inverter as soon as I realized that it was the source of the ground . I may try this again with the inverter isolated from ground while testing the HV battery DC with the oscope.

I started to take a picture of the inverter with the case off and post it but did not want to without the supplier's approval. I remember tracing the wiring from the inverter output to the 120VAC receptacles on the front of the unit. The yellow/green wire from the inverter went to the case ground terminal and to the ground terminals on the receptacles. The L (Hot) and N (Neutral) wires went directly to the back of the receptacle block on a soldering board. I do not remember seeing a connection between N and ground, which is what you would expect to see. You would want to determine when and where to connect N and ground on the load side.

If this is correct, and with the inverter disconnected from a power source, where would the ground connection to the DC- terminal be coming from?

Note: The Hot terminal is actually on the left side of the receptacle picture above, the longer of the two slots

 

Sorry to sound like a worry wart, just watch the Oscope ground and signals. Don't want the Oscope signal to be jumper to ground (guess how I learned THAT lesson!)

So far, the AC side sounds good.

Speculation on my part but powered off, the B- may default to case.

As for the outlet, I'm looking at a local power strip and the photo matches the power strip. The same is true for the office wall outlets, one has the horizontal on the corresponding "Netural" and the others look to have same orientation and size as "Neutral" and "Hot." This is getting a little weird.

You wouldn't have one of those AC testers they sell at hardware stores?

Bob Wilson

 

Good p0int. I learned that lesson the hard way also.

I would like to see a drawing of the unit and know for sure.

This is what I get for being a smart*ss I just checked the wall socket in my office and you are correct. The longer slot is indeed the neutral. Sorry for getting weird on you.

 

I checked my measurements tonight with my multimeter with the exception of current. I remembered why I have not used the current portion of the meter in quite some time. I need to get a replacement fuse !

With the car OFF and the battery DC output disconnected by the main relays, I read from DC+ to chassis ground 2.31 Megohms. DC- read 2.01 Megohms.

With the car READY I read DC+ to ground +108 VDC and DC- to ground -108 VDC.

Also on the AC scale I read ~27 VAC on both.

 

Did you remember to call AIMS today? I have not had an email response from Jack. I may try to call tomorrow if I do not hear anything tonight.

 

Ok, this is much better although the NHW11 Vol. 1 calls out for 10 Mohms.

So it sounds like this ~27VAC may carry the ground fault signal. Oscope may show just garbage.

Bob Wilson