Isolating the Neutral

1 hour ago, Sid Genetry Solar said:

With the inverter running normally (i.e. not connected to a generator), use an AC voltmeter to check the voltage between O-L1 (output L1) and I-N/L2 (input neutral/L2).  If you read 120v, then it's setup for 120v...if you read 240v, it's setup for 240v.  (This is checking the wiring to make sure it's right!)

Next, you can check the last 3 characters of the top line on the STAT screen...it should be "SPB" (240vin) or "SPL" (120vin); this should match the results of the first test.

If these check out correct, then the inverter is setup and wired correctly.  Let me know what you come up with, then we can move to the next step.

Well that bit of info was a good thing to have. Before I went and put 240V into the input.🙂 I did your checks on mine and it is 120V on O-L1 and I-N/L2 and is SPL in the Stat screen so I know it is safe to hook mine up to 120V. So with that, I should be able to hook it up and use the ATS function without my transfer box. And as the internal emergency transfer switch in the house is a manual unit with a 60 Amp feed, the automatic transfer switch should work with it as it is rated at 63 Amps. And remove the bonded Neutral in the shed panel.

Which brings up another question. If the inverter is turned off, does it still feed the power thru to the output terminals from the input terminals?

Rev. A.1 cannot, as the relay requires power to be on in order to stay on.  If the power switch is off, there's no power available to hold the relay on.

Rev. B has the ability, as it uses latching relays (no power required to maintain state)...although this does open a can of worms with loads potentially overloading the transformer, and the inverter being absolutely clueless while the CPU is powered off.  I actually added extra circuitry to the Rev. B control board so the CPU can override the power button...so when you turn the power switch off, the CPU switches the relays off, THEN shuts itself down. 

However, it will be technically possible on Rev. B and newer control boards, for ATS to be active even when the inverter is switched off.

If you're running 120vAC into the inverter, removing a ground-neutral bond does not completely solve the issue: if you're using split-phase grid input, it still only disconnects 1 of 2 expected AC input wires.  If you have L2 connected across from the mains to the output panel, the inverter can only disconnect 1 of 3 wires--still a problem.

If you're just running L1 + N (single-phase 120v) into the inverter input, and L2 mains never gets connected--you'll be just fine.  You can even leave the ground-neutral bonding in the panel; it won't be a problem.  (Keep in mind, the inverter wiring isn't rated for 6kw @ 120v.)

That's the kind of test result I expected...so I'm really confused as to why @chg_coin's inverter is not behaving as expected.  Still trying to figure out what's going on there...

Assuming it's working properly, the inverter should throw a Input Voltage Alarm if 240v is connected to the input (with 120v config)--and refuse to switch.

26 minutes ago, CHG_Coin said:

 

Hmmm... what can I do to test?

<a contenteditable="false" data-ipshover="" data-ipshover-target="/profile/21-chg_coin/?do=hovercard" data-mentionid="21" href="/profile/21-chg_coin/" rel="">@CHG_CoinI'm sorry, something seems quite haywire with the wiring in your inverter.  Not your fault...but we will get this sorted out.

So far I have the following info (assuming the inverter is running normally, 240vAC output, no errors/alarms)...

32 minutes ago, CHG_Coin said:

O-L1 (output L1) to I-L/L1 (input L1) reads 0.062V

1 hour ago, CHG_Coin said:

O-L1 (output L1) and I-N/L2 (input L2) reads 2.85V

I-L1 to Ground reads 122.7V
I-N/L2 to Ground reads 115.6V
I-L1 to I-N/L2 reads 2.971V

Something is seriously mis-wired inside the inverter to get those numbers.  <a contenteditable="false" data-ipshover="" data-ipshover-target="/profile/8-waterman/?do=hovercard" data-mentionid="8" href="/profile/8-waterman/" rel="">@Waterman's tests returned the expected results (which corroborate with my house inverter)...so somewhere along the production line, some wires got mixed up.  No wonder it throws a Transformer Polarity error when you try connecting your generator!

Let's see here...two problems I'm seeing up front:

• why does Ground have such a potential?  There is no grounding terminal on the Rev. A.1 control board, nor any decoupling caps to the AC line.
  • What happens if you connect a 120v lamp (that you hopefully don't mind burning out?--or maybe a 240v lamp!) from either Input terminal to Ground?  Curious whether the Ground terminal has any direct wired potential on it?  Weird as the wiring seems to be right now, I guess I shouldn't be surprised.
  • What do you measure from the Output terminals to Ground?
• NEITHER input terminal has any reference to the AC Output terminals.  This is a problem; for 120v the I-N/L2 terminal should be internally wired to the O-N terminal.  For 240v, the I-N/L2 terminal should be internally wired to the O-L2 terminal.  Where in the world is it connected??!?!?

Trying to wrack my brain for a diagnostic method to figure out the problem without asking a customer to break the warranty seals and open the inverter up.........for something we didn't catch 🤯🙄.  <a contenteditable="false" data-ipshover="" data-ipshover-target="/profile/21-chg_coin/?do=hovercard" data-mentionid="21" href="/profile/21-chg_coin/" rel="">@CHG_CoinIf it comes to that, of course we won't ditch a customer for "breaking the warranty sticker"...

1 hour ago, CHG_Coin said:

O-L1 (output L1) and I-N/L2 (input L2) reads 2.85V

I-L1 to Ground reads 122.7V
I-N/L2 to Ground reads 115.6V
I-L1 to I-N/L2 reads 2.971V

STAT screen displays SPB

2 minutes ago, Sid Genetry Solar said:

<a contenteditable="false" data-ipshover="" data-ipshover-target="/profile/21-chg_coin/?do=hovercard" data-mentionid="21" href="/profile/21-chg_coin/" rel="">@CHG_CoinI'm sorry, something seems quite haywire with the wiring in your inverter.  Not your fault...but we will get this sorted out.

So far I have the following info (assuming the inverter is running normally, 240vAC output, no errors/alarms)...

Something is seriously mis-wired inside the inverter to get those numbers.  <a contenteditable="false" data-ipshover="" data-ipshover-target="/profile/8-waterman/?do=hovercard" data-mentionid="8" href="/profile/8-waterman/" rel="">@Waterman's tests returned the expected results (which corroborate with my house inverter)...so somewhere along the production line, some wires got mixed up.  No wonder it throws a Transformer Polarity error when you try connecting your generator!

Let's see here...two problems I'm seeing up front:

• why does Ground have such a potential?  There is no grounding terminal on the Rev. A.1 control board, nor any decoupling caps to the AC line.
  • What happens if you connect a 120v lamp (that you hopefully don't mind burning out?--or maybe a 240v lamp!) from either Input terminal to Ground?  Curious whether the Ground terminal has any direct wired potential on it?  Weird as the wiring seems to be right now, I guess I shouldn't be surprised.
  • What do you measure from the Output terminals to Ground?
• NEITHER input terminal has any reference to the AC Output terminals.  This is a problem; for 120v the I-N/L2 terminal should be internally wired to the O-N terminal.  For 240v, the I-N/L2 terminal should be internally wired to the O-L2 terminal.  Where in the world is it connected??!?!?

Trying to wrack my brain for a diagnostic method to figure out the problem without asking a customer to break the warranty seals and open the inverter up.........for something we didn't catch 🤯🙄.  <a contenteditable="false" data-ipshover="" data-ipshover-target="/profile/21-chg_coin/?do=hovercard" data-mentionid="21" href="/profile/21-chg_coin/" rel="">@CHG_CoinIf it comes to that, of course we won't ditch a customer for "breaking the warranty sticker"...

Well here is some more check points he can do as mine does seem to be fine. Mind you these are from my unit which may or may not be the same as either of you have.

O L-1 > O N  123V

O L-2 > O N 122.5V

O L-1 > O L-2 244.6V

O L-1 > I L-1 38.4V

O L-1 > I L-2/N 123V

O L-2 > I L-1 34.5V

O N > I L-1 4.6V

O N > I L-2/N .036V

These were taken with just an 18 Watt 120 V load on the unit.

And as you said, there is a swing in the display Voltages as compared to my meters. It also seems to be affected by temp. This morning my meters were reading 122 Volts each ( 244 combined ) and the display was reading 248 Volts. This afternoon, mine reads 245 on the meters combined while the display shows 240-241 Volts. So it looks like a Revision B board may be a good idea in the future.

11 minutes ago, Waterman said:

 

Well here is some more check points he can do as mine does seem to be fine. Mind you these are from my unit which may or may not be the same as either of you have.

O L-1 > O N  123V

O L-2 > O N 122.5V

O L-1 > O L-2 244.6V

O L-1 > I L-1 38.4V

O L-1 > I L-2/N 123V

O L-2 > I L-1 34.5V

O N > I L-1 4.6V

O N > I L-2/N .036V

These were taken with just an 18 Watt 120 V load on the unit.

And as you said, there is a swing in the display Voltages as compared to my meters. It also seems to be affected by temp. This morning my meters were reading 122 Volts each ( 244 combined ) and the display was reading 248 Volts. This afternoon, mine reads 245 on the meters combined while the display shows 240-241 Volts. So it looks like a Revision B board may be a good idea in the future.

Rev. B doesn't fix the issue either...the issue is that optoisolators are horrible for analog signals.  They have a significant (nonlinear) temperature coefficient...which I didn't realize until AFTER the Rev. B boards were ordered.  We actually had to patch some thermistors in-circuit on each control board to reduce the cold-hot voltage swing from 15v to 4v.  Rookie mistake for sure.

This will be fixed on the next revision boards (maybe I should just call 'em Rev. C!), when I switch from optoisolators to voltage sense transformers...which don't have the temperature coefficient.  Plus adding full AC input disconnection (2 relays instead of 1)...and likely auto 120/240v input switch.

EDIT: <a contenteditable="false" data-ipshover="" data-ipshover-target="/profile/8-waterman/?do=hovercard" data-mentionid="8" href="/profile/8-waterman/" rel="">@WatermanYour numbers look perfectly correct.  Really confused as to what's messed up in <a contenteditable="false" data-ipshover="" data-ipshover-target="/profile/21-chg_coin/?do=hovercard" data-mentionid="21" href="/profile/21-chg_coin/" rel="">@CHG_Coin's inverter!

2 hours ago, Sid Genetry Solar said:

Rev. B doesn't fix the issue either...the issue is that optoisolators are horrible for analog signals.  They have a significant (nonlinear) temperature coefficient...which I didn't realize until AFTER the Rev. B boards were ordered.  We actually had to patch some thermistors in-circuit on each control board to reduce the cold-hot voltage swing from 15v to 4v.  Rookie mistake for sure.

This will be fixed on the next revision boards (maybe I should just call 'em Rev. C!), when I switch from optoisolators to voltage sense transformers...which don't have the temperature coefficient.  Plus adding full AC input disconnection (2 relays instead of 1)...and likely auto 120/240v input switch.

 

EDIT: <a contenteditable="false" data-ipshover="" data-ipshover-target="/profile/8-waterman/?do=hovercard" data-mentionid="8" href="/profile/8-waterman/" rel="">@WatermanYour numbers look perfectly correct.  Really confused as to what's messed up in <a contenteditable="false" data-ipshover="" data-ipshover-target="/profile/21-chg_coin/?do=hovercard" data-mentionid="21" href="/profile/21-chg_coin/" rel="">@CHG_Coin's inverter!

Oy,  when you make the rev c boards do all the wires going to be the same if we swap them?  cuase changing the pj board I never did, cause well, the wires was all different location and so I got a perfectly new split phase 8k 12v board sitting in a box.

Well, Rev. B is laid out considerably differently than Rev. A.1...while most of the concepts are the same, a few connectors are different.  All are fully labeled--and remember, you're talking to the designer, so I can certainly help with an upgrade when the time comes.

Seems we're working with "irregularities" at the factory...an inverter on the bench right now has some wires misconnected, where it's 0.062v from the O-L1 (output L1) to the I-N/L2 (input N / L2).  This might well be something close to what's happened in your inverter, and we simply didn't catch it.

If you measure from O-L2 to I-L/L1 (or I-N/L2), what do you get?

Oops, @the-blind-wolf I mistook you for @chg_coin...all the labeling in the world isn't going to help in this case I'm afraid...!

Yes, wires will be in quite different locations.  Sorry.

So sorry for the late reply, I'm really only able to work on this on weekends with the wifey needing internet for work - 

9W Lamp load 

Inputs 
L2 - ground 117V
L1-ground 124.9V

Outputs
L1-ground 124.9V
N-ground 0.01V
L2-ground 124.6V

Inputs / Outputs
OL2 - IL1 249.4V
OL2 - IL2 239.1V
OL1 - IL1 0.055V
OL1 - IL2 2.95V

9W lamp placed on input L1 to ground turns on.

What's next?

Looking at the test results with my tired eyes (so much for the weekend being a time to relax and recharge...NOT!), it appears that your inverter is properly configured for 240vAC input, though the 2 AC input wires are reversed (which we had also noticed on another inverter from the factory).

What I'm not sure is why there seems to be such a potential from the L2 circuit to chassis ground--you most definitely should NOT be able to light a 9W bulb between any of the output (or input!) terminals and chassis ground.

In your tests, are the inverter output terminals connected to anything at all?  (If yes, we're likely dealing with that bane of the Rev. A.1 control board...ground-bonded neutrals and backfeeding power into the inverter.  If no, there's something seriously wrong with the internal inverter wiring.)  Seems there's a ground-bond...for N-ground to read 0.01v.

What's the input impedance on your DMM?  (or...what is the make/model of your DMM?  Will help quantify the numbers I'm seeing...I have a DMiotech Chinese meter with ~1M input impedance.)

Sid I'm very happy with the performance overall of the unit and this is just something I'd like to get going eventually.  Please don't feel obliged to respond on the weekend in kind to my own odd hours!  I'm well enough adjusted to the overall performance of the unit that we can figure out the details whenever optimal.  Thanks though for talking to me on the weekend, that's really special to me and part of what Genetry so unique.
 

To clarify I did not test the light bulb to chassis ground, I tested it with a GFCI outlet hooked up to my sub panel for which I needed to temporarily jumper neutral to earth ground to make it work.  The ground from sub-panel goes straight to earth via my bonded main.
 

L1 and L2 Out are attached to a safety switch, which is in the off position.  The neutral is connected to bonded earth in the main, with no cutoff presently installed.
 

Let me run downstairs and find out in the AM.

THANK YOU!!!!!

Hmm...  I think it is A.1 yes.  What's the recommended course?  Have a good night, Sid, thanks!

It's a Bolyfa 318

Thank you.

-Josef

His has a greater than 10M Ohm impedance.