IMHO, at this point I think you might be chasing your tail. Most LF inverter builds appear to have have some slight noise at idle at random cycling occasions, some other / older designs can be quite noisy in the Toroid and chokes, you will definitely get noise from the Toroids when unbalanced loads pull more power from one half of the AC cycle / cause slight DC injection into the AC line with associated harmonic distortion.

Ok, good. Because I've been dealing with self inflicted balance issues I'm probably being a bit paranoid about there still being a fault lurking somewhere  

If I could ask another question then:

How much difference in frequency is there between 50hz and 50++? The reason being with both of the nanos I've tried the frequency is about 49.87Hz and I'd like it to be a bit higher if possible, but I tried 50++ and didn't notice much change.  Do I just need to try different nanos?

I don't know if it will matter yet TBH as the house is still running on the powerjack (50.002Hz), and it may be there's nothing using it as a reference, but at least one of my old oscilloscopes uses a ferroresonant transformer and needs a supply reasonably close to 50hz.

When you change the frequency in Setup, the change will not be acted on until the Nano controller has been fully powered down and restarted. The exact frequency depends on the Xtal on each Nano board, it varies "slightly" with each Nano  
_
Edited 2025-08-07 12:10 by KeepIS

Update:
The inverter has been working flawlessly for several months now   Although I still haven't wound my big transformer yet .

But I did have a curious failure of the nano module the other day:

I'd turned the inverter off whilst doing some wiring on the feed to the changeover switch, and when I turned it on it was showing 0v for battery voltage, and the "battery low/inverter off" message on the display.

With reference to the schematics, I traced the circuit from the on/off swt through the resistors and q10 to the A3 (Vbat) pin on the nano module and there was 2.5v on A3 as there should be, but still 0v on the display (Vcap was normal).

So I dug out one of the three nanos I'd programmed with Mike's help, dusted it off and popped it in, and the inverters running normally again   .

But I am a bit at a loss as to why the nano A3 input would fail - even if the +48v and 'off' inputs to the WG39 board were somehow seeing something other than +48v and +48/0v, there are surely other components that would be hit 1st? There was zero load on the inverter when turned on and off.

Like most dev boards, the nano has little I/O port protection so it's possible for ports to get burned by ESD.  Adding a small 220 ohm series resistor with a Zener diode to ground is a common cheap protection method for critical ports.

  mab1 said  

Update:
The inverter has been working flawlessly for several months now  Although I still haven't wound my big transformer yet

Great to hear that, I have abused the sh#t out of a number of Nano boards and never had a problem, the V-divider and the cap at the pin should protect it from any transient on the 48V to 60V input, my Nanos should have died a 100 times over, and yes I know an ESD event may not show up immediately etc, etc. Not discounting it, but not when it's plugged into the Nano Controller.

You might try reprogramming that Nano and check if it's really faulty.

I guess from time to time you get a flaky unit, the only other thing I can think of is the Ground/Common to the Nano board was disconnected with around 50V on the input, IE the resistive divider was compromised and the Nano had a path to ground via one of the other IO ports, seem unlikely with this Nano Controller board, I'm just grasping at straws here.

I be surprised if you can repeat the fault

Thanks. I guess a random failure is possible. Esd seems such a long shot on wiseguys control board, and I'm pretty sure it's not been subject to loss of 0v.

Your suggestion that i should try and reprogram the nano is a good one (but i will have to read back to remind myself how i did it):- Because when powering down, it runs off the big cap bank and despite having the 5v reg with monitor wiseguy supplied, I think the 12v switching reg does cycle on and off a few times and the possibility of corrupted code is worth investigating. The nano has your physical mods (removed components).

In that case it will not be corrupted from a bootloader glitch. I'm not sure if you just swapped the Nano out, or tried unplugging it and re-seating it first?

Just a long shot but worth checking.

  KeepIS said  

In that case it will not be corrupted from a bootloader glitch. I'm not sure if you just swapped the Nano out, or tried unplugging it and re-seating it first?

Just a long shot but worth checking.

TBH I didn't try reseating it - I was testing the voltage on the actual nano module, not the control board.  As it was reading absolute 0v, whilst Vcap, temps etc were all normal,  i was anticipating q10 might have failed short. If it was a poor connection leaving that pin floating i would have expected some random low voltage not 0v.

I guess I could connect it to a display and power it up and see what it does with Vbat.

  mab1 said  

I was testing the voltage on the actual nano module, not the control board.  As it was reading absolute 0v, whilst Vcap, temps etc were all normal,  i was anticipating q10 might have failed short. If it was a poor connection leaving that pin floating i would have expected some random low voltage not 0v.

I guess it's possible that corrupted firmware could cause the 0v reading?

To be certain it's a nano HW failure, you could reprogram the nano then test it again standalone with only a pullup resistor from the port to 5v and see if it still reads 0v?


As for potential nano HW failure causes, when q10 is turned on there is an over-current path risk the port to ground if the port was driven high in output mode (even temporarily) somehow (perhaps corrupted firmware)?
Edited 2025-11-30 04:28 by analog8484

Did the simple test - put it on a breadboard and powered up as it is. Only had a 16x4 lcd, but it shows the voltages, and A4 (Vcap) is showing a random value which drops to 0v if you ground A4, A3 just shows 0.0v even when connected to 3v3.



Will have to remember where the usbasp thingy was put to reprogramme it...

I'd put it down to flaky Nano-board and Murphy, sh#t happens