Wiseguy New Inverter Build Nano R6

Author Message


Joined: 21/06/2018
Location: Australia
Posts: 1102
Posted: 09:28am 09 May 2024      

This has been quite a saga to get the Variac to the point where I was happy with it, It also reminded me of the perils of simple design.

The main problems were in the metering circuits, initially in the inverter The AC volts  would normaly live in the 200 - 250VAC region so nominal accuracy only required over this range it was really meant as a voltage monitor not a precision volt meter so I chose to use a peak reading circuit that was calibrated to the RMS value.

Likewise the current meter used a peak reading scheme calibrated as for RMS, again it was a bit of a novelty to have some indication of volts and current - I was well aware that complex loads and half wave and chopped AC would add an error to the readings but for resistive loads it would be very good.

We did graph results of the circuits I used for V & I and the results were impressive but as we were to later find out also reliant on your luck as to what the offsets were in the IC you happened to pick up.
Despite the opamp I chose having very low offsets they could still be problematic.  So the new prototype gave different results to the bread boarded circuit so back to the drawing board for a while.

Given that a typical Variac has usually a 0-100% rotary scale or a 0-260VAC scale that in my experience it is usually somewhere between +/- 10 to 20VAC accurate and never stable, The included metering should be really useful. But the problem was the poor performance at the low end, This would lead to 2 or 3 V showing with the inverter off, the same with Amps.

In the end there was no simple cure and I had to create a negative rail, which turned out to be ~ -3.5V using a LMC555 (CMOS NE555) and a few bits for a charge pump from the 5V supply. Then with 2 x  trimpots between the +5 and -3.5 I could inject a small offset trim to the inverting opamps inputs and set them so that from a positive meter reading you trim down until the 0 just appears, voila done!

As it turns out the Metering now is really good. Accuracy against my bench DVM is within 1V from 4V - 250V (the Variac only has whole volt LCD metering) the current is also really good. At 120mA it is within 10mA and at 3.9A it is within 20mA over the whole range.

So today I sent out another revision which I believe will be the final one and as I had another PCB I needed with a fast turn around it should be here by late next week.

I also need to thank Poida for regular updates to the code as various schemes were tried for all sorts of gremlins but now it is a really good useful device.  The active current limiting is a brilliant feature that Poida added and is something a Variac could never match.

I have an ornery 100W RMS x 2 amplifier that can eat up darlington output devices in mere moments.  Of course the Amplifier is semi smart and needs a minimum AC input before speaker protectors function etc.  Although quiescent looks about right there is not a runaway issue more like a gallop away issue that sees almost instant destructions of the output devices with no warning.  I suspect that I will be able to set the constant current to just above the mains idle current and finally stop the destruction and let me see what is happening to the circuitry.  Its just theory but I reckon it has a chance of helping.    
I think I am ready now to start the list of who wants what - please wait though for a few more posts of what will be available and how they can be configured.  Whilst I am getting the lists together the revised variac board should be here late next week, I will build one up hastily to make sure there are no issues before pulling the pin and placing a quantity order.

Meanwhile KeepIS has proved the inverter controller seems to work well in his build, 400A peak input @ 48V, lets see who can beat that lol.
Edited 2024-05-09 20:21 by wiseguy