I though I had posted this, but no - crap test setup for the last post above.


.

For anyone interested, the Nano controller is now running in my main inverter - I had the property running off grid on it for a few hours this afternoon in the testing phase, only up to 4kW though, I ran out of time today and hopefully I will finish the conversion tomorrow.

I really miss the Peak input current meter that was on the V5 SYM controller. I have a small board I made when designing it initially, and I really need that back in with a connection for Peak DC-OC to the Ext OC pin on the Nano controller board.

I'm pleased to report that noise is again not an issue for the nano controller - all LCD readings are as expected, and the startup shutdown and error tripping is perfect - it actually saved the Inverter when I had one lead back to front - and the LCD error info was 100% accurate, I knew exactly what I had done.

Mike, at the beginning of this thread a schematic for the power board was posted. Now I just trawled through the whole post to see if there is a schematic for the control board as well (the one with the Nano on it) - could not find it.

Was there ever such a schematic posted? I'm curious if and what the difference from the earlier Nano boards of Poida's is to get such a remarkable AC voltage regulation.

I have inverters here, running both, the EG8010 and Poida's early nano for the control but they do not get anywhere close to the AC regulation you posted.

Since PCB design is fun for me and 100x100 boards are cheap, I would like to see if I can replicate your AC regulation results.

Klaus, wiseguy posted them in his thread - Posted: 09:57pm 08 Mar 2024

Here is the Link To Wiseguy post

This should work?

Link to Controller
.
Edited 2024-05-08 19:19 by KeepIS

Thanks Mike, it looks like I'll have to re read WG's post to get the nitty gritty.
That should keep me busy for a while .

The inverter conversion is taking longer than expected - by  the time I redress cabling and verify all connections to LEDS etc.

I tired a few workshop machines with big Induction Motors, then I tried to start the 2.2m high bandsaw with two big heavy flywheels that are rotated up to full speed almost instantly.

The inverter went over current. I knew it would happen as I had the AC trip current set at 35A, I had to set the OC trip to 50A to stop it tripping.

The AC waveform hardly moved - you could see the fast correction for both load and unload events, it kept the amplitude almost constant to within a few volts under the load of the bandsaw.

AC was around 230vac @ 48A. The Inverter DC Input hit a peak of 400A @ 52V.  

It did not faze the Nano controller at all - No noise induced voltage trips, even under that big load.

Beautiful

The only problem I had was with the USB connection to the Nano - I wanted to re-calibrate the AC current under a constant big load, so I has a USB connection. The USB port on the PC would lock up when the inverter powered UP - it had to be RF into the port and cable (think of the earth loop as well) anyway, 3 turns of the USB cable through an RF Noise toriod and - Perfect.

That pretty well seals it - The harsh RF noise working environment does not faze the Nano.

NOTE: The Nano is mounted on a ground plane and the controller PCB ground plane mounting hole is, of course, bonded to the ground plane via a brass standoff.

I showed this towards the end of my previous build - the ground plane made a huge difference to the induced noise into the previous 8010 based Controller Board.
.
Edited 2024-05-10 09:03 by KeepIS

FYI the Nano controller has been running for over 30 hours non stop, not long, but it's been doing some heavy lifting. The interesting thing is AC regulation is much better then previous inverters, and better then the V5 controller and 8010 chip.

I really noticed that this evening, my wife had a larger air fryer running, I only realized it was on when I looked at the extra Inverter load of 2.3kW switching on and off.

Granted it's a resistive load, but the inverter load is switching from 360 watts to  2.6kW and the LED fluro batten lights are not indicating any change.

FYI The Battery and Cap bank are dropping 75mv when the load increases by 2.3kW, no solar, batteries only. The only indication of any load change on the DSO AC sine wave, is one slight "line thickness" change in the 7" full screen waveform amplitude.

The AC voltage on the LCD ACV and the True RMS indicator on a separate Mains voltage monitor are still showing the same AC voltage after the load change. In reality it's a couple of volts as it corrects - but it corrects so dam fast.
   
As usual, there is not a sound from the inverter - but that's normal, even with a 7kw resistive load dumped on this inverter, I assume most inverters would be the same on resistive loads. And as before, the bad loads are virtually silent.

I have my DC input peak meter running, I only need to calibrate the Peak DC current trip for the Nano, a tomorrow job.

The last thing to test is the Nano AC relay control of the ATS (auto transfer) and normal AC load connect, again a tomorrow job.
.

All sounds very exciting Mike, well done on all the hard work. And to the other Mike too, on a major upgrade and new generation of homebuilt inverters.

Looking forward to maybe building one myself as a backup machine.

  KeepIS said  

FYI the Nano controller has been running for over 30 hours non stop, not long, but it's been doing some heavy lifting. The interesting thing is AC regulation is much better then previous inverters, and better then the V5 controller and 8010 chip.

That's great.  Improvement over the eg8010 is easy to understand.  Any idea what changes from the previous Nanoverter version improved AC regulation?  Does the latest version control loop do updates faster?
Edited 2024-05-11 01:32 by analog8484

  analog8484 said  

[  Any idea what changes from the previous Nanoverter version improved AC regulation?  Does the latest version control loop do updates faster?

I dont use an Iron cored transformer and the usual bridge rectifier and electrolytic that was typically used before.

This Nano controller uses the ZMPT107-1 2mA/2mA miniature transformer which is used for AC voltage sense, the secondary feeds a precision full wave rectifier and there is minimal filtering along the feedback path combined with a low pass filter at the voltage feedback input of the Nano.

As far as I know the loop updates are the same but I will leave the definite answer for Poida to confirm.

Thanks for the comments, it's interesting that the only change to the main inverter had been the Nano Controller. The only internal change was to the Controller board connections.

I have a feeling this could be a difference between the 8010 and Nano timing OR algorithms.

There is a visible change in real world feedback, lights not flickering when they were before, these long LED fluro replacements, 12 of them in the workshop, are very sensitive to the slightest change in AC. Then you have the DSO display of AC output, the speed of correction appears like fast a "single step process" correction that holds the AC steady.

I mean, having to stare at the DSO to catch the slightest of movement right at the top of the waveform as I dump 2.3kW onto the Inverter is impressive, so much so that I had to look down to check that the Load had actually switched on - as this Inverter never makes a sound unless you dump very bad complex loads onto it, like a heat gun with a diode in one side of the mains for low heat. You hear a little buzz under that load, a low cost 2kW OZITO SCR variable-heat gun is completely silent.  

I know what the Nano spwm code is doing, and obviously it is not a single step control process. I put the DSO waveform appearance down to the controller speed, algorithm and design changes made by Wiseguy for AC VFB.

Although the Nano controller is "similar" to the V5 8010 controller AC VFB interface, the V5 board was also designed by Wiseguy, it uses the same Symmetrical drive and same footprint as the new Nano controller - This made it easier to swap it over as it screwed into the same location with just connection plug changes needed.

This morning I swapped out the 24v Kilovac and 48v-24v DC-DC PSU that powered it.

The new 48v Kilovac connects straight to the controller and 53V switched input.

That change bought Idle power down a few watts to 23 watts - considering that the lowest load on this inverter is around 280 watts - 23 watts is nothing - And the inverter is either ON or OFF, not idling with no load - It auto starts from OFF if the property is ever switched over to mains priority and the mains fails.  
.

  analog8484 said  

That's great.  Improvement over the eg8010 is easy to understand.  Any idea what changes from the previous Nanoverter version improved AC regulation?  Does the latest version control loop do updates faster?

Just reread your post, the old Nano code that I initially modified is basically the same for the current pwm control code.

I have never used the previous Nanoverter hardware design, so can't comment on the design of the VFB circuit, but as you know, that's critical to the timing of VFB, I believe Wiseguy answered the VFB hardware implementation on the new Nano controller.
.
Edited 2024-05-11 10:36 by KeepIS

I finally fitted my DC peak OC trip to the Nano Controller. Mike (WG) was nice enough to include a steering diode on the controller board, so the only modification to my original peak current board was the addition of a 1k 10 turn pot. I fed 0 to 4 v DC into the peak board until it indicated 550A, then adjusted the 10 turn pot until the inverter tripped.

I also wired up the Nano controllers "AC relay enable" to control a small relay that switches DC inverter input voltage to the existing ATS - this ATS wants to see the correct battery voltage before it will switch (dam ATS includes a low battery dropout setting as well) the ATS is not that accurate. Anyway, this replaced my old AC detect circuit and sense transformer, so as a bonus I got to remove even more stuff from the inverter. Got to love this new controller!!!

The result: If for any reason the inverter shuts down or stops, the AC output relay now switches instantly - It's controlled in the PWM code the instant an OC event is detected, and enabled at the end of the soft start handover in my "Auto mode" setting, or by the desired AC voltage in settings.

EDIT: This means I/we now have the means to control a big Dual Power Transfer Switch directly, I posted a pix of that earlier in the thread - I think?

Ah found it Transfer Switch page  

I have also included Low and High AC voltage setting trips, these can disable the AC relay if inverter AC output drops below a set point for 300ms, or over voltages for the same period, all adjustable in settings.

This controller design is so cool, so impressed with the work Wiseguy has put into the board, and Poidas spwm code engine which the controller is based on.

Kudos to all involved  

BTW, if I don't have the LCD connected, I now feel like I'm blind folded - The first thing I look at is the LCD - and in every single case it has told me exactly what the problem was.

I could never go back to just a LED indicator on an inverter again - I also like that from the moment the inverter is turned on, the LCD starts and indicates the running state of the inverter during power up.

Now I just have to modify the front panel to fit the LCD - that will likely be the trickiest part of the conversion.
.
Edited 2024-05-11 17:42 by KeepIS

  wiseguy said  

I dont use an Iron cored transformer and the usual bridge rectifier and electrolytic that was typically used before.

This Nano controller uses the ZMPT107-1 2mA/2mA miniature transformer which is used for AC voltage sense, the secondary feeds a precision full wave rectifier and there is minimal filtering along the feedback path combined with a low pass filter at the voltage feedback input of the Nano.

Thanks for the clarification.  I have used the ZMPT107 and ZMPT101B and they are both quite good, especially for the price.  The only notable issue I have seen is there can be significant unit to unit variance.  So, each unit needs to be checked/calibrated to ensure consistent results.  I have also used the TV-19 5ma/5ma transformer and it appears to be a bit less noisy and has better unit to unit consistency.  But its cost is a little higher.

I suspect the precision rectifier and filter are also key improvement factors.

  KeepIS said  

Just reread your post, the old Nano code that I initially modified is basically the same for the current pwm control code.
.

Good to know the control code didn't change so the improvement is most likely due to the HW changes.

BTW, The Fan control on the Nano is brilliant, no more twiddling pots to get a close ON-OFF temperature difference to set and run accurately.

I find the mass of a big toriod needs an ON to OFF delta of around 3°c, and that delta is dependent on the Toroid temperature you decide to run your Toriod at.

That big Toroid mass is the reason I have two fans for the Toroid/choke compartment,  there is NO fan on the heatsink - it's so big that it will NEVER need forced cooling.

So I have the option in Code for two Toroid fans ON-OFF settings, utilizing the Heatsink fan temperature control as a second Fan control for the Toroid. The two fans are controlled by the same Toroid temperature sensor.

The Heatsink temperature sensor is still used to trip heatsink over-temp and display the HS temperature on the LCD. The LCD will now display F1 and F2 for Toroid fan on indications, and can display both F1 and F2 as being on at the same time.

Both fans are 240AC, one, the lower temperature fan, is quieter than the second. The second fan is set when the inverter is running above 6kW for long periods.

They are arranged in push pull layout. The lower temp internal fan pulls air in through the Power board and Controller compartment, and pushes it out over, and around, the Toroid and chokes, air exits out through the center of the Toroid.

The high temp fan is mounted below the Toroid on the underside of the cabinet, it pulls air out through the center of the toriod - The Toroid core has never gone much above 45°c.

It's just another small thing that makes this Nano controller such a dream to use and setup - I'm still trying to find a negative - but I can't.

Today is rainy and overcast, the front load washing machine is on a hot wash, drawing 2kW, the workshop has sanding machines, through wall fans and external deducted Vac running, along with all the other housed loads, another 2.8kW, all these loads are turning on and off all the time, I'm been on the Computer and there has not been a flick from the lights in my section of the workshop - WTF!

Only when the inverter draws over 350A peak DC input surges can I get a slight flick of the lights. That requires turning on things like a big Toroidal Power supply, it randomly surge starts to 400A, but only 1 in 12 times that it's powered-on. Typical of some SMPS and equipment with large toroidal transformers. I am unlucky enough to have quite a few of them, and some big SMPS devices.

EDIT: I posted the following a few days back for the Nano controller, I realized that the peak current was more than I had posted, I was looking at the wrong set of recorded data. This is what I should have said:

AC was around 230vac at 48A. The Inverter DC Input hit a peak of 486A @ 51V. I have posted DSO captures of this devices startup in the older thread - Running the same WG Power board and a controller based on an 8010 with the same Symmetrical Drive as the Nano Controller - The point was to indicate the stability of the Nano controller under very high current induction motor startup. Keeping in mind that the whole property was running off grid when that equipment was powered up.  
.
Edited 2024-05-13 15:26 by KeepIS

  KeepIS said  

Klaus, wiseguy posted them in his thread - Posted: 09:57pm 08 Mar 2024

Here is the Link To Wiseguy post

This should work?

Link to Controller
.

Another question if I may ask?
The power schematic at the beginning of this thread shows a 0R (zero resistance) gate resistor for the totem pole drive. I find that hard to believe, can you confirm what size resistor you used there?

My inverter has a 10R from the FOD3182 to the totem pole and a 1R5 from there to the gate. It runs fine with that...

HI Klaus, in my earlier designs I had a 10R resistor between the FOD3182 and the totem pole buffers.  I tried going larger and smaller to see how it affected the actual gate drive at the FETs and with zero ohms it reduced the slew rate time ever so slightly so I left it out.

There is no issue for the FOD3182 as the totem pole transistors are emitter followers.
I don't have a data sheet in front of me but lets assume the transistors have a gain of 50 so if we are driving a peak instantaneous current of 25A into the FETs that were previously off, the current from the FOD3182 is 25/50 or 0.5A. The FOD3182 is capable of 3A so I consider the configuration is fine.

Klaus I apologise in advance for the verbose reply.....war and peace here we go

With regard to your question about what the Variac is, I will assume you have seen and used a Variac before and I would be most surprised if you had not.  For newbies who are maybe not familiar, a Variac is special type of adjustable transformer or a VARIable AC device or VARIAC.

The normal units are essentially a special type of toroidal tranformer that has a single layer winding and for descriptive puproses has the varnish missing from the bottom side. There is a shaft through the centre with an arm and a wiper contact that can touch individual wires as the shaft is rotated. Lets assume it has 100T and we apply 100V across the winding ends. If we meter from one end of the winding as we touch the wiper to the first winding  - 1 Turn - we will measure 1V, and if we rotate the wiper to touch the 20th turn we will measure 20V etc.

Sometimes it is handy to have a variable AC source that we can slowly increase the  voltage to something we wish to test in a safe manner and we can monitor input current against the increasing voltage and stop if things are not looking good.

In the Nano inverter there is a fixed reference voltage within the Nano and it compares the output AC level to the fixed reference and drives the PWM higher until the sample of the output voltage matches the reference level, thereby regulating a constant fixed AC output.

Poida created some code that had no feedback from the output, instead it has the usual feedback pin tied to a variable resistor (5V on the pots top end, ground on the other) and the code creates a PWM that is proportional to the wiper setting. So in theory a 2.5V midpoint pot setting will create PWM of 50% duty cycle modulation.

The only issue with that is the dynamic performance of the Toroid and output stage  will cause the output voltage to sag when a heavier load is applied.  So I created some analog control that compares the output level to the pot setting and if the output starts to sag it increases the control voltage to the Nano slightly to compensate for it.

Next unless you have paid a lot for a Variac it has no volt or current metering and a rough and ready 0-100% or 0-260V scale that is very inaccurate and dependent on the actual mains voltage at any given moment.  The Nano variac has volts & current metering and a settable current limit, which causes the output volts to drop to whatever level is required to maintain the current limit setting.

If you ever want to calibrate something to a fixed AC level, be it current or voltage the Nano Variac is a very stable AC source and stays exactly where it is set so you are not trying to chase the varying mains voltage or current as it wobbles around ad infinitum.

Lastly it operates from a DC source from 16 - 60V but the transformer primary has to be wound to suit the applied voltage.  No doubt 99% of users will either use their 24V or 48V storage supply and suitable Toroid. Of course if there is no mains power and the primary inverter has also died the Variac can be set for 240V and with the small 8 FET Power board should easily supply 2.5kW to bring your fridge to life and turn on a lamp and power your soldering Iron for the primary inverter repair.
Edited 2024-05-14 22:22 by wiseguy

Sorry mike polluted your thread - I thought this was my thread lol.  And now I've written this I cant even delete it so I will copy the Variac portion to my thread.
Edited 2024-05-14 22:26 by wiseguy

  wiseguy said  

Sorry mike polluted your thread - I thought this was my thread lol.  And now I've written this I cant even delete it so I will copy the Variac portion to my thread.

Mike, any information about the Nano or Variac is 110% welcome here. Thanks for the post.

Klaus, R1, R9, R5 and R13 are Zero ohms in both my inverters. Either a link or an R0 resistor.
Edited 2024-05-15 08:25 by KeepIS

Thanks for explaining the gate resistor question. I'll try to short the 10R when I next tinker with the driver board. The 1R directly at the gates is on the power board I will leave that - too hard to get at easily.