Wiseguy New Inverter Build Nano R6


Author Message
wiseguy

Guru

Joined: 21/06/2018
Location: Australia
Posts: 1156
Posted: 11:46am 05 Mar 2024      

To begin, I propose to start at the end. The finish build major parts are a 100mm square controller PCB and a Power PCB 260mm x 108mm and a couple of Bulk Capacitor boards also 100 x 100mm. The controller connects to a 20 character x 4 line display with a 3 wire interface. The inverter can read out VIN, VAC, AAC, temperature of toroid and heatsink and if the 2 associated fans are running.

These nifty LCD features and Nano programming are all made possible by none other than the generous and gifted Poida !! In a weak moment he offered to help we with a few lines of code that grew somewhat in the process.

Now that maybe I have gained interest from some BS artists - we are though, did you see the artwork of RogerDWs inverter on his wall - just brilliant.

So now for some boring background info on what and how the design evolved.

The evolution of my inverter has been somewhat spasmodic and slow due to many asynchronous external interrupts.  The inverse opto-coupled high/low side drive topology has proven to be successful.  A few users are using this topology in their power section, KeepIS is using my Power PCB & EG8010 controller design and has given his creation some real-world scary peak power loads exceeding 20kW.

I do advocate the use of HY5608s for the Power PCB’s, they have proven to be extremely robust, each FET is roughly equivalent to two HY4008’s in parallel.  My Power PCB designs all incorporate 3 low cost Isolated DC-DC converters, two 12 to 12V @ 1W modules for the upper LHS & RHS gate drives. The third 12 to 12V or 12 to 15V @ 1or2W type powers both low sides of the H Bridge.

A 12V output DC-DC converter for lower side FET gate drives works fine, but I prefer and use a 15V output which creates a -3V bias on the lower FETs giving another 3V of protection against glitches from the high side exceeding a lower FETs turn on threshold, causing it to conduct when it should be off.

The opto-couplers and gate drivers are FOD3182s, outputs buffered by an NPN/PNP bipolar transistor totem pole stage which initially used small TO251 packages.  The TO251’s types are now very hard to find but there are some similarly rated SOT223 surface mount transistors that use the same pin spacing as the TO251s and so can have the 3 (short) legs inserted into the existing PCB pad holes and soldered.

LCSC has the SOT223 PNP and NPN versions for the buffers which are ~30c ea and have continuous current ratings of ~5A and peak current ratings of ~ 15/20A so the overall cost is low and the peak current drive worked fine during some testing at 50kHz SPWM.

My latest Nano controller uses a precision rectifier, followed by an adjustable gain section and then an opamp Bessel filter, before the Nano feedback input. On the bench the output voltage is very stable from no load to 50% load (1.5KW) within 0.2V.

The control card can sense 48V low input voltage and the pre-charge capacitor voltage, to hold off Nano start. If the battery voltage falls below low voltage set point (software user setup value) the inverter shuts down, if the pre-charge power bridge capacitors are below the battery voltage by a few percent the inverter running is also inhibited.  These inputs allow an automated High current contactor to automatically bypass the pre-charge resistor and apply full current to the inverter.

There is more to the story but maybe tomorrow. Meanwhile I will post a pdf here of the overview of the inverters main parts.
4KW Nano R6 AutoStart.pdf
If at first you dont succeed, I suggest you avoid sky diving....
Cheers Mike

wiseguy

Guru

Joined: 21/06/2018
Location: Australia
Posts: 1156
Posted: 12:31pm 05 Mar 2024      

In Nano Rev5 the controller PCB could be configured to run as a Variac. This was when Poida offered to help me with some code and the Variac now runs really well.

It has an AC volts and AC current readout as well as DCVin and a programmable AC current limit which operates smoothly to reduce the output voltage smoothly, if the load increases past the current limit point.

The Variac can run from 24V or 48V or 72V just put in the right voltage toroid and a suitable choke and it just works. Perfect for a standby inverter that can run a fridge or two and soldering iron so you can fix the big inverter.

I will use it to fix an ornery 100W RMS per channel amplifier that runs great and then suddenly it stops and the output darlingtons are dead yet again (another $25 for the pair each time) it is not runaway but no warning gallop away. It needs normal mains voltage power to run, but if I set the Variac's current limit just a tad above the operating power required I reckon if the runaway starts the current limit of the Variac will catch it and save the output stage.  If it still kills them its heading for the dump!

To go with this creation I made a smaller power board using 1oz copper it is 100x100mm and uses 8 x HY5608 Fets. Mounted directly above it is the CAP board which also has the FET driver buffers and optos. It is plug compatible with a Nano inverter controller or Variac controller. Power should easily be around 2kW. The cost of the 3 boards required is really cheap. You get 5 sets of PCBs, enough for 5 Variacs for under $20 delivered.

I need to redo the PCB for the Variac as I decided not to compromise the Rev6 inverter controller design for a dual purpose variac/inverter controller. It is a minor task
which will be done in a week or so and the Variac board will be a single purpose PCB.
Edited 2024-03-05 22:35 by wiseguy

rogerdw
Guru

Joined: 22/10/2019
Location: Australia
Posts: 854
Posted: 02:05pm 05 Mar 2024      

That all looks very exciting Mike, congratulations on an amazing effort. It's fascinating to see the design evolve and gain more and more features.

The HY5608s sound like a worthwhile investment as well. I might have to keep a supply here in case I need mosfets at some stage.

And your output voltage tracking is insane. Couldn't ask for anything better.

Now I'm not game to ask what toroid you're going to use  ...  I suppose you've found a suitable one you don't have to rewind. I did wonder if I should offer to wind one for you  ...  in exchange perhaps, for a built up board so I can have a backup inverter for myself as well. I'll probably regret suggesting this.    Do you have a donor toroid to strip?

KeepIS

Guru

Joined: 13/10/2014
Location: Australia
Posts: 1685
Posted: 10:18pm 05 Mar 2024      

Thanks Mike for all the design work you have put into this. Great to see the details  of the new Nano controller board finally released on the forum, and huge thanks to Poida for all the work he has put into programming it.

wiseguy

Guru

Joined: 21/06/2018
Location: Australia
Posts: 1156
Posted: 08:20am 07 Mar 2024      

Hi Roger, I have loaded the car with all my toroids to be wound, will make a trip up there real soon  

A triple like yours is all I want lol. On a more serious note I was going to stack 2 x 3kw aerosharp toroids for my inverter, but now thinking this might be overkill. After one of KeepIS's recent posts I am rethinking something smaller.

My typical loads here probably range from 300W up to 5 or 6 kilowatts but these are like meal times etc and maybe last 5 - 45 minutes.

Mike said a 2.5kW unit would be fine for this type of intermittent use so the 3kW one is probably well suited for it and overnight with the minimal draw will be more efficient.

rogerdw
Guru

Joined: 22/10/2019
Location: Australia
Posts: 854
Posted: 09:42am 07 Mar 2024      

  wiseguy said  Hi Roger, I have loaded the car with all my toroids to be wound, will make a trip up there real soon  

A triple like yours is all I want lol. On a more serious note I was going to stack 2 x 3kw aerosharp toroids for my inverter, but now thinking this might be overkill. After one of KeepIS's recent posts I am rethinking something smaller.


Haha, I'm sure I can accomodate you.  

I did wonder after KeepIS mentioned the power levels he was getting  ...  that using a single 3kW toroid that was fully stripped and then rewound with the ideal number of turns and gauge of wire, may give an even more efficient outcome again. Unless of course what's already on the core is the ideal, which would have to be a massive coincidence.

  Quote  My typical loads here probably range from 300W up to 5 or 6 kilowatts but these are like meal times etc and maybe last 5 - 45 minutes.

Mike said a 2.5kW unit would be fine for this type of intermittent use so the 3kW one is probably well suited for it and overnight with the minimal draw will be more efficient.


I don't have enough experience with household loads yet, but what you described makes sense.

One other thing is that Aerosharp sold 1.5, 2 and 3kW units  ...  so if KeepIs was using a 2kW toroid and you use a 3kW  ...  that's a bigger step up than 2.5 to 3  ...  unless of course his was a 3kW toroid already.

Just out of interest, the girl at the place where we used to get our scrap Aerosharps was adamant that they also made a 5kW version but she had never seen one. I just searched and they did actually make one, though no idea if there were many or even any in Australia.

Anyway, if you have a 3kW core and we can find the right wire, I'm happy to wind one for you.  

KeepIS

Guru

Joined: 13/10/2014
Location: Australia
Posts: 1685
Posted: 10:39pm 07 Mar 2024      

Hi Rodger & Mike, just confirming that my Aerosharp toriod is indeed rated at only 2kW, I still have the spec sticker for the toriod.

I had no trouble starting a 2 meter high heavy duty 2.5hp Bandsaw pulling around 390A peak DC input at startup when I was testing it a few years back.

KeepIS

Guru

Joined: 13/10/2014
Location: Australia
Posts: 1685
Posted: 01:43am 08 Mar 2024      

Mike, I found that I need the Ferrite Beads.

Using a differential Probe and DSO.

Tested unloaded and loaded to around 250 watts resistive load without the beads fitted.

There was quite a lot of ringing on the waveform across the power board output terminals, on both HI and LO sides of the waveform.

I'm not going to waste time describing the various methods, loads, choke variations, Caps and filtering etc that I tried to narrow it down.

The difference between my Big inverter and this inverter is this:

The FET Drain metal facings are bolted directly to the two separate low side heatsinks. There is some expected ringing, but it's minimal compared to what I'm seeing on this board.

Could bridging the FET Drains (4) on each side, "ever so slightly" change the effect of the short Drain lead connection in that build and help to dampen the ringing that I'm now seeing, (no beads fitted).

The new build in contrast, has the FETS fully isolated on both HI and LO side.    

I then fitted the Ferrite beads and retested.

What a massive difference. I also had a very faint very SLOW cycling hum in the Toriod (no load and loaded) a very strange effect that I have not heard before, ear pressed against the toriod. As this faint Hum cycled, the idle current went up and down by around 50mA to 100mA.

The Ferrite beads have completely eliminated 95% of the ringing, and that cycling effect has completely gone.

The beads I used are Electus, from Jaycar - LF1251 or FX1115. They are about 4.2mm diameter x 5mm long the hole in the center is ~ 1.3mm. They seem to change the carried stock form time to time. They have an LF1250 but I don't know the inner dia, it may be suitable?
.
Edited 2024-03-08 13:33 by KeepIS

wiseguy

Guru

Joined: 21/06/2018
Location: Australia
Posts: 1156
Posted: 04:23am 08 Mar 2024      

I follow why you chose not to individually isolate the Mosfets from the heatsink, many others have done this on the BS in the past.  For my builds, where I like to switch the FETs fast, I found out that the beads are worth installing due to the issues you have been dealing with. Fast switching reduces heat as the FETs spend minimal time in the linear region but it comes at the cost of ringing and increased EMI if you don't deal with it.

I am glad the beads (same ones I use) sorted it out for you I fit them without question now, I consider it better to have them and minimise risk of issues.  The FETs run with very low heat I have found so having the heat washers on the FET face does not really cause issues despite a small decrease in  heat transfer.

KeepIS

Guru

Joined: 13/10/2014
Location: Australia
Posts: 1685
Posted: 05:50am 08 Mar 2024      

I totally agree with how little heat is generated in the FETs with this design, interesting though that the big inverter has no beads, and very little ringing, I can only put that down to the bridged Drains on the 4 FETs on either LO side in the big inverter. Nothing else is different, I may simply modify the heat transfer washer position on the big inverter and test it to see if it changes.

Cheers.

-dex-
Regular Member

Joined: 11/01/2024
Location: Poland
Posts: 99
Posted: 09:35am 08 Mar 2024      

  KeepIS said  
The beads I used are Electus, from Jaycar - LF1251 or FX1115. They are about 4.2mm diameter x 5mm long the hole in the center is ~ 1.3mm. They seem to change the carried stock form time to time. They have an LF1250 but I don't know the inner dia, it may be suitable?
.


I used Fair-Rite  from DigiKey, part no: 1934-1460-ND

wiseguy

Guru

Joined: 21/06/2018
Location: Australia
Posts: 1156
Posted: 11:10am 08 Mar 2024      

The new Nano controller board is of course not as simple as fixed resistors and comparators etc to make decisions. After the controller PCB is loaded with the parts it must have a pre programmed Nano inserted into it.

Now we need a laptop running the Arduino environment and a suitable USB cable to suit your Nano USB to Laptop connector and away we go.  You open a serial console window from the Arduino software and type a ? and a menu is presented. First we select "load default values" which gives us a good starting point.

Next we work our way through the menu items one by one and calibrate the various parameters of DC IN Volts, AC OUT Volts and AC Current, Temperatures for Fan ON for the Toroid and Heatsink and a shut down over temperature parameter to give the inverter a spell to cool down.

That is where the Nano comes into its own, all the changeable things we may want to vary are done with the laptop and keyboard, no soldering required. The sorts of functions provided by the Nano include automatically closing a high current contactor after pre-charge has finished. Ramping up the mains, providing drive for an AC output relay so the load is only connected when the mains is up.

It has the 20 character 4 line display for display of VinDC, Vout AC, Amps out AC, toroid and heatsink temperatures and whether the fans are running or not. The controller board has an adjustment for AC output current trip that allows adjustment from ~ 8A to 75A AC.

There is a simple front panel connector to drive 3 LEDs for Mains on (flashing green)
Overcurrent trip Red and Vin Low/Vcap error lamp also Red. There is also a push button to reset the over current latch.  I will post this now before I lose it.
Edited 2024-03-08 21:42 by wiseguy

wiseguy

Guru

Joined: 21/06/2018
Location: Australia
Posts: 1156
Posted: 11:57am 08 Mar 2024      

I apologise in advance for a detailed discussion on controller start and stop.

The controller board although designed for automatic high current contactor closing, is also equally at home with a Manual pre-charge button and DC high current breaker.

I think it is time to present the two different types of operation for Manual and Auto use. I will reiterate why we have pre-charge for any newcomers.

If we close the contactor or circuit breaker with no precharge to the capacitors - it momentarily has hundreds of Amps flowing.  There is a loud splat and chunks are taken from the switch contacts or they become welded together - you only make this mistake once !

So we cause a more gentle current to flow until the capacitors are charged to the battery voltage, then we can make a high current connection, with no splat.

Now to describe the controllers operation for Manual starting up. First the main breaker switch is off, press the pre-charge button, the Nano controller is powered too by the button press but inverter is inhibited whilst the button is pressed. The Run led will light steady during the pre-charge and go out to tell you capacitors are charged.

Ok, still pressing the pre-charge button, close the main circuit breaker to the inverter and then release the pre-charge push button and the magic begins.
The inverter will slow start the mains voltage and the Run Led will flash to let you know mains is happening - be careful, and when mains is up the mains contactor/relay is driven to connect the AC to the load.

Lets look at the batteries becoming flat,  our batteries are now heading below 44V, or whatever low voltage we programmed for the inverter to gently stop.

The mains on relay opens the Run led stops flashing on and our inverter waits patiently.  Now the sun is up the batteries are receiving charge and become higher than batt low shutoff.  The inverter soft starts to bring mains up then the mains up relay is closed and the lights and fridges all start automatically.

If you wish to stop the inverter, press the pre-charge button and the inverter will gently stop - now is the good time to open the circuit breaker when the Run led stops flashing.  With the breaker open it is galvanically isolated from the batteries.

Our wives may not like this inverter much claiming it to be a complicated monster that is scary to use.

Auto Version
The Contactor version has more finesse, it just has a small toggle on/off switch. We need to close the main circuit breaker, is the inverter switch on or off - doesnt matter their will be no splat as the main contactor high current path is open.

If inverter was on, then after Vcap is pre charged, the Nano closes the contactor and ramps up the inverter and then closes the mains output relay/contactor.

If inverter was off nothing is powered, turn switch on, then after Vcap is pre charged now continue as for previous line.

If inverter is running and on off switch is set to off, there is a sequence, first, open the mains relay, ramp mains down to stopped, open main DC contactor and now inverter is off and galvanivcally isolated from the batteries.

Lastly if the inverter shut itself down it is very similar to the Manual recovery as the batteries receive charge. Contactor closes inverter ramps mains up, mains relay is driven to connect the load, lights come one fridges run etc. Wife is happier (maybe...)

Please ignore the previous Auto start PDF latest is here
4KW Nano R6 AutoStart.pdf
The manual overview PDF
4KW Nano R6 Manual Start.pdf
Finally controller schematic
4KW Nano R6.pdf
Edited 2024-03-08 22:00 by wiseguy

wiseguy

Guru

Joined: 21/06/2018
Location: Australia
Posts: 1156
Posted: 12:19pm 08 Mar 2024      

The controller card is powered using an onboard 12V output un-isolated switching regulator.  It accepts from ~ 17V to 75V input and provides 0.5A. There is an option of another regulator type that has similar input voltage span but provides 1.0A

Following that is a small 5V switching regulator fed from the 12V pre-regulator that provides 5V @ 0.5A which is ample for the LCD etc - current requirements for the controller with LCD is less than 0.1A.

The cost for the two regulators is ~ $7.50 (total) from Digikey and probably Element 14 and others.  If you plan to run the 2 fans from the onboard regulator depending on their current, budget for 12V fans with a maximum of no more than 400mA each and select the 1A 12V regulator.

With this controller you can run a 24V or 36V or 48V inverter with essentially no changes to the electronics. The Toroid and the choke are the only parts that need to be changed to suit the DC input voltage and Power output requirement.  Of course the Nano low voltage shut down will need to be modified to suit the input used.

The FET Power board is galvanically isolated from the controller card high power circuitry as far as the isolated Gate power supplies and opto drives go. There is a ground connection and a VCap sample output on the connector to the controller so the controller knows the state of the Caps voltage during pre-charge.

The following mainly refers to 48V operation:
Due to the Galvanic isolation and that the Power PCB & FET drives are powered from the controller card has some neat advantages. You can Power the controller card from ~15-16V onwards. Meanwhile the FET power boards high current connectors can be powered from 0 to 50V. As the SPWM drive from the controller will be at 99% (maximum from software as AC feedback is too low to regulate) it will still produce a nice sine wave even from 0.5V supply to the main FETs. This makes fault finding and testing before applying full battery voltage very simple and pyro technic free.

Or the whole inverter is low voltage friendly so without changing they wiring, any voltage from 16V up will produce a nice clean sinewave albeit starting from about 80VAC for 16V in.  Note the controller card has a Jumper (J11) for this type of testing and operation so it ignores the low voltage input that would normally keep it in shutdown until above the Vlow threshold.
Edited 2024-03-09 09:49 by wiseguy

wiseguy

Guru

Joined: 21/06/2018
Location: Australia
Posts: 1156
Posted: 11:50pm 08 Mar 2024      

Fixing the Power Board:

In the event of FET failure it it is wise to remove the Bulk capacitor boards and use your "spare" capacitor boards fitted with a 220/330/470uF cap each, I dont usually even solder them in, a small tweak to the legs if they are at all loose when inserted "snapped in" to make them tight  usually suffices for low current testing after replacing FETs.

In my failures to date usually a high side FET meets its maker and goes shorted, of course when the upper FETs should be off and the lower FETs turn on "Surprise M..F.." and takes them all out in an instant. How do you find a dead FET with 4 in parallel, there is a way that has worked for me.

With a current limited supply set to 1A and a few volts apply the voltage between the high side supply and the transformer connection node. This is between the Drains and Sources of the bank with a dead FET. With an accurate DVM measure each FET (Drain to Source) at the wide shoulders as close to the FET as possible usually 3 FETs will be a couple of millivolts but all exactly the same, the faulty FET will be slightly less,    that is the one to desolder before checking if the others are now OK.  If a difference was hard to pick turn the current up to 2 or 3A and measure again.

I have never replaced the other 3 when only one has died, despite the theory that others may have suffered I have not found that.  The lower FETs though are very unlikely for any to have survived, so usually all have to be replaced.
..
If anyone has any questions or comments please feel free to post away - I have been looking at this for so long now I probably take things for granted and leave important bits out.
Edited 2024-03-09 09:52 by wiseguy

wiseguy

Guru

Joined: 21/06/2018
Location: Australia
Posts: 1156
Posted: 12:07am 09 Mar 2024      

At this point in this topic I want to ask if anyone out there reading this thread is interested in building one of these inverters? If you are please indicate - if there are more than 5 yes pleases I will continue to expand this post and no more replies are really required.

There is still the bill of materials, test procedures for the newly finished PCBs and the Variac to present. If there is no interest its fine but it will save me from a heap of time and effort.  And for Poida too who was considering to make a video of how to program a nano, calibrate the controller etc.

I believe that it is a pretty robust unit with some runs on the board and a lot of "bang for the buck". The inverter PCB set would probably cost around $30 to $40 posted. I would be happy to throw in 2 new current sense transformers and a couple of LCD/Nano interface boards that make adding an LCD simple with no wiring needed just some soldering.

Another option might be for the PCBs to be ordered in lots of 5 and shipped directly to each buyer. I would order the lots of 5 boards and create a new temporary shipping address with the PCB supplier for direct shipping.

I dont want to make gerbers etc available simply because I don't want anyone to capitalise too easy on the hard work for creating the design. Like wise the software will probably not be posted on the Forum but will be available by email request or similar method.  We could possibly even supply pre programmed Nanos for units to also safeguard all the effort Poida has gone to in creating the code.
Edited 2024-03-09 10:20 by wiseguy

KeepIS

Guru

Joined: 13/10/2014
Location: Australia
Posts: 1685
Posted: 12:43am 09 Mar 2024      

Hi wiseguy, first of all, I wanted to point out something to anyone new to building inverters.

The Auto start and Manual start circuits and Block diagram shown, are typical of all and any inverter you might build. They all have the same startup requirements, but the use of the Nano and the way Wiseguy and Poida have designed and programmed the setting changes in the Nano is really neat, and it will make setup and changes a breeze, even if, at first, it may not appear that way.

Mike, I of course would love to see your setup post and Poidas video, and of course the Variac information (Adjustable 240VAC supply with adjustable current trip)

I will also post a complete build of the Autostart wire up, this is almost the same as my current big inverter, just from a safety standpoint, that is absolutely the way to go - seriously!        

Way back, I actually purchased two other boards for an inverter build. Once I got hold of the Wiseguy board and design, there was just no contest in it for the size, quality, stability and relative simplicity of building Wiseguys design and power board.

IMHO, in combination with the New Nano R6 controller, it's simply unbeatable.

Mike, you likely already have me down for a couple of boards including the Variac controller.

Wiseguy If you see any errors or omissions in my thread, pleas feel free to correct me. I'm trying to keep it simple, but some meaning get lost or miss understood when I try to do that.
.

rogerdw
Guru

Joined: 22/10/2019
Location: Australia
Posts: 854
Posted: 03:37am 09 Mar 2024      

  wiseguy said  At this point in this topic I want to ask if anyone out there reading this thread is interested in building one of these inverters? If you are please indicate - if there are more than 5 yes pleases I will continue to expand this post and no more replies are really required. .


Yes please thanks Mike, I'm certainly interested  ...  just as a spare backup inverter.


  Quote  I have never replaced the other 3 when only one has died, despite the theory that others may have suffered I have not found that.  The lower FETs though are very unlikely for any to have survived, so usually all have to be replaced.


Very handy to know, thanks Mike. I probably tend to worry about the "what-if" and so change them all  ...  but it's nice to know it isn't always neccessary.

Regarding your way of determining the shorted mosfet, it won't really help too many others, but I have a Polar 850 benchtop tester that reads down to .1 milliohm and that is very helpful in determining where shorts are located.

Of course you need to regularly sharpen the probes and push hard to make it effective.

Bryan1

Guru

Joined: 22/02/2006
Location: Australia
Posts: 1364
Posted: 03:53am 09 Mar 2024      

Hi Mike I do remember back when you visited you did show me your inverter and I did say wow that's small now on reading here it's for sure size doesn't matter when it comes to power it's the thinking behind the scenes.

So yes please put me down for some boards and they can be my next builds  

I have been thinking of getting a new air compressor and plasma cutter which I'm real sure will test this board out to the max.

Regards Bryan

disco4now

Guru

Joined: 18/12/2014
Location: Australia
Posts: 906
Posted: 03:55am 09 Mar 2024      

  wiseguy said  At this point in this topic I want to ask if anyone out there reading this thread is interested in building one of these inverters?

Mike,
I would like to build an inverter (plus a spare) and the variac.

Gerry
Edited 2024-03-09 13:57 by disco4now