Notice. New forum software under development. It's going to miss a few functions and look a bit ugly for a while, but I'm working on it full time now as the old forum was too unstable. Couple days, all good. If you notice any issues, please contact me.
Forum Index : Electronics : Inverter building using Wiseguys Power board and the Nano drive board
Page 18 of 23
Author
Message
KeepIS Guru Joined: 13/10/2014 Location: AustraliaPosts: 1679
Posted: 09:36am 28 Aug 2024
Copy link to clipboard
Print this post
I've just had a look at the AC waveform to see how those mains filters worked, I described them a few posts back, they are brilliant, for the first time I cannot see any 20khz SPWM on the AC waveform, either in real time zoomed, or zoomed in on a 1 cycle capture, DSO bandwidth wide open at 380Mhz, it's just a clean line across the whole cycle.
I'll get the AC distortion analyzer on it tomorrow if I have time.It's all too hard. Mike.
wiseguy Guru Joined: 21/06/2018 Location: AustraliaPosts: 1156
Posted: 02:25pm 28 Aug 2024
Copy link to clipboard
Print this post
Maybe I am tired but I had to read this twice to get its full meaning, for anyone else who might be pondering this I would like to clarify - if I am still wrong I'm sure KeepIS will straighten me out.
The 48W is consumed when both Toroids were driven but the mains outputs were isolated from each other, but both producing mains voltage. Then the mains outputs were joined together in phase and in parallel and the idling power remained the same which proves there were no mismatches in turns or phase etc between the two huge toroidal transformer stacks.
If the PWM drive to the second Toroid is stopped as per when the stage 2 switching is not required by the load being below the threshold, then I expect the idling power will reduce to around 24W which I am guessing will be confirmed in the next exciting post. I have been so tempted over the last few days to post "Just power it up NOW already" I want to know it works and cant stand the suspense.... lol - I dont blame you for the checking/double checking, it is a bit of a beast. Great news also about the very clean output waveform after the filters. Edited 2024-08-29 00:31 by wiseguyIf at first you dont succeed, I suggest you avoid sky diving.... Cheers Mike
Cpoc Regular Member Joined: 28/05/2024 Location: PortugalPosts: 78
Posted: 04:04pm 28 Aug 2024
Copy link to clipboard
Print this post
I called it back as well at 50 watts idle for both toroids. 48 watts is even better. Amazing work to both wiseguy and KeepIS. This is going to be one beast of an inverter. Now to see the second stage board working in action.
KeepIS Guru Joined: 13/10/2014 Location: AustraliaPosts: 1679
Posted: 07:02pm 28 Aug 2024
Copy link to clipboard
Print this post
Wiseguy is correct, the magnetizing idle power with one stage is 24w to 25w and 48w to 50w with both, the predicted and expected result.
Cpoc, your guess is almost spot on, idle can vary 1 or 2 watts depending on AC output setting and final DC input.
There is also a slight rise and fall in idle power in the unloaded inverter in either 1 or 2 stage mode, I've noticed this for a long time now in the single inverter.
If you listen to the Toroid in a very quite room, your ear next to the toroid, you can hear a slight hum slowly rise and fall, put a small 12 watt load on the inverter and it stops.
From the initial tests, I know that 2 stage circuit is going to work, I will test it though once I have the safety of all current limit circuits in place.
I think in my situation I'll be running this as a 2-stage inverter full time.
So bear with me while I get all safety circuits running
Hopefully by the end of the day _ Edited 2024-08-29 06:29 by KeepISIt's all too hard. Mike.
Cpoc Regular Member Joined: 28/05/2024 Location: PortugalPosts: 78
Posted: 10:52pm 28 Aug 2024
Copy link to clipboard
Print this post
Mine was not a guess. I had spoken to an electronic engineer a while back ago and asked him what was the best case for a 12 KW inverter idle power. He you told me 50 watts best case and anything lower than that would be really good. I still have no idea how he came up with that number so I take no credit in my guess.
KeepIS Guru Joined: 13/10/2014 Location: AustraliaPosts: 1679
Posted: 12:11am 29 Aug 2024
Copy link to clipboard
Print this post
IMHO He likely calculated magnetizing current/loss for a typical 12kw "Toroid" core mass.
If he had done that for a typical iron core transformer it would be higher, even if it was specifically designed for better efficiency, and really high with with a standard two limb (shell) iron core transformer.
Also we don't know the quality of the core material we are all using, it's a ballpark guess, and you can go all out and reduce core losses even further, but I think that core material would be quite expensive.
Of course it's just my opinion and could be way off the mark. Things have advanced so much since my days of learning, and I think I've forgotten most of it anyway.
BTW: I can lower the idle current by using higher inductance chokes, but IMHO from all my testing under big loads - you can create other nasty problems in the inverter doing that.
I will select chokes for the best performance and waveforms under very bad and very high loads, over chokes that give lower Idle current and slightly better waveforms under very light loads - any day! _ Edited 2024-08-29 10:19 by KeepISIt's all too hard. Mike.
phil99 Guru Joined: 11/02/2018 Location: AustraliaPosts: 2135
Posted: 01:07am 29 Aug 2024
Copy link to clipboard
Print this post
The difference in performance between toroids and EI cores is toroids can take advantage of Grain Oriented Silicon Steel. GOSS has a very high magnetic permeability along the direction of the crystal grains, which have been greatly stretched out by cold rolling.
In a toroid the flux is always along the grains whereas for EI cores it must also cross the grain direction. That part of the magnetic circuit is significantly worse than random grain silicon steel, which is cheaper so that is what they often use.
Large 3⌀ transformers get around this by using separate strips of GOSS for each of the 3 limbs and the top and bottom bridges. For the outer limbs the ends are bevelled at 42° and 48° (not 45° so alternating layers can overlap a little) The centre limb has a 90° point at each end, (tilted 3° to one side for the same reason). The bridges have bevels and V notches to match.
KeepIS Guru Joined: 13/10/2014 Location: AustraliaPosts: 1679
Posted: 02:19am 29 Aug 2024
Copy link to clipboard
Print this post
Thanks Phil, yes and that's where the cost seems to get very eye watering if someone wants to get lower losses by applying those construction materials and methods to a DIY inverter.
I think if we are lucky enough to get toroids from a high quality older GT inverter, then were doing pretty well for the price - That's if were not being ripped off by second hand solar equipment sellers _It's all too hard. Mike.
KeepIS Guru Joined: 13/10/2014 Location: AustraliaPosts: 1679
Posted: 08:16am 29 Aug 2024
Copy link to clipboard
Print this post
I made up two Peak DC input current boards and calibrated the Peak current trip points at 430A on each power stage, a total of 860A before the dual inverter will trip.
FYI: Peak meters both worked first go, as did current trip, trip indicator and reset button
Note- I could have set each stage to 600A for a total of 1200A, and yes, a single power board has been tested at 600A for months.
I didn't get my separate AC trip setup finished, however I set the combined AC sensor to trip the Controller at 78A AC - I wanted to go higher but I would have to change a resistor value on the Nano - there will be no need once I get the other AC Trip boards running, so it's set at 78A AC for some testing.
About 30 minutes ago, I connected it to the LiFePO4 banks, gulp
I ran the inverter at over 6kW for about 20 minutes, that's all the load I could couple up, it's not connected for off Grid running, and it won't be until I get everything finished.
Anyway, the result was that it ran cold, all DC connections were cold, a very slight warmth felt on the power boards, but did not raise Heat sink temperature, so completely uneventful.
However - there is a difference in power output between the two power stages and is obviously more noticeable as the power increases.
At 6kW it was about 300 watts, however there appears to be some difference even at lower power levels.
I need to get some accurate measurements and get an exact value, I also need to recheck all AC connections to the Toroids.
One thing I can rule out: There is NO loss or difference in the DC supply to each Power board.
Finally - The 6kW+ test was very impressive for how effortless the inverter sounded as each 2kW+ load was switched on and past 6kW. The toriods and chokes were virtually cold.
That's it for me for today, I'm going to relax this evening More fun tomorrow. _It's all too hard. Mike.
mab1 Senior Member Joined: 10/02/2015 Location: United KingdomPosts: 209
Posted: 12:29pm 29 Aug 2024
Copy link to clipboard
Print this post
All very interesting.
That's a really tidy build - i suspect mine will look like a used rats nest when i get it going
On the load ballance between the stacks: i would expect some different when it starts using the 2nd transformer as the full time (1st) transformer/powerstage have been running and must be slightly warmer than the 2nd, giving the copper a higher resistance. If that theory is right it would suggest the extra 300w would be from the cold 2nd stack? But should level out over time. The good news of course is that the copper losses ought to be self leveling as it's a positive temp coefficient.
I do wonder how much intrinsic imbalance is to be expected given that there might be small variations between the 6 salvaged torroids that were presumably from different production runs? Perhaps 5% is to be expected?
KeepIS Guru Joined: 13/10/2014 Location: AustraliaPosts: 1679
Posted: 09:31pm 29 Aug 2024
Copy link to clipboard
Print this post
Thanks, I have yet to verify the exact power difference in output.
The Inverter is currently wired as a full dual power stage inverter, no stage switching.
I've been very careful to keep the DC paths and the AC paths as duplicated as possible, there is one slight difference though - the chokes are different between the two Toriods.
Both are the same inductance, saturation profile and number of turns, but two are slightly longer: 2 x 6 stack toroid rings verses 2 x 9 stack and they are wound with 25mm cable for a very specific reason that I've mentioned before.
What I intend to do is swap one across, so both sides have one of each choke.
However that should not make any real difference at low power, and this is the testing I need to do today.
Exactly, and every one is years apart in build dates, I always expected to see some difference, especially at high power, how much remains to be seen.
It's for this reason that I included separate AC current trip and DC input current trip circuits for each power stage.
I'm not really worried if one stage is doing a bit more work, it's still just loafing along compared to running with only one power stage.
IMHO This is a must when protecting each stage against very high power transient switching and AC motor start surge loads.
I can't test the surge capability until those trip stages are all installed, this is also why I incorporate separate overload trip indicators - I want to know which power stage tripped and what tripped - AC or DC. _ Edited 2024-08-30 07:32 by KeepISIt's all too hard. Mike.
KeepIS Guru Joined: 13/10/2014 Location: AustraliaPosts: 1679
Posted: 10:37pm 29 Aug 2024
Copy link to clipboard
Print this post
A good update, I just added two AC output meters, one on each AC output, the difference in AC load sharing:
LOAD Delta
1.0kw 30w
2.0kw 67w
4.0kw 126W
5.8kw 180w
I will now accurately re-calibrate the Peak DC input meters as they were pretty close to being correct.
Then see why the digital average input Power meters are reading a bit off?
This is looking really good
I'll also swap two chokes over and see if that makes any difference at high power, I doubt it, but at least the toroid drive current paths will be balanced. _ Edited 2024-08-30 08:54 by KeepISIt's all too hard. Mike.
phil99 Guru Joined: 11/02/2018 Location: AustraliaPosts: 2135
Posted: 10:57pm 29 Aug 2024
Copy link to clipboard
Print this post
As you mention the imbalance is small compared to maximum output, and likely to get smaller when they have the same primary inductors so you could leave it at that. However if you really want perfect balance it is possible by adding a relatively small 1:1 autotransformer to the output. As the voltage difference is a fraction of a volt 1 or 2 turns through a 300VA toroid should ensure the current is shared equally.
I have seen it done when extra chillers were being added to a 3.3kV buss. The existing 22kV/3.3kV transformer was too small so another was added in parallel. The impedance of the system was so low that a fraction of a volt difference resulted in a current imbalance of almost 100A / phase. The 1:1 3⌀ autotransformer cured it.
The chillers were quite large - 140A @ 3.3kV.
KeepIS Guru Joined: 13/10/2014 Location: AustraliaPosts: 1679
Posted: 11:29pm 29 Aug 2024
Copy link to clipboard
Print this post
Very interesting idea, but as you say, it's a relatively small difference.
Even if I can't improve on what it is now, I'm absolutely stoked with the result as it is, I mean, 6.5kw is just idling to this beast, I can't even move the toriod temperatures above room temperature.
The extra room around the toriods and the silently assisted natural airflow is allowing complete air flow around, through and over every mm of both toriods.
I'm just amazed at how well this beast is working, it looks like I finally managed to get it right this time, it only took me 3 different builds and 2 years though It's all too hard. Mike.
KeepIS Guru Joined: 13/10/2014 Location: AustraliaPosts: 1679
Posted: 02:59am 30 Aug 2024
Copy link to clipboard
Print this post
The news just keeps getting better:
All of 3 minute to swap two chokes to balance the SPWM drive path - WOW!
Load Delta
1.75kW 16w
2.10kW 18w
3.56kW 43w
3.8kW 45w
5.8kW 79w
The power delta has dropped from 180 watts down to 78 watts at 5.8kW
The AC current delta was between 60ma and 300ma @ 5.8kW on the AC meters- which track power and current identically.
This tells me those six toroids are pretty well made and matched.
I put this result down to some inductive differences in toroidal ring cores and the slight extra cable length with two taller choke stacks on one side.
First lesson in a two stage inverter: All cable lengths must match very closely on the SPWM outputs and chokes should be as close to identical as possible on each side - which I could not do (cores not available), but I ended up doing by using one long and one short on each power Board and Toroid stage.
The DC peak meters are reading exactly the same now - No I did not make a mistake.
The noticed that the peak DC meter decay times were completely different between the two:
I tracked it down to a slight typo on wiseguys 2 stage board, R22 and R21 values are swapped on the circuit.
Swapped the resistors around and they now track perfectly _ Edited 2024-08-30 13:01 by KeepISIt's all too hard. Mike.
wiseguy Guru Joined: 21/06/2018 Location: AustraliaPosts: 1156
Posted: 03:39am 30 Aug 2024
Copy link to clipboard
Print this post
You are correct, schematic was right but somehow the silkscreen numbers were reversed ie R21 = R22 whoops, big difference 2.2M to 10K lol. Looking forward to hearing about the auto 2 stage switching.
Also looking forward to efficiency tests with Stage 1 @ 1-6kW and then the same efficiency tests with both toroids running for 1-6kW loads, maybe just a few % better at 6kW. But I expect the Toroid & choke heating would be halved with both Toroids drive for the same load.
Re the better balance between the toroids now it probably comes down to minor differences between the Toroidal steel. Re balancing the two halves, the 48V 10KW 3phase power supply I bought off ebay had 3 internal Power PCBs each one driven from a different phase and outputs in parallel. They went to a lot of trouble to keep the AC leads exactly the same length for the same reasons. The longest cable set the length of the others and they were doubled over etc to neaten up the excess. Edited 2024-08-30 13:46 by wiseguyIf at first you dont succeed, I suggest you avoid sky diving.... Cheers Mike
phil99 Guru Joined: 11/02/2018 Location: AustraliaPosts: 2135
Posted: 03:48am 30 Aug 2024
Copy link to clipboard
Print this post
An excellent result.
This shows the very low impedance of the large toroids requires everything to be matched to get good current sharing.
The autotransformer idea might still be useful for anyone that does not have identical toroid stacks, but clearly not needed here.
KeepIS Guru Joined: 13/10/2014 Location: AustraliaPosts: 1679
Posted: 03:48am 30 Aug 2024
Copy link to clipboard
Print this post
The Inverter is now running the property off Grid.
Since I have DC two current trips and an AC trip working, I just had to give it a test at starting a big piece of workshop machinery.
BTW The Aircon is running at 2kW (very warm day here) as well as everything on the property, and then I started the machinery.
I videoed the Peak meters as I'm home alone and had to wander to the other end of the building.
Instead of the peak DC input surge reaching almost 600A on the single inverter, both meters tracked perfectly and each power board saw exactly 280A on this 2 stage Inverter.
_It's all too hard. Mike.
KeepIS Guru Joined: 13/10/2014 Location: AustraliaPosts: 1679
Posted: 03:50am 30 Aug 2024
Copy link to clipboard
Print this post
Thanks, very happy
Almost forgot A big shout out to wiseguy for his brilliant Inverter designs.
BTW I just unplugged the control panel-controller housing and placed on the workbench to fit the extra trip boards, the two small boards on the bottom are the Peak DC meter drive boards, once the AC trip boards and the over-current latch and indicator boards are fitted, they will all be mounted, and then I'll be shortening cables and fitting colored cable ties correctly.
This is looking a bit messy because of all the temporary cable ties as I'm still building and laying it out.
But you can see how easy it is to work on, again, literately seconds to unplug and remove this cabinet controller section, which also completely opens up the toroid cabinet, power board, high current DC input and kilovac & precharge section.
_ Edited 2024-08-30 16:14 by KeepISIt's all too hard. Mike.
Cpoc Regular Member Joined: 28/05/2024 Location: PortugalPosts: 78
Posted: 06:53am 30 Aug 2024
Copy link to clipboard
Print this post
Now get the second stage working…. I cannot wait anymore….most of us don’t even know what the board looks like. So far this project has exceeded my expectations. Not only is this inverter a true beast but the design and build quality is far superior than even the commercial stuff out there at over 6 times the price. All I can say to all the devs well done sirs I tip my hat to you all. Edited 2024-08-30 16:54 by Cpoc
Notice. New forum software under development. It's going to miss a few functions and look a bit ugly for a while, but I'm working on it full time now as the old forum was too unstable. Couple days, all good. If you notice any issues, please contact me.
Print this post
More fun tomorrow.
- i suspect mine will look like a used rats nest when i get it going 
