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 22 of 23
Author
Message
wiseguy Guru Joined: 21/06/2018 Location: AustraliaPosts: 1156
Posted: 01:17am 22 Sep 2024
Copy link to clipboard
Print this post
Further comment re Bret76. re reading your post a number of times I think we are saying similar things but with the translation a bit more difficult to fully interpret.
You did allude to, with the 20kHz periods and your diagram that there was a difference between where we are (without the heavy load) where we are with the heavy load and where we need to be with the heavy load.
I believe that using a DSP or similar really fast number crunching ability yes we could make corrections on a 20kHz cycle by cycle basis, so we are correcting for load changes on the very next on/off period, but I know my limitations and that stuff is on a whole different level.
Yes there is a definite tradeoff compromise between the inductor, mains capacitor, acceptable ripple on the output voltage waveform, idling current etc.If at first you dont succeed, I suggest you avoid sky diving.... Cheers Mike
KeepIS Guru Joined: 13/10/2014 Location: AustraliaPosts: 1679
Posted: 02:31am 22 Sep 2024
Copy link to clipboard
Print this post
I should have replied to bret76 on how I finally settled on the Choke values and the compromise.
I literately spent months over the past few years winding chokes and testing saturation levels. Looking at low and high power load waveforms for various inductance and core material, until I found the sweet spot, and that was spread across building three different Inverter.
Interestingly, the final chokes turned out to be the best choice for each of the different electronic designs of Inverter power-stages and controllers when they were retested - I still have them.
As wiseguy said, it is always going to be a tradeoff.
The tradeoff was between low power efficiency and perfect idle AC waveforms, OR making the Inverter live under ridiculous levels of startup loads and running high power loads.
I chose the latter
Sometimes you just have to do the hard boring repetitive work of just testing the dam things under the conditions that you hope/want it to survive
99% of the time the waveforms are identical and good, and then that 1% shows up like that waveform, for me the 99% rule applies.
BTW: For those who followed the two failures I had, I am now 100% positive that both were due to the garbage Mains/Inverter changeover Switch.
It took having Inverter AC and Mains AC arcing together to eventually cause a FET to blow.
Since changing that switch out, there has been a complete absence of strange spikes in Inverter current at changeover, they have vanished - even with running loads of 4.8kW. EDIT and that change was with the previous single stage Inverter. _ Edited 2024-09-22 13:18 by KeepISIt's all too hard. Mike.
@wiseguy : we agree. Maybe for a short time after the current spike, the tuning is not optimal (but it works, so perfect). What I mean is that with a no-load tuning, adding all the 20Khz steps together, we will get a nice sinusoid. But changing the current without changing the tuning, the sinusoid will be different.
And if a too small energy is stocked in the L+C, will be flat before endding the 50µs, so an inductive the load can try to push current back to the inverter -> big oscillations ? New add on my manual oscilloscope
We are here in particular conditions : big step in power (in a normal house, step will be lower because of multiple loads and lower loads), long time before PWM actualisation (like keepIS say : the only one like this).
@KeepIS : I'm not trying to question this inverter or the work implemented, but rather my understanding. I spend a lot of time reading the hours you and others spend in testing, but still sometime not clear to me / like : "ok this value worked for me" (but why ????? How dit you find it ???).I don't have your experience to develop an inverter, and I'm not sure I'll get there yet. So I have to try to understand at least a little bit how it works. Can you confirm : for 1 stage (one toroid) you have something like 2x23µH + 2µF or 4.7µF If possible, can you also put me the PWM values of a similar conditions (low load / big load). Steady state is ok.
I have the feeling that the sinusoid may be a little distorted - at least when passing 0V (not a problem, just to confirm my interprétation).
nickskethisniks Guru Joined: 17/10/2017 Location: BelgiumPosts: 458
Posted: 09:40am 22 Sep 2024
Copy link to clipboard
Print this post
If you want to calculate something, what you can do is use 20% ripple current and 5% reactive power to calculate inductor and capacitor as a startpoint. Unless we have special simulating tools or specifications to make analysis there is no calculation involved. Maybe try something in excell and play with different parameters. I found some youtube stuff that would be interesting for you I posted the link before, I will try to find it. Then it's up to realworld testing like keepis and others did.
Like we all agree you will allways have distortion on each step of the sine with this kind of technology, you can only try to minimize it and look for a compromize. We also use lf transformers and like wiseguy stated in the past it is important to have symetrical pulses so you don't create flux wandering trying to correct something.
Asking this kind of questions I feel you are ready to make one prototype and do some experimentation with it. Comfirming some theory and find there is allways more to learn and understand. Edited 2024-09-22 19:42 by nickskethisniks
KeepIS Guru Joined: 13/10/2014 Location: AustraliaPosts: 1679
Posted: 10:35am 22 Sep 2024
Copy link to clipboard
Print this post
I indicated how I arrived at choke values in the previous post, tested with different toroid transformers, different types of controllers and different power stages.
The results were always the same, two chokes wound close to 21uH, or single choke close to 42uH, wiseguys SPWM drive needs 2 chokes, 8010 based only need 1).
The dual chokes must be wound on 6 to 8 small stacked Sendust toroids with 4 or 5 turns, or equivalent, needed for high saturation in the hundreds of amperes, this can give a near perfect sinewave with a no load THD of around 1%.
For all toroids using these chokes, the secondary cap value I use is between 3.8uf to 4.4uf, whatever I have on hand. This also gave the best performance with respect to ringing and filtering.
The PWM value is dependent on the transformer turns ratio for a given regulated output:
PWM is around 73% at idle (considered low I think), and around 78% @ 5kW on the dual Inverter, I would need to check the exact numbers though.
AC output regulates at 240V from 44V to 59V DC input with my winding ratio.
The Inverters low voltage cutoff is set at 52.4V - I have a lot of head room. _It's all too hard. Mike.
Solar Mike Guru Joined: 08/02/2015 Location: New ZealandPosts: 1138
Posted: 11:08am 22 Sep 2024
Copy link to clipboard
Print this post
>> The dual chokes must be wound on 6 to 8 small stacked Sendust toroids with 4 or 5 turns, or equivalent, needed for high saturation in the hundreds of amperes, this can give a near perfect sinewave with a no load THD of around 1%.
Do you have a link to source those cores, they used to be quite inexpensive a couple of years ago, now the ones on Aliexpress are all over $10 each with shipping.
Thanks Mike
wiseguy Guru Joined: 21/06/2018 Location: AustraliaPosts: 1156
Posted: 01:15pm 22 Sep 2024
Copy link to clipboard
Print this post
I need to correct something that is propagating on the BS about a requirement for two chokes that must be used for wiseguy's inverters. This is not really accurate or true, a single choke will work, but will it perform best? I also feel that the inverters are not really mine per se, I just tried to perfect already existing inverter circuitry and topology to be more robust by doing it my way. As I kicked for the invisible goal posts maybe I overshot slightly which is not a bad thing. Credit is also due to Oztules, Madness, Poida and the chief of test to destruction KeepIs.
I do like the concept of symmetry, however there are some more compelling reasons behind the two chokes solution. To help avoid the inverter becoming a powerful EMI noise generator to the outside world, putting a choke on the 2 output nodes of the main power PCB outputs helps to ensure the the very fast switching edges of the Power FETs, which are generating harmonics from DC to Light (well almost.....) are not presented raw to the Toroidal tranny.
If a single choke is used then one side of the Toroid is exposed to the fast edges with all the harmonics and these are then being coupled to the output winding, if there is not a shield between the windings and then radiated to the outside world via all the house wiring etc. Noise is better to be killed off as early in the chain as possible, this is often much more preferable to adding stuff to kill excessive noise later.
If you put a CRO on either node of the Toroidal primary and the choke feeding it you will see a nice sinewave. With a single choke you see a nice sinewave at the one choke & toroidal node but on the other side of the toroidal primary connecting to the H bridge switching node there is just a hash of square waves.
I reasoned at the time I chose to use 2 chokes, that a single 10nF at the 2 choke to toroidal primary nodes, with the other side of the 10n's connected to ground would help to kill any fast edge noise to ground if it managed to get through the chokes before it got to the toroid. However it was almost impossible to pick the difference with or without the caps, so I did not fit them in the end.
I apologise if this was interpreted as a "must" do rather than a should or good to do. Edited 2024-09-23 08:26 by wiseguyIf at first you dont succeed, I suggest you avoid sky diving.... Cheers Mike
KeepIS Guru Joined: 13/10/2014 Location: AustraliaPosts: 1679
Posted: 10:38pm 22 Sep 2024
Copy link to clipboard
Print this post
You know I was almost going to reword the "needs two chokes" and say "works best with two chokes", but since you had drawn them on your wiring diagrams, I though why confuse the issue and spend another for lines of text elaborating on that.
Perhaps BS is an interpretation of unintended meaning.
Of course it works with one, but I ALSO find the output is cleaner with two, and two chokes with 4 to 6 cores each, is smaller than a single choke wound on 8 to 12 cores to get a similar power handling and saturation levels, I have tried both.
I also tried caps at those positions, and also found it made no visible difference. _It's all too hard. Mike.
KeepIS Guru Joined: 13/10/2014 Location: AustraliaPosts: 1679
Posted: 10:47pm 22 Sep 2024
Copy link to clipboard
Print this post
Unfortunately I'm having the same trouble, I has to get slightly different cores for the last two chokes I wound, the originals are very expensive and hard to find now.
To make matters worse, I needed an extra two cores for each choke to get the same performance values, cost me even more.
Some of the really nice larger cores, only need two, as used by someone on the forum are just way above my cost limit to get over here _It's all too hard. Mike.
phil99 Guru Joined: 11/02/2018 Location: AustraliaPosts: 2134
Posted: 11:36pm 22 Sep 2024
Copy link to clipboard
Print this post
An experiment for anyone with a spare toroid and spare time.
Add both choke windings (in opposite directions) to the toroid, leaving all other windings open circuit. Depending on its size one or two turns may be enough.
If it works a second hand toroid might be cheaper than a bunch of Sendust cores.
KeepIS Guru Joined: 13/10/2014 Location: AustraliaPosts: 1679
Posted: 12:58am 23 Sep 2024
Copy link to clipboard
Print this post
I had toyed with trying that, but even the smallest toroid I have is to big to fit.
I would love to have tested the ones posted by dex, but I can't find any way to get them to AU for a reasonably price, and they would just fit in the housing.
It's all too hard. Mike.
wiseguy Guru Joined: 21/06/2018 Location: AustraliaPosts: 1156
Posted: 02:08am 23 Sep 2024
Copy link to clipboard
Print this post
Most of the sendust type cores KeepIS used were type number 0077109A-7. They are available from Digikey for $7.85 when ordered in quantities of 10 and the total price exceeds their free shipping threshold so shipping is free. If you wanted to duplicate KeepIS's choke values (but less Isat) 6 of these cores stacked with 5 turns according to my calcs is ~ 23.4uH with a saturation current ~ 185A (~9KW).
I just noticed zero stock - for the first time! - they have always had stock and I am sure will already be on backorder. But there are also other similar cores on Digikey including KoolMu, MPP, Hi Flux etc, have a look around. Edited 2024-09-23 12:08 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: 02:14am 23 Sep 2024
Copy link to clipboard
Print this post
Thanks for for testing out the core it seems like a perfect fit. Do you recommend 50mm2 for the primary size wire ?
That would be a AWG 1/0 which is 53.5mm2
I was planning on using AWG 2/0 which is 67.4mm2 would that be ok with that core ? Edited 2024-09-23 12:19 by Cpoc
KeepIS Guru Joined: 13/10/2014 Location: AustraliaPosts: 1679
Posted: 03:35am 23 Sep 2024
Copy link to clipboard
Print this post
Thanks for posting that info, those 0077109A7 are like the last units I used for the final two chokes, which were KS225-125A.
These have a larger diameter hole, so very easy to wind thicker cable compared to the original chokes I used, those were MS-226125-2. _It's all too hard. Mike.
KeepIS Guru Joined: 13/10/2014 Location: AustraliaPosts: 1679
Posted: 04:12am 23 Sep 2024
Copy link to clipboard
Print this post
Hi, that was tested by dex, I just copied his post.
I personally would not use anything bigger than 2G 35mm2 for the chokes, there have been discussions about this going back years, but not everyone will agree and I'm sure someone can give you a technical calculation that says I'm wrong.
I like a small Resistive buffer between the FETS and a very large Toroid.
It may surprise some that I'm only using a quality 25mm2 cable for the chokes, these are fine for 12kW in my inverter given the amount of time that it would run at that level - like never!.
They handled 8kW for 30 minutes without any problems in a single stage Inverter.
Perhaps you could do some initial testing with 35mm2 and see how you go. _It's all too hard. Mike.
nickskethisniks Guru Joined: 17/10/2017 Location: BelgiumPosts: 458
Posted: 07:56am 23 Sep 2024
Copy link to clipboard
Print this post
I think I tried this a few years ago, but I think it saturated to quickly, I should do it again to confirm but with a small airgap this would certainly work
From your experience, do you know if the inverter survives a BMS (or Kilovac) trip? (say with an inductive load). Maybe it never happened.
David
KeepIS Guru Joined: 13/10/2014 Location: AustraliaPosts: 1679
Posted: 09:28pm 23 Sep 2024
Copy link to clipboard
Print this post
Hi David, that has happened countless times over the past year in testing, with the Controller tripping the Kilovac due to over current, usually right at peak current loading and/or 5 to 30 cycles into spinning up large Inductions motors.
This was something I was also concerned about, but there has been no issues with the Inverter. I've even disconnected the DC input with the inverter under 4kW of mixed load and the DC current meters do not spike, they simply drop.
I use the Nano Controller AC enable line to signal the Inverter<->Mains transfer switch, if used completely off grid, an AC enable controlled relay could disconnect the Load or control another backup power source.
With sudden DC disconnect, the Nano quickly drops the AC enable Control Line, and/or enables an Instant powers off below a set voltage, SPWM halts and in my case, the Inverter load is normally transferred back to Mains AC.
Should DC be reconnected, and since the Inverter power switch is still on, the Inverter will power up, first automatically pre-charging the capacitor Banks, then when AC output has reached 240Vac (adjustable in setup) the controller enables the AC control line, this signals the transfer switch to transfer back to Inverter AC, but only if I've set it up to transfer back automatically. _ Edited 2024-09-24 10:54 by KeepISIt's all too hard. Mike.
nickskethisniks Guru Joined: 17/10/2017 Location: BelgiumPosts: 458
Posted: 11:29am 24 Sep 2024
Copy link to clipboard
Print this post
I didn't have problems with the ozz inverter (without low voltage protection) either at least 5 times it happened the dc was interupted under 7kW (mostly resitive load but it's never completely resistive). Even inputvoltage rise to 72V did not gave problems (battery disconect suddenly and with solar panels connected directly). I use 100V mosfets, because I feel more comfortly with them. Edited 2024-09-24 21:31 by nickskethisniks
KeepIS Guru Joined: 13/10/2014 Location: AustraliaPosts: 1679
Posted: 10:19pm 24 Sep 2024
Copy link to clipboard
Print this post
The well made high power China Inverter Boards I have were similar to the ozz in many ways, they were fine as well.
The Nano Controller I use has a setup setting for Over voltage cutoff, I set mine to 59V, the inverter stops and disconnects the Kilovac, tested numerous times, quite often when testing with an adjustable supply when I turn the voltage knob too fast.
Instead of having the LCD display "Over Voltage", I should just program it to display "!!! MORON !!!"It's all too hard. Mike.
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
