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.
Revlac Guru Joined: 31/12/2016 Location: AustraliaPosts: 1024
Posted: 07:48am 24 May 2023
Copy link to clipboard
Print this post
Thanks for the picture and info Mike, that resistor fits in there nicely. Cheers Aaron Off The Grid
analog8484 Senior Member Joined: 11/11/2021 Location: United StatesPosts: 108
Posted: 05:39pm 24 May 2023
Copy link to clipboard
Print this post
No need to apologize. I personally think your explanations and discussions are way more valuable than the actual inverter build documentation which is also very valuable. Much appreciated.
I agree the inductances/reactances can dynamically change to some degree in the system. However, that still doesn't seem enough to explain the severe effects of certain (mainly non-linear?) loads. Learning more about transformers, I see all the factors (e.g. winding stiffness, etc.) you mentioned do affect the transformer inductances, resistances and capacitances but they are all relatively static properties in a built system. So, the question I want to get answers for is which ones can really have sudden severe effects under non-linear loads? And what is the process?
After some further research, my theory now for the likely root cause is specifically the interaction between the primary choke inductance and the transformer mutual inductance. The primary voltage is positively dependent on the transformer primary inductance and current change rate and negatively dependent on the mutual inductance and secondary current change rate. Under non-linear loads, the secondary current change rate can suddenly increase drastically. The effect of this increase is that it will have large negative effect on the primary voltage that will in turn drive large increase in the primary current rate of change. The large change in the primary current will be resisted by the primary choke. Therefore, if the primary choke inductance is too high the result is the observed sudden severe effects. For me, this also explains the observation that a relatively small non-linear load can have similarly severe effects as a large non-linear load. It's the current change rate that matters rather than the current magnitude.
If this theory is accurate then a likely key design limit for the choke inductance is the max secondary current rate of change. I wonder what is the max secondary current change rate (100A+/us?) when your air-con is running.
KeepIS Guru Joined: 13/10/2014 Location: AustraliaPosts: 1679
Posted: 04:22am 25 May 2023
Copy link to clipboard
Print this post
Another ramble, but more to give an idea of where I'm coming from mentally at this stage of the inverter build, which is virtually completed. I'm moving on to make a few Solar regulators, and make a few of my Hot Water system controllers for friends who want to connect some reclaimed old solar panels to run their existing 240vAC HW units, price of power is literally going through the roof for some of them. Then I want to make a backup set of Controller and Power boards for the inverter.
The bad loads, it depends on the definition of severe, the sound definitely is, but it doesn't take much to make a Toriod talk. The Inverter AC waveforms may simply have a small kink in one side, sometimes with a bit of ringing, and as you are aware, the value of choke to quieten every load may not allow the same Inverter to survive very long under repeated transient 25+kW loads. BTW, running basic household devices will see over 12kW DC peaks into an inverter, that's not a guess, that's a fact that I constantly monitor. It takes very little to do that. You may also end up chasing your tail trying to get consistent meaningful capture of these AC loads conditions, especially with the equipment that only a few of us may have available to actually do that clearly, and usually that equipment will be in the lower cost range of test gear.
If these inverters were factory produced, they would all behave the same, the only variance would be the DC supply, cabling and Loads, any value of components chosen to solve a problem in one, would apply to all.
If you take all the relevant factors you bought up, and appreciated BTW, then add even more interactions in the magnetic interplay of widely varying toriod builds, the current and drive requirements for both the effect it has on the magnetics, and the effects all of this has on the problem of keeping the FETS we use, or can afford to use, from self destruction, and then throw into the mix, hand wound and user constructed toriods from junk inverters of various types, welders, variacs, etc, and winding transformers that are different in so many ways.
I really feel that trying to find a definitive answer that is constant and applicable to all hobby built inverters, with the equipment we hobbyists have, and while there is still debate on the best drive system, gate resistance, board design, layout, FET, resonance point, current protection, failure modes, and on and on, is something I don't really have the time or equipment to do.
Within the limited resources available, I think most of us try to document our own build observations in the hope that others, some with real understanding, experience and even training in various fields (some quite complex) of thermal, electronic and magnetic theory and likely ones I've missed, which all combine to make a powerful inverter, can either shed light on what we observe happening in our builds, or that our observations may help new builders to at least have an insight into the way these inverters are constructed, and the problems encountered in building one.
Once again, and right on cue, another spanner is thrown into the works.
The EMF produced by my inverter in my HF radio equipment has now increased from nothing to pretty dam bad - what changed - I changed chokes, in particular that small Ferrite core.
So now I have to drag out my spectrum analyzer, see what I'm looking at, hope splitting the big choke for each leg reduces it somewhat, and then likely set about designing an RF filter while finding the exact mode of injection or radiation paths into the external communications equipment, which is something akin to herding cats.
As I sit here typing, there have been five 500A peak DC load startups, each a full 2.5 seconds of solid 500A, followed by another half dozen 290A DC startup loads each time, the UPS is silent! the inverter is silent!
How do I know they started ? Exactly the same way when running on Mains AC, the "slightest" flick in the LED fluorescent lights above, and slightly in front of me, if I was walking around the work area I likely wouldn't even notice.
BTW, Inverter has been running nonstop for days now without a single fan startup, and the Toriod temperature managed to reach a staggering 30.6 deg C a few minutes ago, and that's running inside a warm man cave. .It's all too hard. Mike.
KeepIS Guru Joined: 13/10/2014 Location: AustraliaPosts: 1679
Posted: 04:34am 25 May 2023
Copy link to clipboard
Print this post
Spent a bit of time over the past week making up a picture wiring and large component layout, the area on the left is pretty close to the Top control interface deck in the Inverter housing, it really makes finding where a terminal goes so easy in the future, anyway I like it, I've made these in the past and really glad I did as the memory gets a bit hazy over time when you want to repair or modify something.
. Edited 2023-05-25 18:51 by KeepISIt's all too hard. Mike.
Revlac Guru Joined: 31/12/2016 Location: AustraliaPosts: 1024
Posted: 12:01pm 25 May 2023
Copy link to clipboard
Print this post
Good ramble though, I totally agree with it, Its always good to read a build thread and all the experiences along the way good or bad, it helps others to see if they are on the right track when making there own. The basic designs work, its the variables (parts) we throw into the mix that make it unique.
I like your pictorial diagram, I recently tried to make one for a friend doing an off grid solar setup, used pictures from the net but it was not quite what I liked, should have took some photos, cropped and drawn it up. Need to learn how to use a photo editor and drawing program. Cheers Aaron Off The Grid
analog8484 Senior Member Joined: 11/11/2021 Location: United StatesPosts: 108
Posted: 04:53pm 25 May 2023
Copy link to clipboard
Print this post
It's now very clear that the primary choke design is all about balancing multiple conflicting objectives.
Perhaps adding back a small ferrite or two would help?
I think your build has really demonstrated the value of a high saturation current primary choke.
analog8484 Senior Member Joined: 11/11/2021 Location: United StatesPosts: 108
Posted: 04:57pm 25 May 2023
Copy link to clipboard
Print this post
Just curious, which components besides the transformer cores were the most expensive? Choke cores? Caps? MOSFET's?
KeepIS Guru Joined: 13/10/2014 Location: AustraliaPosts: 1679
Posted: 12:42am 26 May 2023
Copy link to clipboard
Print this post
The Mosfets were really low cost, I purchased 50 X HY5606W from LCSC.COM, wiseguy and others users have suggested them as being about the best reliable source. After going through all the parts I purchased from them, it's pretty obvious they are the real deal.
I didn't even bother matching the FETS, or even checking then. Everything came in the correct package, FETS in the correct long plastic casings, not loose in bag or wrapped in block of tape like other sources, and cost no more then the usual crap, they also supply most of the devices used in this build, including the littler DC-DC supplies. Nice when you can get almost everything from one source.
The most expensive items were the CAPS, partly because I got them from a big AU electronics supplier, almost every other source had no stock of Screw CAPS in the values I wanted, these units has full data sheets including real current ratings and temperature, voltage testing under various conditions.
I had bought some name brand CAPS prior to these, after testing them I would not trust those to filter a 1.5v transistor radio supply, OK.. slight exaggeration, but you get the drift.
IMHO the FETS, obviously, and the CAPS are BOTH critical to the reliability and life I expected from this inverter build, built it once, build it right.
Toriods were scrounged from old Grid-tie sunny Boy inverters, I paid $150 for 3, but the ones I wanted are getting really hard to find and the price for one is sometimes twice what I paid for 3, I didn't want to chance waiting and miss out on them.
Toriod Choke Cores from the Grid-tie units and one other Grid-tie, can't remember names but I scrounged around and found a few for free, I stripped them all for parts. But those Toriod Choke cores are a low price from China, I think I purchased two extra from there.
A very expensive overall system component though was 2G, 0G and 00G cable, I was gifted some 2G and 0G, but the shortfall (I didn't bother with 00G) cost me $200.00 - I was desperate -- what can I say crazy prices. . Edited 2023-05-26 10:44 by KeepISIt's all too hard. Mike.
KeepIS Guru Joined: 13/10/2014 Location: AustraliaPosts: 1679
Posted: 12:54am 26 May 2023
Copy link to clipboard
Print this post
Thanks, I have a very old version of Photoshop, still works great. I believe there are free programs available now that are similar.
The layout and connection diagram I posted was really just chucked together. I wasn't going to waste much time on it, I just wanted to get that Control section on the left very close to the actual build while it was still fresh in my mind. Using photos of the connectors and parts will make it feel familiar 2 years from now when I have to find a fault, components eventually fail, especially little bypass caps. .It's all too hard. Mike.
wiseguy Guru Joined: 21/06/2018 Location: AustraliaPosts: 1156
Posted: 09:31am 26 May 2023
Copy link to clipboard
Print this post
There is a good chance that when you split the large choke to 2 smaller ones it will help to reduce the RF emissions. If further reduction is required, try 2 x 10n capacitors, 1 each from the transformer/choke nodes to ground (input side gnd).
Can you please describe the operation of the auto inverter / 240V mains input switching that you have incorporated (if previously described a pointer is fine). I understand it auto switches as required back and forth ?If at first you dont succeed, I suggest you avoid sky diving.... Cheers Mike
KeepIS Guru Joined: 13/10/2014 Location: AustraliaPosts: 1679
Posted: 10:06am 26 May 2023
Copy link to clipboard
Print this post
I was going over the peak DC input current testing I did the other day, I noticed something interesting when comparing the recordings of two chokes compared to one choke:
One Choke: Notice the different peak current in each cycle of DC current [Yellow Trace]
The inverter is drawing slightly more current producing one half of the AC cycle compared to the other, this is likely due to having only one choke in one side.
The smooth ramp up in current to 520A is likely due to the 10 Stack toriod choke and it not saturating even at 520A, it still has some inductance past 600A.
There is also more current being drawn over the first few cycles with only one choke, there is now half the inductance and some winding resistance (R-LOSS) removed for good measure.
Below - Two Chokes: A small amount of variation, the Hybrid choke E-Core is totally saturated past 100A and the 16uH Toriod ring choke is close to full saturation, there is less initial startup current for the first few cycles, more initial inductance of two chokes and some extra R-Loss thrown are the reasons.
I need to test this correctly when I split the 10 stack choke into one half in each leg. I have decided to use 7 cores in each half, I can use a turn or two less and have slightly higher saturation as well, meaning a little more inductance in reserve at 520A "plus" current peaks. .It's all too hard. Mike.
KeepIS Guru Joined: 13/10/2014 Location: AustraliaPosts: 1679
Posted: 10:30am 26 May 2023
Copy link to clipboard
Print this post
The Auto operation is as follows:
There is an ATS - automatic transfer switch on the Solar wall.
It switches to Inverter input when it senses a voltage greater than 53V (adjustable) on it's Control Line.
The ATS is dumb:
It will switch to Solar even if the Inverter has no AC output. It will switch back to mains even if there is no Mains AC.
I modded it:
I have an external switch on the Solar wall that is for Solar Priority or Mains priority, it also provides 24VDC from a tiny transformer to indicate Mains AC present or not.
In Mains priority, if there is Mains 240vac the tiny transformer in the box with the priority switch enables a 24v relay in the Inverter top deck, which keeps the inverter off when the inverter is in Auto position.
If mains drops and the Inverter is set to Auto:
The Mains powered 24V relay in the inverter Auto control opens, switching the inverter ON/OFF control to ON.
The inverter starts as normal, The Capacitors are automatically charged and the inverter soft starts as normal.
When the inverter AC output reaches over 200VAC, sensed by a 240v to 24v toriod transformer in the inverter, another relay is enabled and this sends 54 volts to the ATS control line.
The ATS has a 1.5 delay on its control line (under voltage verify) before switching to inverter power.
This setup seamlessly switches back and forward, turning the inverter off and on as needed, but only if I set it to Auto.
If set to Solar priority:
The inverter switches on and stays running until manually switched off, or Inverter AC output drops or the DC control voltage drops for any reason, then it switches back to mains AC supply instantly, regardless of the Mains AC status..
It has worked faultlessly on the previous HF inverter, and in this inverter with the China Driver and obviously works perfectly with your Driver and Power board, after all it is independent of the Driver & Control board.
Getting tired, heading off for a cuppa and an early night, so I hope that makes sense.
Edit: Made this post a bit clearer then it was before I quit last night, at least I think I did.
BTW I only noticed your post after I had posted the Choke startup difference I had been thinking about. . Edited 2023-05-27 05:52 by KeepISIt's all too hard. Mike.
Godoh Guru Joined: 26/09/2020 Location: AustraliaPosts: 458
Posted: 12:33am 27 May 2023
Copy link to clipboard
Print this post
Thanks for the setup advice on over current. I will find time sometime soon to do some adjusting on mine. Reading about your automatic transfer switch, I know that the supply authorities are pretty touchy about the chance of solar backfeeding when their grid is down. Is the setup you have used approved by the supply authority? Pete
KeepIS Guru Joined: 13/10/2014 Location: AustraliaPosts: 1679
Posted: 02:33am 27 May 2023
Copy link to clipboard
Print this post
As nothing is fed through the Mains Switch box nothing could ever backfeed.
A heavy shed feed we had installed from the main fuse box is connected to a separate Fuse panel on the shed, that feeds the input to the ATS, the ATS can only switch it's output between Inverter AC input and that shed AC feed input, again, nothing can ever back feed.
The inverter would instantly trip or be destroyed if that could ever happen.
The ATS output feeds the Shed via a small breaker ELD box which splits the feed for: 1: Powering the shed and 2: A separate isolated emergency circuit running in the house, that circuit has no connections whatsoever to the main fuse box.
There is a separate existing Solar panel and "feed in" Inverter on the house, it's completely independent of the shed solar array, batteries and inverter.
I had an approved "Generator Input" change over switch fitted into the Main fuse box.
That won't be used unless we start having sustained power cuts, or they stop our feed in tariff, in which case the Property will then be permanently disconnected from the stinking Mains supply. . Edited 2023-05-27 13:46 by KeepISIt's all too hard. Mike.
KeepIS Guru Joined: 13/10/2014 Location: AustraliaPosts: 1679
Posted: 04:46am 27 May 2023
Copy link to clipboard
Print this post
Wound two chokes, each a 7 stack comprising 5 x black and 2 x blue toriod rings each, it's what I had, there is not a lot of difference between the T-rings.
Still same 4 turns each, that's all that will fit, 6 cores got me 23uH instead of 16uH for 4 cores:
Inductance is 23uH each, both measured identical of course.
The saturation of one could be enough, but the two combined are passable I guess.
BOTH images are for only ONE choke.
The HF noise reduction with wiseguys design was the first thing I noticed when I swapped the inverter boards over from the China board, it was virtually silent across the band. Nothing else was changed when the boards were swapped.
Running with only one choke bought it back to a similar noise level as the China board. After splitting the 37uH choke into two x 23 uH chokes, it was almost silent again, just needed two RF toriods rings over the AC feeds from the transformer to remove one or two slight birdies on the band.
Started every load and now have some DSO captures to go over. . Edited 2023-05-27 14:52 by KeepISIt's all too hard. Mike.
KeepIS Guru Joined: 13/10/2014 Location: AustraliaPosts: 1679
Posted: 06:04am 27 May 2023
Copy link to clipboard
Print this post
I'm not going to go over again why the chokes stay cool with the cable used, or why I use that cable, one obvious reason and one personal belief reason, So:
1: The chokes are wound with 4G cable.
The current is definitely higher again now, possibly in part due to:
A: The total cable used has been halved since remaking the chokes, less R-Loss.
B:The interplay of lower inductance chokes on each side, however, combined, a higher inductance by an extra 9uH compared to the single 37uH choke used on one side.
Current peaks measured from the base dc offset to the Peak DC current swing is now reaching 568A [26.7kW DC peaks] Measured a few times to confirm and its always the same.
The 1st cycle startup surge is now between 300A and 404A peak, depends on where in the AC cycle that the Compressor takes a dump on the inverter.
Nearly 400ms of huge startup surge before it starts to taper off.
Yellow is Current, Blue AC.
Below: 1st cycle at 404A peak.
The startup AC waveform varies depending where in the cycle it's forced to take the load. VFB in the mix as well? . Edited 2023-05-27 16:13 by KeepISIt's all too hard. Mike.
KeepIS Guru Joined: 13/10/2014 Location: AustraliaPosts: 1679
Posted: 06:24am 27 May 2023
Copy link to clipboard
Print this post
Mike, I thought you may be interested in these. Its the ATS switching the LOADS over to the Inverter, that's a 62A peak DC pulse, not much but obviously the change over is not synced so anything could happen playing Russian roulette, but no blowups with the inverter, Toriod doesn't even grunt at me. Seems like the AC distortion is not causing any real current or instability issues with the toriod or inverter, switched loads don't seem to care much.
EDIT: You know, that AC spike and ripples on the following leading slope look like the result of a brief muted ringing in the toriod.
Below: Inverter Halt [off]. Triggered by SPWMEN enable disable. Wondering if I should fit the all the extra diodes across the FETS, just in case.
For someone new, Blue is Inverter AC out and Yellow is the trigger point on the SPWMEN line to the 8010 when the drive waveforms are disabled at power off.
. Edited 2023-05-27 18:06 by KeepISIt's all too hard. Mike.
analog8484 Senior Member Joined: 11/11/2021 Location: United StatesPosts: 108
Posted: 05:14pm 27 May 2023
Copy link to clipboard
Print this post
How critical is having caps with screw mounting?
Wow ... I forgot about the primary cables. They cables for the chokes probably cost as much as the choke cores.
Thanks for the info as always.
KeepIS Guru Joined: 13/10/2014 Location: AustraliaPosts: 1679
Posted: 10:12pm 27 May 2023
Copy link to clipboard
Print this post
I initially wanted screw mounting caps to make replacement easier, however wiseguys design incorporates two removable cap boards comprising four caps on each board.
I still wanted the caps for higher ripple current ratings that "usually" comes with quality screw in caps. So wiseguy quickly redesigned the CAP boards for me.
I looked at the Caps as a one off purchase, knowing they would not be rebadged fakes and another possible source of Inverter disaster.
Most inverters use solder in caps and preform perfectly.
You've seen the DC input currents and machinery that I need to start on an almost daily basis, it was an easy decision for my particular circumstances.
The Cables:
Chokes don't use much cable, off cuts can usually do them, but the Toriod is another matter, 14 turns through 3 stacked cores is quite a few meters, add that to the battery bank interconnect cables and DC supply cables feeding to the inverter . Edited 2023-05-28 08:13 by KeepISIt's all too hard. Mike.
wiseguy Guru Joined: 21/06/2018 Location: AustraliaPosts: 1156
Posted: 10:17pm 27 May 2023
Copy link to clipboard
Print this post
The chokes each side should be considered as 1 choke, it is a series circuit and the current through any part of the series circuit will be the same at any given time. The only advantage I saw for splitting the choke was for symmetry when looking at transformer primaries with a CRO and reduction of fast switching edges applied to the transformer so less coupled noise.
The current peaks being higher on one side should have zero to do with the choke being on one side. When compared to normal 240V behaviour with regard to current peaks are they identical ? Have you put a heap of resistive loads (all stove elements turned on) for comparison of the current peaks too?
If the symmetry issue of current peaks is definitely caused by the inverter my guess would be possibly the dead time capacitors across the LED's of the opto coupler drivers not being matched for each side drive. If there was unequal dead time that could rob one side of energy throughput for the side with the greater dead time, another possible reason is slight asymmetry of the EG8010 output waveforms for each half.
Last comment, is it possible that increasing the inductance from 37uH to 46uH has slightly reduced the efficiency at high loads resulting in higher current required to start the same load ? I only ask these questions and make these comments in the interests of being as nice to the inverter as we can be.If at first you dont succeed, I suggest you avoid sky diving.... Cheers 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
I totally agree with it, Its always good to read a build thread and all the experiences along the way good or bad, it helps others to see if they are on the right track when making there own.
I think your build has really demonstrated the value of a high saturation current primary choke. 
crazy prices. 
Measured a few times to confirm and its always the same. 
