 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 41 of 41 |
|||||
| Author | Message | ||||
KeepIS Guru  Joined: 13/10/2014 Location: AustraliaPosts: 2272 |
Something I did for fun. No VFB 20V DC supply. A: Test Inverter, Power board with 34,000uF. B: DC voltage DOES NOT CHANGE, monitored with A DSO. C: No AC Feedback connected, besides it's below VFB activation in any case. D: Inverter powered by regulated 20Vdc supply with a max current of 21 Amperes. Pink: Primary Current. Blue: LT AC output. Capture 1: Idle.  Capture 2: 180w Restive Load. A: 11.5A Primary current via Clamp meter. B: 12A vPrimary current via TLS76L1.  Capture 3: Heat Gun. A: 34A Primary current via AC clamp meter. B: 35A Primary current via TLS76L1. C: DC current switches between 11A and 20A every few seconds. D: Buzz in Toroid does not change with that swing.  . NANO:Inverter V 8.4ks - Linux AvrDude GUI script V4.1 |
||||
wiseguy Guru Joined: 21/06/2018 Location: AustraliaPosts: 1317 |
Great image shots and good idea to remove feedback as Nicks also recently did. The yellow drive waveform appears to be phase reversed compared to the output - it just made it a bit harder to interpret at first. Maybe my analysis below is wrong or questionable, I am not trying to speak with authority, its just a theoretical comment and I invite comments for or against. The peak voltage for the positive unloaded 180 degrees is ~ 2 divisions, The peak for the loaded negative 1/2 wave is ~ 1.5 divisions. The result of the asymmetrical waveform results in the positive unloaded current but with saturation peaks. This is why I described our toroids as "puny" compared to the mains. With a distorted waveform such as this applying feedback might make it even worse as it tries to drive harder for the lower average output? My suggestion of 180 degree measurements and applying them to the second 180 degree waveform - not the next immediate period should try to prevent or at least partially mitigate the issue of the low negative toroid drive voltage. I am unsure but suspect that changing the phasing of the feedback might only make a small difference on the idle hum. I am not telling you or anyone to write the code but I can see an issue that we can either accept or have a further play with the code to see if we can get a better result. I am happy to offer up my inverter as a sacrificial unit for testing if it makes any difference. Edited 2026-09-21 18:07 by wiseguy If at first you dont succeed, I suggest you avoid sky diving.... Cheers Mike |
||||
| KeepIS Guru Joined: 13/10/2014 Location: AustraliaPosts: 2272 |
I'm currently laying a partial false floor to finish the Workshop extension, so still no time to do much on the Bench. I intend to capture the Heatgun on the Main Inverter in a real world off-grid situation, I'm sure that 1kW or more of normal running loads when the Heatgun is connected, will considerably dampen the effect on the Inverter. Years ago I remember reading many posts on various forums bemoaning the destruction of Inverters when a stupid load like this was connected. Again, it has never caused a problem for my Inverters. I do understand the desire to see if the effect can be reduced as a technical exercise, I'm just not sure I want to start changing 100% stable Inverter code to accommodate a piece of crap that has no right to be sold or used with Inverter power, or Mains for that matter. If you want to play with that and no one else comes forward, we can certainly do that in the future. Unfortunately, as soon as I'm finished the Shed extension, I will be heading down to inspect the Toroids that have finally arrived, and winding two for a second complete Dual Inverter backup, I then have 6 solar panels to add, cabling to run, an MPPT controller to build, and I still haven't got around to cleaning up my part of the workshop  On another note, and the main reason for a second Dual backup Inverter, I see the cost to just have Mains power connected as a backup is expected to be around $1200 a year or more!! And I can see more control laws coming to prohibit us from saying NO and disconnecting from the grid.   : Edited 2026-09-22 13:39 by KeepIS NANO:Inverter V 8.4ks - Linux AvrDude GUI script V4.1 |
||||
| wiseguy Guru Joined: 21/06/2018 Location: AustraliaPosts: 1317 |
Hope your flooring punishment is almost done by now. After a few days of thought I am coming around to your way of thinking. Whilst I will remain curious about whether it could behave better with poor loads, when it is presented with typical daily loads including welding etc it behaves just fine. I think I was falling into a trap of trying to fix a perceived problem rather than a proven problem. I guess I wanted the inverter to behave really well no matter what crap was plugged into it which is probably overthinking a solution for someone elses poor choice. After replacing 16 FETs maybe better choices will be made  .If at first you dont succeed, I suggest you avoid sky diving.... Cheers Mike |
||||
| mab1 Senior Member  Joined: 10/02/2015 Location: United KingdomPosts: 287 |
I've been following this mostly in the hope of understanding the come-&-go hum at idle (i will try swapping the v feedback polarity, but haven't yet), but i am fascinated by the torroids reaction to a recified load: it's as if having passed current (or more like coulombs - is the area under the curve equal i wonder?) on one half cycle, it needs to pass an equal number of coulombs in the other half cycle, but, being unable to pass them through the load on the secondary, the 'pressure' builds on the primary until the H-bridge voltage drops towards zero, letting them all through on the end of the cycle. And yet the secondary voltage doesn't show this 'pressure'. And then having reset to 'no load' current for the zero crossing, it has to build the load current up again on the new cycle, indicated by the 'sag' on both voltage and current in the rise from zero. Not really planning on experimenting with asymmetric loads myself, but i sometimes think i should scope the waveforms when charging the EV as i get some quite loud transformer noise from that - but then it is a quite powerful ac-dc convertor (up to 22kW) being modulated down to just 2kW. Edited 2026-09-25 01:49 by mab1 |
||||
Revlac Guru Joined: 31/12/2016 Location: AustraliaPosts: 1308 |
This is also present in an old 700w 24v Selectronic inverter, still works, who knows how many others do the same thing.  Cheers Aaron Off The Grid |
||||
| nickskethisniks Guru Joined: 17/10/2017 Location: BelgiumPosts: 493 |
Well I'm "close" (maybe by the end of the weekend) to finishing my first "baby" WG inverter and then I will have a play with assymetric pulse modulation. But not with feedback I think, just to see if this let the transformer hum or cancel it, and then to see the effect on an assymetric load, just for fun.  Here is a teaser:  I need to make thinner plates for the inductor as it is to close to the transformer. But the layout is somewhat made now. It will charge the EV or heat the hotwater tank, it should run 3kW nominal for the heater. But should run the house when I'm converting the ozz inverter to a dual WG inverter. I mostly charge the car at 1700W wich is the lowest setting, but this does not upset the inverter it looks like a nice resistive load. ( it's a E5008 and only capable of 11kw on 3 phase, maybe a next project, a hf 3ph inverter, someday )I'm gonna try to use the car as generator the coming winter, in combination with an eltek psu. Edited 2026-09-25 16:38 by nickskethisniks |
||||
| KeepIS Guru Joined: 13/10/2014 Location: AustraliaPosts: 2272 |
Would you believe that I just finished and had my release pass stamped 5 minute ago, I had stupidly mentioned that the old interlocking vinyl plank sheeting was stacked in one of the sheds, idiot!! another two days added, and because I did such a great job (I'm so gullible) she suggested that her wonderful new extension would be perfect if I could just make a new work bench and some extra shelving for her With respect to your Inverter design, obviously my thoughts are the same, it's sitting about 3 meters behind me as I type, in a completely quite workshop, I walked over and opened the door to the Dual Toroid & Choke compartment, pressed my ear against each toroid, and they were silent, power level was only 700 watts, a few minutes later the loads started to climb past 7kW as the oven, stove, and other kitchen items started running, there was just a hint of hum at that point, IMHO it's perfect. NANO:Inverter V 8.4ks - Linux AvrDude GUI script V4.1 |
||||
| KeepIS Guru Joined: 13/10/2014 Location: AustraliaPosts: 2272 |
That looks really cute NANO:Inverter V 8.4ks - Linux AvrDude GUI script V4.1 |
||||
| analog8484 Senior Member Joined: 11/11/2021 Location: United StatesPosts: 226 |
Regarding the idle hum, has anyone observed it with varnished transformers? Just wondering if varnish could help. In any case, the idle hum seems harmless. |
||||
| analog8484 Senior Member Joined: 11/11/2021 Location: United StatesPosts: 226 |
I imagined similar instability issue when I looked at the nanoverter control loop a while back. However, I think one thing that helps is the very low proportional gain in the PID loop but it also means it's very slow to adjust to load caused AC voltage distortions. That may work for half wave rectified loads but in general the control loop still does not seem well suited for significant non-linear loads. The control loop is fine for linear loads and it's also okay for relatively small non-linear loads where the inverter capacity is much larger than the non-linear load as Keepis has demonstrated. But when the non-linear load is large relative to the inverter capacity then you can get serious transformer saturation as nickskethisniks's scope capture shows. Unfortunately, it's not practical to significantly oversize inverter capacity for me. I need an inverter to AC couple with a large (relative to the inverter capacity) array of grid-tied inverters that have very aggressive anti-islanding mechanisms that can mimic large non-linear loads. They are worse than the a-hole designed heat guns that skip cycles that you imagined. The anti-islanding disturbances can easily distort the AC voltage waveform which if not compensated for quickly (within 1 or 2 AC cycles) would cause the grid-tied inverters to stop generation then repeat the cycle leading to total system shutdown. This observation is based on testing with a couple off the shelf low frequency inverters with slow control loops. All this is a long way to say the control loop limitation is a key reason why I have been hesitant to take on the effort to build one so far. So, I second your request and would appreciate any control loop and associated HW improvements to handle non-linear loads more smoothly. Also, smooth frequency shifting for GTI curtailment would be the other feature desired for my needs. I have looked at some inverter reference designs that compensate for non-linear loads in real time and they all have AC current sensing as an essential control input. The sampling frequency is much higher than mains frequency, usually same as the PWM switching frequency. The control update bandwidth is also typically much higher than mains frequency. So, I wonder if the existing inverter HW components can really meet the challenges of non-linear loads. |
||||
| KeepIS Guru Joined: 13/10/2014 Location: AustraliaPosts: 2272 |
This is a DIY inverter for off grid AC supply. It handles any nonlinear load that has been thrown at it, the standard single power stage version handles the same massive surge loads that the Dual Inverter is now running, but with half the current in each Power stage. Builders use it with Back Feed, it wasn't a design criteria, if someone has such nonlinear loads that this Inverter will fail, then they are going to spend "big time" for a massive special Industrial Inverter and not build a DIY powerful off-grid power Inverter. Wondering if this Inverters HW components can really meet the challenges of non-linear loads is is is ??????? a bit like navel gazing IMHO . Edited 2026-09-27 11:54 by KeepIS NANO:Inverter V 8.4ks - Linux AvrDude GUI script V4.1 |
||||
| Godoh Guru Joined: 26/09/2020 Location: AustraliaPosts: 707 |
Analog, i have varnished transformers in my 8010 based inverters, i dipped them in polyester varnish a couple of times and air dried them. They run very quiet. The only time the transformer gets noisy is when like Mab, I am using it to charge our EV. So when i charge the car i now switch to a 3kw latronics inverter which the GTI inverters can backfeed through without becoming noisy. I would think that varnish would make the transformer last longer as there is a lot less vibration in the windings that could rub off the insulation after time causing internal shorts. Pete |
||||
| KeepIS Guru Joined: 13/10/2014 Location: AustraliaPosts: 2272 |
Only if you're sure it's the toroid core or secondary winding really vibrating, and mine are varnished. Yes the idle hum is an absolute non event, a few mA of current if that. Transformer based GTIs could sometimes give a loud metallic bang and a loud hum in service, and forced the relocation of the GTI in some installations, from what I have read and heard from owners that was a given back then. You may have had a different experience? : Edited 2026-09-27 12:18 by KeepIS NANO:Inverter V 8.4ks - Linux AvrDude GUI script V4.1 |
||||
| Page 41 of 41 |
|||||
 Print this page |
 |
The Back Shed's forum code is written, and hosted, in Australia. | © JAQ Software 2026 |
