 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 : Hopefully? Another 48vdc-240vac Toriod Inverter build.
  Page 25 of 26   |
|||||
| Author | Message | ||||
Murphy's friend Guru  Joined: 04/10/2019 Location: AustraliaPosts: 648 |
Yes, I tend to agree with you about the 75Hz static resonance. This only became an item after warpspeed mentioned this was the optimum harmonic killer. But, as you say, it keeps changing all on its own there is not much point pursuing it. I'll just leave the 3.6uF cap I put there and keep an eye on the waveform to see if it gets worse over time. Perhaps not at my place as the inverter is running non stop and the heaviest start load is a metal chop saw occasionally. 70 mosfets? what are you building? My warpverter uses only 40 mosfets. But wait until you see my next inverter  . Yes, I'm still at it, will have half a dozen then - too much fun building them  |
||||
KeepIS Guru Joined: 13/10/2014 Location: AustraliaPosts: 1679 |
A variation on that old advert about a laxative: "I don't need 70 FETs, but if I have 70 FETs, then I'll never need 70 FETs  . Edited 2023-08-28 07:33 by KeepIS It's all too hard. Mike. |
||||
poida Guru Joined: 02/02/2017 Location: AustraliaPosts: 1418 |
I have a version of the picoverter code that makes a variable output frequency (while keeping the output voltage constant) Freq ranges from 33 Hz to 99 Hz. Put a 0-5V signal into A4. 2.5V approx = 50Hz Would this be useful to you in your work tuning the transformer/capacitor combination? Maybe see what happens when the output freq = resonant freq.. I use it to take the toroid into varying degrees of saturation It might be interesting to see the asymmetrical primary current with different freqs and a remenant magnetism in the core. To use it, you will just take the two PWM outputs from the nano and feed into Wiseguy's low/high side with dead time circuit. No need for IR2184 gate drive ICs wronger than a phone book full of wrong phone numbers |
||||
| KeepIS Guru Joined: 13/10/2014 Location: AustraliaPosts: 1679 |
I do remember a post from you discussing saturation and using that code, it may help indicate the resonance under idle conditions, I would be very interested in trying that, maybe on the next build where the complete inverter will be out of the cabinet and running on the bench. I would be very grateful if you could point me to the code download and possibility any relevant build information. Thanks. Edited 2023-08-28 15:59 by KeepIS It's all too hard. Mike. |
||||
| poida Guru Joined: 02/02/2017 Location: AustraliaPosts: 1418 |
No problem, attached is the code I used today when playing around a bit. It has the "perfect symetry" code that ensures the best waveform it can produce is output. I don't know what you will need to drive the inverters you are building. (I suspect you have one that takes a single PWM signal and creates the low and high side drive signals with dead time. This is built with logic chips along the ideas Wiseguy has been working on) But if that is not the case, use something like the IR2184. The high side outputs float on top of the voltage on Vb And Vs is the high side "ground" reference. In the past I have hooked up Vb to Vcc and Vs to Vss when testing if these chips were dead. This gives low and high side outputs going from GND to Vss (usually 15V). I like them. $8 each, a couple caps and a diode and you have proper gate drive signals. A minimal inverter control that uses this code needs only a Nano and a bit of wiring. Vfb needs to be somewhat smoothed out via a LP filter. See the picoverter schematic. 2.75V DC output from this needs to correspond with your desired AC output voltage. Feed Vfb into nano's pin A0 Pin D8 is the inverter run control. Pull it down to stop the inverter. Pull It up to 5V to run it. Pins D9 and D10 are the PWM outputs. These can drive 20mA each. D9 will be for one of the half bridges, D10 for the other. In the picoverter I use IR2184 ics to generate the low and high side drive signals. D9 and D10 are similar (the same, but with some dead time) as the high side gate drive signals. On boot, the Nano can let pins go HighZ so pull D9 and D10 down to ground with 10K resistors and there will be zero spurious PWM signals. I like the soft start function since it lets me bail out and stop the inverter before any damage occurs. I would mount the run switch in a robust way so you can get to it easily. Put the wiper from a 1K pot on pin A4, and the other 2 pins 5V and ground. When in mid-point it will be 50Hz. All signals both in and out are relative to the nano's GND. Pin D7 will output a square wave at the AC output frequency. This is ideal for DSO triggering. Pin D5 is used to drive the IR2184 shutdown pins on the picoverter. On boot, D5 is low AND we need to pull it down with a 10K. It is pulled up to 5V when the inverter is running, including soft start. It is also up during soft stop and then it's pulled low when soft stop is complete. I think WG's gate drive design has a shutdown control for the outputs. nano_1_v7_no_bessel__wider_vfreq.ino.zip wronger than a phone book full of wrong phone numbers |
||||
Revlac Guru Joined: 31/12/2016 Location: AustraliaPosts: 1024 |
Great,  will have a go at it, Thanks.Cheers Aaron Off The Grid |
||||
| poida Guru Joined: 02/02/2017 Location: AustraliaPosts: 1418 |
Aaron, this code is a plug-in replacement for the picoverter. Except it has no LCD support and no LV cutoff etc., etc. It just runs the inverter. But it's amazing hearing the toroid hum more or less as you change the freq. Look also at the idle DC input current change. I think the toroid manufacturers have the 240V winding & core designed at a small amount of saturation at 50Hz. At 55 it's lower input current. At 45 Hz it's a lot more. wronger than a phone book full of wrong phone numbers |
||||
| analog8484 Senior Member  Joined: 11/11/2021 Location: United StatesPosts: 108 |
That makes sense. I doubt it's possible to maintain near 75Hz under various loads without a controller that performs harmonic compensation in real-time. It looks like significant residual magnetism can happen easily. Perhaps the inverter should try to demagnetize the transformer at startup? |
||||
| KeepIS Guru Joined: 13/10/2014 Location: AustraliaPosts: 1679 |
Well, we already soft start the inverter via SPWM, however I don't believe that has any bearing on running state, except to help with lessen the effect of residual flux that can cause a very high current surge as startup if the SPWM is switched on full, even with slow start you can still get a decent surge and those strange growls that fade away as the toriod settles into what I call, a stable running state, but the growl is not there all the time and neither is the surge. I think that what happens at startup is likely only transient, what happens when the inverter is running under complex continually varying load conditions. Power cycling a VARIAC is the best example of this startup surge effect, even the AC Mains will occasionally drop a CB with a big Variac. . It's all too hard. Mike. |
||||
| KeepIS Guru Joined: 13/10/2014 Location: AustraliaPosts: 1679 |
Thanks, looking forward to having this neat testing device in my DIY Inverter tool box. It's all too hard. Mike. |
||||
| analog8484 Senior Member Joined: 11/11/2021 Location: United StatesPosts: 108 |
I was thinking the controller could go through a few soft-start/soft-stop iterations (at lower voltage?) before the full ramp up and closing the output relay at startup. |
||||
phil99 Guru Joined: 11/02/2018 Location: AustraliaPosts: 2135 |
As KeepIS has shown as little current as a 1.5V cell produces (a couple of amps perhaps) is enough to change the residual, so when running every half cycle will finish with a new residual. What the starting state was is of little importance after just one half cycle of normal running. |
||||
| KeepIS Guru Joined: 13/10/2014 Location: AustraliaPosts: 1679 |
Hi, I'm sure you can find a way to get the transformer back to the same state (whatever that may be) by presenting it with the same start and shutdown sequence X number of times. That keeps a little oscillator and CRO happy by seeing a peak at around 75Hz each time you measure it- but are we still really measuring the true operational resonance The startup sequence is irrelevant. The instant you close the OUTPUT startup relay, you drop a complex load across the secondary, all of that soft start and stop effort is effectively blown into the weeds as an initial Peak DC input surge from 20A to over 100A or more is pulled into the Power stage and Transformer. We have been using a small sinewave oscillator voltage and a capacitor across the secondary of the Toriod transformer that is isolated from the inverter, does the static resonant frequency hold when the Transformer is then driven by an inverter connected to complex loads that likely present some DC components back to the transformer secondary. Now I would like to test this, but for no other reason then an interest understand what is happening I was testing some filtering ideas yesterday, and in between I had the inverter running with four different value of secondary capacitance. I had already tested 3.3uF, so I tried 1.5uF, 6uF and 10uF across the secondary. Four channel DSO looking at DC input current waveform, AC output, across the Inverter output input, and once side of the Toriod primary REF to ground. Difference under the same complex loads was basically zero. Difference in Toriod sound was hard to really notice. A slight difference with part rectified AC (2.5kW Heating element in series with a diode) a slight effect on ringing at the abrupt knee in the AC waveform, but that varied when other loads were in an idle state of off, and importantly, caused no increase in DC input current. The only real measurable difference with the values of Capacitance placed across the secondary was a "slight" increase in ripple on the AC waveform from less filtering of the SPWM drive with lower capacitance. This immediately disappears under any load, the THD changes by around 0.04% THD jumps to 18% with that 2.5kW Heating element in series with a diode, drops to 13% when I install an AC filter of 3.5mH and increase the output capacitance of the AC filter from 1uf to 5uF. Changing the AC filter output to that high value makes zero difference to the AC output idle noise, and makes zero difference to any ringing from the bad loads, and made minimal difference (0.5%) to normal load THD. I have measured and posted in the past that the ringing, visual AC waveform imperfections and now THD, appear to be influenced by the complex interaction of varying non-linearity and inductance of different Chokes, SPWM drive, Toriod and Inverter Loads. I'm sure someone with more knowledge of Magnetics and AC distortion products could Correct or shed (or both) more light on those observation. It's all too hard. Mike. |
||||
| KeepIS Guru Joined: 13/10/2014 Location: AustraliaPosts: 1679 |
Hi, I got called away while replying and missed your post, seems I also missed seeing a tree for the forest, an obvious big tree I used a really fine thread of wire to give the briefest of brush stroke on the edge of a terminal, watching the resonance jump from 46Hz to 220Hz and everywhere in between was more then I was expecting, and an eye opening  moment.. It's all too hard. Mike. |
||||
| analog8484 Senior Member Joined: 11/11/2021 Location: United StatesPosts: 108 |
If so, why did the transformer resonance get stuck high and required the cap change to restore the 75Hz resonance and normal looking output waveform without clipping? |
||||
| phil99 Guru Joined: 11/02/2018 Location: AustraliaPosts: 2135 |
KeepIS has done a lot of careful testing and what I believe he has discovered through it is the magic 75Hz always changes after any run, not just a fault. It wasn't "stuck" as just a couple of amps changes it. He also showed the 75Hz is of no importance when driving a load, its only value is in getting pretty pictures on the scope. The output capacitors only real role is to filter out the high frequency components. The no-load resonance frequency is just a minor consequence of that. Edited 2023-08-31 08:50 by phil99 |
||||
| KeepIS Guru Joined: 13/10/2014 Location: AustraliaPosts: 1679 |
Getting stuck is what it appeared to be, however I believe the reason was due to the difference in SPWM drive from the modified output of the 8010 chip in this new controller leaving the Toriod in a fairly consistent state. The new controller shutdown state changed the resonance up to around 110Hz to 120Hz with the same cap value that made the standard China inverter 8010 output resonance around 75Hz. Knowing what I know now after the last few days testing, I might have realized what was actually happening. Yes, there was a waveform change under High starting currents which was not reported by others in the description of change after some Inverter FET faults. However, I do have the reported change in Toriod sound and so far that appears to be a permanent change. Thinking back a while I remember that I had changed out a big Resistive shunt on the DC input, I replaced it with a Hall effect current sensor, the shunt was heating up above 4kW continuous output and at 5kW I had to shut it down. With that in mind during the re-installation of the Power board and slightly modified controller, I had taken the opportunity to really beef up the DC input common terminating plates and all other connections, again because of the noticeable warmth running at 4kW to 5kW, this would have had a large effect on High current DC input current waveforms as the supply would now be very stiff at 490A to 580A. Adding more capacitance across the transformer secondary made a difference to the high current DC current waveforms, but as I look at the effects of Harmonic distortion on wave shape, I'm beginning to think that just like the apparent clipping of the AC output, The DC current waveform clipping is more likely an effect of Harmonic distortion, as under certian load conditions that clipping sometimes disappears under high current with lower value of Cap. I think the Diode failure has been a good learning experience, and acquiring more advanced diagnostic equipment had revealed waveform imperfections that I had not really noticed before. I look back at some of the earlier assumptions from day one of this journey and I see quite a lot of thoughts that were off base. One of the reasons I wanted to document apparent findings while still fresh in my mind, and with so many "often needless" words in this thread, was to allow me to go back and understand the journey. I do thank you and others who continue to questioned these thoughts and assumptions, and along with these different perspectives, I feel, at least for myself, that I have removed a few veils from some of the quirks/black magic surrounding inverter building that were obscure to me. So once again, the learning process continues. . It's all too hard. Mike. |
||||
| Revlac Guru Joined: 31/12/2016 Location: AustraliaPosts: 1024 |
So, One size fits all then? Perhaps later, if anyone would like to start another thread and test if the resonance has changed since there inverter was built, many different transformers out there so I expect results to very some. Edited 2023-08-31 18:21 by Revlac Cheers Aaron Off The Grid |
||||
| KeepIS Guru Joined: 13/10/2014 Location: AustraliaPosts: 1679 |
The testing I carried out shows that if the Inverter shuts down in the same manner each time then a simple static resonance test may not change much, +- 10Hz. My China inverter was like that. The new controller will also shut down to within a similar range of resonance most of the time, but if using the same cap and transformer as the China unit, it's now at a different resonant frequency, so it needs a bigger cap value. What are we trying to achieve with this static resonance test if the resonance can change with controller design, and a shift in flux in the transformer that is a result of the inverter power down state, and therefor likely to be completely different when the inverter is running, let alone under load. I have two completely different toriod transformers now, so when I get the little project from Poida running, I hope to detect and measure transformer resonance in a running inverter under load, but it will depend on how far I can shift the SPWM frequency before the transformer(s) says NO! . It's all too hard. Mike. |
||||
| Solar Mike Guru Joined: 08/02/2015 Location: New ZealandPosts: 1138 |
Perhaps with your scope switched to view Fourier transform of the AC output sinewave with a variable inverter load; that would show up the best "Cap" to use. I would just go for whatever value has the least amount of distortion harmonics and leave it at that. Cheers Mike |
||||
| Page 25 of 26 |
|||||
 Print this page |
