 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 9 of 26   |
|||||
| Author | Message | ||||
Revlac Guru  Joined: 31/12/2016 Location: AustraliaPosts: 1024 |
The link posted earlier is not working for me, I haven't measured peak current before, just the typical 2 second reading with a clamp meter when stating anything up to 2Hp motors or a little more, if the peak currents are so high I should have Murdered the 200A BMS already, there is still time to do that.   The 900w bore pump, with a locked rotor was around 2.3Kw for a few minuets before the breaker popped, it is very easy to run that. Would like to try testing the peak current one day  .Cheers Aaron Off The Grid |
||||
KeepIS Guru Joined: 13/10/2014 Location: AustraliaPosts: 1679 |
Hi Aaron here is the full link and hopefully it does not change. These are unidirectional DC Current. Normal 4V output @ rated Current for all units, but they will go higher, like to over 11 volts for the 12V supply units, I have not measured the 24v supply units, I have one of those as well. Very easy to use but the only document I have found is the Web page. Not a problem but if you are interfacing to a 3.3V micro you may want to have a series resistor and a diode clamp to limit current and voltages above 4v. Link to Current Sensor It's all too hard. Mike. |
||||
| analog8484 Senior Member  Joined: 11/11/2021 Location: United StatesPosts: 108 |
Understood. I was just wondering if the AC power factor is low as that can lead to higher DC current draw. It's a bit surprising the dust extractor has such high power factor. Perhaps the transient startup phase has lower power factor. |
||||
phil99 Guru Joined: 11/02/2018 Location: AustraliaPosts: 2135 |
"It's a bit surprising the dust extractor has such high power factor. Perhaps the transient startup phase has lower power factor." Yes, usually 0.8 to 0.85 at full speed with full load, less at lower load. Around 0.3 at startup. |
||||
| KeepIS Guru Joined: 13/10/2014 Location: AustraliaPosts: 1679 |
The Power factor changed UP to that reading with the DE running, but with everything else running as well, the house and workshop from the inverter. I cannot just isolate the Dust Extractor. However you have to consider the accuracy of a combined AC display unit for Voltage, Power, kWH, Frequency, PF and Running time. In any case the massive Air compressor draws more, the Log cutting band saw a bit less, there is obviously nothing wrong with any of the equipment. It is what it is. It's all too hard. Mike. |
||||
| KeepIS Guru Joined: 13/10/2014 Location: AustraliaPosts: 1679 |
I wanted to clear up the impression that this 3HP Dust Extractor is drawing abnormally high current and power. I said many times that I wanted to determine the Peak currents that an Inverter is subject to under various loads. Somewhere along the way it was assumed that the D-Extractor was drawing high power, if you had a quick read of my disjointed ramblings then I can understand why one would think that. 1: The Peak AC Current from the inverter is around 42A @ 220V (normal Inverter AC out is 230V), so power consumed by the 3HP DC at startup, having No start winding, No start Caps, No start Cnt-F switch, only a Run Cap, is around 8.7kW. Over the years many people who have installed this single Phase 3HP DC have had to get the workshop wiring upgraded, or modify it to run from a VFD to reduce the stall startup current. My Inverter powering the Dust-EX: AC Output 8.7kW @ 42A DC Input 9.4kW @ 200A Peak current pulses are 500A @ 100Hz (distorted half sine wave pulses), 1 Pulse for each half of the Inverter sine wave. The Peak current switched by the FETs when starting the DE is 500A every 10ms for two seconds. If I use sine wave RMS calc then the power input is too low for the AC output power, closer inspection of the pulses indicate a flattish rise with some narrow peaks and other distortions so I changed the conversion slightly (fudge). Shocking, unscientific, something else wrong. Really NOT interested. I don't have a true RMS meter that can read that high to get an accurate indication, I'm not sure I could capture it by using a meter at the Sensor output due the brief startup time. Frankly, I'm only interested in finding the peak current pulses that an inverter has to handle. This paints a clearer picture "at least for me" of the Inverter design for wiring, choke saturation and toriod design for high load low DC input and other build parameters that need to be considered when making an inverter "almost" bullet proof. Finally, consider the Peak DC current I see from powering on a smallish Toriod Mains transformer. When the Toriod transformers residual magnetic flux is in a bad place, the current surge is 250A at power on, otherwise it's virtually nothing. You know when it spikes as the Inverter Toriod gives a big Thump and the inverter DC input voltage has a good downward spike on the Graph. I touch the Peak instantaneous graph overlay button and there is that DC 250A spike. BTW to recap: I capture instantaneous DC input LOW voltage, right at the inverter board + - terminals, and Peak Load Current for the Graph Recordings overlay. The DC Load voltage used for Calculations and display is always the lowest voltage sag. It agrees 100% with a DSO across the inverter board terminals, even with the highest transient load pulse. . Edited 2023-04-09 12:49 by KeepIS It's all too hard. Mike. |
||||
| phil99 Guru Joined: 11/02/2018 Location: AustraliaPosts: 2135 |
Regarding the inrush when switching on a toriod transformer only a small part of it is due to residual magnetism. The GOSS core material has quite low residual as residual magnetism causes power to be converted to heat on every cycle (hysteresis loss) and it is designed to minimize that. The main culprit is the point in the cycle at which it is switched on. An engineering text on the theory of AC machines will explain it better with graphs and equations than I can, but here goes. To start assume it is an ideal inductor and the power is already on. The current waveform lags the voltage waveform by 90 deg. so when the voltage is at its peak the current is at 0 and begins to rise. As the voltage falls the current rises, when the voltage is at 0 the current is at max. As the voltage goes negative it pulls the current back down. When the power is off the inductor current is 0. Switch it on at the peak of the voltage cycle and exactly the same thing happens, as that matches the normal state. Switch it on at 0 V and as the voltage rises so does the current, slowly at first reaching it's normal maximum peak as the voltage reaches its peak. as the voltage falls the current continues to rise reaching double the normal peak as the voltage reaches 0. It can only fall when the voltage goes negative. The current has had a full half cycle to rise instead of the normal quarter cycle. In the real toroid that may push the core into saturation, reducing its inductance. That causes the current to rise even further. A similar, though smaller thing happens with induction motors. . Edited 2023-04-09 13:57 by phil99 |
||||
| KeepIS Guru Joined: 13/10/2014 Location: AustraliaPosts: 1679 |
When a transformer is de-energized, the magnetic core material will contain a certain amount of residual magnetic flux. This flux may intensify saturation and lead to high transient in-rush current upon remagnetizing. Published in: IEEE Transactions on Applied Superconductivity ( Volume: 24, Issue: 3, June 2014) It's all too hard. Mike. |
||||
| Revlac Guru Joined: 31/12/2016 Location: AustraliaPosts: 1024 |
Thanks Mike, that paints a better picture to understand (for Me at least), I fixed up the start switch on a friends large air compressor as it was not starting, just stalling his genny and blowing breakers all the time.Since it is fixed I checked the starting surge with a clamp meter 2 second inrush setting and it was reading 42A at 230Vac, he wants to use this compressor off grid at some stage, and so will need half decent setup to run it. I Also have a large isolation transformer I use to run the dishwasher so it starts at a higher voltage, and refuses to run at 230vac like it used to, When I plug in the isolation transformer I quite often hear a loud thump but not always. Cheers Aaron Off The Grid |
||||
| KeepIS Guru Joined: 13/10/2014 Location: AustraliaPosts: 1679 |
I also have a big 30 year old Isolation transformer, around 2.5kW, does the same thing and it will sometimes drop a normal mains breaker at power on. So that compressor startup would also be around 500A peaks on the 48v inverter  . It's all too hard. Mike. |
||||
| KeepIS Guru Joined: 13/10/2014 Location: AustraliaPosts: 1679 |
After learning more about the nature of the input currents of the inverter, I found an hour to finish a Modification I wanted to try, and it's getting back to the boring subject of Chokes. I have a 2 Ferrite E-cores joined together and wanted to build a high current version with lower inductance and heavy cable. The idea was to run it series with a 3 core Toriod ring choke, Why? As noticed by others builders, the saturation curve of the E-Core is quite wide and reasonably flat until the saturation point. The Toriod chokes have much higher Saturation but the curve is little steeper, further, the Ferrite E-Core seems to be better at choking some very sharp HF transients around the FETs. I wanted to have the best of both worlds and have nice wide flattish saturation slope up to the Saturation point, which should be a little less abrupt. I think it was Poida who noticed less spikes around the FETs and a better looking drive signal with the E-Core compared to the same inductance value of Iron-Core choke. The result of a gaped E-Core crammed with 5 turns of heavy cable was a 16uH choke that has a flat and gentle slope to around 220A. The smaller Toriod Core choke also starts to climb at around 200A. NOTE: The smaller 3 ring Toriod choke is 37uH. The Combined inductance of both was 53uH, I would have liked the 37uH choke to be a bit less inductance but I will test it as is. The combined Saturation point is now around 400A and it's a nice curve. This combined Choke is placed in the HI Freq side of the Inverter drive. A 600A choke is in the LO side to help the FETs when starting those really big 500A+ DC current peak loads . The Mod really made no noticeable difference to the Inverter, all heavy loads start as normal, but at least I have that nice wide slope at higher saturation to boot, and maybe happier FETs. The Yellow Trace is voltage across the Resistive current Shunt: 20mV = 80A The Blue is a Current Clamp: 500mV = 62.5A, so scales are slightly different. This is the Smaller Toriod China Choke around 180A.  This is the Ferrite Choke @ around 210A:  Combined: The Blue trace continues climbing off screen and curves at the same value as the yellow, I wanted to keep the 3 photos at the same settings. The voltage is over 110mv just off the edge of the screen before the it starts to really climb, that's over 400A, but keeping in the linear part @ 80mv it's 330A.  . Edited 2023-04-10 16:41 by KeepIS It's all too hard. Mike. |
||||
| nickskethisniks Guru Joined: 17/10/2017 Location: BelgiumPosts: 458 |
Hi, did you measure those spikes @no load? How high are they? Capacitance of the inductor and transformer can also play a role. Which "probing technique" do you use, it might effect the measurements. I 'm finding this interesting, never thought those spikes were that high. |
||||
| KeepIS Guru Joined: 13/10/2014 Location: AustraliaPosts: 1679 |
Hi, the measurements were described a few posts back, and the reason for the High DC pulses every 10ms. I'm not sure what spikes you are referring to? If you mean the 500A+ Pulses (100Hz DC Sine wave pulses into the inverter) that is covered a few posts back. Obviously the Current changes with Load. Like all DC switching circuitry there is some noise spikes on the DC input, but these are in the 50 to a few hundred millivolt range depending on the Load. The AC supply is very clean. There is no probe, just a DSO and various isolated low noise current clamps, the main one rated at 500A with good accuracy to 700A and usable to over 1000A. I was also surprised at the peak currents, and questioned them when I found out my previous measurements at 390A were low and compromised by DC Clamp limitations, including a 400A commercial unit. I'm across building Toriod transformer, Chokes and the design and experimentation has fully tested all variations in this test inverter along the way. It's all too hard. Mike. |
||||
| KeepIS Guru Joined: 13/10/2014 Location: AustraliaPosts: 1679 |
I changed the 37uH Chinese choke out for a 4 turn 3 Core toriod at 15uH. The combined Ferrite and 15uH Toriod choke is now 30uH which is much better from past testing, they have the same combined current rating and maintain that slower transition to saturation that I was after by virtue of the Ferrite E-Core in the HI side. The Inverter:  The Cabinet door opened showing 15uH 3 stack toriod on the right in the lower part of the cabinet (I made this one nicely as I knew I would be keeping it. It's in series with the Ferrite E-Core on the HI side of the inverter drive.  The 4 turn 10 stack toriod on the left is 37uH with an almost linear saturation curve past 800A, but I still downgrade that to the straight part at 600A. It's in the LO side, it's there to help the FETs switching 525A peak currents from seeing a piece of wire caused by a saturated choke, instead, the FETs see an inductive resistance provided by this Choke. (I know it's not the correct term, but it might make the intention clearer to someone new reading it) The Choke is not pretty, but I threw it together for a test and it worked perfectly so I left it there. I'm going to rewind it tightly like the 3 stack on the right and cover and mount the 10 cores correctly when the new inverter driver is built. A lot of tiding up of the Toriods and wiring to do also.  The Top lifts of the Cabinet for access to all interface connections. On the Left rear is the High Current High Voltage Contactor for automatic cap charge and DC isolation when the inverter is in the off position, there is a 250A fuse, it takes time to open at continuous RMS currents, peak 500A pulses will not blow it.  There is no back on the interface connection shelf, it's left open, and when fished the AC section on the right will be enclosed. This makes it so easy to test things and check the unit during the build and long run in and testing process.  As it is now there is just one 3 position knob, Off (and DC isolated as well), ON (Standby) and RUN (Inverter AC Running). Never have to worry about switching the battery banks off or reconnecting. Absolutely fool proof, sounds like it's made for me (I'm getting old)   . Edited 2023-04-11 18:08 by KeepIS It's all too hard. Mike. |
||||
| analog8484 Senior Member Joined: 11/11/2021 Location: United StatesPosts: 108 |
Great looking system. That's a lot of chokes. The term you are looking for is "reactance". |
||||
| KeepIS Guru Joined: 13/10/2014 Location: AustraliaPosts: 1679 |
Thanks, I know the correct term, but someone new to electronic may not understand reactance in this context, which is why I went to the trouble of pointing that out and use the (technically incorrect) term inductive resistance, to differentiate from a saturated choke with only the resistance of the winding wire left in circuit. It's actually only two chokes. One a 31uH hybrid of Ferrite and Silicon iron, and the other a big non saturating 37uH choke in the low side, my extensive testing proved this made a big difference to this "Toriod" in the inverter under difficult complex loads. This inverter is for all intentional purposes, Silent, even when running at 5kW for an hour. The only thing that gets a very slight Hum from it is the Half cycle load from the Heat Gun on LO setting (diode in series with the AC element on LO). The simple idea of using a Toriod Ring choke in series with a Ferrite E-Core had never dawned on me before. I really liked the Ferrite E-Core choke in the early testing phase but the low saturation point was a disappointment, now I have the best features of both in this hybrid design. BTW the Toriod transformer is around 190mm high X 190mm wide, 210 X 210 if you include the primary winding. Looks smaller because of depth of field. . Edited 2023-04-12 13:56 by KeepIS It's all too hard. Mike. |
||||
wiseguy Guru Joined: 21/06/2018 Location: AustraliaPosts: 1156 |
Just out of interest Mike where did you source all the toroidal choke rings from ? Are they the Blue sendust type material ? Do you have a part number or specification for the material ? I know I could probably work it out from the number of turns final value no of rings etc but I think as I get older I try to find the easiest ways first It looks the part - and I like the separate sections for the various stages. If at first you dont succeed, I suggest you avoid sky diving.... Cheers Mike |
||||
| Revlac Guru Joined: 31/12/2016 Location: AustraliaPosts: 1024 |
Mike, I was going to ask about the black ones, but answers are on page 3, I found 8 of the blue ones in an Aurora inverter, not sure about the actual size yet. Edited 2023-04-12 21:05 by Revlac Cheers Aaron Off The Grid |
||||
| wiseguy Guru Joined: 21/06/2018 Location: AustraliaPosts: 1156 |
Thanks for the pointer Aaron, I haven't had the pleasure of stripping an Aurora yet lol. All I seem to get are the standard steel type c-core inductors. I have said before that I much prefer the sendust type toroidal cores for chokes. Mike has used a combination of a Sendust type choke and a ferrite gapped core both in series. My first comment is that in a series circuit it shouldn't really matter much which side of the network goes to the HF side as the currents etc in a series circuit are the same everywhere. But with regard to noise transference in the Toroid I would prefer to have the larger inductor towards the HF leg. The next comment is that in my experience when a ferrite is driven into and running saturated it usually gets very hot - however Mike has stated that nothing gets hot. So why is that ? I am surmising that the saturation currents are not as severe due to the sendust taking over and continuing to have a reasonably linear current characteristic. I can see that I will also be having more fun playing with chokes later when I finally finish my Inverter MK2. But sadly - for me - that will be much later as there will be a fairly big gap in my following and contributions to the BS in about 5 weeks. We are taking the caravan on an extended 5-6 month tour starting off Westerly. Edited 2023-04-12 23:03 by wiseguy If at first you dont succeed, I suggest you avoid sky diving.... Cheers Mike |
||||
| Revlac Guru Joined: 31/12/2016 Location: AustraliaPosts: 1024 |
Yes, interesting how its working, so far I have not seen any commercial inverter use chokes on either transformer primary,but this might have changed in recent years. There are a few green coloured Toroidal rings as common mode chokes, but I had assumed this was a different material. I will do some more experimenting on a working inverter, to see how noisy the primary is.Have a good safe trip. Cheers Aaron Off The Grid |
||||
| Page 9 of 26 |
|||||
 Print this page |
