 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 : Nano Power Inverter - Roll Your Own Style
  Page 8 of 16   |
|||||
| Author | Message | ||||
wiseguy Guru  Joined: 21/06/2018 Location: AustraliaPosts: 1156 |
You are quite correct Tony, when I had the extra 3 Phase outlet fitted I was thinking unidirectional (feed only) and 10+KW for bench testing stuff seemed more than adequate. But my thoughts began to wander "what if the power goes out tonight, how could I easily utilise the inverter power for the house" ? - I know, I'll use my new 3 phase connection along with gee I wish I had put in 32+A cable  .I'm contemplating installing a contactor in the main switchboard to accomplish an automatic changeover. I can use the nano phase compare to keep in step with the mains so it minimises any upset when this occurs. Maybe I could just run a suitable rated single phase cable back to the switchboard if the inverter stays in my workshop and common up the phases at the contactor. The last detail to be worked out is how to ensure the 3PH solar inverter never tries to feed my inverter and always has the mains supply (if its available) to dump power to. If at first you dont succeed, I suggest you avoid sky diving.... Cheers Mike |
||||
| wiseguy Guru Joined: 21/06/2018 Location: AustraliaPosts: 1156 |
Noneya, thanks for the fuse & fuse-holder info. Are the fuses rated for DC - they looked to have a ~ next to their current. Not being picky but I have seen first hand the results of a >50V DC ARC and it was scary, they appear to be ceramic which would be a plus. Mark, yes I agree that once powered up the pre-charge is not an issue but in this testing phase where it is on and off multiple times, you need to keep well focused on what sequence you need to meet. Re-Wiring the house would work but I want to try to avoid that and work with what is - a contactor that links all phases after disconnection from mains should work ok. If at first you dont succeed, I suggest you avoid sky diving.... Cheers Mike |
||||
| noneyabussiness Guru Joined: 31/07/2017 Location: AustraliaPosts: 512 |
Unsure of where the ~ is, however i have been assured by a reputable friend these are more than needed. And have worked twice now with accidental shorts with no bang / flash personally. I think it works !! |
||||
| wiseguy Guru Joined: 21/06/2018 Location: AustraliaPosts: 1156 |
The ~ is the little mark next to the 500V ~ & 600V ~ with their associated fault current interruption ratings (120kA & 50kA ?) . The Main current rating ie 150A does not appear to indicate AC or DC. I do accept that they have worked ok with DC without disaster. If at first you dont succeed, I suggest you avoid sky diving.... Cheers Mike |
||||
| wiseguy Guru Joined: 21/06/2018 Location: AustraliaPosts: 1156 |
Time for another update - this inverter build is taking a very long elapsed time.... I bought the kilovac solenoid rated at 900V 500A a few weeks ago. I intended to use a simple toggle switch to power up the inverter including an automatic enable of the high current path after the pre-charge is finished. The solenoid has now been sitting on my shelf patiently staring at me for about 3 weeks now, time for some fun. The first obstacle to overcome was that I ordered a 24V coil solenoid as they were ~$100+ cheaper than the 48V or 12V units - why I have no idea - maybe the 24V units have high sales volumes. I always intended to find a simple drive mechanism to operate the 24V unit from 48V. The next feature I liked is that the hold in current is much lower so my circuit had to hit the unit with 24V and then transfer to hold at ~ 8V. I developed a circuit that can produce two separate PWM periods the first has a duty cycle of ~ 50% the second has a duty cycle of ~ 17% the frequency is ~ 2kHz. The operation is, when enabled, first apply the 50% duty cycle, the solenoid pulls in with ~ 22.5V on a 45V supply, after a ~0.5 second delay the PWM lowers to 17% the coil voltage is ~ 8V, a FET buffers the pwm to drive the solenoid which has a catch diode across it which acts as a flywheel diode to keep the current flowing during the off time. I used a whole 2 schmitt trigger inverters and a FET to do all of this. The supply current (from 45V) is ~ 220mA during the 22.5V phase this then drops to ~ 30mA whilst maintaining the holding 8V. My start up & stop sequence will be: 1) Turn small toggle switch on - this pre-charges the caps via a 100 Ohm resistor 2) A comparator waits until the Vcaps is ~95% charged then enables the solenoid high current path. 3) Nano now begins to soft start Inverter to supply backup mains power. 4) Use mains power for stuff 5) Turn toggle switch off, signals low voltage to low batt detector circuit 6) Nano is commanded to stop - and enters soft stop period & off timer started 7) After selectable 2,4,8 second timer delay (for soft stop to finish) solenoid is commanded off, goodbye world. 8) Lets do it all again...... Circuit of solenoid drive scheme if interested: solenoid drive.pdf D1 stops oscillator & holds Fet off initially when its anode is high. When D1 Anode is set low oscillator starts to pwm FET & solenoid for 24V. For a delay set by [R1,C1 & U1B], D3 shunts second feedback path to ground. After the delay, U1B goes high, reverse biases D3 & second timing path via D2 & R3 is enabled to reduce pwm high period & sets solenoid to 8V. It works a treat. The Nand latch is to set the solenoid on or set it off with a low trigger to either the set or reset inputs. When triggered to run it stays running even if trigger input is gone, until triggered to stop which is a whoa camel stop - power is removed so needs a restart. Edited 2020-08-25 03:07 by wiseguy If at first you dont succeed, I suggest you avoid sky diving.... Cheers Mike |
||||
| Solar Mike Guru Joined: 08/02/2015 Location: New ZealandPosts: 1138 |
One of these will do the same as your circuit DRV103 I use them quite often to lower the power to relay coils, especially when confined in small cases's to prevent over heating. Those Kilovac relays normally have a pwm driver associated either internally or externally, not sure how they will work being also driven by a PWM source. Mike |
||||
| wiseguy Guru Joined: 21/06/2018 Location: AustraliaPosts: 1156 |
Thanks for the information about the DRV103 series, it is almost OK but unfortunately the absolute max operating voltage is only 32V, I need about 55V. The other difference is that it applies the full applied voltage to the relay - in my case up to 55V to a 24V solenoid during the pull in period. The max voltage for the 24V solenoid is 30VDC so not a good plan. My circuit PWMs the 48 to 24 for the initial solenoid pull in and then PWMs to 8V to hold it in. Maybe there is another version of the DRV103 around but it cant do the same job as my circuit. The kilovacs that I could find, 41 to be precise there were only 8 versions which have electronic pwm circuits so < 20% is not quite "normally" and of those only 2 were suitable for 48V use but then the other issue, they were $230 & $246 $AUD. Mine with a standard coil was $111.40. A FET still had to be used to switch the relay so why not modulate it with ~$2.00 worth of bits. A failure of their pwm circuitry may render the expensive device as junk. Do you really think I would try to drive a PWM circuit with a PWM circuit - come on.... Edited 2020-08-25 18:55 by wiseguy If at first you dont succeed, I suggest you avoid sky diving.... Cheers Mike |
||||
| Solar Mike Guru Joined: 08/02/2015 Location: New ZealandPosts: 1138 |
OK my normal is not the same as yours, all 12 of the various ones I have here have the factory PWM economizer fitted, either internally or externally. And yes two have stopped working, the economizer's have failed; and I cannot locate a circuit anywhere for them.  ; your circuit is pretty good, at least its repairable.  |
||||
| wiseguy Guru Joined: 21/06/2018 Location: AustraliaPosts: 1156 |
Recently I redesigned my earlier isolated DC/DC circuit for lower cost and simpler design. I decided to post this if it may be of use to anyone else who wants a simple isolated supply/supplies. Dont be put off my use of a pcb winding - I am basically lazy and find it tedious, boring and a time waster to wind these little transformers by hand. By using 2 PCbs we have a primary drive winding & 3 isolated secondaries, 1 for the 2 x common low side legs and 1 each for the LHS high and RHS high legs. By adding another winding pcb we can have another 2 isolated 12V supplies if needed (5 x isolated supplies in total). The purpose of this circuit is to supply adequate isolated DC gate drive bias for the opto-coupler and FET buffer/drive circuitry in mosfet H bridge power stages, and should be quite adequate for 20kHz switching with 4-8 Fets per leg. A few pages back I designed an RM6 PCB winding to simplify construction. It did not provide a lot of isolation and had to run at >300KHz due to the 7 turn primary winding. The UCC3808 was also a bit expensive, harder to obtain and probably overkill for what was needed. The new version has a 4 layer board for the windings with more creepage/clearance and a 12 or 13 turn primary (jumper selectable). By placing a 1nF cap in parallel with the (12T) primary it resonates at ~ 160kHz and removes switching glitches and noise when tuned properly (adjusting R1). The parts list is a Cmos 555 ie TLC555 a couple of resistors caps and and a C6003 60V 3A mosfet in a TO251 through hole version (~12c ea from LCSC). The circuit is about as simple as it gets with just enough to make it all work. The gate drive resistor is 100R which slows down the fast switching edges eliminating noise glitches and ringing almost entirely. The circuit uses ~ 8mA from 12V unloaded, and when delivering ~1W runs at ~ 80% efficiency. It is not a flyback converter, essentially a forward converter, so the input to output voltage is directly related to primary and secondary turns. With 12.1V & 12Turns for the primary supply and 13Turn secondaries with schottky rectifiers, output voltage is ~ 11.9V @ 30mA. When lightly loaded (~0.5mA) output voltage only rises ~ 0.5V (to ~ 12.5V). Note only one of 3 output winding/rectifier/filter Caps is shown for simplicity. Apologies for crappy image but print to pdf is playing up.  Yellow trace is Q1 drain Light Blue is Q1 gate drive Tuned for best performance  Osc Freq too low  Osc Freq too high  Now its documented I can clean this phase off my bench for some more space I have about 5 concurrent things I'm working on..... Edited 2020-09-13 20:19 by wiseguy If at first you dont succeed, I suggest you avoid sky diving.... Cheers Mike |
||||
poida Guru Joined: 02/02/2017 Location: AustraliaPosts: 1418 |
this 12V isolated supply of yours is looking good. I like the use of the 555, you could even get some form of closed loop control with it, maybe a cheap opto and a TL431. But that would be too much for something that just needs to make a few mA at 12V sort-of-ish The lazy bastard in me thinks "just use a RB1212S or RE1212S" and forget about it. But this project is future-proof. wronger than a phone book full of wrong phone numbers |
||||
| BenandAmber Guru Joined: 16/02/2019 Location: United StatesPosts: 961 |
hello poida i have a nano control board if you ever need me to send it some where it would be nice to help in some way have a blessed day :-) be warned i am good parrot but Dumber than a box of rocks |
||||
| Warpspeed Guru Joined: 09/08/2007 Location: AustraliaPosts: 4406 |
The 555 is a very nice touch. If planning some closed loop voltage regulation, it will need to have an averaging choke and flywheel diode in the secondary, and a higher voltage secondary winding. The way it is now, with just a single diode peak rectifier, the output filter capacitor will always charge up to the peak voltage, so duty cycle control (in theory) should not be able to change the dc output voltage, as the peak value does not change. It would very likely work in practice, in a fasion, but not very well. Best to do it properly if some deliberate duty cycle control is to be used. Anyhow, its a great little circuit with very few parts. Cheers, Tony. |
||||
| wiseguy Guru Joined: 21/06/2018 Location: AustraliaPosts: 1156 |
Ok time to wipe the dust off this old post and give a belated progress report. Since I last posted I finished the inverter complete with Kilovac circuit breaker & a 63A chinese DC breaker. There is a 20A AC breaker on the output as well as an over current shutdown, current transformer wired back to the nano. Recently we received an sms notice that there was maintenance in the area and there would be no power between 8.30am and 3.00pm in a few days. We had friends coming for lunch that day - no problem I had made the 3phase plug with L1 L2 & L3 shorted so I was all prepared & waited for the lights to go out. When they did I disconnected the street (supply?) and turned off the 3Phase solar grid feed inverter, plugged in the 3 phase shorting plug and started the inverter. We heated stuff in the microwave and oven, used the toaster, 2.4kw kettle, lights, the gas space heater fan and 3 fridges and stereo. Finally at ~ 3.00 we were notified that the power was back on so I reverted back to normal. The inverter was only a few degrees above ambient when I checked it after lunch, after all its hard work to make lunch. Guests were suitably impressed - and so was I - I had decided if it goes bang I will need to improve it. I had bypassed the over current sense though as the first time I tried to start it, it shut down when I closed the AC breaker. A couple of weeks later I had a visitor and decided to show off the inverter. I plugged in what I had described as one of the worst loads (my compressor) and said watch this, I turned on the inverter switch, heard it finish soft start with its usual grunt followed by a loud "Bang" and flash and the compressor never uttered a sound. "That was neat can you do it again" was my friends comment. Ok what happened? well the end result is one of the 4 FETs on the high side was dead short (which saved the other 3) and the 4 FETs on the low side were all dead & shorted. So obviously the top side FET died saving the other 3 and the first time the lower 4 FETs tried to turn on they uttered an expletive and promptly died. I did 2 things wrong, I had previously disabled the AC over current latched shutdown with out re-enabling it and the compressor and AC were all turned on from go. I surmise that during soft start the compressor turned a little to a compression stroke and with no flywheel effect to carry it over TDC it just stalled drawing more and more current as the AC ramped up until something gave. Its DC impedance is 3.5 Ohms so when stalled it must draw an impressive amount of current. There is also something that has always happened when the inverter is first started that I now need to look into further. At the end of ramp up there is a loud audible hum that becomes more like a grunt for about 0.5 - 1 second before settling to a soft hum. If I power down the inverter that then invokes soft stop and re enable it before the mains is totally gone it starts up again but always with the gentle hum and never growls. It was during the "grunt" that it went bang I have never liked the "grunt" but being software challenged accepted the situation. I am not saying it is definitely software but I am pretty confident it is - regardless when I find the culprit I will post the results - it might be my design....nah not possible lol. It sounds like VFB is ignored until the instant soft start is finished or ignored for maybe for a cycle or two as it transfers from soft start to controlled causing an overshoot but the CRO hasn't revealed it so this cant be the case - needs more investigation. Can anyone suggest how I can link to an audio recording of the "grunt" at start up and a comparison hum during a run interrupt & immediate restart. I would be interested if anyone else has a similar audible effect at the end of soft start. I have tried to capture various nodes and the output amplitude on the CRO but havent cracked it yet. - It does not appear to be an AC output overshoot due to a Vfeedback delay so I am going to fiddle with some parameters in the code to see what influences it. By the way I fixed the inverter and when I test powered it from a bench supply limited to 1A it started the micro - hang on the switch is still off, did I make a mistake in the wiring? After a few measurements it was obvious that my 900V 500A solenoid was shorted - oh crap. With nothing to lose I got a small hammer and tapped the panel it was mounted on and heard a "clunk" internally and normal behaviour has now returned. The only parts required was the 4 x HY4008 FETs. I am guessing that during the blow up the breaker might have opened and then closed again welding the contacts - I will look further into how to protect it for next time - hows that for confidence - I am feeling a bit more humble now. If at first you dont succeed, I suggest you avoid sky diving.... Cheers Mike |
||||
Revlac Guru Joined: 31/12/2016 Location: AustraliaPosts: 1024 |
Quite often things go this way, all excited to show something to someone, and disaster, had it happen a few times, with other things.  Repaired an air compressor for a friend and start up current was checked with a meter, started it with a generator to test it. 42.5A at 240vac and 7A running before pressure builds up, so yes, they are heavy to start, but not really over the top. I will do some more readings one day when I have things sorted a bit better. As for start up sound, using the eg8010 it starts quiet and just increases to a gentle hum, if close enough to Actually hear it. You could attach a sound file in a zip file, if that is a safe way to do it. So far I haven't blown up a single HY4008 fet yet in any build.....plenty of time for that to happen.  I haven't tried the nano driven inverter yet, just had some PCB's made. Cheers Aaron Off The Grid |
||||
| wiseguy Guru Joined: 21/06/2018 Location: AustraliaPosts: 1156 |
Me too - but it was 5 total unfortunately, not a single one haha (Note my last post said I replaced 4 - it was actually 5) Thanks Aaron I will try the zip file for the audio file- good idea. Inverter Hum.zip There are 2 files, VLC works fine for playback - the audio level is not real high sorry. On my list of things to try one day is to use the EG8010 to create the PWM on 2 outputs and use my gating that I use with the nano to provide all the high and low drives from the 2 outputs just as an exercise of curiosity. Interested to hear if anyone else playing with nanos has a similar sound on start up. If at first you dont succeed, I suggest you avoid sky diving.... Cheers Mike |
||||
| Revlac Guru Joined: 31/12/2016 Location: AustraliaPosts: 1024 |
Put it through Audacity and amplified that section, Can see it now.  Yes it is a little different than I expected, will have to compare with others. I remember a friends selectronic did sound a little like that, and had some reverberation, louder and softer every 10 or 15 seconds. Cheers Aaron Off The Grid |
||||
| wiseguy Guru Joined: 21/06/2018 Location: AustraliaPosts: 1156 |
Poida, sorry to hear about your back, I do hope its on the mend and a lot better now, I always follow the doctors advice to drink plenty of liquids when you are ill. When Port D8 is set high to start the inverter, soon after, Port D5 (gate drive enable) is set high, the inverter code then begins a soft start and I get the "grunt" every time on a restart, but I do need to wait until D5 is set low at the end of soft stop before setting D8 high to restart, to get the grunt. If I turn off the inverter (set pin 8 low) but then set it high again before soft stop finishes and D5 is set low, I see the decaying sinewave instantly truncated to zero at the instant D8 is set high, I then see a complete soft start cycle begin again but this time it always finishes as a gentle hum, no grunt. What bit of code or initialising is bypassed or runs differently between waiting for D5 to set low after a complete soft stop cycle and then a full restart on D8 set high, compared to momentarily toggling D8 low, then interrupting the soft stop code by toggling D8 back high again to begin a new soft start, but before D5 is set low ? I have looked at your code and explanations to the best of my ability but it just eventually gives me a headache, software is definitely not my forte! If at first you dont succeed, I suggest you avoid sky diving.... Cheers Mike |
||||
| Warpspeed Guru Joined: 09/08/2007 Location: AustraliaPosts: 4406 |
Ah, that sounds like a classic flux doubling problem in the transformer. What can happen if you suddenly switch the power off to a transformer, the steel in the core can remain magnetized. If you then suddenly switch the power back on, if the ac cycle happens to be in the same direction as the stored residual flux, the ac flux swing adds to the residual magnetism, and it can easily drive the core into horrific magnetic saturation. In theory the flux can actually double ! Its never quite that bad, but it can still be a real problem. There will sometimes be a massive inrush current spike at switch on that can blow fuses or pop circuit breakers (or mosfets). How bad the problem is, depends on the original design of the transformer. The residual magnetism will be far less with a laminated E and I transformer because of the small unavoidable air gaps between the edges of the laminations. Its much worse with a wound toroidal transformer. It also gets worse if the toroid is designed to run at a fairly high flux density, which many commercial transformers are. What can be done ? First line of defense is to design your monster toroid to run at a lower than usual flux density (one Tesla recommended). That will not only greatly reduce inverter idling current, but at startup the flux could actually safely double to two Teslas without causing much of a surge. The surge will usually be far less, and the surge problem no longer exists. If you are using a toroid that runs at a higher flux density, you need to be very careful. There is a way around the problem, but its a bit of an inconvenience. The trick is firstly, never just totally switch off the power to the inverter while under heavy load. To minimise the possibility of stored residual flux, remove all load from the inverter output first, then allow the dc voltage to the switching bridge to fall gradually while the mosfet gates are still fully driven. What happens is, the voltage on the big electrolytics slowly dies, and the ac amplitude at the transformer slowly reduces to zero amplitude. That very effectively de-magnetises the core. The ac voltage does not suddenly dissappear mid cycle, leaving residual flux to be stored in the transformer core. This may all happen naturally during inverter switch off anyway, and you may be lucky 99 times out of 100. The other thing is to soft start the inverter with no load on the output, and then switch on the load only when its up to full voltage. Most of us do this naturally as well. But just occasionally some unusual circumstance either during inverter shut down, or start up, may provoke a potentially destructive surge if your toroid runs at a high flux density. If you sometimes get a short groan, or a chirp, or buzz at inverter start up, BEWARE !!! Cheers, Tony. |
||||
| wiseguy Guru Joined: 21/06/2018 Location: AustraliaPosts: 1156 |
Hi Tony, the soft start that the nano uses begins from zero and rises to normal amplitude over a few seconds. It seems to have no problem defluxing a magnetised core and keeping it away from saturation, it slowly pulls it back from saturation as the flux swings slowly increase. I am well aware of the remanence problem and massive inrush currents caused by trying to flux it up further if you happen to reconnect power at the wrong part of the sinewave. When the soft stop is interrupted and soft start begins again - this scenario is where the nano starts my inverter just fine. Where it doesnt start fine is from a cold start when the soft stop/deflux has continued down to zero and the nano is halted before doing a restart - then I get the grunt at the end of soft start. The nano essentially does exactly as you describe but without reducing the DC voltage to zero it alters the sine pwm duty cycle bit by bit which if you watch on the CRO gives exactly the same effect as reducing the DC whilst a static pwm is maintained. Apart from trying to start my compressor when I killed it, my inverter is unloaded for all the testing and recorded grunts etc that I am now trying to resolve. Edited 2021-10-18 11:01 by wiseguy If at first you dont succeed, I suggest you avoid sky diving.... Cheers Mike |
||||
| Warpspeed Guru Joined: 09/08/2007 Location: AustraliaPosts: 4406 |
If it both soft starts and soft stops, that is ideal. It should not grunt ever at startup, if its working as described. Is it possible that gate drive is interrupted somehow, defeating the soft start/stop ? Cheers, Tony. |
||||
| Page 8 of 16 |
|||||
 Print this page |
