 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 : Turning a Toyota Prius into a "portable" generator
  Page 2 of 14   |
|||||
| Author | Message | ||||
| noneyabussiness Guru  Joined: 31/07/2017 Location: AustraliaPosts: 512 |
Add a dozen turns or so to the 200v winding...or am i missing something?? I think it works !! |
||||
| Warpspeed Guru Joined: 09/08/2007 Location: AustraliaPosts: 4406 |
That is certainly an option. Even if the primary is the inner most layer, that does not prevent any extra primary turns from being added right over the top of everything else. Cheers, Tony. |
||||
Murphy's friend Guru Joined: 04/10/2019 Location: AustraliaPosts: 648 |
No need to hit it with anything to loosen the core. Let gravity work for you. Place housing upside down onto carpet or foam piece. Lift the end where the core is a bit and let it drop. Repeat if it does not loosen by lifting a little higher. But don't forget to remove the 10mm nut & washer in the center first  |
||||
Haxby Guru Joined: 07/07/2008 Location: AustraliaPosts: 423 |
Next question is about capacitors; Ballpark wise, how many uF should I have around the IGBT's? If it matters, I'm using 3x IGBT modules that each have 3x half wave bridges inside. They are designed for driving 3 phase motors. Each module is a little lower in current handling as the previous one, since the transformers get smaller and smaller. I have a 150A one, a 50A one and a 30A one. |
||||
| Warpspeed Guru Joined: 09/08/2007 Location: AustraliaPosts: 4406 |
Stored energy is proportional to voltage squared, so a very rough ball park figure might be: 48v 5Kw inverter 160,000uF 100v 5Kw inverter 40,000uF (I used three 12,000uF which worked out fine) 200v 5Kw inverter 10,000uF 200v 2.5Kw inverter 5,000uF Something like three 1,500uF or 2,200 uF screw terminal low ESR type, one bolted as close as possible across each module would be my choice. There will be a lot of circulating energy flowing back and forth between the inverters, and a nice clean dc power rail will produce a nice clean ac waveform. Cheers, Tony. |
||||
| Haxby Guru Joined: 07/07/2008 Location: AustraliaPosts: 423 |
Got it. Another question: Previously you said not to use a single common power source for the gate driver ic's Since I am using modules that contain 6 IGBTs each, any failed transistors will require changing the whole module. Given that, is it still that important to drive them with different independent supplies? Maybe a compromise of one set of power supplies per IGBT module? |
||||
| Warpspeed Guru Joined: 09/08/2007 Location: AustraliaPosts: 4406 |
That is a problem with those big modules. If one small internal part fails, the whole thing has to go into the wheelie bin. And a replacement module bought new is very likely to be horrifically expensive. No problem though if you have a few replacement modules safely stashed away. Its a case of playing Devils advocate, thinking right through the possible effects of a blow up or random failure. If X goes up in smoke, will Y or Z be damaged too through being joined at the hip. Noise immunity is quite a different question. Switching 25 amps at 200v (5Kw) is a very different thing to switching 100 amps at 48v, also 5Kw. You are going to be much better off, and can get away with rather more than the 100 amps+ guys. Cheers, Tony. |
||||
| Haxby Guru Joined: 07/07/2008 Location: AustraliaPosts: 423 |
I'm a bit stuck on the 15v power supplies. The little 15V Switchmode supplies have vanished... I could get these for about $6: 15v psu Or I was thinking about getting these and running them from a common 5v supply: isolated psu Is 1W enough or should I use the 3W modules? Any other suggestions? |
||||
| Warpspeed Guru Joined: 09/08/2007 Location: AustraliaPosts: 4406 |
For some peculiar reason all the 15v supplies became totally extinct about a year ago. I suppose nobody is making 15v dc plug packs anymore, all very strange. Since then those vertical 15v supplies in your first link have become available, but only from one source when I originally looked at them, which made me nervous. The smaller supplies in the second link need to be selected with very great care, including reading all the fine print in the specifications. Some are not voltage regulated and some have a minimum specified load which can be fairly high. Others are regulated but blow up if the output is ever shorted even for an instant, and the prices are more than two dollars each which makes them expensive compared to the Chinese mass produced plug pack boards. The efficiency and no load power is also not as good as the plug pack boards. If you look at the whole picture with all the fine details and various limitations, including cost and robustness, those Chinese plug pack boards always seem to come out way in front of anything else. Anyhow, the original 12v Chinese plug pack supplies are still readily available, and that gives two options with those. 1/ Use the 12v supply as is, along with a gate driver chip that has an appropriate undervoltage lockout operating range such as a FOD3182 (10v to 30v). https://au.mouser.com/new/on-semiconductor/onsemi-fod3182-gate-drive-optocouplers/?gclid=EAIaIQobChMIx-zlzIX-6gIVSw4rCh2odw6mEAAYBCAAEgKN_PD_BwE The original gate drives I used were specified to work between 15v to 30v. The under voltage lockout range has a worst case specification of needing greater than 11.5v to turn on the chip. That is getting a bit close to a 12v supply that might be a bit low. 2/ Modify the 12v supplies to provide +15v. This is very easy to do by changing a resistor in the voltage feedback. These are flyback supplies and provided the load current is low, gaining the extra voltage is no problem. You will find a voltage divider across the dc output that has a 2K4 resistor. Change that to 1K8 and that will produce a 15v output. When I built my own inverter, I needed sixteen of those 15v supplies and found that I only had fourteen. So I modified two 12v supplies and used those, and they are still in there working perfectly. If I was doing this again I would probably just run 12v to all the gate drive circuits and use the lower voltage FOD3182 which are compatible in every other way. See what the application notes for your big IGBT modules say about supply voltage limits to the gate driver section. If it needs 15v then its not difficult to provide. Altronics have the 1K8 surface mount resistors. Edited 2020-08-03 13:57 by Warpspeed Cheers, Tony. |
||||
| Solar Mike Guru Joined: 08/02/2015 Location: New ZealandPosts: 1138 |
These 3 Watt 15v ones are available on EBay, note pcb size is 4 x 1.5 cm. Mike |
||||
| Haxby Guru Joined: 07/07/2008 Location: AustraliaPosts: 423 |
I ended up buying 20 of the Hi-Link PM15 power modules from Ali express. $2 each and I like that they are in a plastic case. I'm considering scrapping the IGBT modules I have and going with half bridge modules like this part: CM600DY-12NF It's 600A but only 600V. Warpspeed, I saw your IGBTs are rated 1200v. Is that just because you have them on hand or is the extra huge headroom desirable? Also at what point will the optoisolated drivers fail to quickly turn on a 600A IGBT? The data sheet says they can do up to 100A but I presume that that is mostly frequency dependent, and since we are switching at slower speeds, we can get away with more? |
||||
| Warpspeed Guru Joined: 09/08/2007 Location: AustraliaPosts: 4406 |
Yup, already had a bunch of those 1200v IGBTs, had them for many years in fact. Only too glad to finally find a use for them. Six hundred amp IGBT half bridges will be perfect. IGBTs are very robust anyway, much more so than equivalent rated mosfets. An ordinary curve C circuit breaker on the ac output will trip long before the IGBTs are anywhere near reaching their current ratings. My IGBTS are rated to carry two hundred amps rms continuous, but are also rated to carry 1,000 amps of fault current for a full 10mS half mains cycle. I had one fail open circuit though, and that appears to have been just an internal random component failure, definitely not failure through overload. The opto isolators will turn on any sized IGBT, they only have to charge up the gate capacitance, and switching speed for us can be ultra slow. The difference in speed between 50Hz and 20Khz is greater than the difference between walking speed, and the speed of a fighter jet going supersonic. There are very large factors involved here, and definitely in our favor. Cheers, Tony. |
||||
| Warpspeed Guru Joined: 09/08/2007 Location: AustraliaPosts: 4406 |
Just looked up the data sheet for those CM600 IGBTs: https://pdf1.alldatasheet.com/datasheet-pdf/view/305120/MITSUBISHI/CM600DY-12NF.html They are much faster and of a much more recent technology than I am using. Date on the data sheet is 2009 and they are about four times faster than my ancient old devices, not that it really matters. Excellent choice if you can get them at a reasonable price. Chinese clones of my devices are around seventy dollars each on e-bay. Also found these, a 200 amp version, which is cheaper and should surely be enough. https://www.ebay.com.au/itm/1PCS-MITSUBISHI-CM200DY-12H-Module-New-100-Best-Service-Quality-Guarantee/282299830801?_trkparms=aid%3D1110006%26algo%3DHOMESPLICE.SIM%26ao%3D1%26asc%3D228418%26meid%3Da6692113c58c48399dcf133359d34371%26pid%3D100005%26rk%3D6%26rkt%3D12%26mehot%3Dsb%26sd%3D320884565008%26itm%3D282299830801%26pmt%3D1%26noa%3D0%26pg%3D2047675%26algv%3DSimplAMLv5PairwiseWebWithBBEV2bDemotion&_trksid=p2047675.c100005.m1851 Edited 2020-08-11 10:25 by Warpspeed Cheers, Tony. |
||||
| Haxby Guru Joined: 07/07/2008 Location: AustraliaPosts: 423 |
My weekly update: I've settled for some IXGN IGBTs. Part number IXGN320N60A3 for the big transformer, then I'll use 2 X 3phase Mitsubishi igbt modules for the other 3 transformers. I'm waiting impatiently for the dc-dc Hi-Link PM15 modules to come in. As I wait, I'm thinking about anything else I need. Since the warpverter doesn't have soft start or shutdown, is it worth having a contactor on the 240v output that disconnects sensitive appliances before the inverter shuts down? I imagine the giant transformers will have significant kickback if they lose power abruptly? |
||||
| Haxby Guru Joined: 07/07/2008 Location: AustraliaPosts: 423 |
Also any pointers for a pre-charge relay? |
||||
wiseguy Guru Joined: 21/06/2018 Location: AustraliaPosts: 1156 |
Are you asking about a relay to pre-charge the caps or the one that shorts out the pre-charge resistors ? If the latter, I recently bought a Digikey A101172-ND (LEV200A5NAA) which is good for 500A & 900VDC. It was $111.43 so not cheap but should outlive me. It has a 24V coil & I am perfecting a simple discrete circuit that will pull in the coil at ~ 24V and then hold it in at ~ 6V circuit when wired to a 48V supply. The 48V coil types are nearly double the price. I figured if we are driving it with a FET anyway then PWM the FET with 2 suitable fixed PWM's add a flywheel diode and we're in business. From memory around 300 - 500Hz gave good results in a similar app. Edited 2020-08-22 11:48 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 |
I use a two pole 63 amp dc circuit breaker between the battery and the inverter. In parallel with the circuit breaker are a pair of relay contacts with a 30 ohm 200 watt soft start resistor on the positive side only. There is a push button which operates the relay, and that brings up dc power to the inverter and its 36,000uF capacitor bank. Within about two to three seconds I can tell that the 230v ac output is up, and I can then safely close the dc circuit breaker. Its then a case of closing the 20 amp ac circuit breaker on the inverter output. Switching off is the reverse. Ac breaker off first, then dc breaker off. The dc circuit breaker has a shunt trip coil which energises if the battery voltage goes beyond either a high voltage limit or a low voltage limit. There is no electronic overcurrent protection, apart from the two circuit breakers. The 20 amp curve C circuit breaker on the output always trips first on any overload. The only other feature is a 230v powered digital LED heatsink temperature display, and controller, to turn the big 230v powered air blower on and off. In almost three years of operation it has never come on. So basically push the soft start button, when the LED readout of heatsink temperature lights up, close the dc then ac circuit breakers. All takes less than about three seconds. Never had any issues over almost three years of operation. Cheers, Tony. |
||||
| Haxby Guru Joined: 07/07/2008 Location: AustraliaPosts: 423 |
If the inverter loses power abruptly on the input side, is there a spike on the output side due to the inductance of the transformers? Should I add some MOV's or a contactor that is mechanically linked with the DC circuit breaker? |
||||
| Warpspeed Guru Joined: 09/08/2007 Location: AustraliaPosts: 4406 |
No, there is no stored energy. As with any transformer, the secondary voltages mimics the primary voltage. If the whole beast suddenly gets its power source switched off, the electrolytics lose voltage, either very quickly, or very slowly, depending on the inverter load. The amplitude of the ac output voltage correspondingly falls. The various switching power supplies die at perhaps +35v ?? At which point everything shuts down. You can watch this happen on an oscilloscope. It slowly dies in the bum, and stops. No transients, no drama. Its all been thought through very carefully. And testing and experience has borne this out. Cheers, Tony. |
||||
| Warpspeed Guru Joined: 09/08/2007 Location: AustraliaPosts: 4406 |
The only REALLY lethal failure mode that can happen, is if the crystal oscillator fails to start up, or stops. That can leave diagonal pairs of IGBTs permanently switched on. Both Klaus and I have had big blow ups from this particular problem. That is one reason I decided to use a costly and inconvenient crystal oscillator module. These start up at only 2.0v and work up to +6.0v. These oscillator modules MUST be soldered in, as the skinny wire legs are just not compatible with normal IC sockets. An oscillator module soldered in has proven to be very reliable and solved these particular problems. Another advantage of these oscillator modules is the excellent frequency stability. The inverter can run for months and any synchronous mains clocks or timers will stay wonderfully accurate. A typical module might be specified as 50ppm, a really good one 10ppm. A million seconds = 16,666 minutes = 277 hours = 11.57 days Fifty seconds gain or loss in 11 days is not too bad. Its about 27 minutes per year absolute worst case. A decent oscillator module will stay well within that. Edited 2020-08-22 17:29 by Warpspeed Cheers, Tony. |
||||
| Page 2 of 14 |
|||||
 Print this page |
