 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 : Choke rules of thumb?
Page 1 of 3   |
|||||
| Author | Message | ||||
Haxby Guru  Joined: 07/07/2008 Location: AustraliaPosts: 418 |
I'll be playing with a few EG8010 boards that have been gathering dust over the next week or so. Just tinkering really. So the 48v portable pack that i built in the other thread is working great. It's using a smaller choke from an aerosharp inverter. I unwound all of the turns on one side, and half of the turns on the other side. I also increased the gap by a couple of mm. The label on it said 2.67 mH so to get down to 4.7uH I knew I had to remove a lot of windings. Without a choke tester, I just winged it. It's about half the size of the 500va transformer I'm using, so it's a bit too big, but hopefully that will come in handy when starting up stubborn loads. In any case, the sine wave looks great and the standby power is fine. Now I'm wondering how the commonly recommended 4.7uH figure was calculated. Or is it a rule of thumb that just seems to work? Is it mathematically tied in with the 23khz switching of the EG8010? Or is it due to some characteristics of the transformer size? Let's say I wanted to build a 200vdc input inverter with the same topology. Would 4.7uH still be the aim? |
||||
| Godoh Guru Joined: 26/09/2020 Location: AustraliaPosts: 378 |
Hello Haxby. I can't answer the question on how many microhenries the chokes need to be. The inverters I have are all using either the chokes from the wrecked powerjacks that I used the transformers from. ( they had small torroid chokes with 2 or 3 turns around them about75mm from memory. Or they have 65mm E core chokes that were first suggested by Oztules when he was still about. My memory is that they have 3 or 4 turns on them. My three 8010 based inverters run fine with them and have fairly low idle currents. I don't have an inductance meter so just went with Oz's suggestions. Just wondering whether you are going to modify the 8010 board to delete the 393 chip and the other mods that Oz suggested or whether you think the later boards are fine as they are. Good luck with the inverter Pete |
||||
| Haxby Guru Joined: 07/07/2008 Location: AustraliaPosts: 418 |
I will be deleting the 393 chip. I'm not a fan of the shunt feedback circuitry. I experienced a lot of EMI problems with some previous designs, so I'm super cautious of things spontaneously oscillating. |
||||
poida Guru Joined: 02/02/2017 Location: AustraliaPosts: 1387 |
(this is not for Haxby, it will be telling him how to suck eggs. It's more for any new or random readers) the 23kHz switching freq is not important in respect to the LC filter design. I have experimented with all sorts of PWM switch freqs using the same LC filter and not a lot changes. (apart from the PWM freq and the wobbles in the AC voltage output, all minor "issues" but excellent for thinking about) The PWM input into the transformer is a very strong squarewave. Lotsa current and lotsa harmonics (3x,5x, 7x etc of the PWM freq) We need to smooth this crap out. And so we use the LC filter on the primary winding. I have found for 20kHz or so PWM and moderate power levels, such as up to 3kW something like 40 or 50uH is good for the choke. But what is vital, is that the choke does not saturate at peak power levels so that means we need to test them to see the current levels that cause saturation. (Saturation of the choke will have it no longer being a 50uH choke, it now is far less than that.) The core of the choke needs to work at these high freqs too, including the harmonics. I like ferrite cores but we have had a lot of success with the Aerosharp fine layered Iron laminate cores too. You have to test them to prove they are OK. The L in the LC filter is the choke plus stray inductance of the toroid which is usually small, of the order of 5uH or so. The C is the capacitor across the inverter board's outputs. This can be placed across the primary or it can be put across the secondary winding where you get the turns ratio magnification of the capacitance. We usually put it across the secondary. I have found I blow inverters with a 10uF on the secondary. This had a self resonance of about 17Hz. This is far too low a self res freq. I use something like 2.2uF now and the self res is about 40Hz. Only after you try some combinations will you see what the self res freq of the LC + toroid combination is for you. It might be that only 1uF is needed to get the freq above 50Hz. Some people here insist that you aim for 75Hz. I have not done this and no inverters have died with self res freq of 40-50 Hz in my builds. The critical things are 1 - no saturation at peak loads 2 - self res freq about 40Hz or more And be mindful of the voltages that are driven BACK INTO THE gate drive ICs Their output drive stages are not very robust and are easily destroyed by out of spec voltages on their output pins. IR2110 output pin spec say no less than -0.4V to supply ground (or VS). wronger than a phone book full of wrong phone numbers |
||||
| Haxby Guru Joined: 07/07/2008 Location: AustraliaPosts: 418 |
So you're saying suck eggs and build an inductance tester?  Fair enough. Or maybe I'll just try a few inductors and monitor the current in the primary using a LEM sensor on a scope? That should give me an idea as to what's going on? The main thing I'm getting from this thread so far is that 50uH is the general aim no matter what the input voltage nor frequency. |
||||
Murphy's friend Guru Joined: 04/10/2019 Location: AustraliaPosts: 580 |
If you are messing about with chokes then an inductance saturation tester is very handy. I have a spare PCB for that here (my PCB layout, Tony's idea) you can have for the price of postage. You will need a DC power supply, maybe 5-6A up to 30V adjustable, it must have really big output capacitors with a high ripple current rating (I use 66,000uF/40V @ 24A ripple). You also need a thru hole hall sensor and a CRO. Those devices were discussed here at length some time ago. |
||||
| Haxby Guru Joined: 07/07/2008 Location: AustraliaPosts: 418 |
A PCB would tip me over the edge to make a tester. Thankyou for that. I'll send you a private message now. |
||||
| Haxby Guru Joined: 07/07/2008 Location: AustraliaPosts: 418 |
Klaus, I just tried private messaging you but I came up against a problem with the thebackshed website. I get in an endless loop selecting your username. Can you send me a message and I'll reply? |
||||
| noneyabussiness Guru Joined: 31/07/2017 Location: AustraliaPosts: 506 |
AU $26.48 | LCR-TC1 1.77inch Colorful Display Multifunctional TFT Backlight Transistor Tester https://a.aliexpress.com/_mKZe4kq You could use one of these to test inductance... they obviously wont test saturation, but are incredibly handy... I think it works !! |
||||
| poida Guru Joined: 02/02/2017 Location: AustraliaPosts: 1387 |
seeing the current through the inductor when in circuit is the best way to prove things. You can prove that it's not saturating and so has enough Henries. Also you can try others and see if they are OK or not. I did this in the mppt development process and it was well worth the effort. this is some testing I did The LEM sensors can be found on the Aerosharp control boards. Usually there are 3 of them. Got any dead Aerosharps lying around? It does not have to be LEM, Allegro sensors are also suitable but are a bit more noisy if I recall. wronger than a phone book full of wrong phone numbers |
||||
| Haxby Guru Joined: 07/07/2008 Location: AustraliaPosts: 418 |
Yes I have a lot of the LEM current sensors. Sounds like I have a good week of testing to come. |
||||
| Murphy's friend Guru Joined: 04/10/2019 Location: AustraliaPosts: 580 |
PM sent. |
||||
| Haxby Guru Joined: 07/07/2008 Location: AustraliaPosts: 418 |
One thing I haven't seen any discussion on is the emphasis on the placement of the inductor on the SPWM side of the inverter. What I mean is that one side of the transformer is driven by a 50hz square wave, while the other side is doing the (far shorter) pulse widths of forming the sine wave. It's easy to think that it doesn't matter what side of the transformer the choke is on, but the transformer has capacitance to the core and to the secondary, so I think it's important to wire the choke as physically close to the SPWM side as possible. I'd imagine it would decrease EMI and RFI too. Thoughts on this? |
||||
| Godoh Guru Joined: 26/09/2020 Location: AustraliaPosts: 378 |
Hi Haxby, all the inverters I have made and also the powerjack inverters have the choke on the primary side of the transformer. It makes sense to me to filter the waveform before dumping it in the transformer. The 8010 boards come with a capacitor and a diagram ( if you are lucky) on where the capacitor and choke go. No actual description of the requirements of the choke though. Just wondering if a multimeter with inductance measuring capabilities is up to the job of measuring choke inductance. I saw one in a Jaycar catalogue that had inductance and capacitance measuring built in and it was only $60 from memory. I would imagine if the choke were on the secondary side of the transformer then it could be much smaller and have lighter windings, but I would think that the spikes caused by the switching transients would be pretty wild to filter out. Pete |
||||
| Haxby Guru Joined: 07/07/2008 Location: AustraliaPosts: 418 |
Pete, I'm talking about what side of the primary winding the choke should be on. Put another way: There are two screw terminals for the transformer on my Sunyima inverter board. You would think there is no polarity as such, but one screw terminal outputs a 50 hz square wave, while the other outputs the variable pulse width modulation. I'm suggesting strongly that the choke should be directly connected to the SPWM output. Get what I mean? So the transformer will see something closer to a sine wave after the choke. |
||||
Revlac Guru Joined: 31/12/2016 Location: AustraliaPosts: 961 |
Generally the choke was mounted between the high side (high frequency side) of the inverter power board and the transformer primary, most of us built these inverters with the transformer quite close to the power board, about 300mm give or take, and fitted the choke within that space, uni polar needs one choke as was indicated in the manual. This was talked about a lot in the past, but is buried deep in the forum notes. Cheers Aaron Off The Grid |
||||
| mab1 Senior Member  Joined: 10/02/2015 Location: United KingdomPosts: 149 |
Interesting discussion. I was thinking along the same lines as Haxby and thought it a valid point - but reading what revlac said made me realise that the pwm is on the high side (top) of the h-bridge and 50hz on the low (bottom). But to achieve what Haxby wants the pwm would have to be on one side of the h-bridge - i.e. left or right and 50hz on the other. So as it is the choke is on the pwm side for 1/2 the ac cycle, and 50hz side for the other 1/2 cycle. |
||||
| Haxby Guru Joined: 07/07/2008 Location: AustraliaPosts: 418 |
EG8010 based boards that follow the recommend circuit will have one terminal output 50hz square wave, while the other terminal will be SPWM for both the high and low side of the sine wave. Putting the choke on the SPWM side, as close as possible to the fets, will decrease EMI and decrease the capacitive coupling issues in the transformer. It's a shame they are not labelled. Anyone buying these boards would need a scope to see which is which. On a side note, the fets that do SPWM will likely run hotter than the others, due to higher switching losses. So a highly optimised design will have more cooling on that side of the circuit. |
||||
| Godoh Guru Joined: 26/09/2020 Location: AustraliaPosts: 378 |
Thanks for the heads up Haxby, I did not realise that there were two different outputs from the FET board. I will have a look at mine and see what side I have the choke on. I know that on one inverter I put two chokes in, one on each leg. It would be great if they actually marked the outputs like you suggest. Then again not much is marked at all, on my boards all the writing is in chinese so figuring out what is what as far as adjustments go is not simple. Pete |
||||
| mab1 Senior Member Joined: 10/02/2015 Location: United KingdomPosts: 149 |
Oh, ok, so i got it the wrong way round: it is a left - right split for the eg8010 boards. Is it the same for the nano and other backshed diy pwm inverters? Or are they highside lowside split? |
||||
| Page 1 of 3 |
|||||
 Print this page |
