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.
Warpspeed Guru Joined: 09/08/2007 Location: AustraliaPosts: 4406
Posted: 10:53am 31 Jan 2017
Copy link to clipboard
Print this post
Yes that is correct. It either passes through the hole and counts as one full turn, or it does not.Cheers, Tony.
Tinker Guru Joined: 07/11/2007 Location: AustraliaPosts: 1904
Posted: 09:55pm 06 Feb 2017
Copy link to clipboard
Print this post
Warpspeed, I'd appreciate if you could look at the schematic below and confirm I'm on the right track with those high current gate drivers.
In particular my arrangement of ground and gate drive return connections.
The 5V and 12V grounds are all connected together at the battery negative rail. Oztules has these connected to a common ground plane on the PCB, I'm thinking the COM (lo side return)connection is better separated from the logic ground?
Thanks.
Edited by Tinker 2017-02-08Klaus
Madness Guru Joined: 08/10/2011 Location: AustraliaPosts: 2498
Posted: 10:19pm 06 Feb 2017
Copy link to clipboard
Print this post
Won't you need a charge pump on the high side?
I am just feeling my way around in the dark here so I may well be wrong but this is my guess.
Edited by Madness 2017-02-08There are only 10 types of people in the world: those who understand binary, and those who don't.
Warpspeed Guru Joined: 09/08/2007 Location: AustraliaPosts: 4406
Posted: 11:17pm 06 Feb 2017
Copy link to clipboard
Print this post
Yes, that is it. Both TC4452's just share the same power and grounds as the output sides of the IR2110. The diodes across the gate drive resistors are not required. Cheers, Tony.
Warpspeed Guru Joined: 09/08/2007 Location: AustraliaPosts: 4406
Posted: 12:22am 07 Feb 2017
Copy link to clipboard
Print this post
One thing you will need to do before you make a circuit board is to prototype the whole gate driver part, and check that there is sufficient dead time at the mosfet gates.
Its vital that one mosfet turns off before the other one turns on. There absolutely must not be any conduction overlap, but a short rest interval of maybe about 100nS where both are off together.
That may be done in software, or in hardware before the IR2110, I do not know. But it must be checked before you finally power up the main H bridge.Cheers, Tony.
Tinker Guru Joined: 07/11/2007 Location: AustraliaPosts: 1904
Posted: 12:45am 07 Feb 2017
Copy link to clipboard
Print this post
Thanks Tony, I'm so glad I asked. Its easy at my age to do something silly and at the price of these IXTN320s it can be expensive to get it wrong.
To make double sure, here is the amended schematic:
Klaus
Tinker Guru Joined: 07/11/2007 Location: AustraliaPosts: 1904
Posted: 12:51am 07 Feb 2017
Copy link to clipboard
Print this post
The EG1810, where all the magic originates, has provision for 4 different dead time settings. I have set it to 300nS. this seems to work fine with a simple 4 x HY4008 Mosfet bridge.
Madness selected a longer dead time with his 16 x IRFP4110PBF Mosfets. Klaus
oztules Guru Joined: 26/07/2007 Location: AustraliaPosts: 1686
Posted: 12:52am 07 Feb 2017
Copy link to clipboard
Print this post
Klouse,
My take on this......this is a digital circuit, and both linear feedback circuits are floating above ground anyway, and temp is inconsequential, (as are the other two).
Ground noise in this case is immaterial as far as I can tell with this class of circuit.... linear would be a different matter entirely, and proper grounding techniques would be necessary.... but I can't see how noise in the earth lines would effect any part of the operation of the 8010, it is all pretty low impedance stuff.
I have found you can kick it , beat it, stick your fingers on any part of it.... and it makes no difference, so it is pretty immune to noise from my tests with it.
I did discuss this with an Electronics Engineer in South Australia.... and he told me not to get too carried away with the grounding either for the same reasons.... He wasn't even overly concerned with ground loops on the circuit board.
Maybe Warp has a different take on this.
...........oztulesVillage idiot...or... just another hack out of his depth
Warpspeed Guru Joined: 09/08/2007 Location: AustraliaPosts: 4406
Posted: 11:43am 07 Feb 2017
Copy link to clipboard
Print this post
The final circuit as drawn is correct, and its good to know the 8010 board definitely provides the required dead time. That is not always the case with some PWM sources, but it is something that is a necessity.
Grounds are interesting, and not only from the noise aspect but getting it wrong can blow things up. The IR2110 has three common or ground pins that need to be treated correctly, the recommended way is as follows.
The first is pin 13 Vss which should come from the ground on the 8010 board and the five volt dc supply that powers both the 8010 and input side of the IR2110.
Second ground is pin 5 Vs which is a fully floating common that must be connected directly to the source of the upper mosfet. This can be anywhere between zero and up to several hundred volts positive. It flies up and down freely as the output mosfets switch.
Third ground is pin 2 com which is also a floating common that must go directly to the source of the lower mosfet. But unlike Vs it can only float up and down at most by two volts or the IR2110 will be damaged.
The five volt ground side must also be connected to the power negative of the H bridge. That should probably best be done at the source of the lower mosfet. That may not always be possible. For instance if there is a current shunt in the negative side of the H bridge.
But however the main system grounding works, there must never be more than two volts between pins 2 and 13 of the IR2110. Its best to tie both back to the source of the lower mosfets.
Normally the main system grounding would be at point X, but if there is a current shunt in the negative side it would be at point Y as drawn.
Cheers, Tony.
Tinker Guru Joined: 07/11/2007 Location: AustraliaPosts: 1904
Posted: 01:38am 08 Feb 2017
Copy link to clipboard
Print this post
Warpspeed, below is a screenshot of the EG1810 & IR2110 PCB.
You can see my grounding arrangement here. The COM return (pin 2) connects to the ground at the bottom (20 pin connector). The motherboard, where this plugs into, has its ground plane connected to the negative battery rail.
This PCB works in a slightly different version (no 4 x 3 pin connectors) with my test inverter.
The 4 x 3 pin molex connectors at the bottom connect the TC4452 PCB boards which are screwed directly to the IXTN320 terminals. I have not done these yet.
Klaus
Warpspeed Guru Joined: 09/08/2007 Location: AustraliaPosts: 4406
Posted: 11:16am 08 Feb 2017
Copy link to clipboard
Print this post
That's going to make a very neat and compact inverter.Cheers, Tony.
Tinker Guru Joined: 07/11/2007 Location: AustraliaPosts: 1904
Posted: 12:40am 21 Feb 2017
Copy link to clipboard
Print this post
So I had my IXTN320 blocks for some days now, time to report some progress with these.
They are neat little Mosfet blocks but there are a few things worth mentioning:
They have no markings as to which terminal is which. Only way to tell is to refer to the drawing on the specs sheet. There one can see a mounting slot, the other end has a hole. Still a bit confusing for me so I made some little labels for the gate and drain terminal.
Connecting 4 of these on my flat back heat sink (ex Aerosharp 3KW) takes a bit of thinking to achieve an economical layout with minimum interconnections.
I started with drawing full size on graph paper:
You might also note the full scale perspex template I made, very useful to work out where the terminal screws locate.
Some considerations with respect to the heatsink were: I do not enjoy drilling and tapping blind holes so all mounting holes had to fall within the valleys between the fins. These are 10mm apart. Fortunately the IXTN320 has its mounting holes 30mm apart so this can be made to fit nicely. The mounting screws can only accept a small diameter head, I used allen head screws which work very well.
The placement of the capacitors separate the 2 pairs of Mosfet blocks a fair bit.
The little PCB boards which hold the gate driver were made to screw onto the gate and one of the source terminals to locate them. I used a simple single sided board, easy to etch at home, to hold the few components. I think a double sided board is way overkill for this and a ground plane may actually be detrimental by introducing unwanted capacitance at the gate.
A left and a right handed PCB board version was required to have the 3 pin molex terminal facing outward.
You may notice there is now a back plate on the heat sink.
This converts the whole heatsink into a module which can be fully assembled with all the parts and then inserted as one unit from the inside, the heat sink fins passing through a cutout in the cabinet. As the mounting screws for that plate are now on the outside of the cabinet, the job of getting this unit in or out is very quick. Just unplug 3 molex connectors and disconnect the power cables.
I had planned to use my tinned copper strips for connections but these were far too 'bendy' for my liking. So I used 25 x 6mm alu strip instead. Very rigid to hold all those capacitors securely and at 150mm sq plenty of conductor area.
These power 'rails' required to be above the Mosfet block terminals to clear the little PCB's. I used 10mm diameter alu rod and turned some spacers from them. 10mm is about as big as can be fitted to the terminals. The blocks come with M4 screws but these are only just long enough to fix the little PCB on. For the spacers I got longer M4 screws and machined the spacer lengths to suit stock screw sizes.
The spacers can just be seen on the above picture.
To prevent poor contacts due to the alu oxydising I brush cleaned the connecting surfaces with a rotary plastic bristle brush first. Then I used a thin smear of an electrical jointing compound for alu (Alminox).
So that is as far as I got today, the electronics board is finished but mounting it and testing will have to wait until next week when I get back from a sailing break..
Klaus
Warpspeed Guru Joined: 09/08/2007 Location: AustraliaPosts: 4406
Posted: 06:13am 21 Feb 2017
Copy link to clipboard
Print this post
Excellent Klaus. Its all a bit fiddly to make, but once the bits are fabricated it will all go together and comes apart in seconds without any soldering, and it all makes for a very neat job.
Start out at low voltage and very low power, monitor everything, and very gradually work it up to flat out.Cheers, Tony.
oztules Guru Joined: 26/07/2007 Location: AustraliaPosts: 1686
Posted: 10:24am 21 Feb 2017
Copy link to clipboard
Print this post
Keen to see how this goes for you Klous. Looks the goods.
........oztulesVillage idiot...or... just another hack out of his depth
Grogster Admin Group Joined: 31/12/2012 Location: New ZealandPosts: 9308
Posted: 02:25pm 21 Feb 2017
Copy link to clipboard
Print this post
I too am very interested to see how you get on with this, as this big "Big-block V8" MOSFET method is how I also plan to build mine - your's will be the test subject, as you will have yours up and running before me, so I can find out how you get on.
Thanks for being the test subject. Smoke makes things work. When the smoke gets out, it stops!
johnmc Senior Member Joined: 21/01/2011 Location: AustraliaPosts: 282
Posted: 08:34pm 21 Feb 2017
Copy link to clipboard
Print this post
Good day Klaus,
Most impressed with the excellent work and the description of your inverter build. johnjohnmc
Tinker Guru Joined: 07/11/2007 Location: AustraliaPosts: 1904
Posted: 09:47pm 26 Feb 2017
Copy link to clipboard
Print this post
Thanks for the kind comments, I appreciate them. I would also appreciate if oztules spells my name right, its a touchy subject to me.
Anyway here is a picture of everything assembled on the power module:
You'll notice two little fans. The one on the left is a 24V fan running on 12V all the time, its right under the top exhaust grille. The other fan is controlled by the EG1810 logic. Most of the heat is dissipated by the fins located outside the enclosure. There is another, larger, fan blowing straight onto the toroid. It has its own temperature controlled relay to turn it on.
Now should follow my report how this power module performed.
I took warpspeed's advice and started with a lower voltage, my 24V battery bank. Everything powered up normal from this through a big 30 Ohm resistor in series.
The AC waveform looked fine on the scope, just a very slight flat topping at the peak of the trace which I think is caused by the turns ratio being for a bigger primary voltage. The series resistor was shorted at this stage.
The dead time was set to 1us BTW.
Then I connected it to my 48V battery bank. Sadly it did not get past the switch on stage. A slight 'pop' was heard at the instant of switching on (with the series dropping resistor) and the DC rail voltage dropped to zero.
One half of the bridge (2 blocks) was found short circuited when I disassembled it.
Now I'm thinking I have to start testing this in a different way, those Mosfet blocks are not cheap.
Can I fool the circuit into thinking its switching a Mosfet when I connect a 2200pF capacitor to the gate drive chip output?
I can simulate the VFB (voltage feed back) to the 1810 chip easily enough with an adjustable DC supply.
Then I could measure as long as I wished with my 25MHz dual trace oscilloscope and check what's happening with that circuit
In particular what signals are expected on the pins of a good IC to see if any chip had gone south as the Mosfets went there.
I think I'll have to understand more fully what's going on here, rather than the suck it and see method used so far.
I would appreciate any tips as to what wave forms should be where.
Thanks, Klaus.Klaus
Warpspeed Guru Joined: 09/08/2007 Location: AustraliaPosts: 4406
Posted: 11:13pm 26 Feb 2017
Copy link to clipboard
Print this post
Oh dear, that is really bad.
This requires some very cautions testing before the next power up.
You could try it first with four much smaller mosfets and run it up under no load.
There are several possibilities before you have another try. First check the gate drive waveforms with no dc power connected to the bridge. The two upper mosfet isolated gate drivers will not have any dc power to operate, unless the charge pumps are operating. So a couple of 12v batteries might be one way to supply temporary power to the upper two gate drivers.
Check for nice clean fast rise time gate waveforms, and that there really is 1uS of dead time at the gates. And everything looks o/k.
Next test would be to unplug the four gate driver chips, and short all four gates to their respective sources. Connect just the bridge (without the large electrolytics) to an adjustable current limited power supply and set the current limit to just a few milliamps. Slowly increase the voltage up to some safe maximum, say 70v for a 48v inverter. If one of the mosfets breaks down, they are probably Chinese crap. If the current limit is set very low, the breakdown may not be destructive. But at least you will then know.
The next step would be to power the whole thing up, again from a current limited source under no load, and very slowly increase the dc voltage to the bridge while monitoring the current directly at the mosfets with an oscilloscope. There should only be a very few amps, but if there are very large current spikes, discover why and fix it.
Apart from being extremely timid and creeping up on the problem very slowly, there are no real ways to be certain of anything.
Just going a tad off topic here...
I have been running three potentially 2Kw home brew MPPT solar controllers at only 1Kw for almost a year now, without any problems. The power is limited by heat sinking capacity. So I bought ten much larger mosfets that had much higher ratings. These cost me $300.00 for ten. Original mosfets 600V 48A 100 milliohms. These new ones 750v 75A 33 milliohms. Only one third the heat, which will allow at least double the power.
Put three of them in at night with the current limit still set to 1Kw (this is a scary high voltage system). When the sun came up all three solar controllers blew up. Replaced one more mosfet and that blew up instantly as well. Put the original three back in and everything runs fine again.
Moral of the story, those F&%*g Chinese counterfeit parts got me once again. And they were not cheap counterfeits either, damned expensive ones. Nothing wrong with the circuit its been running for many months and is running now. But I have just been ripped off $300.00 and there is nothing I can do about it.
This has happened to me before, but usually the low price is a pretty good indicator of fake. But not in this case.
So Klaus, a faulty circuit is not the only reason why things sometimes go bang. Sometimes its some thieving foreign devil. And this counterfeit parts problem is getting worse.
Cheers, Tony.
Tinker Guru Joined: 07/11/2007 Location: AustraliaPosts: 1904
Posted: 12:34am 27 Feb 2017
Copy link to clipboard
Print this post
Thanks Tony, much appreciated. I printed your suggestions out and will slowly work my way through them before connecting more expensive Mosfets. I do have a stack of salvaged smaller (TO220)ones to experiment with. Never thought of using a variable voltage source for testing, I have a variable power supply to do that.
Sorry to learn about your not so good parts purchase, very annoying isn't it? Klaus
Madness Guru Joined: 08/10/2011 Location: AustraliaPosts: 2498
Posted: 12:42am 27 Feb 2017
Copy link to clipboard
Print this post
It's sort of comforting to know I am not the only one.There are only 10 types of people in the world: those who understand binary, and those who don't.
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.
Print this post
