Author Topic: testing the egs002 inverter board  (Read 147632 times)

0 Members and 1 Guest are viewing this topic.

Offline peter

  • Newbie
  • *
  • Posts: 19
  • Karma: +6/-0
Re: testing the egs002 inverter board
« Reply #315 on: June 06, 2016, 09:22:16 pm »
An easy way to disable the LM393 on the EGS002 is to remove R7 and C21 then short C21 contacts.
This pulls down to ground both negative inputs into the 2 opamps on LM393 as well as the 8010 IFB.
This now ensures the gate drive chips will never get a high input on the shut down pins, nor let the 8010
receive a SPWMEN low to disable output.

Doing this stops the sudden stop/start that sometimes occurs, stressing the bridge mosfets sometimes to failure
when IFB is connected to a meaningful input signal. Now IFB is rendered useless.
Only when controlling the +5V supply is the soft start (and graceful stop) available.

Offline oztules

  • Moderator
  • Hero Member
  • *****
  • Posts: 1177
  • Karma: +105/-8
  • Village idiot
Re: testing the egs002 inverter board
« Reply #316 on: June 07, 2016, 01:27:24 am »
Not quite right from my experience with this thing.
The IFB works very well IF you control it ... and don't allow anything else to do it either....AND you totally insure that the driver chip inhibit does not ever ever work.

The IFB on it's own is very useful routine, and always soft starts ( if you have the settings for soft..... hard wired as well). It is the driver chips driven by the 393 that causes the problems... best get rid of all of it (393).

Control of the soft start is accomplished with
IFB  or TFB or SPWM (pin6)

All work equally well, so now I use IFB for current set points ( this morning set it by trial and error to 6kw.. so it is now well tested at high power now)

TFB works well, and is useful for stop start except for the flashing light irritation. It does not interfere with temp control either, as it is switch open when running, so open only to the NTC sensor

SPWM control is not so well researched by me at this time.
I am using it from today onwards, and will see if any anomalies arise... so far I found that pull down resistor of 1k4 and a make to +5v switch works very well.

Any one of these operations will result in soft start ... but you must.... must  get rid of any chance of the driver chips being able to initiate... as they have NO soft start sequence.... they will blow the fets up easily if you have a decent sized torroid connected... it will be a dead short as far as they are concerned.

Not sure I would use the 5v supply as a power switch.... would prefer boot up to be not under pressure.

Today, forced start up under 6kw of load... no sag, no fuss. This was using the SPWM pin, then changed the IFB trim pot, and it stopped the unit at 5kw, then restarted under the same load, and cycled off again.... did this a few times expecting the worst.... never happened.

I have thus far seen no evidence of a graceful stop.....both the literature and observations say pretty much the same thing.... swpm just stops in the off position ( high or low depending on how you have set it up)... gate pulldown does the rest if drive is disabled for any reason ( holds gate to gnd).

Hope that helps to clarify what I have found works and what seems not to work so well.


.........oztules



Flinders Island...... Australia

Offline peter

  • Newbie
  • *
  • Posts: 19
  • Karma: +6/-0
Re: testing the egs002 inverter board
« Reply #317 on: June 07, 2016, 07:18:26 am »
I agree the shut down pins on the gate drive chips must not be driven by anything since this results in huge and bad things
happening in the mosfet bridge.
The quad opamp on the inverter board is supplied with +12V. The outputs of these - at 12V - are too high to directly drive the EC8010 IFB pin or SWPMEN.
So I need the 393 on the EGS002 board to convert levels from 12 to 5V.
 Instead of removing R7 and shorting C21, I may need to cut the traces running to the shutdown pins on the IR2110 driver chips
and, and....stuff it, I may as well make my own driver along the lines of what you have done!
You certainly have things working really nicely.

When I say "graceful stop", it is not a 3 second stop. It takes about 1  50 Hz cycle to stop power going into the
primary of the transformer. No sudden transients. I had a look at the PWM applied to the gate drives and it seems
to me the 8010 is quickly ramping down the switching duration during this final 1/2 cycle. Not just stopping within a few of the 23.4 kHz PWM cycles.
But then, this is from the inverter board controlling the device via +5V supply only, not IFB or SPWMEN.
here is a capture of the output voltage after shutdown (+5V supply pulled to ground)

Don't know what to do now, I don't trust the original design of the inverter board over current limit. We can't have the gate driver chips shutdown by anything. Time to have another think I suppose.


Offline oztules

  • Moderator
  • Hero Member
  • *****
  • Posts: 1177
  • Karma: +105/-8
  • Village idiot
Re: testing the egs002 inverter board
« Reply #318 on: June 07, 2016, 09:29:42 am »
5801-0
You can see from this that the 393 does not have 12v... the only thing with 12v on the entire board is the driver chips output stage.. the rest is 5v... so don't worry about that aspect.... ie vcc of 393 is +5v

Drive the IFB pin with the current transformer, and it will activate at 0.5v and pull outputs down to zero instantly.... no messing about.
Drive the TFB with 5v and it will stop instantly as well.. both are soft start recovery with pin 21 high ( soft start select)
And of course pin 6 is soft start and instant stop if pin 21 high as well.


By turning off the 5v as you suggest, it goes through the under voltage lock out protection I think... thats where the lag is I suspect... I have not tried it... bit dangerous to do as a control I expect. Anything may happen as the voltage degenerates, I fully expect the under voltage lock will work out fine... but....it is doing this with a decaying voltage on the drivers too, and not sure if that will always be a good result.

Providing you ground the pin 13's of the 2110, you will have no problems. I did that by cutting the 393 off, shorting pin 1 to 4.. there are probably more elegant ways to do it, but that works.... then you can use the current sense as you would expect to have done.... ie a CT and rectifier to pin 1 of your interface.

pin6 spwm will now be pulled high by the 10k r30. I found that on my 8010 board, it was best to make it more definite and lower impedance, in fact I use 5v for on and a pull down resistor of 1k5 or thereabouts to shut it off by default if you want a stop start switch, thats the way to go I think. The temp is as good as any without getting to the 002 board direct.

 I know this works.

But I don't use the 002 boards anymore... so I have 10 or more spare..... I like the 8010 boards instead.

It does not matter which board type you use, you will need an isolated AC feedback to achieve consistent AV out voltage if there is any chance of the battery side sharing  an earth ( high impedance as well as low) with the ac side ( off gird house with MEN earth system etc). Can be opto or transfromer... I used the transformer, as I have plenty of 1w 14v type to burn up that I haven't used in over a decade or more.


I think I'm the last person on the planet with an old fashioned  Cathode Ray Oscilloscope now   ...... feeling very old.....

......... oztules




Flinders Island...... Australia

Offline ClockmanFrance

  • Full Member
  • ***
  • Posts: 171
  • Karma: +11/-0
Re: testing the egs002 inverter board
« Reply #319 on: June 07, 2016, 10:02:08 am »
Hi peter,

If it helps, I have just posted up the latest OzControl Board No8, with the latest amendments.

http://www.fieldlines.com/index.php/topic,149017.msg1037736.html#new

http://www.thebackshed.com/forum/forum_posts.asp?TID=8706

Sorry Anotherpower, I just can not post images here. (For some reason this site is blocked by the French Authorities, and I have to use a VPN site in Germany, its free but limited bandwidth, times etc).

At present, I am not quite Man enough to go the Pin 6 route, so for the time being I have stayed with the ON/OFF on the Over Temperature system, and it still works as an Over temperature with a thermometer probe/sensor attached.


Offline rossw

  • Senior Moderator
  • Hero Member
  • *******
  • Posts: 879
  • Karma: +35/-0
  • Grumpy-old-Unix-Admin
Re: testing the egs002 inverter board
« Reply #320 on: June 07, 2016, 03:57:21 pm »
I think I'm the last person on the planet with an old fashioned  Cathode Ray Oscilloscope now   ...... feeling very old.....

Feel comforted, Oztules... you're not alone.
I still have my old BWD 539C from when I was a teen, and a Tek 24445A, and something else I can't recall that I haven't used in ages. The little Tek TDS210 (DSO) is my knockabout cro for quick-and-dirty, because it's small, light and fairly tough.

Offline frwainscott

  • Newbie
  • *
  • Posts: 21
  • Karma: +0/-0
Re: testing the egs002 inverter board
« Reply #321 on: June 08, 2016, 01:44:20 am »
Anyone looking for torroids in the US.....

 Here is a link to 2kw  torroids -

http://www.ebay.com/sch/i.html?&_from=R40&_trksid=p2054502.m570.l1313.TR0.TRC0.H0.TRS5&_nkw=Toroidal%20transformer%202500VA%20240V%20secondary%20153V%20primary%20with%20screen

The dude says he has 10+ on hand they have a 230 volt secondary and two 153 volt primary. As is .5 volt/wrap... they are $50.00 USD......

Frank...

Offline MarNet

  • Full Member
  • ***
  • Posts: 60
  • Karma: +0/-0
Re: testing the egs002 inverter board
« Reply #322 on: June 08, 2016, 03:44:29 pm »
Hi all
Antman, did you seen the pwm library?
I mean this one
http://forum.arduino.cc/index.php?topic=117425.0
I don't have the oscilloscope to test it but there some guyz said its ok.
My mppt will be based on simple buck converter schematic...
I'll see how Tim Nolan 's algorithm works
I only had a look on that, i seen that 20uH, on my knowledge its 50khz... For 100khz there is 10uH delay but its not easy on 10 bits
My mppt will be for an 60A 24v axial wind power generator
I have difficulties with output current sensor....  I'm thinking about reading a shunt resistor at this stage.
I'm based on Dublin /Ireland but my farm is not here...

Offline Antman

  • Full Member
  • ***
  • Posts: 51
  • Karma: +0/-0
Re: testing the egs002 inverter board
« Reply #323 on: June 09, 2016, 04:42:35 am »
Hey Mar

I had a programmer look at the programming and the 8 bit arduino isn't capable of more than an 8 bit PWM at over 62.5Khz. I've ordered some of the STM32 boards as recommended by Frackers

http://www.aliexpress.com/item/STM32F103C8T6-ARM-STM32-Minimum-System-Development-Board-Module-For-Arduino/32326304541.html?spm=2114.01010208.3.1.I0X3mZ&ws_ab_test=searchweb201556_0,searchweb201602_5_10037_10017_507_10032_401,searchweb201603_6&btsid=076970e2-7f8b-4eac-aba9-5fd2311095e4

I use the INA271 and a 10 milli ohm shunt for accurate current reading in most of my projects.

Offline Dr_Zogg

  • Full Member
  • ***
  • Posts: 109
  • Karma: +0/-0
Re: testing the egs002 inverter board
« Reply #324 on: June 09, 2016, 06:57:39 am »
hi Oz,
would you happen to have a component overlay of the new pcb?

Cheers,
           Zogg
Of all the things I've lost, I miss my mind the most

Offline jdevine82

  • Newbie
  • *
  • Posts: 1
  • Karma: +0/-0
Re: testing the egs002 inverter board
« Reply #325 on: June 09, 2016, 09:36:30 am »
Hi all
I have just looked back into all this after following Oztules advice with a powerjack conversion which has been running by house for a year or so. I am wondering if the gerber files or dxfs are available? Just makes it a bit easier than pdf for those not printing out the pcb layouts. Thanks to everyone for the information, truly the opensource spirit makes dreams come true for many people.
Jason
Greetings from Sunny West Wyalong

Offline oztules

  • Moderator
  • Hero Member
  • *****
  • Posts: 1177
  • Karma: +105/-8
  • Village idiot
Re: testing the egs002 inverter board
« Reply #326 on: June 09, 2016, 06:02:30 pm »
Clockman has difficulty processing some stuff to this site for reasons to do with the French system...
So here are the files posted on the sites he can.

* 2016-06-07_105638_OzContrlBrd8Etch7thune16a.pdf (479.65 kB - downloaded 478 times.)

* 2016-06-07_105821_PCBOzContrlBrd8List.pdf (355.39 kB - downloaded 651 times.)
or
5815-2

5817-3

Are they the ones your after Zogg.

Jason, I will put up a pcb file for protel. You can extract the gerbers, or alter it etc. It is not identical to the clockman one, it uses pin 6 for on/off... so electrical almost the same.
 
It will be later today when I have time to check it.

Do you have a protel pcb program PFW for windows is what it will be, but you can download the latest altium, and it will run it as well.
Most board manufactures across the world can run protel files.

...........oztules

Flinders Island...... Australia

Offline Dr_Zogg

  • Full Member
  • ***
  • Posts: 109
  • Karma: +0/-0
Re: testing the egs002 inverter board
« Reply #327 on: June 09, 2016, 07:51:09 pm »
i grabbed those ones from the other site but i have etched my board using the last pdf that you put up in this thread and your layout looks a bit different to clockmans

Cheers,
           Zogg
Of all the things I've lost, I miss my mind the most

Offline oztules

  • Moderator
  • Hero Member
  • *****
  • Posts: 1177
  • Karma: +105/-8
  • Village idiot
Re: testing the egs002 inverter board
« Reply #328 on: June 09, 2016, 10:37:51 pm »
Do you mean the one ?
"Here is the tentative pdf.
Will build one today all going well.
If no changes then you can action it.
Don't use this except for perusing at the moment.
"
you weren't supposed to use that one, I will have to go back to it and check changes from then.. I think they are mostly convenience changes ( slightly different part size to match what I have on hand etc.

Let me know which one you etched.. all will work, but some may require a track nip and tuck...


..............oztules
Flinders Island...... Australia

Offline Dr_Zogg

  • Full Member
  • ***
  • Posts: 109
  • Karma: +0/-0
Re: testing the egs002 inverter board
« Reply #329 on: June 09, 2016, 10:49:31 pm »
$#!+ balls lol. looks like i have made the wrong one. is it still workable or would i be better off making a new one?

Cheers,
            Zogg
Of all the things I've lost, I miss my mind the most