testing the egs002 inverter board

[Phaedrus]

Just had a look at your SCR latching circuit.Once the SCR is triggered,will there be 5v at the IFB pin?Atleast in the board diagram it looks like the red top traces are merging with the 5V net.As per the datasheet,the IFB pin seems to be rated for 0.5v TYP,so will 5v there be fine?
So unless there is a power Off and On,the inverter will not start again after an over current incident.
I am using the egs002 board with the 17pin connector for my first trials and then will try to spin my own board if needed.Do we will still have to do the mod on the egs002 board with the LM393 after the SCR latch?I just don't want to desecrate that lovely SMT board.

[oztules]

The egs002 board's native current limiter is not a good thing. Perhaps use a CT and rectifier bridge and load resistor to control the TFB pin. That never seems to cause trouble, but he ifb on the egs002 seems to have a train of spurious pules at shutdown.. be aware.

All I/O pins are rated at 5v for 5vcc. the .5v is the trigger level. I hit it with 5v or more to make sure it turns off, and as an extra precaution, so do the 2110.

Only a cold start guarantees that the 2110 is turned fully on *before* the 8010 soft starts.... other way around is death.



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

[tinyt]

It  took me several times to read this great thread during my limited spare time and would like to give this a try. To understand the concept, I tried to trace the manufacturing gerbers shared by oztules to create a schematic and it is attached. I kept part symbol placement approximately the same as in the gerbers and also kept the zero ohms resistors. I don't know how accurate I have made it.

I have limited finances and will probably start with something below 1kw. I don't know if I can get a toroid core for the transformer. I bought a 1kva EI transformer from ebay a while back and would like to know if I can use. Any guidance will be appreciated.

[Phaedrus]

Hi tinyt,
Thanks for the circuit diagram.
1)The temperature probe part seems to have an error.R11 is floating and probably should be grounded.R3 should go to the TFB pin from the junction of R11 and the thermistor.
2)D2 from the SCR cathode to the LED seems unnecessary.
3) I dont know why there is an addition zener D6 (6.5v) to the anode of the SCR. Probably 5v supply should have been enough.Maybe Oztules,might clarify that.
Just some superficial observations.Am sure someone will correct me if I am wrong.

[tinyt]

Hi Phaedrus,

Thanks for checking. You are correct, one leg of R11 is floating in my schematic (was sleepy when I was doing the schematic) it is now corrected in the attached. I assume that since R3 is 100 ohms and if a thermistor value of 10K @25C is used, with R11 being 10K. TFB will read around 2.5V @25C.

I think you are correct about D2, but it is in the gerber, picture attached also.

[oztules]

Yes there are some legacy things in there that are inexplicable.
For some reason I can't find now, I had the idea that the SD or the scr I was using  needed more than 5v to trigger,... hence the 6v supply.. Probably can be the 5v rail instead. When something is working this well, you don;t mess with it I suspect was my thinking later on.

I must have had other designs in mind with the floating supply on the led for the o/current, It can probably go too.

The 10k to ground (R11)
The R7 is now 1k not 4k3
The C8 is far more stable with 22n not 1n
Connector J8 needs to be a three pole socket. Pin 6 of the 8010 needs to go to the center pin, and released from the 5v as it is shown. Then off on will work as advertised. It originally  ( I originally) had trouble with smooth operation of the spwm pin 6, and so I abandoned it. Held at 5v, the unit stays on, and we turned it off with temp control. Later I changed back to the pin 6 shut down.

On my boards, I simply put the pins on the adaptor board UP for pin 6, so when it plugs in, there is a pin sticking up to take a plug that goes to the spdt on/off switch. It would be better to change the board to do this with a three pin socket, instead of the 2 pin and stick up pin.... if that makes sense.... ie pin6 to ground is off, pin 6 to vcc is on ( or the other way around)....

Another change I would do is put a 2 pin socket near R12 ( 120r) that shorts the interface of the tip35 and the R12 to ground for reset. I just solder on a few leads from gnd to that point on R12 to the spdt switch to do the same thing....

R25 needs to be changed to match the voltage of the transformer to be used. Mine was 14v and so R25 was 7k5. If you have 12v or some other voltage, it will need to be changed.

You could use a trim pot at the tmp probe to control the temp you want also.



Good work there. I like it.
So thats what it looks like in schematic form.



..........oztules

[oztules]

The 1kw EI tranny will work just fine. Heavier idle current may be the only thing... usually twice the torroid for the same size.

I will be doing a full how to for this inverter now it is mature, and has proved itself.
I would like to use your schematic when I come to doing that if I may?
What display credit would you like in that instance.. eg tinyt on anotherpower, or something else.

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

[tinyt]

Hi oztules,

Yes you may use the schematic and I don't deserve any credit for all the hard work you have done including the other contributors here. Maybe just put from the team in anotherpower or something.

In the attached updated schematic, I have incorporated your comments. please review and post any changes/corrections you want. Do you also want to remove all the 0 ohm resistors.

Thanks also for the EI transformer advice.

[oztules]

Yes, thats looking like what I build. The only change that needs to be made is the reset switch.
1. it is momentry 2 pin
2. it only shorts the rail to ground.

You have depicted it in a more sensible arrangement, where you don't directly short the rails, but turn to ground and then to rail voltage.

But in real life, this safer better way actually make the chance of blow up much worse. The reason for the 120r in the first place with the 100uf cap, is to have a situation where the voltage rise on the tip35 is slower than direct 40-60 volts. This gives it time to regulate properly.

Without the time lag due to 120r and 100uf, the tip is unable to establish 5v instantly, and will likely blow the zeners and possibly the chip.... so I know it looks brutal, but shorting the rail to ground ( after the 120r) is the safest and has proven to be fool proof.... as it always results in a soft start for the tips.
Straight 60v to the tips  caused me a few blowups.

You will note that in your rendition, the 100uf cap gets to charge up to full potential before make .... so it makes it worse still.

Other than that, yes the 0r can go. It was only for the bathtub boards, which were single sided.... and so I used 0r resistors, as it is easier than links for me.

Thanks for your efforts.

I have to apologize for my boards and no circuits, I just don'lt tend to do them, and this means my boards do not have the correct component description sometimes, as I was not interested in that part... just wanted to make the board in the bathtub so to speak.

When I have to modify that to get it manufactured, I must tidy up the component descriptions etc which were non existent.... and I don't get it right, or it is a series behind. So you have done other folks a good deed.


........oztules


John Tulloch

[tinyt]

Ok, thanks for the review. I have removed the 0 ohms and changed J9 to 2 pins to ground the R12 output. And thanks for the very clear explanation. Somebody like me who only draws schematics often times does not know the consequence of changes until it gets tested like what you have done extensively.

[oztules]

looks ok


......oztules

[tinyt]

Looks like I missed one important connection. Pin 9 of the EG8010Adapter should be grounded. It is now corrected in the attached. I also re-annotated to get rid of the missing R numbers when I removed the 0 ohms resistors. Unfortunately it is a mass re-annotation, so the capacitors, etc. probably got re-annotated. Sorry about that. I hope I made no more mistakes.

[Phaedrus]

1) Any reason why there are two connectors for the temp probes viz. J1,J2 ?
2) R4,100E as it is does not seem needed ?

[eidolon]

That 100 ohm is a noise filter. May not be necessary, but at those power levels there will be a lot of magnetic fields.

I'm a little iffy on the two temperature sensors. As there will likely be two heat sinks and a fan for each they need to be monitored separately. So the 10K resistor likely gets about halved. Given that they are added, one sensor could be high and the other low. It likely doesn't matter much but an ideal solution would be a dual op amp that sends the highest voltage to the board.

[oztules]

The fan sensor of the 8010 is utter rubbish like most inverter fans. They are just a bang bang thing. I would not use them for cooling purposes, any more than using the native ones on most inverters I have been familiar with. They all come on too late, and they all run too fast when they don't need to.

If you want your bulk caps to survive, you want to keep them in the <40c range.. so you want them as cool as possible.... eg a good capacitor will get 2000hours from a 105C environment, but will get 30 years at 35c. With a sensible fan driver system, this is very possible in my area, and 45 is possible in most areas. This is assuming 2-4kw continuous. You simply cannot do that with bang bang junk.... so use the temp pins at your peril. They do very good soft starts.

They are used by me only as a catch all in case the external power supply or the fan itself fail, and they will then shut the unit down before it cooks. There are two sockets, only so that you can monitor the heat sink for shut down, ... ie if things get up into the 60c range... then shut down, else leave it to the dedicated fan circuits to control things. he second is a legacy from using it as a shut down point.

It happened recently that a 10 amp buck converter that I use for the fan power, did fail. The TFB  caught it for me.... so it does what it is supposed to do.... catch the faults.

The 100r is purely a 0r resistor plus a 100 ohms... is it was for a single sided board, and I left it there.. legacy.... look at the board to see why.... it was not for noise, but just for the hell of it.

Ok.......now we have a circuit, everyone will want to change things a bit.

There have been a lot of these things built now, and another 3000 in Africa as well apparently, and I can guarantee that if they use those circuit boards, and that layout, they will work just like mine.

I can also point to the backshed, ( http://www.thebackshed.com/forum/forum_posts.asp?TID=8542&PN=1&TPN=1 )where smarter people than me have done the circuits, and then gone freestyle... only to find a year and a half later, that they have blown hundreds of fets, and finally come back to the same design as I use..... and people like Clockman use the same design, and have perfect success with units even bigger than mine...... there is a warning in there.

I might not know what I'm doing, and I maybe the village idiot, but mine work perfectly... ** as they are **....... I  was just lucky I guess.

It is heart breaking to see the work some folks put into these things, only to see it blow to bits every time they try it out....

It is simple.  If you want it to work perfectly well out of the box, then... copy mine exactly.... then experiment, as you will have a reference to work against.
You will build more than one, as they work so darn well, you will want to make a bigger one or a spare...... but one never seems enough....... now experiment with it. At least you will have one that works properly if the cunning changes you make don't work as advertised.


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