Use for the Tesla Switch

[citfta]

Diode type

Hi Garry,

I see Matt hasn't answered your question about the diodes yet so I guess he is busy or out of town again. Those diodes appear to be there to supply a negative connection for the controller back to whichever battery has the highest voltage. I think you can use almost any diode as long as it has a voltage rating of 50 volts or higher and a current rating of a half amp or higher. Matt will correct me if that is not right but I am pretty sure you will be safe with what I have said. I built mine using the opto coupler circuit so I didn't need those diodes. Good luck on your build.

Carroll

[Turion]

The Good (UPS) Fairy left me a package late last night that I just found! All my parts!! Finally!!! Now I have everything to build all three small ones and the larger one (except the transformer for the larger one) unless of course the transformer from my battery charger will work so I don't have to buy one. I also bought extra parts, so breaking something here or there won't slow me down again. Hope to have this up and running later today. Finally, Finally, Finally!

David

[citfta]

Hi Turion,

Glad you got your parts at last. If I remember right you were going to use a modified transformer like I used and also one you had rewound like Matt has shown us. I would be very interested to see how they compare to each other when used in the same circuit. I suspect the rewound one will work better but it will be interesting to see that for a fact. Don't waste time reading any more. Get to work now! Good luck.

@Sobakin, That is a very impressive build. Congratulations on getting it to work. I seems to me that you are putting a pretty heavy load on such small batteries so I am really curious how long they will hold up with this new TS design. Please keep us posted of your results.

Carroll

[Matthew Jones]

Garry.
Carroll is correct. Any standard diode will work 50v or above.

@Sobakin

Very nice build. I am surprised by the frequencies. One thing to watch....
Normally everyone wants to watch the power why the batteries are running. This is the way I am encouraging testing of the switch againts the baseline test.
But really you also want to measure the batteries before you run and after they run and rest.
Your going to see that even after 8-10 hour runs most of your power is still in the battery.
I ran a 100 watt loads for weeks, and the batteries appeared to be dropping in Voltage. But after I stopped and let them rest they returned to just below the voltage they started. Now granted the power moved around but it was still there.

And don't be scared to move on and run bigger loads, even on small setup 50 - 100 watts will show you something. Your pulling off the ground side in a potential based system C20 and all that really is not applicable.
I have never damaged a battery and I really try to push all the power I can through and out of the system. I have also watched the internals of the batteries while big loads are running and I have yet to see anything boil in this setup.

Cheers All
Matt

[Sobakin]

So first test is finished.
Batteries discharged in ~4 hours: 2 hours before swap and 2 hours after.

I am surprised by the frequencies.

I think it's because of transformer's small size and section.
It work good at 200-300 Hz too, but output voltage is 0.05 - 0.1V less than at 900 Hz. At about 2 kHz effectiveness start to go down dramatically, iron can't work at such frequencies.
But while transformer power is 7W at 50Hz, at 1000 Hz the same core with rewind can do up to 140W, if wires could handle >10A, of course.
When I lower frequency to 100 Hz, output voltage raise by ~1v, but battery voltage starts to drop (raised input current) and transformer became heating, so effectiveness go down abruptly (transformer core go to saturation).

Also I've rewound bigger iron transformer, former 60W 50Hz, I've try it, but it too loud while working, while my load and batteries so small, I don't need such big transformer now.

One thing to watch....
Normally everyone wants to watch the power why the batteries are running. This is the way I am encouraging testing of the switch againts the baseline test.
But really you also want to measure the batteries before you run and after they run and rest.
Your going to see that even after 8-10 hour runs most of your power is still in the battery.
I ran a 100 watt loads for weeks, and the batteries appeared to be dropping in Voltage. But after I stopped and let them rest they returned to just below the voltage they started. Now granted the power moved around but it was still there.

After high load and some time for resting battery voltage will go up in any case, with switching or without it, but in my case it raised from 10.5 to 11.5V, not to initial 13V, and then I put them in work after 2 hour of resting, they will work no more than some minutes and again discharged
Maybe I miss some key?
Scheme of my replication attached.

And don't be scared to move on and run bigger loads, even on small setup 50 - 100 watts will show you something. Your pulling off the ground side in a potential based system C20 and all that really is not applicable.
Matt

I'll try to load them harder, I have 21W bulb, with anonter 10W it will be ~30W, but I need some hours to charge batteries.

[garrypm]

Hi Matt & Carroll,

Thanks for the replies.

Where may I find the schematic that uses the optos. I guess for a little
more effort, one would feel safer, divorcing the logic from the power.

Thanks, Garry

[Bit's-n-Bytes]

Hi Matt & Carroll,

Thanks for the replies.

Where may I find the schematic that uses the optos. I guess for a little
more effort, one would feel safer, divorcing the logic from the power.

Thanks, Garry

http://www.energeticforum.com/157633-post3650.html

Bit's

[citfta]

Hey Bits,

Good to see you in here again. Are you building one of these too? It is a pretty impressive operating TS. Keep in touch.

@ Garry, you can find the full drawing of the TS with optos in the PDF Matt made. It is on page 25 I think and is the next drawing down from the one you are looking at.

One of my batteries took a big nose dive suddenly dropping to 9 volts or so. I guess it was sulfated worse than I thought from sitting around for the last few months. When I get it rejuvenated I'll get back to my testing.

Carroll

[Bit's-n-Bytes]

Hey Bits,

Good to see you in here again. Are you building one of these too? It is a pretty impressive operating TS. Keep in touch.

@ Garry, you can find the full drawing of the TS with optos in the PDF Matt made. It is on page 25 I think and is the next drawing down from the one you are looking at.

One of my batteries took a big nose dive suddenly dropping to 9 volts or so. I guess it was sulfated worse than I thought from sitting around for the last few months. When I get it rejuvenated I'll get back to my testing.

Carroll

Hi Carroll, yes I have been working behind the scene on the "Big Kahuna" to give this setup a try. The transformer I am configuring is a 1500va. I have designed a PCB to switch IGBT's, 2 to 4 per switch and should have that all together by next week. I'll be using the PICAXE 18M2. Had to take a bit off due to a Massive melt down with the heart. Much better now though.

Thanks

Bit's

[citfta]

Hi again Bits,

Well I am sure glad you are doing ok now. I have learned so much from you and Matt and some of the other guys over the last couple of years that we have been working on this. I'll be looking forward to your posts.

I saw the other day where you were making some suggestions to someone about the programming for the Pic. I am still using the 18x since I am pretty comfortable with it now. I did discover a problem the other day. When I was using the pulsout command I noticed that with short pulse times I had a pause after 2 pulses. Then it would repeat. 2 pulses and a pause. I added a short pulsout command for an unused pin between my 2 pulses and that evened things out so I had the same short pause between each pulse. I assumed the pause was caused by the program jumping back to the top and starting over. I had tried raising the clock speed to 8 mhz which is the max for the 18x. That didn't really help that much but the added pulse seemed to be the best solution. My transformer also started working better with the pulses firing evenly. Is there a better solution or is this just a problem we have to live with when using the Pic?

Thanks again for all your help,

Carroll

[Bit's-n-Bytes]

Hi again Bits,

Well I am sure glad you are doing ok now. I have learned so much from you and Matt and some of the other guys over the last couple of years that we have been working on this. I'll be looking forward to your posts.

I saw the other day where you were making some suggestions to someone about the programming for the Pic. I am still using the 18x since I am pretty comfortable with it now. I did discover a problem the other day. When I was using the pulsout command I noticed that with short pulse times I had a pause after 2 pulses. Then it would repeat. 2 pulses and a pause. I added a short pulsout command for an unused pin between my 2 pulses and that evened things out so I had the same short pause between each pulse. I assumed the pause was caused by the program jumping back to the top and starting over. I had tried raising the clock speed to 8 mhz which is the max for the 18x. That didn't really help that much but the added pulse seemed to be the best solution. My transformer also started working better with the pulses firing evenly. Is there a better solution or is this just a problem we have to live with when using the Pic?

Thanks again for all your help,

Carroll

Send me your code and I'll take a look. Sounds like it is doing just as you are describing it.

Thanks

Jeff

[citfta]

Ok Bits,

Here it is:

setfreq m8

main:

pulsout 2 , 200
pulsout 4 , 2
pulsout 3 , 200

goto main



The pins I am actually using are pins 2 and 3. Pin 4 is just to even out the pauses between pulses.


Carroll

PS: I see you are also helping Bizzy on another thread so don't jump right on this if you don't have the time right now. I have it working for now. I was just curious if there was a better way.

[Bit's-n-Bytes]

Ok Bits,

Here it is:

setfreq m8

main:

pulsout 2 , 200
pulsout 4 , 2
pulsout 3 , 200

goto main



The pins I am actually using are pins 2 and 3. Pin 4 is just to even out the pauses between pulses.


Carroll

PS: I see you are also helping Bizzy on another thread so don't jump right on this if you don't have the time right now. I have it working for now. I was just curious if there was a better way.

Try this;
setfreq m8

main:

pulsout 2 , 200
pause 2 ;adjust as needed
pulsout 3 , 200
pause 2 ;adjust as needed


goto main




Thanks

Jeff

[citfta]

Thanks Bit's,

I'll try that as soon as I get my batteries back in shape.

Carroll

[Sobakin]

Try this;
setfreq m8

main:

pulsout 2 , 200
pause 2 ;adjust as needed
pulsout 3 , 200
pause 2 ;adjust as needed


goto main




Thanks

Jeff

I was never program microcontrollers, but I feel that this variant will be correct too:
---------------
main:
pulsout 2, 200
goto _delay ; this delay is for symmetric waveform
_delay:
pulsout 3, 200
goto main
---------------
As we actually don't know how many tacts or microseconds use "goto" instruction, we simply add another same instruction.