Agentgates´s TPU setup with strange wavehill hump

[broli]

For the first time in history this is getting scary intresting 
There has always been a dead time in between pulses comming from 555 and or 4017 circuitry because there is no overlapping.
This shuts the field down between each step and this methode can never work.

If you guys suceed in proper overlapping, thus switching coils in time before the previous ones are switched off so the field never drops out, things can become real.
Nobody tried that in all these years, and i have always been thinking this was most important in order for the fields to interact properly.

NT.

It's pretty easy to have overlapping signals. Instead of using the delay function and letting the whole microchip do nothing for a while, you check if a certain time limit has passed if not you ignore the code block and move on to the next. Here's yet another example:

http://arduino.cc/en/Tutorial/BlinkWithoutDelay

An arduino is really a must next to an oscilloscope and a singal generator in this field, it allows for fast prototyping and feedback without too much headaches and it's cheap. I know tony wanted to keep things simple, hence the simplified decade timer, but an arduino offers much more control and is very simple to adjust on the fly thus ideal for finding the right parameters.

[Bruce_TPU]

Hello ALL,

A fellow member of this forum, JDO300, asked me to post this for him.  He is busy working on his new sine wave generator. 

"Hi Bruce,

Here's the schematic I mentioned to you. Rather than using the 4017, this circuit uses a 74HC164 Shift register. That in combination with a NOR Gate produces the same effect as the 4017 (and the circuit could easily be modified to use the 4017 instead). But the useful thing here is that the signal source's pulse width is used to control the output pulse widths of the three channels simultaneously through the use of the three AND gates.

I didn't bother to redraw the circuit with the 4017 since I already had it in my stash so I thought I would pass it along. The signal source can be a 555 timer or any other suitable clock source." 

Any questions can be directed towards Jason.  Thank you.

[Mannix]

I have as best I can attempted to get the results with the first coil setup.

I have varied the pulse width frequency, drive and powersupply. 0 -32v
I notice a solid resonance around 1.5 megahertz
I notice the coil gets quite hot as does the massive heatsink on my three igbt's
I do not see a dc level rise on the out put coil at any time

The heat being dissipated although not measured does not seem excessive for the wattage supplied

I have put various voltage globes on the output coil which is 5 turns and the 12 volt tail lamp gives the best response.

My pn devices have not blown up yet

There are Many things that I may not have  exactly correct so I wait for more specific  info on this first test please .
I have 3 phase controller etc etc etc  ready to go .

But need the start point

@Tony or any body who has run this up
could you of somebody else please help me and others get to the first described results    first..


Thanks

still mostly optomistic,

lindsay

[agentgates]

Here's some info on port registers for the Arduino. Being a software dev I'm sure you will have no problems getting the Arduino to do what you like. This is just for future reference should you need the functionality;

Thanks, that looks handy.

Have you seen this example:

http://arduino.cc/en/Tutorial/Fading

You could use a single analog output to keep it on high and fade it out afterwards.

I am not looking to fade for now. Those were digital ports.

It's pretty easy to have overlapping signals. Instead of using the delay function and letting the whole microchip do nothing for a while, you check if a certain time limit has passed if not you ignore the code block and move on to the next. Here's yet another example:

http://arduino.cc/en/Tutorial/BlinkWithoutDelay

An arduino is really a must next to an oscilloscope and a singal generator in this field, it allows for fast prototyping and feedback without too much headaches and it's cheap. I know tony wanted to keep things simple, hence the simplified decade timer, but an arduino offers much more control and is very simple to adjust on the fly thus ideal for finding the right parameters.

Thanks. Yes I agree. Actually Arduino will also keep it simple.


@Nikola Tesla

Yes I did try these things a year ago and worked but used flyback diodes and thicker for primary thus blown my PSU, later PG and other instruments so I gave up at that time.

@Lindsay

1. Set the PG to 1kHz/50% and leave it as a constant reference signal, that should always work. And BTW it is not sensitive for frequency like a radio device. Think of the importance of frequency like at a switching PSU. Nothing more.

2. Make sure the IGBTs turn on properly.

3. Check the primary if it's not too thick. Mine is 0.38mm.

4. Check the left/right handed winding thing on both primaries and secondary. Perhaps it also counts. Take it as a reference:
http://www.samsonium.org/sites/default/files/P1060253.JPG
(That is equivalent with 1 primary)

5. Make sure the signal is square wave, not sine or "flatter". (it works correctly with poor squarewaves as well)

6. Remove the bulb and let it to run without load for now.

7. Make sure the gaps are correct between the turns. If they are too close you really need the juice to see something on the scope.

8. If you still have problems please don't hesitate to send photos or we can chat on Skype if you'd need further help.

UPDATE

9. Make sure the secondary is tight enough. No need to force it just make sure it's not loose. That gives much worse output.

[ramset]

Tony
Something is different in this thread!
Feels very good.

Thanks Tony
and @all

Chet