My new approach

[Login to see usernames]

The drain of the igbt is creating high voltage spikes up to 600 volts

It see the proximity of the snubber to the circuit may be causing the noise

So I need to shield it with copper foil for now…

You don’t do the same mistake…

I will try to change the ground point to check any improvement too


Also the lock in capacitor and the filter need to be changed when we change the dividers stage

That’s why a 3 way 4 channel switch one is for the dividers and one for the filter capacitor and one for the lock in capacitor

This is the correct manner

I don’t have many capacitors here so I didnt made it yet like it should … also would be too many wires to put

On the pcb I wil make it soldered on the board so will be easy to have all this

I I’ll make the capacitors easy to change too not smd

One think i suggest is to put a small speaker on you secondary output so you will get a feedback of what’s going

It’s boring sound but it keeps you working and anyone else’s far away hahahaha

After we make it work we can make a challenge to who can make it with protoboards hahahaha

[Login to see usernames]

I just burned an other igbt

That’s why I keep saying to my self is so important to have a safe circuit!

I was just testing the scanning circuit with some power on

Not much but it didn’t like it

I may need to add more tvs around specially in paralllel with it

This raise the capacitance in parallel too working as a snubber tio

Is really bad to lose so much money on components during tests


I pay like 8 dólars each

[Login to see usernames]

The worst part of burning a component is that  sometimes is hard to tell why it burned

Improvements I can try
Make the primary coil interleaving to improve the coupling such that the second primary capture better the residual energy and clamp it more effectively

Add some capacitor resistor snubber on the tvs too to reduce DV/dt

Add tvs in parallel with the igbt too

I already have 420kohm in parallel with the igbt for taking the voltage for labview on the 20kohm portion

Maybe I could reduce this to half resistance to try improve the static discharge

And if nothing works maybe adding more capacitance
Y

The load being matched also may be fundamental

The components were around 45 degrees Celsius when burn

Why you think it burned?

What would you do?

[Login to see usernames]

I found a cheap rail to rail opamp for the scanning circuit

Also got 2kg of wire to make the coils

It’s the lmv358. I was using the lm358 that I took from a arduino module and so it was not working nice

As there are two in the package I use one as voltage follower and the second as amplifier

Now the circuit work very nicely after changing the igbt now it sounds more clear on the speaker

I added a overkill irg7ph50k10d

It seem like the other igbt was faulty or that there arrangement of cables and repositioning of the ground kind of solved the noise issue.. I also added a varistor in parallel with the tvs and a capacitor in paraellel with the resistors

Seems all I’m doing for now

Maybe the noise was coming from the fact that the current feed back resistor was the point of connection of ground if the circuit

I did that because the tc4420 caused spikes on the current sensor

Anyway is still possible to see the wave form in oscilloscope but as there are this short spikes the measurement function don’t show the values correctly … maybe my oscilloscope is not very good too

The current waveform show the peak current that is much different than average or rms

Specially when using the duty cycle

1 amp peak may show as 100ma depending on duty cycle of gate

The capacitor that is charged from the other primary is getting up to 100v for 20v input so that Tesla the statuary of the pulse have 100v exponential decay on top of the square

[Login to see usernames]

After a long day of work many things going

I have had the luck that I created a business that runs while I’m doing other stuff this is priceless

Now I get the wire and it seem like is the 28awg I need to confirm because the diameter measured seems like .33mm not .286

Maybe is just a double coat

That will change a little the numeber of turns

Now is Time to get a coffee and play more …

I posted new videos shorts on YouTube go check it and comment and leave a thumbs up!

[Login to see usernames]

I installed the lmv358 works great now! I added a trim so now the scanning frequency can be changed, taking the diode off the waveform is converted from saw toth to triangle the nice thing about the way I did is you can change the scanning frequency and the waveform keeps constant

In Stan circuit he uses a integrator that is totally not needed and make the frequency fixed I guess

I adjusted the values of the capacitors on the pll now it has overlapping ranges so I can give it a bandwidth regulation with a trim

For example the offset make the frequency go from 10khz to 170khz and I can have the scanning circuit or the man button to adjust from 5 to 20% bandwidth or more fo

[Login to see usernames]

I did the first two tests

To resonate the coil in parallel and in series

It seemed there is a voltage signal mostly triangular


It resonates as predicted

[Login to see usernames]

 i am jelious 
It sounds like that you are having all the fun 
Lets see if it delivers what you expect, Fabio!
Please keep on going.
If you want to evaluate or talk...You know i am available, if you want.

cheers my friend!