Transformer signal help

[vonsproken]

Hi, I'm building a circuit that must send a 15V signal to IGBT but must be isolated. My first part of the circuit has a 300V difference between the second part and therefore cannot share the same ground. This signal must not have more than a 300ns delay to the output. I thought of using a transformer - any other ideas will be welcome.

I built an oscillating circuit - this is my signal source.
I then power a TC4428A chip to give an AC signal to drive a transformer, this chip has an inverting and non inverting output.
The transformer has 20 turns primary to 10turns secondary wound on a ferrite core.
I manage to get a perfect squarewave out of the transformer.

This is my problem - When I connect it to a comparator I get an output but when I connect this output to anything else I get a huge amount of noise coming from the comparator - I have tried 393 and 311's and they do the same. I'm not sure what part is wrong. The output needs to be able to drive a TC4422A mosfet driver chip.

I can provide pictures if neccessary, I cannot have any more that a 300ns delay between the signal input and the output. I do not want to use high side driver chips as the voltage difference in my circuit may become 700V in the future - I'm not aware of any cheap high side driver chips that can do this, please notify me if there are any.

 

[ericgibbs]

hi,
Have you considered an opto coupler, for example this type.

 

[vonsproken]

I first went for optocouplers but was struggling to find cheap ones without buying in bulk + I bought a load of stuff for transformers a while ago but didnt use it so I had a go with it.

That optocoupler is what Im looking for although I really could do with it to run at 15V due to all of my circuits running at this + I dont want to have to modify my power supply, can you think of any other ones - I could resort to these voltages in the end but dont really want to.

thanks

 

[ericgibbs]

hi,
The problem is the low rise/fall times of most opto's in the order of microsecs,, I see you are looking for 400nanoSec times.

EDIT:
Look at some options on this link.
https://www.google.co.uk/search?q=f...t&rls=com.yahoo:en-GB:official&client=firefox

 

[vonsproken]

That was the biggest problem, I remember them having 1 - 2us rise/fall times which is far too slow, when I found the fast ones I couldnt get them and the voltages were a little low. What I'm doing is driving an IGBT, I have seen many people drive them directly with a transformer but I find voltage spikes quickly kill them. So I thought of using the transformer to drive a comparator to then drive a IGBT driver chip which drives the IGBT. Im having a lot of trouble with the signal from the transformer driving anything. I think my problem may be that I have too many turns on the transformer and at these speeds 50-100kHz the impedance is too high and therefore not enough current coming form the secondary. What do you think?

 

[ericgibbs]

That was the biggest problem, I remember them having 1 - 2us rise/fall times which is far too slow, when I found the fast ones I couldnt get them and the voltages were a little low. What I'm doing is driving an IGBT, I have seen many people drive them directly with a transformer but I find voltage spikes quickly kill them. So I thought of using the transformer to drive a comparator to then drive a IGBT driver chip which drives the IGBT. Im having a lot of trouble with the signal from the transformer driving anything. I think my problem may be that I have too many turns on the transformer and at these speeds 50-100kHz the impedance is too high and therefore not enough current coming form the secondary. What do you think?

hi,
Would you post details of the home wound coupling transformer.?

 

[vonsproken]

Here are the cores that I am using FERROXCUBE|TN25/15/10-3E25|FERRITE CORE, TN25-15/10-3E25 | Farnell United Kingdom

a 15:15:15
and a 20:10:10

 

[ericgibbs]

Here are the cores that I am using FERROXCUBE|TN25/15/10-3E25|FERRITE CORE, TN25-15/10-3E25 | Farnell United Kingdom

a 15:15:15
and a 20:10:10

hi,
Its possible that the ferrite has a frequency limit less that the required 2mHz+ rise time.?
Look at this pdf.

 

[vonsproken]

I think it could be possible but the grade of the ferrites that I have are 3E25 and unfortunately in the table there are no values for the operating frequency at this grade. I do get an output on the transformer and the wave looks exactly like the input so I'm guessing its within the operating frequency, I will do a few tests later to see how much current I'm getting from it. I'm just wondering if the comparators arent liking their input voltages because they are a little on the high side - about 18V. I will step the transformer down further to see what I get.

 

[mneary]

Google search for '3e25 ferrite' and you'll find a lot more frequency data. The datasheet on rcgroups.com seems pretty comprehensive.