signal generator with very low rise and fall times

[Snorri]

Hi Forum,

I'm a physicist at an Austrian University, and would greatly appreciate your help with a circuit I planed (shown in "Fig. 1 - circuit") - especially with one component where I'm not sure how to integrate it.

What I need is a signal generator that supplies a relatively high capacity (~200 V at ~1 A), while having as low rise and fall times as possible. Therefore I want to use an IXYS IXZ4DF12N100 (https://www.electro-tech-online.com/custompdfs/2010/04/ixz4df12n100.pdf) as a driver and MOSFET combination.

Now my questions:
1. Will the output of the ADUM correctly switch my IXYS?
2. How important is the coil prior to IN-VCC? (data-sheet --> Fig. 15)
3. How important are all the electrolytic capacitors? (data-sheet --> Fig. 15)
4. May I leave IN-GND unconnected, or do I really have to connect it with ground over a choke coil? It is internally connected with DGND anyway.

Is there anything else I could improve?

Kind regards,
Snorri

 

[MikeMl]

At what frequency are you driving your load?

At what Amplitude?

 

[Warpspeed]

Hello Snorri, I am located in Melbourne.

We need to design this from the load backwards.

First question, what is the nature of the load ?
Is it predominantly resistive, capacitive or inductive ?

Any power driver stage needs to be designed to suit the required characteristics of the load, so that is where we need to begin.

 

[Snorri]

First off many thanks for your fast replies!

You both tell me to start my considerations at the load. So, my load is an oscillating circuit. I have an electrolytic capacitor for performing measurements on different electrolytes, and a variable coil to match the impedance.
What we want is maximum voltage over the capacitor at minimum current – what means that we will have to adjust the inductivity to the frequency.
The frequency should sweep from some hundred – or at least some thousand - Hz to about 1 MHz. (This is vital as the goal is to observe different electrolyte behavior at different stimulation frequencies.)
Now we can vary the tuning coil - according to the frequency change - only manually in steps; and between these steps, I guess, my load is changing from capacitive to inductive; am I right?

The amplitude on the load-side is about 200 V; maybe more, when we get this working.


Looking forward to your kind help!
Snorri

 

[Warpspeed]

First observation is that the load will be approximately resonant, but fairly lossy, and both the resistive and reactive components may vary widely depending on the nature of the particular electrolyte under test.

Rapid rise and fall times will be impossible to achieve in a resonant circuit. Resonant circulating energy always tries to revert to a sine wave. Driving a resonant circuit with a square wave could cause the cell voltage to be either square, sine, or some unpredictable distorted version of both, depending on the quality of the resonance.

The required drive level power of 200 watts is fairly significant power that must be provided, and almost all of that is going to go into heating up your electrolyte.

In order to get maximum voltage across the electrolytic cell at resonance, the circuit will be easier to drive with a series resonant circuit. In other words, the tuning inductor needs to be placed in series with the cell and tuned for maximum cell voltage.

The power amplifier then will be working into a relatively low impedance at resonance, which should be about equal to the resistive losses in your cell.

At this stage, I might suggest you seriously consider driving your load with a sine wave instead of a square wave. You will then always be measuring true sine wave voltages and currents at the load and the measurement results will be far more accurate and repeatable.

Attempting to accurately measure highly distorted waveforms that change shape with tuning and frequency changes as well as resistive cell changes, will add a level of complexity and uncertainty to this experiment that may best be avoided.

Still, you are obviously doing something rather special here, and you may have some particular special needs and requirements that conflict with the above........

 

[shortbus=]

> snip<
Still, you are obviously doing something rather special here, and you may have some particular special needs and requirements that conflict with the above........

Hmm, is that HHO in the air?

 

[Snorri]

Thank you, Warpspeed, for your deep thoughts concerning my setup!

First observation is that the load will be approximately resonant, but fairly lossy, and both the resistive and reactive components may vary widely depending on the nature of the particular electrolyte under test.

Rapid rise and fall times will be impossible to achieve in a resonant circuit. Resonant circulating energy always tries to revert to a sine wave. Driving a resonant circuit with a square wave could cause the cell voltage to be either square, sine, or some unpredictable distorted version of both, depending on the quality of the resonance.

The required drive level power of 200 watts is fairly significant power that must be provided, and almost all of that is going to go into heating up your electrolyte.

Unfortunately for my project, you are right with every word you wright.

In order to get maximum voltage across the electrolytic cell at resonance, the circuit will be easier to drive with a series resonant circuit. In other words, the tuning inductor needs to be placed in series with the cell and tuned for maximum cell voltage.

That's exactly what my setup looks like – sorry for my inaccurate description of the load.

At this stage, I might suggest you seriously consider driving your load with a sine wave instead of a square wave.

So, if I'm going to follow your suggestion – what surely is a good idea - and use a sine wave instead of a square wave, and want to use my described setup to achieve high frequencies, how do I have to implement the IXYS driver/MOSFET-combi?


@ shortbus=:
With some used electrolytes, HHO might/will be produced; but that's a circumstance we already considered. Fortunately, that's not a problem for us, as we have a good flue in our laboratory.
But many thanks for your advise!

 

[Warpspeed]

So, if I'm going to follow your suggestion – what surely is a good idea - and use a sine wave instead of a square wave, and want to use my described setup to achieve high frequencies, how do I have to implement the IXYS driver/MOSFET-combi?

The IXYS driver is just not suitable for generating a sine wave output.
It switches a single MOSFET on and off very fast.

What you really need is a linear power amplifier.

You may be able to find an off the shelf commercial audio or ultrasonic power amplifier that has sufficient output voltage and upper frequency range,
It does not need to have an absolutely flat frequency response, because you will be adjusting the output amplitude anyway.

The voltage measured across your cell will be considerably higher than the output voltage at the amplifier when the cell is tuned to resonance.

It may be possible to hire an amplifier to try all this out, or perhaps arrange to get an amplifier for evaluation from a supplier prior to purchase.

A further possibility is to drive two power amplifiers 180 degrees out of phase and bridge the load between the two outputs. This doubles the effective output voltage.

 

[Snorri]

Dear Warpspeed,

let me tell you that I never got that exhaustive answers as yours are in a forum till today! So please don't think that I don't appreciate your effort if I don't take up every suggestion you make – even if I'm sure that everyone of them is well elaborate.
Also, please forget my answer concerning the sine wave through the IXYS; that was really embarrassingly stupid of me.

So, for whatever reason, please let me stay at the square wave signal. We already work with this signal form – but until now with two separate components: a driver (TC4451V) and a MOSFET (IRFP260N). And it works out quite well, just not fast enough. That's why I want to use the IXYS. Furthermore, we got everything else working, and just need this component integrated to give it a try.

So, if you don't mind helping me anyway...