Zero Voltage Transition Problem

[kaosad]

Hi,

I am new to ZVS. I am wondering what circuit is able to sense zero voltage crossing. In one of my application a voltage can spike up to 400V.

Any suggestion.

[cerrem]

A comparator.....You need to set the reference on the comparator for higher than zero to make up for the latency... It can also be a comparito modified for one-shot output with a quick reset upon a rising edge....

Chris

[janneman]

First decide what kind of ZVS signal you need. A pulse at exactly ZV? Do you want a square wave that switches at ZV? In de last case, an opamp running without feedback is all you need.

Jan Didden

[kaosad]

A pulse for each ZVC, because I need it to be repeatable. Can an opamp able to handle high voltage and work in the range 100-200kHz?

[cerrem]

A few things to keep in mind....
You don't want to directly sense the high voltage....Your sense should be a fraction of the high voltage such as using a divider...
Using an open-loop Op-Amp for comparator operation will be unstable and ring.....

Chris

[janneman]

Well, of course you would use a bit of pos fb to get some hysteresis, but I thought that would be obvious...

So, an open loop opamps to make a square-wave with transitions at the ZC, followed by a 555-type one-shot to make it a pulse?

Jan Didden

[kaosad]

Yes, I can use voltage divider, but I have like a spike of 400V and then an oscillation of say 5V peak-to-peak which crosses zero point. Would the opamp be sentitive enough to detect that if I scale down the voltage too much?

[kaosad]

Hi again,

I figured out an idea, but still facing some issue. I use 12V zener diode to clamp the voltage. However, I must not affect the original signal in a direct way by leaking too much current. To fix this I put a large resistor in series with the zener diode. So, when the voltage goes pass 12V, the zener breaks down but would still leave the original signal intact, thanks to the resistor. This seems to work. But too large resistance distort the voltage across the diode (seen when the original voltage oscillates within 0-12V range) perhaps due to parasitic capacitance of the diode. A resistor greater than 10K Ohm distorts the voltage across the diode. I wanted to use 1 Meg Ohm or greater.

Anyone can recommend me a low parasitic capacitance 12V zener diode?

[janneman]

Originally posted by kaosad
Hi again,

I figured out an idea, but still facing some issue. I use 12V zener diode to clamp the voltage. However, I must not affect the original signal in a direct way by leaking too much current. To fix this I put a large resistor in series with the zener diode. So, when the voltage goes pass 12V, the zener breaks down but would still leave the original signal intact, thanks to the resistor. This seems to work. But too large resistance distort the voltage across the diode (seen when the original voltage oscillates within 0-12V range) perhaps due to parasitic capacitance of the diode. A resistor greater than 10K Ohm distorts the voltage across the diode. I wanted to use 1 Meg Ohm or greater.

Anyone can recommend me a low parasitic capacitance 12V zener diode?

You can use a normal small diode in conduction, then you can use a large series R. The voltage across the diode then looks as a about 0.6V DC with valleys at the zero crossings. Now put that into an open loop opamp + input (-input to gnd) and you have a beautifull zero xing square wave without loading the original signal too much.

You may need to put, say, 10mV DC at the -input from a supply divider to make sure it switches cleanly.

Add a 555 as one-shot after the opamp and you have your zero xing pulse. There are good apps in all 555 data sheets.

Edit: depending on the exact signal, you may not need the 555, the opamp itself may have a clean pulse. Try it.

Jan Didden

[kaosad]

Thanks janneman. That is a great idea!

I used silicon diode (in series with 1 Meg Ohm resistor) and I got smaller phase shift and good response time, but the the voltage won't go down below 200mV when the voltage dropped, not sure why. When I changed to Schottky diode, the phase shift was so much smaller and it dropped to slight negative. That is great!