High voltage analog signal splitter

[jimmykk]

Hi
I am looking for A circuit which could split my 400khz analog high voltage sine signal to 8 outputs i.e. the same input signal goes out of these 8 outputs simultaneously
and it should also work the other way which means when signals are received at output it could be seen at inputs too.

 

[betwixt]

400KHz at 20KV?
400KHz at 100KV?
8 outputs each need 100A?
...or is it 500A?

Please give us a clue. We can't help when you give such vague details.

Brian.

 

[jimmykk]

Hi Sorry for little info.
Voltage would be +/-200V and current would be max. 4A total
So each output would have much lesser current

 

[KlausST]

Hi,

8 relays?
an 8 position rotary switch?
a drawing?

Klaus

 

[jimmykk]

Would not it just connect one output at a time to the input. I want all outputs simultaneously connected to the input without any daisy chain/series connection.

 

[KlausST]

Hi,

Would not it just connect one output at a time to the input. I want all outputs simultaneously connected to the input without any daisy chain/series connection.

It all depends on your unclear (until now) requirements.
I still have no clue what your circuit should do... and I don´t want to ask again ... I guess you will have your reason why you leave us in the dark.

Klaus

 

[jimmykk]

Let me put more details:
I have 5 channels which give out analog signals at 400khz high voltage and there is 60 degrees phase difference between the outputs.
Suppose at Ch1 my analog signal has to be connected to various elements(8) of a piezoelectric phased array transducer. So i need to split that signal to 8 different elements of that transducer. So it is important that signal reaches to all these elements simultaneously and without any delay. I am looking for a kind of SP8T switch if it makes sense. Hope it helps. Assuming each element has 80 ohm impedance at 400khz.

 

[d123]

Hi,

That's 2.5us/period, 2,500 nanoseconds. With a 50/50 duty cycle like a square wave that gives 1.25us, 1250ns for (using BJT language) delay, rise, storage and fall times. No delay is physically impossible, much to our chagrin... You need to look for devices that meet a realistic calculation of what amount of delay the application can tolerate. All I can generically think of are buffer gates suitable for > 8 fanout fed from avoltage divider and feeding back into level shifter stage such as MOSFETs.

 

[albbg]

I'll try to summarize. You want to split an analog signal in 8 ways.

1. the input signal dynamics is 400Vpp (+/-200V)
2. the input signal frequency is 400kHz
3. the delay between any two outputs should be zero
4. the output load of each output will be 80 ohm at 400kHz.
5. the input maximum current is 4A (RMS ?)

Is it the signal a pure sinewave or it is a modulated signal ? In this case what about the bandwidth ?
Of course you cannot reach exactly zero delay. You shoud, instead, foreseen a delay compensation network. In my understanding the absolute input-output delay is not very important.
If the current has to be drawn from the input signal the maximum voltage at each output will be 80ohm*4A/8=40V (56V if the current is RMS) considering no losses (ideally). This means, unless you don't accept a quite low oputput voltage, you need an active circuit that draws the necessary current from a power supply.

 

[BradtheRad]

I am looking for a kind of SP8T switch if it makes sense.

If ordinary relays are not suitable there are solid-state relays. However 400kHz may create a challenge for them.

Also triacs might be able to do the job but you must apply a gating signal continuously for as long as you want a triac to conduct.

Your control circuitry must be able to tolerate high voltage.

A triac conducts AC in either direction although it makes a difference which terminal is referenced to your gating signal. You may even need two triacs for every channel, to allow versatility as to gating.

 

[betwixt]

If the source is current compliant, you are asking for all the outputs to be in parallel so why not just wire them across each other. If you need to switch them all on or off simultaneously so they can be fed from a different source, I would suggest using MOSFETS as low side switches. You seem to imply that the output would be the same voltage but current shared between outputs (post#3), the easy way to do that is to use a resitance in series with each output to balance them but you would need to start with a higher voltage.

You mention switching between waveforms with different phases, wouldn't that be more easily achieved by switching it at low level before the driver stage?

Brian.

 

[Easy peasy]

There are coax relays for just this - they usually switch medium power RF - so 8 relays required - should suit your app - however if the source needs to see a load ( i.e. it will blow up if completely unloaded ) then you need to time a bit of overlap on the relay drives ...

 

[FvM]

According to previous posts, the OP wants to connect 8 piezo transducers with 80 ohms impedance each to a single 200 V/4A driver. He's talking about simultaneous drive, but also mentioning a phased array which would involve different magnitudes and phases driven to each piezo. First contradiction to be solved.

Secondly, input impedance of paralleled tranducers is 10 ohms in contrast to 200V/4A = 50 ohms driver output. We would expect a transformer or impedance transforming matching network.

It would be helpful to know, how the receiver is connected to the piezo array. Do we have pulsed operation with TX/RX switching?