Flexible square wave generator

Hi everyone,

I would like to hear some opinions and ideas of to face the design of the following square wave generator:

It must be capable of:

1. Having a rise time in the order of 25ns
2. the higher value must be able to go from 0 up to 42v
3. the lower value must be able to go from 0 up to -5v
4. The higher and lower values must be programmable.

The hardest point that i found is to fulfill point number 1)... I'm also afraid that it will be no easy to fulfill the range... that's why I wanted to hear about some with more experience in this field.

Any comment or ideas will be very welcome!!

I don't think that it's hard to achieve, depending of course on a complete specification (particularly load impedance), available parts - and your analog design skills.

25 ns rise time corresponds to about 15 MHz linear bandwidth.

The problem is that with a video amp you can get your 25nS and 47V, but if you want use it anywhere then you will have to connect it by a bit of cable, which is usually 50 ohms impedance. So now you are talking about a wideband power amplifier. Indeed if you want 40+V at the far end of the cable then you have to generate 80 + volts behind a 50 ohm impedance. it can be done, high definition CRT monitors use these sort of amplifiers using VHF power transistors.
Frank

Hi chuckey and FvM

Many thanks for the prompt answer !

FvM said:

I don't think that it's hard to achieve, depending of course on a complete specification (particularly load impedance), available parts - and your analog design skills.

25 ns rise time corresponds to about 15 MHz linear bandwidth.

Click to expand...

Loads impedance are always an non fixed RC load. The max freq to be used is around 20khz.
I'm able to generate the values for the high and low level, the problem comes when I try to build the signal by combining these two values. I do that with a mosfet configuration, but the rising/falling time is always around 100ns. Maybe there are better ways to solve it... could you guys point me out some documents or datasheets so I know in which other direction can I focus my development?



Thanks in advance!

Of industrial equipment comes closest to you
Trek Model 2100HF High Frequency High-Voltage Power Amplifier


Output Voltage: 0 to ±150 V DC or peak AC
Output Current:0 to ±300 mA
Slew Rate:Greater than 2 kV/µs
Large Signal BandwidthC to greater than 2.6 MHz (-3dB)
Gain:50 V/V Fixed

I would not call this a simple design.

Loads impedance are always an non fixed RC load.

Click to expand...

Put in "non-fixed RC load" and get unknown rise time in return.

There must be minmum R/ maximum C specification to start with a design.

Of industrial equipment comes closest to you

Click to expand...

The present specification doesn't decide yet between a linear HV amp output stage or a variable voltage supply with fast switches. Both are thinkable solutions - in principle. If you add other requirements like exact reproducable or possibly adjustable rise/fall times, a linear amp approach might become obvious. A "simple" switch approach can achieve 25 ns rise time, depending on the load capacitance.

Hi FvM,

I'm sorry that I wasn't clear from the beginning. The rise/fall time is required to be in the range of 25ns without load.
When I connect an RC load, I should be capable of measuring the RC constant by "observing" the new rise/fall time. This generator is made to characterize the RC value of the load connected to the generator. I hope it helps to clarify the requirements..

"RC load" means parallel or series circuit? What's the R and C range, what the intended generator output impedance?

FvM said:

"RC load" means parallel or series circuit? What's the R and C range, what the intended generator output impedance?

Click to expand...

Hi FvM,

The RC load will be a serial circuit and it is going to be connected directly to the output of the generator. I assume that the output should have a high impedance if I want to be capable of measuring the RC constant.
RC charge will adopt different values, in the range of 500 ohms up to 100 K and the C loads will always be smaller than 1 uC.