10V 100mA fast edge pulse amplifier...what are the options ?

[chipeater]

Hello,

By coincidence, I came across the documentation of this AOP : AD8009.
Among the applications, it talks about pulse amplifier.

It made me remember a subject which consisted in synchronizing old clocks by 10ns, 10V pulses (10V out from the chip but 5V into 5 Ohm load ) with short 1ns rise time (fall time was less important).
Pulses are generated by a FPGA or a microcontroller and it should also be the shortest possible propagation delay.

Except with this kind of AOP, with which type of schematic can we realize this kind of pulse amplifier ?

I would have thought of using a fast comparator but I can't find any reference able to be powered in 10V and able to output a current >= 100mA
And it is almost the same story with logic gate, I can't find any references with both "high voltage" and fast enough rise time.

Do you have any other solution to propose or to assemble like logic gate or comparator + transistor ?
If you could put me on the track it would be great
Needless to say, I'm not very comfortable with electronics yet so please bear with me

Thanks in advance for your help

[Benta]

It made me remember a subject which consisted in synchronizing old clocks by 10ns, 10V pulses (10V out from the chip but 5V into 5 Ohm load ).

Could you elaborate here, please? It's not really understandable.

[jonpaul]

who or what is AOP?

Read the excellent Tektronix Circuits Concept book Oscilloscope Vertical amplifiers
https://www.w140.com/tekwiki/index.php/Concepts_Series

Jon

[TimFox]

Just a note about technical word choice:
Ordinarily, an "amplifier" is a linear (or quasi-linear) device where the output magnitude depends on the input in some way.
You want a "buffer" or "driver", which is probably non-linear, where the input and output are binary levels.

[Benta]

who or what is AOP?

Prolly "Amplificateur OPerationelle" or something like that. In French, it's necessary to reverse things, eg, NATO/OTAN.

[SiliconWizard]

This kind of opamps would really be your best bet if your goal is to design something practical in a reasonable amount of time, it's not even that expensive.

But if you want to design something more discrete just for kicks, you have a number of options.
One relatively "common" approach is to use transistors in their avalanche region.

Some notes:
https://arxiv.org/pdf/1712.04819
https://www.analog.com/media/en/technical-documentation/application-notes/an94f.pdf

[MathWizard]

So if Intel's latest CPU's are doing around 5.6GHz without too much trouble, that's ~180ps. So what can that be used for, by the time it would get to the outside world ?

Or if you work at Intel, what would be the best pulse or clock signal they could get to the outside world ?

[jbb]

I think they’re going towards something like 12.5 GHz with PAM4 (Pulse Amplitude Modulation, 4 level). So that’s like 100 Gbits/sec over a single differential pair. (Less some overhead for encoding; I think encoding 64 real bits into 66 bits on the line is in the works)