Generating a square waveform 2Gbps
Hi list:
I have reading this list some months, but always in "read-only mode".
This is my first post.
My work is a bit strange: I make endurance test on electronic
components, comparing characteristics
of the samples before and after the tests.
My work now is with a module that needs differential input signal
(LVDS) at 2Gbps.
Normally we use a commercial pulse pattern generator with differential
capabilities to test only one module each time,
but now we need to generate 10 modules simultaneously.
I have searching integrated circuits working as oscillators able to
generate signals up to Gbps, but
I can only find commercial circuits up to Mbps.
Could you give some good indication, please?
Best regards
AlBundy
Hi,
As someone who considers anything above 100MHz as Black Magic, I'm
afraid I can't really help you.
Most Volt-Nuts are bound to be more DC oriented anyway, so I would
advise you to post your query
on [1]the Time-Nuts list. Lots of folks with multi-GHz experience
there.
Do you actually require those ten signals to be individually
controllable/configurable, or would some sort
of distribution amplifier, driven from your generator, do?
Best regards,
Rob.
AlBundy schreef:
Hi list:
I have reading this list some months, but always in "read-only mode".
This is my first post.
My work is a bit strange: I make endurance test on electronic
components, comparing characteristics
of the samples before and after the tests.
My work now is with a module that needs differential input signal
(LVDS) at 2Gbps.
Normally we use a commercial pulse pattern generator with differential
capabilities to test only one module each time,
but now we need to generate 10 modules simultaneously.
I have searching integrated circuits working as oscillators able to
generate signals up to Gbps, but
I can only find commercial circuits up to Mbps.
Could you give some good indication, please?
Best regards
AlBundy
volt-nuts mailing list -- [2]volt-nuts@febo.com
To unsubscribe, go to [3]https://www.febo.com/cgi-bin/mailman/listinfo/volt-nuts
and follow the instructions there.
References
On Sun, Apr 18, 2010 at 5:35 PM, Rob Klein rob.klein@XXXXXXX wrote:
Hi,
As someone who considers anything above 100MHz as Black Magic, I'm
afraid I can't really help you.
Most Volt-Nuts are bound to be more DC oriented anyway, so I would
advise you to post your query
on [1]the Time-Nuts list. Lots of folks with multi-GHz experience
there.
Hi Rob,
First, soory the delay in the answer. Thank you very much for your help.
I didn't know Volt-Nuts was restricted to low-frequency topics.
I have read my own question, and I think it is not very well composed
for the Time-Nuts list level. I will try to recompose my question and
will post it in Time-Nuts.
Do you actually require those ten signals to be individually
controllable/configurable,
No, i don't have this requirement. I can use the same generator for the
ten devices (directly, or with 10 buffers if the generator is not capable
of deliver the signal).
My problem is I would prefer the cheapest solution, with integrated circuits;
I can not afford to buy (or to rent) ten commercial laboratory generators.
or would some sort
of distribution amplifier, driven from your generator, do?
It would be an interesting solution. The problem is that I can not find LVDS
integrated-circuit drivers in Gbps range.
Thanks
Best Regards
AlBundy
AlBundy schreef:
Hi Rob,
First, soory the delay in the answer. Thank you very much for your help.
No problem, I know thing can get busy sometimes :-)
I didn't know Volt-Nuts was restricted to low-frequency topics.
Whoa! I certainly didn't mean to imply that you question was in any way
off-limits. I merely pointed
out that when we talk "Volts" on this list, it's most commonly high
accuracy DC and that Time-Nuts
would therefore be more likely to get you answers (a few of which might
even be useful ;-) ), that's all.
My problem is I would prefer the cheapest solution, with integrated circuits;
I can not afford to buy (or to rent) ten commercial laboratory generators.
Mmh, somehow, I can well understand that ;-)
It would be an interesting solution. The problem is that I can not find LVDS
integrated-circuit drivers in Gbps range.
Well, black art or no, your question did pique my interest and I had a
look round. ON semi have 1:10 clock
distribution circuits that can go up to 3 GHz. They're PECL out, so
you'd probably need a lever translator,
like TI's [1]SN65LVDS20. It seems like there is stuff out there,
question is if you can get it "in here".
Good luck and best regards,
Rob.
References
On Mon, May 3, 2010 at 9:08 PM, Rob Klein rob.xxxx@xxxx wrote:
[...]
Well, black art or no, your question did pique my interest and I had a
look round. ON semi have 1:10 clock
distribution circuits that can go up to 3 GHz. They're PECL out, so
you'd probably need a lever translator,
like TI's [1]SN65LVDS20. It seems like there is stuff out there,
question is if you can get it "in here".
Good luck and best regards,
Rob.
Hi Rob:
I have searching these days and found one 1:10 LVDS repeater (Maxim MAX9150)
and others 1:N repeaters of other manufacturers, but the problem is
always the same: maximum data rate is 400Mbps, 800Mbps in other
manufacturers. I need Gbps.
LVDS standard was initially designed for Mbps rates, and all
commercial chips are designed in this range.
You idea of generating faster clocks with technologies different from
LVDS, and adding a translator is becoming the most logical solution.
Thank you for your help. I will post another mail in a pair of months
with our final solution.
--
alBundy