How to make a super fast sampling head?

[Jan Panteltje]

J Larkin was saying his super sampling head costs 750$.
I was thinking how can you make one yourself, what else is needed?

 

[Joerg]

J Larkin was saying his super sampling head costs 750$.
I was thinking how can you make one yourself, what else is needed?

Usually a transistor that avalanches nicely, a chunk of good coax and
blazingly fast diodes, preferable a matched quad.

 

[Jan Panteltje]

Usually a transistor that avalanches nicely, a chunk of good coax and
blazingly fast diodes, preferable a matched quad.

Any suggestion for diodes?

 

[Joerg]

Any suggestion for diodes?

I have a stash from Microwave Associates which has become M/A COM. You
can find the really hot stuff here but you might want to sit down when
the quote comes in:

http://www.macom.com/psc/jsp/ListParts.jsp?dataFile=mixing_detector_diodes.txt

 

[John Larkin]

Usually a transistor that avalanches nicely, a chunk of good coax and
blazingly fast diodes, preferable a matched quad.

I did one with a step-recovery diode impulse generator, feeding a
2-diode sampler. Got 70 ps risetime, roughly 5 GHz, and apparently
good waveform fidelity. The parts cost is tiny, but it's labor
intensive. The timebase and trigger stuff is more work than the
sampler.

The older Tek sampling stuff, the S-series heads and their 7000-series
plugins, are dirt cheap on ebay. But not nearly as nice and
quantitative as the 11801-series stuff.

I doubt you could make a head that would work in an 11801 frame; the
interface is undefined.

It would be fun to do a sampling scope/TDR as a USB dongle.

John

 

[Phil Hobbs]

I did one with a step-recovery diode impulse generator, feeding a
2-diode sampler. Got 70 ps risetime, roughly 5 GHz, and apparently
good waveform fidelity. The parts cost is tiny, but it's labor
intensive. The timebase and trigger stuff is more work than the
sampler.

The older Tek sampling stuff, the S-series heads and their 7000-series
plugins, are dirt cheap on ebay. But not nearly as nice and
quantitative as the 11801-series stuff.

I doubt you could make a head that would work in an 11801 frame; the
interface is undefined.

It would be fun to do a sampling scope/TDR as a USB dongle.

John

I just got John Mulvey's book, "Sampling Oscilloscope Circuits", that
goes through the operation of the 1969-vintage Tek samplers in truly
gory detail. If you can find a copy, that would be a great place to
start. For instance, they used a tunnel diode for the triggered
sweep--one diode controlled both the trigger and the reset. It also
covers the importance of getting the sampling loop right--you feed back
100.0% of the previous sample to the sampling head. This helps with
linearity and (interestingly) speed variations.

Cheers,

Phil Hobbs

 

[Jan Panteltje]

I did one with a step-recovery diode impulse generator, feeding a
2-diode sampler. Got 70 ps risetime, roughly 5 GHz, and apparently
good waveform fidelity. The parts cost is tiny, but it's labor
intensive. The timebase and trigger stuff is more work than the
sampler.

The older Tek sampling stuff, the S-series heads and their 7000-series
plugins, are dirt cheap on ebay. But not nearly as nice and
quantitative as the 11801-series stuff.

I doubt you could make a head that would work in an 11801 frame; the
interface is undefined.

It would be fun to do a sampling scope/TDR as a USB dongle.

John

I had some idea (not had much sleep, and then barriers do not
function on my ideas) to make a small PIC scope, say one PIC,
one graphics display (found one dirt cheap), and because PIC will
not digitise faster then a few kHz perhaps, add a sampling head just for fun.
I think the main problem is in the timebase, to do it digitally, and have
it set so it actually produces a steady picture on a repetitive signal,
could be very difficult.
And there a is sample memory size issue too...
It is just an idea, I have several other wish projects, but seems fun,
and would give me some more experience with high frequencies (I top out
at about a GHz atm).
Do I see it right that you have to adjust sample frequency gradually
to get a good 'view', is this perhaps done with some PLL?

 

[Jan Panteltje]

I have a stash from Microwave Associates which has become M/A COM. You
can find the really hot stuff here but you might want to sit down when
the quote comes in:

http://www.macom.com/psc/jsp/ListParts.jsp?dataFile=mixing_detector_diodes.txt

mm 80Ghz! wow.

 

[Jan Panteltje]

I just got John Mulvey's book, "Sampling Oscilloscope Circuits", that
goes through the operation of the 1969-vintage Tek samplers in truly
gory detail. If you can find a copy, that would be a great place to
start. For instance, they used a tunnel diode for the triggered
sweep--one diode controlled both the trigger and the reset. It also
covers the importance of getting the sampling loop right--you feed back
100.0% of the previous sample to the sampling head. This helps with
linearity and (interestingly) speed variations.

Cheers,

Phil Hobbs

Thank you Phil.

 

[Joerg]

mm 80Ghz! wow.

Good stuff. But check you bank account before ordering

 

[Joerg]

I did one with a step-recovery diode impulse generator, feeding a
2-diode sampler. Got 70 ps risetime, roughly 5 GHz, and apparently
good waveform fidelity. The parts cost is tiny, but it's labor
intensive. The timebase and trigger stuff is more work than the
sampler.

The older Tek sampling stuff, the S-series heads and their 7000-series
plugins, are dirt cheap on ebay. But not nearly as nice and
quantitative as the 11801-series stuff.

I doubt you could make a head that would work in an 11801 frame; the
interface is undefined.

A lot in this field is a trade secret or barely defined. And then old
Leroy goes into a nursing home and one day takes that knowledge with him
into the grave.

Same on my main turf (medical ultrasound). Lots of stuff is never
disclosed in detail other than what's required for biocomp testing. To
avoid potential risks some isn't even written down. So if someone would
knock me over the head and walk away with the whole office contents it
wouldn't do them any good.

It would be fun to do a sampling scope/TDR as a USB dongle.

Sure would be fun. But I assume the market size would be rather paltry.
Especially in view of the ever smaller number of young lads who would
know what to actually do with such gear.

 

[Joerg]

I had some idea (not had much sleep, and then barriers do not
function on my ideas) to make a small PIC scope, say one PIC,
one graphics display (found one dirt cheap), and because PIC will
not digitise faster then a few kHz perhaps, add a sampling head just for fun.
I think the main problem is in the timebase, to do it digitally, and have
it set so it actually produces a steady picture on a repetitive signal,
could be very difficult.
And there a is sample memory size issue too...
It is just an idea, I have several other wish projects, but seems fun,
and would give me some more experience with high frequencies (I top out
at about a GHz atm).
Do I see it right that you have to adjust sample frequency gradually
to get a good 'view', is this perhaps done with some PLL?

Many ways to do it. Hint: You can build delay lines where the delay can
be set in almost infinitely fine steps. And yes, from a processor if you
want to and you can also servo that so you don't have to spend hours
aligning and calibrating. 'nuff said ;-)

 

[John Larkin]

I just got John Mulvey's book, "Sampling Oscilloscope Circuits", that
goes through the operation of the 1969-vintage Tek samplers in truly
gory detail. If you can find a copy, that would be a great place to
start.

Got it, and most of the other Tek Concept Series books. Publishing
stuff like that would be unheard of nowadays.

For instance, they used a tunnel diode for the triggered
sweep--one diode controlled both the trigger and the reset. It also
covers the importance of getting the sampling loop right--you feed back
100.0% of the previous sample to the sampling head. This helps with
linearity and (interestingly) speed variations.

I loved tunnel diodes; too bad they're gone. The fabrication process
was insane. I bet an amateur could make them.

Cheers,

Phil Hobbs

Do you have Mark Kahr's paper on the history of samplers? The
technology dates back to the 1800's.

http://www.caip.rutgers.edu/~kahrs/books/sampling.html

John

 

[John Larkin]

A lot in this field is a trade secret or barely defined. And then old
Leroy goes into a nursing home and one day takes that knowledge with him
into the grave.

Same on my main turf (medical ultrasound). Lots of stuff is never
disclosed in detail other than what's required for biocomp testing. To
avoid potential risks some isn't even written down. So if someone would
knock me over the head and walk away with the whole office contents it
wouldn't do them any good.



Sure would be fun. But I assume the market size would be rather paltry.
Especially in view of the ever smaller number of young lads who would
know what to actually do with such gear.

If it did TDR, you could sell bundles of them to PCB houses, and to
engineers and QC people who care about trace impedances. Especially if
it were cheap and interfaced to a PC, so it could document pcb test
coupons.

I'd design the fast stuff if somebody else handles the USB part and
the PC software. I have a slick deconvolution algorithm that would run
on the pc side and really beautify the TDR response.

John

 

[John Larkin]

Good stuff. But check you bank account before ordering

Skyworks SMS7621-079, 0.25 pF schottky. I have a reel of them, 23
cents each.

John

 

[Joerg]

Skyworks SMS7621-079, 0.25 pF schottky. I have a reel of them, 23
cents each.

Neat! That's a great deal.

Jan, in Europe you can get this brand via BFI Optilas. In case you want
to look these up:
http://www.skyworksinc.com/products_display_item.asp?did=798

 

[Joerg]

If it did TDR, you could sell bundles of them to PCB houses, and to
engineers and QC people who care about trace impedances. Especially if
it were cheap and interfaced to a PC, so it could document pcb test
coupons.

Really? We did lots of RF things in the companies I worked for but never
used much TDR. Ok, it wasn't GHz gear.

I'd design the fast stuff if somebody else handles the USB part and
the PC software. I have a slick deconvolution algorithm that would run
on the pc side and really beautify the TDR response.

Hey, you always want to do the fun stuff ... ;-)

 

[Joerg]

J Larkin was saying his super sampling head costs 750$.
I was thinking how can you make one yourself, what else is needed?

Another part you might want to look at are those blazingly fast
transistors from your companies over there, Infineon and NXP. For
example the BFP620. Ok, it's a bipolar transistor but it sports a
whopping 65GHz ft and is very cheap. Just don't push it much past 2V.

 

[Phil Hobbs]

Do you have Mark Kahr's paper on the history of samplers? The
technology dates back to the 1800's.

http://www.caip.rutgers.edu/~kahrs/books/sampling.html

John

No, I didn't, but I'll download it, thanks. The 19th century example is
the dynamo waveform plotter, I expect. Nice piece of work.

Looking down the list of references, I know a couple of those guys: Mark
Rodwell and I were grad students together--he's now at UCSB trying to
make 1-THz InP bipolars; Dave Bloom was on my PhD committee, and has
retired to Montana after making a pile in the telecom boom.

One missing person is Sadeg Faris, the guy who started Hypres--they had
a SQUID-based sampling scope with (iirc) 70 GHz bandwidth back in the
late 80s. It was a cool gizmo--the input card was about 2 inches long,
and went from room temperature at one end to liquid helium at the other.
Another pair are Dan Grishkowsky and Jean-Marc Halbout, who pioneered
picosecond plasma-optical sampling back in the mid-80s.

Cheers,

Phil Hobbs

 

[[email protected]]

I had some idea (not had much sleep, and then barriers do not
function on my ideas) to make a small PIC scope, say one PIC,
one graphics display (found one dirt cheap), and because PIC will
not digitise faster then a few kHz perhaps, add a sampling head just for fun.
I think the main problem is in the timebase, to do it digitally, and have
it set so it actually produces a steady picture on a repetitive signal,
could be very difficult.

I've done it twice - once for real, with an 800MHz clock and GaAs
counters for the coarse intervals and a built in time to voltage
converter to work out when the trigger arrived within the the 1.25nsec
clock period. My 800MHz clock was poor and we had about 60psec of
jitter on the sampling edges.

A few years later I redesigned the delay-generating part of the
circuit around a 500MHz clock (which would have been crystal
controlled with less than a picosecond of jitter) and ECLinPS counters
for the digital part of the delay, and I was planning on using the
MC100E195 for the fine delays - it offered 2nsec of delay range and
20psec resolution. The MC100EP195 looks even nicer.

The delays through the MC100E195 are temperature dependent, and the
design did depend on recalibrating these delays against the crystal
controlled clock every few minutes, by using the system to set up a
pulse-width modulated waveform and digitising the DC level to find
what each delay really was. We should have been able to run through
all 128 discrete delays within a millisecond or so.

The schematics had been passed on for printed circuit layout when the
customer backed out.

The MC100E196 offers infinite resolution, but would need even more
calibration.