Maker Pro
Maker Pro

Narrow pulse generation

  • Thread starter Anthony Fremont
  • Start date
J

Joerg

Jan 1, 1970
0
5nsec pulses don't sound too hard - we got 0.5nsec FWHM at 5V into 50R
and 7.5V for anything longer.

We did use a fairly expensive HP RF transistor as the output stage.

On the way there I got 800psec FWHM out of a bunch of dead cheap BFR96
wide-band transistors.

Ye olde BFS17 can also swing around in that speed class. Lots of stock
everywhere and I always keep a stash ready.
 
A

Anthony Fremont

Jan 1, 1970
0
Joerg said:
Anthony Fremont wrote:

For example, you can build an LC or RC section that scoots your bridge
trigger time sliver by time sliver. You can use a varicap in the C or
make the R variable. The latter is what I usually do. Early on with
the SD5400 but when that got expensive I used dual gate FETs for TV
tuners. Those cost only pennies but have very low inherent
capacitances. Low stray capacitance is very important here. You can
easily scoot a few hundred psec. Personally I avoid digitally
adjustable delay lines for noise reasons.

<woosh like blue angels fly by> I'm pretty dense here, so it's not yet
becoming obvious to me how something like this works. I can't begin to
fathom how a quad analog switch will let you generate thousands of precisely
different time delays (different R values) to trigger the ADC.
That SD5400 was very handy. Since contains four matched devices I
could servo it an literally steer resistance. Very smooth. But all
good things will some day end <sigh>.

BTW the scope connection in Hi-Z that John mentioned needs one word of
caution added: Some of the glitzy "modern" digital scopes have the
nasty habit of reducing their bandwidth when cranking up the channel
gain.

I'll be doing this with my old analog scope, probably only at night. ;-)
 
J

John Larkin

Jan 1, 1970
0
I can understand that, you just build up the information over many pulses
one piece at a time. My question then becomes, how are you "walking" your
samples so precisely? What is generating the time delays?

Usually it's just a linear analog ramp feeding a comparator, with the
other side of the comparator driven from a dac to set the delay. With
reasonable layout and such, jitter of 1 part in 50,000 of max delay
isn't hard.

High-end sampling scopes use a triggered oscillator to tick off coarse
delays, with a ramp/comparator for fine delays. TDR really doesn't
need that.

John
 
A

Anthony Fremont

Jan 1, 1970
0
John said:
Usually it's just a linear analog ramp feeding a comparator, with the
other side of the comparator driven from a dac to set the delay. With
reasonable layout and such, jitter of 1 part in 50,000 of max delay
isn't hard.

Ding, the light goes on. :) I see, so you just trigger the ramp when you
"step" the output signal. Pretty slick. Ok, I think I can envision how
this works and how to drive it with a cheap micro. I know the reflected
signal can come back with the opposite polarity, so how do you get the ADC
to cope with that. Just float the circuit up from ground with a couple of
diodes so that the ground ref for the ADC is below ground level for the
signal? Sorry for making you design an entire product, but it has always
intrigued me how they can build a hand-held TDR with a numeric digital
display for so cheap. I think I see now. :)
 
J

Joerg

Jan 1, 1970
0
Anthony said:
<woosh like blue angels fly by> I'm pretty dense here, so it's not yet
becoming obvious to me how something like this works. I can't begin to
fathom how a quad analog switch will let you generate thousands of precisely
different time delays (different R values) to trigger the ADC.

It's used as controlled resistors. But that chip has become nearly
unobtanium so you'd be pretty much back to single FETs. If you happen to
have a fast small FET you can try it out. You'll need something like a
BF998 or another from that series. In a pinch they can be found in TV
tuners. Who needs TV anyhow...

For simplicity use a L-R where the FET is the R. 330nH or something
around that value connected to the output of the inverter where the fast
transition comes from. Drain to other side of the cap, source to GND,
then place a variable DC voltage onto both gates (or to the one if it's
a single-gate FET). Input of another inverter to the drain. Now connect
probes to both digital outputs and see one of them scoot when you vary
the gate voltage. I don't know whether your Hitachi scope is able to do
that though, I use a 1GHz scope for that kind of stuff.

Regular switching FETs such as the 2N7000/7002 aren't very suitable. Too
much capacitance.
I'll be doing this with my old analog scope, probably only at night. ;-)

If it's one of those 100MHz Hitachis don't expect much in time
resolution. Other than that they are good scopes.
 
J

Jim Yanik

Jan 1, 1970
0
That would certainly mean some paralleling or transformer-based
combining so you can have fast logic and boost up the voltage. But as
Rich already mentioned TDR is usually done with a transition, not a
pulse.

TEK 1502 long range TDR uses a 1/2 sine pulse,the TEK 1503 TDR (short-
range)uses a step pulse.Examining those units schematics would help you
with your TDR design.

You can use a L-pad to change source Z. The output amplitude drops,though.
 
J

Jim Yanik

Jan 1, 1970
0
Should. The source impedance of the tdr step should be 50 ohms. Put
several gate sections in parallel to get a stiff source, and then put
a series resistor, 47 ohms maybe, to make a net 50 ohm source. Run
that to a tee connector at the scope input (assuming a hi-z scope) and
the other tee port is the tdr thing.


Dunno. The old Tek 1503 (?) handheld used a classic diode-bridge
sampler. Nowadays you could just trigger a fast adc, once a tdr step,
with the adc trigger slowly walked in time to make the range sweep.

John

Minor nitpick;

TEK1503 was hardly "handheld",more accurately;portable.
It's size was about 12" wide,~4" high,and ~15" deep.


I serviced 1502s and 1503s for a short time at TEK.
 
D

Day Brown

Jan 1, 1970
0
Hmmm. could you use a pair of matched tank circuits that were 180 out
of phase, and put a negative pulse on an antenna for zero and a pos
for one? Since a stream of bits never alternates evenly from one to
zero very long, ordinary tuners would never latch on to it. It aint a
carrier wave.

I can see where you mite needa put the pulse generator right on the
antenna dipole to avoid resonances on the coax. Guess a center tapped
tank would keep the two polarities on time, but is an FET fast enough
to do a couple hundred meg pulses/sec and have enough voltage on the
steep part of the 'wave' for useful power output?
 
J

Jim Yanik

Jan 1, 1970
0
Hmmm. could you use a pair of matched tank circuits that were 180 out
of phase, and put a negative pulse on an antenna for zero and a pos
for one? Since a stream of bits never alternates evenly from one to
zero very long, ordinary tuners would never latch on to it. It aint a
carrier wave.

the step WILL generate a lot of harmonics.
I can see where you mite needa put the pulse generator right on the
antenna dipole to avoid resonances on the coax. Guess a center tapped
tank would keep the two polarities on time, but is an FET fast enough
to do a couple hundred meg pulses/sec and have enough voltage on the
steep part of the 'wave' for useful power output?

I dn't think you should TDR-pulse an antenna;a fast-rise step will xmit a
LOT of harmonics,way up in freq. =Interference.
If you want to match an antenna to a cable,use a VSWR meter,TDR-test the
cable by itself.
 
D

Day Brown

Jan 1, 1970
0
the step WILL generate a lot of harmonics.




I dn't think you should TDR-pulse an antenna;a fast-rise step will xmit a
LOT of harmonics,way up in freq. =Interference.
If you want to match an antenna to a cable,use a VSWR meter,TDR-test the
cable by itself.
I'm not too fussy, but I only expect to need a few watts, and at that
power, could just as well mount the emitter on the antenna, and not
worry about matching the cable.

I dont see that harmonics will be a problem out in the boonies where
I'm going to use it. They'd all be higher frequency, and the whole
reason I need a new deisgn, is that nothing with a wavelength shorter
than pine needles gets anywhere.

But this thing is so simple, I dont see the circuitry to support
harmonics. Just a center tapped tank with an FET (or maybe even a
mosfet?) on each side triggered by the data stream to feed a +/- pulse
to the antenna. And if the Yagi is tuned, as it would havta be to
punch thru the canopy, then where, if there are any harmonics, can
they go?
 
V

vasile

Jan 1, 1970
0

Very nice design (and nice web pages too). The avalanche is the best
methode for generating fast rising edge pulses, but the problem is
syncronisation with an external trigger which is almost impossible .

The load variation is changing dramatically the slew rate, so a
request like load from open circuit to dead short, 5nS pulses (which
will be the rise time ?) and 10V amplitude...

best wishes,
Vasile
 
V

vasile

Jan 1, 1970
0
5nsec pulses don't sound too hard - we got 0.5nsec FWHM at 5V into 50R
and 7.5V for anything longer.

Sorry, I'm sure the other readers (except myself) knows what FWHM
means:
http://en.wikipedia.org/wiki/Full_width_at_half_maximum

The TDR needs a very short rise time and as high amplitude as
possible. The pulse lenght could be higher than 5nS.
http://en.wikipedia.org/wiki/Time-domain_reflectometer

A toy doing some "TDR" is here:
http://www.epanorama.net/circuits/tdr.html

Vasile
 
J

John Larkin

Jan 1, 1970
0
Very nice design (and nice web pages too). The avalanche is the best
methode for generating fast rising edge pulses, but the problem is
syncronisation with an external trigger which is almost impossible .

Avalanche transistors are easily triggered, with picosecond jitter.
Generations of sampling oscilloscopes, from Lumatron up through the
Tek 7000 series, used triggered avalanche transistors to drive their
sampling gates, directly or through an srd. Nowadays, they mostly use
SRDs or shock lines driven by non-avalanche stuff.

HP tended to not use avalanche transistors in their samplers, at least
after the srd was discovered (with, ironically, an avalanche-based
sampler!)

John
 
J

Jonathan Kirwan

Jan 1, 1970
0
Avalanche transistors are easily triggered, with picosecond jitter.
Generations of sampling oscilloscopes, from Lumatron up through the
Tek 7000 series, used triggered avalanche transistors to drive their
sampling gates, directly or through an srd. Nowadays, they mostly use
SRDs or shock lines driven by non-avalanche stuff.

HP tended to not use avalanche transistors in their samplers, at least
after the srd was discovered (with, ironically, an avalanche-based
sampler!)

Interesting. I just learned something. Found this link to an early
HP paper on the general subject:
http://www.hp.woodshot.com/hprfhelp/5_downld/lit/diodelit/an918.pdf

Thanks,
Jon
 
J

John Larkin

Jan 1, 1970
0
Interesting. I just learned something. Found this link to an early
HP paper on the general subject:
http://www.hp.woodshot.com/hprfhelp/5_downld/lit/diodelit/an918.pdf

Thanks,
Jon

The story (I have it in an old HP Journal somewhere) is that some guy
at HP was experimenting with a diode-based frequency multiplier, and
found he was getting far stronger harmonic output than theory
predicted. They had just introduced the HP185 sampling scope (700 MHz
bw, using an avalanche transistor in the sampling gate) so dragged it
over to see what was happening, thus discovering the SRD. Shortly
after that, the 185 used an srd instead of the avalanche transistor,
increasing its bandwidth to 1 GHz. A little later they introduced the
187 plugin, with almost 4 GHz bandwidth, still largely tube-based.
This all started about 1962.

I have the Journal article, and the 185 manuals, and an HP185 scope
itself; historical and very, very ugly.

John
 
Top