Author Topic: EEVblog #306 - Jim Williams Pulse Generator  (Read 294394 times)

0 Members and 1 Guest are viewing this topic.

Offline T3sl4co1l

  • Super Contributor
  • ***
  • Posts: 21606
  • Country: us
  • Expert, Analog Electronics, PCB Layout, EMC
    • Seven Transistor Labs
Re: EEVblog #306 - Jim Williams Pulse Generator
« Reply #600 on: October 11, 2015, 08:57:40 am »

So you are saying the scope measurements (shown with my generator and scope, and with mtdoc's generator and scope) are or are not accurate measurements?

They are what they are, of that there can be no doubt.  But the impulse response is not the step response, and so cannot be used to measure it.

Tim
Seven Transistor Labs, LLC
Electronic design, from concept to prototype.
Bringing a project to life?  Send me a message!
 

Online Electro Fan

  • Super Contributor
  • ***
  • Posts: 3161
Re: EEVblog #306 - Jim Williams Pulse Generator
« Reply #601 on: October 11, 2015, 09:27:30 am »
EF:

Simply that because the pulse width is so short, well before your scope's vertical could even reach the real voltage peak, the pulse generator was already heading south. By extending the pulse width, it gives time for the scope to reach the actual voltage peak.

Therefore, I'd say you weren't measuring 90% of the peak voltage, just 90% of how far the vertical made it before the input had dropped back.

Ok, that seems understandable and plausible - but how does it explain the mtdoc's results posted below?  Seems like mtdoc's 350MHz scope caught the peaky pulse (and the wider pulse didn't yield a much different result).  So it would seem that a 400MHz scope should have also caught the peaky pulse.... unless there is something notably faster about my generator, which would be difficult thing to determine at the rate I'm going.... :)
« Last Edit: October 11, 2015, 09:33:43 am by Electro Fan »
 

Offline T3sl4co1l

  • Super Contributor
  • ***
  • Posts: 21606
  • Country: us
  • Expert, Analog Electronics, PCB Layout, EMC
    • Seven Transistor Labs
Re: EEVblog #306 - Jim Williams Pulse Generator
« Reply #602 on: October 11, 2015, 11:19:34 am »
Like I said, these things need not be consistent, because the response to different generators isn't.

Tim
Seven Transistor Labs, LLC
Electronic design, from concept to prototype.
Bringing a project to life?  Send me a message!
 

Offline Jay_Diddy_B

  • Super Contributor
  • ***
  • Posts: 2731
  • Country: ca
Re: EEVblog #306 - Jim Williams Pulse Generator
« Reply #603 on: October 11, 2015, 11:45:44 am »
Hi,

Like Tim said, I think the pulse from your pulse generator is too short. Here is a simulation of a filter being tested with a short pulse and a long pulse:



You would think the short pulse indicates higher BW, but the filters are identical.

Try adding about 18 inches of 50 Ohm coax in parallel with the capacitor in your pulse generator. This will length the pulse to about 4ns.


Regards,

Jay_Diddy_B
 

Offline mtdoc

  • Super Contributor
  • ***
  • Posts: 3575
  • Country: us
Re: EEVblog #306 - Jim Williams Pulse Generator
« Reply #604 on: October 11, 2015, 05:59:48 pm »
I just picked up on eBay one of these Tek scope demo boards. It includes a fast, <200 ps pulse output for checking scope rise time. It will be interesting to see what it looks like on my 2467.

FWIW my best offer of $30 was accepted though I specifically said I did not need the manual included (it's online). As of now, it looks liike there is one left.
 

Online Electro Fan

  • Super Contributor
  • ***
  • Posts: 3161
Re: EEVblog #306 - Jim Williams Pulse Generator
« Reply #605 on: October 11, 2015, 08:01:49 pm »
Hi,

Like Tim said, I think the pulse from your pulse generator is too short. Here is a simulation of a filter being tested with a short pulse and a long pulse:

You would think the short pulse indicates higher BW, but the filters are identical.

Try adding about 18 inches of 50 Ohm coax in parallel with the capacitor in your pulse generator. This will length the pulse to about 4ns.


Regards,

Jay_Diddy_B

Jay_Diddy_B, Thanks.

Ok, these questions are going to officially get me labeled as a complete rookie.... :-[

(I didn't build the pulse generator - I'm just using it.  Except for the transistor pretty much everything on the board looks like SMD; so it's kind of small stuff and I'd prefer not to wreck an otherwise nice pulse generator. :scared:)

I see a 1M Ohm resistor (connected to the Collector?) and a 10k resistor (connected to the Base) near the transistor; instead of one 50 Ohm resistor connected to the Emitter I see two 50 Ohm resistors next to each other - not sure why there are two there.

There are only two capacitors that I see near the transistor (and also reasonably near the BNC connector) - one that is considerably larger and further from the transistor and another that is relatively smaller and a little closer to the transistor.  The one that is larger is pretty much the same physical size as two other capacitors also on the board - all three measure about 420nF.  The one that is smaller and closer to the transistor and BNC is difficult to measure; sometimes I measured it as being about 3-5pF and sometimes around 60pf??  Any chance it is this smaller capacitor that is the right capacitor to connect to the coax?

Assuming I have the right capacitor (the one that measures 3-5pF and occasionally 60pF), any chance you could post a diagram showing where to add the 18" of 50 Ohm coax (from where to where)?

Thanks

Howard, I saw your suggestion:  "You can extend the pulse with a foot of RG174 open ended stub in parallel with the cap."  I wasn't sure if "open ended stub" meant just solder the length of the coax onto one side (either side?) of the capacitor and let the other end of the coax dangle free, or something else.  Any clarifications or suggestions on how not to mess up a working generator would be very welcome.  Thx! EF
« Last Edit: October 11, 2015, 08:37:11 pm by Electro Fan »
 

Offline Howardlong

  • Super Contributor
  • ***
  • Posts: 5315
  • Country: gb
Re: EEVblog #306 - Jim Williams Pulse Generator
« Reply #606 on: October 11, 2015, 08:54:33 pm »
Coax shield to ground plane, inner to the collector, directly across the cap, but keep the exposed coax as short as possible, a mm or two, best done under an illuminated magnifier at least.

Open ended stub means the far end of the coax is left open rather than shorted (or anything else), mine's about 15" or so.
 

Offline EV

  • Frequent Contributor
  • **
  • Posts: 525
  • Country: fi
  • Aficionado
Re: EEVblog #306 - Jim Williams Pulse Generator
« Reply #607 on: October 14, 2015, 07:51:47 am »
Here are 2 pictures about rise time of my 2465B scope.

In first picture the pulse is from Tektronix type 284 pulse generator. Its rise time is faster than 35 ps.

In second picture the pulse is from JW pulse generator (MK-2 rise time tester Nr 32). I have used it with 10 dB attenuator. Its rise time is 190 ps.
 
The following users thanked this post: electrolust

Online Electro Fan

  • Super Contributor
  • ***
  • Posts: 3161
Re: EEVblog #306 - Jim Williams Pulse Generator
« Reply #608 on: October 14, 2015, 08:29:57 pm »
Here are 2 pictures about rise time of my 2465B scope.

In first picture the pulse is from Tektronix type 284 pulse generator. Its rise time is faster than 35 ps.

In second picture the pulse is from JW pulse generator (MK-2 rise time tester Nr 32). I have used it with 10 dB attenuator. Its rise time is 190 ps.

Very cool - Thanks for posting.

This would seem to confirm that the rise time of the scope is roughly as specified (875ps) and that Jay_Diddy_B's formulas are working :), and that your JW Mk-2 is very fast and that your 284 is blazingly fast.   :-+ :-+
 

Online Electro Fan

  • Super Contributor
  • ***
  • Posts: 3161
Re: EEVblog #306 - Jim Williams Pulse Generator
« Reply #609 on: October 16, 2015, 11:49:42 pm »
A progress report.

First, thanks to everyone here for the ongoing education  :-+

A bit of extra thanks to mtdoc for mentioning the Tek demo board earlier in this thread.  It happens to have a <200ps rise time signal which provides a much wider (~20ns) pulse than the very "peaky" pulse produced by the JW pulser I've been trying to measure.  As noted in the photo below the Tek demo board produces a rise time that is measured by the 2467B as being 720ps - which seems reasonably consistent with others results shown in this thread.

A second thing/further thanks to EEVers.  This forum is really, really helpful.  The other thing that is helpful is to just keep plugging (haha) with more experiments and gaining more experience.  When I received the demo board I tried powering it with a 9V battery but that had it's limitations.  So, next I tried using the leads from the plug-on cap for the battery as attachment points for the alligator clip leads on my DC power supply.  For some reason it wasn't working - turns out that while I recognize the larger 9V battery terminal as the minus and the smaller battery terminal as the plus, when you look at the cap the connector sizes are of course the opposite (inside the cap the larger connector goes over the smaller plus terminal on the battery, and vice versa on the other connector/terminal).  So by habit I clipped them backwards, and .... no power on the board.  Duh.

Between trial and error and EEVblog it's a doable and very enjoyable learning journey  :)
« Last Edit: October 17, 2015, 12:10:14 am by Electro Fan »
 

Offline Jay_Diddy_B

  • Super Contributor
  • ***
  • Posts: 2731
  • Country: ca
Re: EEVblog #306 - Jim Williams Pulse Generator
« Reply #610 on: October 17, 2015, 01:20:23 am »
Hi,

Can you reverse engineer this section of the Tektronix demo board? Or put a high resolution picture of this part of the board on the forum?

Thanks !!

Jay_Diddy_B
 

Online Electro Fan

  • Super Contributor
  • ***
  • Posts: 3161
Re: EEVblog #306 - Jim Williams Pulse Generator
« Reply #611 on: October 17, 2015, 01:59:33 am »
Reverse engineer?  I can hardly forward engineer. :-DD

(It's a nice product - I'd recommend one for purchase.  Since it's in the public domain I think it's A-OK to show it / look at it.)
« Last Edit: October 17, 2015, 02:03:52 am by Electro Fan »
 

Offline mtdoc

  • Super Contributor
  • ***
  • Posts: 3575
  • Country: us
Re: EEVblog #306 - Jim Williams Pulse Generator
« Reply #612 on: October 17, 2015, 02:08:38 am »
Hi,

Can you reverse engineer this section of the Tektronix demo board? Or put a high resolution picture of this part of the board on the forum?

Thanks !!

Jay_Diddy_B

Manual with schematics is available HERE. Pulse generator schematic on A-11.
 

Online Electro Fan

  • Super Contributor
  • ***
  • Posts: 3161
Re: EEVblog #306 - Jim Williams Pulse Generator
« Reply #613 on: October 17, 2015, 03:05:30 am »
Question:

On the test board the spec says it can run between 6 and 12 Volts and it wants a 5 Watt transformer.  I calculated that 9 Volts would support 0.555 Amps (and stay within 5 Watts).  It turns out that at 9 Volts the board draws less than 0.4 Amps.  Given that the spec says the range is 6 to 12 Volts, is there any reason to think that supplying the board with more voltage (beyond 9V but less than 12V) and lowering the current, or supplying less voltage (below 9V but at least 6V) and raising the current would somehow enable the board to produce a cleaner pulse and/or a faster rise time?  Or within the range of 6-12 Volts the pulse and rise time performance will not be impacted?
« Last Edit: October 17, 2015, 03:10:12 am by Electro Fan »
 

Online Electro Fan

  • Super Contributor
  • ***
  • Posts: 3161
Re: EEVblog #306 - Jim Williams Pulse Generator
« Reply #614 on: October 18, 2015, 09:44:08 am »
Some more measurements and observations.

I found a SMA to BNC connector which allowed me to directly attach the Tek demo board to the scope (with no cable).

With the automeasure function the scope shows 850ps.  With 30" of RG223 the automeasure shows the same 850ps as with the directly attached board.  I tried with a 5' RG174 cable and the automeasure shows 960ps.

With the directly attached board using the cursors the scope shows 740ps.  I think part of this is due to my uncertainty of where to line up the 0% and 100% lines.  I can get a pretty good line on 0% by tracing backward on the waveform well before it begins to ascend, but I'm not so sure about what constitutes the 100% peak.  Even so, the automeasure seems to provide significantly more conservative (slower) results.

Conclusions:

1. The automeasure is consistently (always) slower (more accurate?) than the cursor measurements.
2. With the directly attached board and the 30" of RG223 I think the automeasure is bumping up against the threshold of the scope's rise time.
3. 5' of RG174 is enough to add some delay.
4. If you want to get a more impressive rise time use the cursors instead of the automeasure :)
5. YMMV
« Last Edit: October 18, 2015, 05:38:54 pm by Electro Fan »
 
The following users thanked this post: electrolust

Online Electro Fan

  • Super Contributor
  • ***
  • Posts: 3161
Re: EEVblog #306 - Jim Williams Pulse Generator
« Reply #615 on: October 22, 2015, 06:32:02 pm »
Still trying to measure the peaky JW type pulse gen.  Here is an attempt with a Tek 7A19 (500MHz).  The timebase readouts are correct - the Voltage readout is not correct.
« Last Edit: October 22, 2015, 06:48:53 pm by Electro Fan »
 

Offline mtdoc

  • Super Contributor
  • ***
  • Posts: 3575
  • Country: us
Re: EEVblog #306 - Jim Williams Pulse Generator
« Reply #616 on: October 24, 2015, 06:59:10 am »
Finally got around to playng with the Tek Demo board fast rise output.  Fastest rise time I get is with board directly connected to my 2467 (50 ohm input).  Rise time of 0.78 - 0.79 ns is a bit slower than what i get with my JWP.
« Last Edit: October 24, 2015, 07:09:24 am by mtdoc »
 

Offline notadave

  • Contributor
  • Posts: 48
  • Country: de
Pulse Generator on the Grid
« Reply #617 on: October 24, 2015, 06:13:52 pm »
I build my own version that is powered by an unregulated 12V AC wall adapter (just a Transformer).
I got it to work after I increased the capacities in my charge ladder from 220nF to 22uF. To my surprise the capacity had a great impact on the voltage. Especially the lower capacitors have to be bigger than I thought.
By using 6x 22uF capacitors in only 3 steps I got a peak voltage that is so much higher than the breakdown voltage, that the breakdown occurs, but so low that it will happen only once per grid cycle.
It is good enough to measure a cable of 9ft, or less if it is shorted. Since I do not use a delay line I get a normal discharge shape.
I designed a version with a 0.2m coplanar PCB 50 Ohm line that is not yet build.

My 2N3904 breaks at 112V.
 

Offline max-bit

  • Frequent Contributor
  • **
  • Posts: 668
  • Country: pl
Re: EEVblog #306 - Jim Williams Pulse Generator
« Reply #618 on: October 27, 2015, 06:51:01 pm »
OK
To dispel doubts
Based post :
https://www.eevblog.com/forum/buysellwanted/pulse-generator-kits-or-ready-to-use/
and my topic:
https://www.eevblog.com/forum/projects/new-version-pulse-generator-(based-jim-williams-pulse-gen)/msg775033/#msg775033
used oscilloscope:
LeCroy LC584AL 1 GHz 8GS/s (rise time ~330-350ps)
Pulse the load 50Om


Pulse the load 1MOm

jitter trigger pulse

 

Offline max-bit

  • Frequent Contributor
  • **
  • Posts: 668
  • Country: pl
Re: EEVblog #306 - Jim Williams Pulse Generator
« Reply #619 on: October 27, 2015, 08:47:18 pm »
An interesting generator with a small rise time 35ps
Basen on IC ONET8501
http://www.ti.com/lit/ds/symlink/onet8501p.pdf
Link project
https://www.kickstarter.com/projects/1855991221/10-ghz-usb-oscilloscope
 

Offline LY1

  • Contributor
  • Posts: 10
  • Country: lt
Re: EEVblog #306 - Jim Williams Pulse Generator
« Reply #620 on: January 15, 2017, 05:40:27 pm »
Where I can bay fast pulse generator?
 

Offline luisprata

  • Regular Contributor
  • *
  • Posts: 58
  • Country: br
Re: EEVblog #306 - Jim Williams Pulse Generator
« Reply #621 on: April 01, 2017, 08:21:31 am »
Hi,

Anybody knows if I can use the board bellow confiured as 60MHz clock generator with 200ps rise time as a pulse generator?
https://www.digikey.com/products/en?keywords=LMK61PDEVM

Thank you.
Luis.
 

Offline David Hess

  • Super Contributor
  • ***
  • Posts: 16545
  • Country: us
  • DavidH
Re: EEVblog #306 - Jim Williams Pulse Generator
« Reply #622 on: April 01, 2017, 11:16:44 am »
Anybody knows if I can use the board bellow confiured as 60MHz clock generator with 200ps rise time as a pulse generator?

You could but:

1. It is not 200ps even in the worst case.  The specification is for a 20/80 transition time.
2. There are no specifications or even graphs showing the output transient response.  Texas Instruments does not even provide a simplified schematic of the output structures.
3. The output is not ground referred.

So you would just have to buy one and test it.
 
The following users thanked this post: luisprata

Online tautech

  • Super Contributor
  • ***
  • Posts: 28136
  • Country: nz
  • Taupaki Technologies Ltd. Siglent Distributor NZ.
    • Taupaki Technologies Ltd.
Re: EEVblog #306 - Jim Williams Pulse Generator
« Reply #623 on: March 03, 2022, 06:31:03 am »
Well how things have changed in a few short years when we can now get one of Leo Bodnar's pulsers shipped complete with a spec sheet showing a 30ps result from a 40 GHz sampling head.

Now you might wonder why I'm here....after a few hours Facetime chat to an EEVblog member in the US I found it was he that gave this design to Jim Williams whom was a good friend.  :-X
Apparently the last time they chatted was to arrange a lunch date for the following week but Jim never made it.  :(
Avid Rabid Hobbyist
Siglent Youtube channel: https://www.youtube.com/@SiglentVideo/videos
 
The following users thanked this post: HighVoltage


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf