Author Topic: Yet another fast edge pulse generator  (Read 169829 times)

0 Members and 7 Guests are viewing this topic.

Offline Leo BodnarTopic starter

  • Frequent Contributor
  • **
  • Posts: 801
  • Country: gb
Re: Yet another fast edge pulse generator
« Reply #125 on: July 26, 2017, 06:39:23 pm »
Received mine today as hoped.
Tested on Rigol MSO2302A (300MHz, upgraded...) and Tektronix 2445A (150 MHz). Results as awaited, perfect!
The software rose an error when launched: Error '13' uncompatible type (Windows 10).
Cheers,
I am looking into the problem - looks like a decimal comma/point issue.
Leo
 

Offline Leo BodnarTopic starter

  • Frequent Contributor
  • **
  • Posts: 801
  • Country: gb
Re: Yet another fast edge pulse generator
« Reply #126 on: July 29, 2017, 12:19:29 pm »
I have modified the configuration software to fix the decimal comma/point issue for locales that use decimal comma.

Please try downloading new version again: http://leobodnar.com/files/pulser_configuration.exe

Thanks
Leo
 

Offline Gixy

  • Regular Contributor
  • *
  • Posts: 232
  • Country: fr
Re: Yet another fast edge pulse generator
« Reply #127 on: July 29, 2017, 03:33:58 pm »
Hi Leo,
Works perfectly well now, including calibration. Thanks a lot  :-+
 

Offline maukka

  • Supporter
  • ****
  • Posts: 107
  • Country: fi
Re: Yet another fast edge pulse generator
« Reply #128 on: July 30, 2017, 07:34:25 am »
I have modified the configuration software to fix the decimal comma/point issue for locales that use decimal comma.

You should tell Rigol how to do that...

They say it's impossible to fix their PSU control software.

Quote
we can't change it. All of our test devices needs point separator for communication
 

Offline hs3

  • Regular Contributor
  • *
  • Posts: 73
Re: Yet another fast edge pulse generator
« Reply #129 on: July 30, 2017, 07:55:09 pm »
I tested a few Rigol scopes with the pulse generator.

DS1102E





DS1074Z 100 MHz





DS2072 300 MHz with 1M input





DS2072 with 50R through terminator





DS4014 500 MHz





One thing that could be noticed when testing several Rigol scopes one after another was that as the series number gets higher the user experience can be quite different and feels much snappier. Well the hardware under the hood is quite different too.
« Last Edit: July 30, 2017, 08:02:46 pm by hs3 »
 

Offline hs3

  • Regular Contributor
  • *
  • Posts: 73
Re: Yet another fast edge pulse generator
« Reply #130 on: July 30, 2017, 08:06:11 pm »
And then a few other scopes too

Keysight DSOX3014T





LeCroy WavePro 7300A


« Last Edit: July 30, 2017, 08:07:55 pm by hs3 »
 

Offline awallin

  • Frequent Contributor
  • **
  • Posts: 694
Re: Yet another fast edge pulse generator
« Reply #131 on: July 31, 2017, 07:57:32 am »
I think it would be nice to collect the results into an online spreadsheet and a chart similar to this one - any volunteers  ???

with an estimated Tpulser ~50ps the rise-times should be accurate to 10% up to 3 or 3.5 GHz - if my chart is roughly right..
 

Offline H.O

  • Frequent Contributor
  • **
  • Posts: 807
  • Country: se
Re: Yet another fast edge pulse generator
« Reply #132 on: July 31, 2017, 09:29:23 am »
I just started a Google spreadsheet and added what's been posted in this thread so far. See attached screenshot.
Any suggested changes before I try to make it public (never made a shared one before)? What should we call it?


 
The following users thanked this post: awallin, lukier, Tom45

Offline grumpydoc

  • Super Contributor
  • ***
  • Posts: 2905
  • Country: gb
Re: Yet another fast edge pulse generator
« Reply #133 on: July 31, 2017, 09:33:23 am »
I just started a Google spreadsheet and added what's been posted in this thread so far. See attached screenshot.
Any suggested changes before I try to make it public (never made a shared one before)? What should we call it?
I'd make the default sort order manufacturer and then model.
 

Offline awallin

  • Frequent Contributor
  • **
  • Posts: 694
Re: Yet another fast edge pulse generator
« Reply #134 on: July 31, 2017, 01:52:37 pm »
I just started a Google spreadsheet and added what's been posted in this thread so far. See attached screenshot.
Any suggested changes before I try to make it public (never made a shared one before)? What should we call it?

awesome  :-+
I'd make the X-axis the manufacturer specified BW of the scope.
maybe there's a way to make a shaded background between say 0.35/BW and 0.45/BW - where we'd expect the measurements to fall.
 

Offline Tom45

  • Frequent Contributor
  • **
  • Posts: 555
  • Country: us
Re: Yet another fast edge pulse generator
« Reply #135 on: July 31, 2017, 03:54:39 pm »
Here are some of my scopes:

Tek 2246                  100MHz   Analog     50 ohm*   2070  2280
Tek 2465                  300MHz   Analog     50 ohm     1080  1260
Tek 7904/7A26          200MHz   Analog     50 ohm*   1650  1700
Tek 7904/7A24          400MHz   Analog     50 ohm      700   800

Rigol 1074Z                 70MHz   1GS/sec   50 ohm*   3940  4570
Rigol 1074Z                 20MHz   1GS/sec   50 ohm*  14630  15910  (20MHz bandwidth limit)

Siglent SDS1202X      200MHz   1GS/sec   50 ohm    1750 1790

OWon PDS8102T        100MHz   1GS/sec   50 ohm*   2040  1960

*= 1Meg input with external 50 ohm pass through terminator
 

Offline HighVoltage

  • Super Contributor
  • ***
  • Posts: 5453
  • Country: de
Re: Yet another fast edge pulse generator
« Reply #136 on: July 31, 2017, 04:10:21 pm »
Any suggested changes before I try to make it public (never made a shared one before)? What should we call it?

Yes, add a column with Leo's calculation for the scope's rise time
see below..






Now my last scope for this test, a Keysight MSO-X-6004A, 6 GHz, 50 Ohm
Rise time is 88 ps
Assuming total system risetime Tsystem2 = Tscope2 + Tpulser2
It checks out reasonably well:
Tsystem = 88ps, Tpulser = 50ps
From there Tscope = 72ps
MSO-X-6004A specification lists 75ps risetime.

Pleasant UI on that Fluke.
There are 3 kinds of people in this world, those who can count and those who can not.
 

Offline Gixy

  • Regular Contributor
  • *
  • Posts: 232
  • Country: fr
Re: Yet another fast edge pulse generator
« Reply #137 on: August 01, 2017, 05:21:27 am »
Good idea for the spreadsheet and good suggestions to improve.
Two remarks for my lines :
- first scope is in fact Rigol MSO2072A upgraded in MSO2302A (that's what appears in the system info after upgrade)
- second one is Tek 2445A, not 2445
Gixy
« Last Edit: August 01, 2017, 05:29:05 am by Gixy »
 

Offline H.O

  • Frequent Contributor
  • **
  • Posts: 807
  • Country: se
Re: Yet another fast edge pulse generator
« Reply #138 on: August 01, 2017, 11:12:53 am »
Hi,
I've fixed the mistakes pointed out by Gixy, added the datapoints provided by Tom45 and added a column with the calculated scope risetime taking the risetime of the pulser into consideration. The calculated bandwidth numbers are then based on the calculated risetime.

I've opened the spreadsheet up for viewing (at least I think I have):
https://docs.google.com/spreadsheets/d/1uknvUdL4gNuTyuK7MNAkCj95GFSsDtfgr7nObVppFiE/edit?usp=sharing

I've yet to figure out the sorting stuff and I don't quite understand what graphing we should do. If anyone wants to help out just drop me a PM and I'll try to set you up with editing rights (or should I just allow full editing rights?)
 

Offline mk_

  • Regular Contributor
  • *
  • Posts: 225
  • Country: at
Re: Yet another fast edge pulse generator
« Reply #139 on: August 01, 2017, 01:08:00 pm »
MDO3054 from TEK @ 50 Ohm.



 

Offline awallin

  • Frequent Contributor
  • **
  • Posts: 694
Re: Yet another fast edge pulse generator
« Reply #140 on: August 01, 2017, 01:17:27 pm »

I've opened the spreadsheet up for viewing (at least I think I have):
https://docs.google.com/spreadsheets/d/1uknvUdL4gNuTyuK7MNAkCj95GFSsDtfgr7nObVppFiE/edit?usp=sharing

now we really need someone with a >10gig scope to confirm the intrinsic pulser rise-time  ;D
https://docs.google.com/spreadsheets/d/1W7VlSqhM8XUgPErtco0ETyDvSDprG7elEpLsuc_W-FA/edit?usp=sharing
(I wanted a solid line for the Tscope+Tpulser prediction, but not sure how to do it in googledocs..)
 

Offline lukier

  • Supporter
  • ****
  • Posts: 634
  • Country: pl
    • Homepage
Re: Yet another fast edge pulse generator
« Reply #141 on: August 01, 2017, 01:29:34 pm »
I've requested the edit access, can update the details on my entries.

Also, please bring back the 0.45 factor. 0.35 is useless nowadays except the analog or low-end scopes, i.e. not most scopes on the list.
 

Offline Tom45

  • Frequent Contributor
  • **
  • Posts: 555
  • Country: us
Re: Yet another fast edge pulse generator
« Reply #142 on: August 01, 2017, 01:50:07 pm »
I've requested the edit access, can update the details on my entries.

Also, please bring back the 0.45 factor. 0.35 is useless nowadays except the analog or low-end scopes, i.e. not most scopes on the list.

The last three columns are 0.45/risetime, user, and comment. Perhaps you need to scroll over to see them?
 

Offline lukier

  • Supporter
  • ****
  • Posts: 634
  • Country: pl
    • Homepage
Re: Yet another fast edge pulse generator
« Reply #143 on: August 01, 2017, 01:56:25 pm »
The last three columns are 0.45/risetime, user, and comment. Perhaps you need to scroll over to see them?

I was replying to awallin and in reference to his copy of the sheet.
 

Offline mk_

  • Regular Contributor
  • *
  • Posts: 225
  • Country: at
Re: Yet another fast edge pulse generator
« Reply #144 on: August 01, 2017, 02:48:19 pm »
and here from an Agilent DL6104 @ internal 50Ohm

 

Offline Howardlong

  • Super Contributor
  • ***
  • Posts: 5315
  • Country: gb
Re: Yet another fast edge pulse generator
« Reply #145 on: August 01, 2017, 05:43:55 pm »
57.6ps on a 20GHz HP 54120B/54121A setup, but as this uses 3.5mm/SMA connectors, I very strongly suspect the limitation is my professionally terminated BNC-SMA cable. The best I could get out of the cable from the internal <33ps rise time TDR output of the 54121A test set with that cable (and a further BNC/SMA transition) was about the same. I tried a number of BNC/SMA interseries connectors of assorted qualities (none that I'd call professional) as well as the 12" RG223 SMA-M/BNC-F cable, and the cable gave the best results.



 

Offline H.O

  • Frequent Contributor
  • **
  • Posts: 807
  • Country: se
Re: Yet another fast edge pulse generator
« Reply #146 on: August 01, 2017, 06:47:39 pm »
Hi,
Data from mk_ and Howardlong has been added, so has my own 2465B:



It does not quite live up to it's specification or I'm doing something wrong with it using the 10x mag

 

Offline Tom45

  • Frequent Contributor
  • **
  • Posts: 555
  • Country: us
Re: Yet another fast edge pulse generator
« Reply #147 on: August 01, 2017, 07:12:30 pm »
Hi,
Data from mk_ and Howardlong has been added, so has my own 2465B:

It does not quite live up to it's specification or I'm doing something wrong with it using the 10x mag

It looks like your 500ps timebase is OK.

You need to adjust the vertical gain so that the square wave top is at the solid line above the top dotted line, and the square wave bottom is at the solid line below the lower dotted line. Then the two dotted lines will be at 10% and 90%. Set the two cursors where the trace crosses the 10% and 90% dotted lines.

That should give a better result.

Also, I've noticed on my analog scopes, the shape of the trace varies some between channels. Is that because the analog scopes are old and need some TLC? Or was there that much variation when they were new? I don't know.

You might want to try 1 and 2 to see if there is a noticeable difference.
 

Offline lukier

  • Supporter
  • ****
  • Posts: 634
  • Country: pl
    • Homepage
Re: Yet another fast edge pulse generator
« Reply #148 on: August 01, 2017, 07:19:09 pm »
You need to adjust the vertical gain so that the square wave top is at the solid line above the top dotted line, and the square wave bottom is at the solid line below the lower dotted line. Then the two dotted lines will be at 10% and 90%. Set the two cursors where the trace crosses the 10% and 90% dotted lines.

His measurement is correct. Waveform's top and bottom should be on the dotted lines(!) and then the next solid line is the 10% and 90%. This is even marked on the overlay in the photo.
 
The following users thanked this post: Tom45

Offline Tom45

  • Frequent Contributor
  • **
  • Posts: 555
  • Country: us
Re: Yet another fast edge pulse generator
« Reply #149 on: August 01, 2017, 08:36:35 pm »
You need to adjust the vertical gain so that the square wave top is at the solid line above the top dotted line, and the square wave bottom is at the solid line below the lower dotted line. Then the two dotted lines will be at 10% and 90%. Set the two cursors where the trace crosses the 10% and 90% dotted lines.

His measurement is correct. Waveform's top and bottom should be on the dotted lines(!) and then the next solid line is the 10% and 90%. This is even marked on the overlay in the photo.

You are right!  The few times I've ever needed to measure rise time I never bothered to look at the scale. I just went with my faulty memory of what I once knew.   Thanks for setting me straight.

Now I need to go back and redo my Tek analog scope measurements.

Tom
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf