Author Topic: Super simple "fast-edge" pulse generator for scope-based basic TDR  (Read 26115 times)

0 Members and 1 Guest are viewing this topic.

Offline w2aewTopic starter

  • Super Contributor
  • ***
  • Posts: 1780
  • Country: us
  • I usTa cuDnt speL enjinere, noW I aR wuN
    • My YouTube Channel
Here's a video I posted today that shows a very simple oscillator / pulse-generator that has about 2ns edge speed and the ability to drive 50ohm lines.  This makes it a nice, low-cost circuit for use as a very basic TDR (time domain reflectometer) for checking cable length and impedance.  Enjoy!




YouTube channel: https://www.youtube.com/w2aew
FAE for Tektronix
Technical Coordinator for the ARRL Northern NJ Section
 

Offline Rufus

  • Super Contributor
  • ***
  • Posts: 2095
Re: Super simple "fast-edge" pulse generator for scope-based basic TDR
« Reply #1 on: April 22, 2013, 02:01:30 am »
Heh, photo of my version attached.

A dead bug 74FCT octal bus driver (fastest thing in my junk box at the time) and 7 x 330R resistors. The first buffer drives the other 7 and I feed it with a signal generator for more flexibility. You can't see it but there is an 0805 ceramic cap soldered between the ground plane and chip VDD pin.

I don't have anything fast enough to properly measure its rise time. Would the spectrum from a SA give you a meaningful indication?

One reason I made it was as a source for hf compensation trimming on scope probes.
 

Offline babysitter

  • Frequent Contributor
  • **
  • Posts: 896
  • Country: de
  • pushing silicon at work
Re: Super simple "fast-edge" pulse generator for scope-based basic TDR
« Reply #2 on: April 22, 2013, 07:22:24 am »
Kindly allow me to refer to the charged line pulse generator that has some properties making it interesting for the TDR theme.
I made a little jig from a ebayed mercury wetted relais, if wanted I can post something later.

Hendi
I'm not a feature, I'm a bug! ARC DG3HDA
 

Offline ice595

  • Regular Contributor
  • *
  • Posts: 98
  • Country: ca
Re: Super simple "fast-edge" pulse generator for scope-based basic TDR
« Reply #3 on: February 04, 2014, 03:18:47 am »
I just designed and etched a little PCB for this pulse generator circuit
using a laser printer with the toner transfer method


Since I didn't have any SMD resistor or caps at the time so I had to do this the "rural area" way  ;)


I also added a high performance oscilloscope probe socket as an after-though to make accurate measurements.


With my vintage HP 54502A 400MHz 400MS/s oscilloscope I able to measure <4ns rise time with averaging turned on and equivalent time sampling on


since my Rigol DS1102E doesn't have a 50ohm input option, the measured rise time is WAY out of wack.
and no I don't have any 50ohm terminators at the moment so I can't just use a T connector to make the input 50ohms

Anyways, I tried to measure a 75ohm TV coaxial cable and the calculation doesn't match the actual length of my coax cable
according to my calculations, my coax cable should be around 8 meters but the actual length is about 10 meters.  :-\
is it because my TV coax cable doesn't have a velocity factor of 0.66?
also, should the cable's 75ohm characteristic impedance matter too much? no right?




attachment contains EAGLE CAD schematic and board layout if anyone is interested in etching their own PCB
note: click the rats-nest tool to create the ground plane, EAGLE doesn't seem to save my ground plane  ???
« Last Edit: February 04, 2014, 03:53:38 am by ice595 »
 

Offline jwm

  • Regular Contributor
  • *
  • Posts: 64
  • Country: us
    • NaN
Re: Super simple "fast-edge" pulse generator for scope-based basic TDR
« Reply #4 on: February 04, 2014, 03:47:55 am »
Nice. I made one with a 74ac14 and a repurposed junk pcb not long ago and was pleasantly surprised at how well it worked. I used one of the six inverters as an oscilator to generate the pulses and the other 5 with 220 ohm resistors on the other five inverters in parallel, looks like pretty much the same circuit as you.

Experimenting with a TDR

     John

John Meacham - http://notanumber.net/
 

Offline ice595

  • Regular Contributor
  • *
  • Posts: 98
  • Country: ca
Re: Super simple "fast-edge" pulse generator for scope-based basic TDR
« Reply #5 on: February 04, 2014, 04:03:12 am »
what was the rise time for your pulse generator?
 

Offline w2aewTopic starter

  • Super Contributor
  • ***
  • Posts: 1780
  • Country: us
  • I usTa cuDnt speL enjinere, noW I aR wuN
    • My YouTube Channel
Re: Super simple "fast-edge" pulse generator for scope-based basic TDR
« Reply #6 on: February 04, 2014, 04:07:26 am »
I just designed and etched a little PCB for this pulse generator circuit
using a laser printer with the toner transfer method


Since I didn't have any SMD resistor or caps at the time so I had to do this the "rural area" way  ;)


I also added a high performance oscilloscope probe socket as an after-though to make accurate measurements.


With my vintage HP 54502A 400MHz 400MS/s oscilloscope I able to measure <4ns rise time with averaging turned on and equivalent time sampling on


since my Rigol DS1102E doesn't have a 50ohm input option, the measured rise time is WAY out of wack.
and no I don't have any 50ohm terminators at the moment so I can't just use a T connector to make the input 50ohms

Anyways, I tried to measure a 75ohm TV coaxial cable and the calculation doesn't match the actual length of my coax cable
according to my calculations, my coax cable should be around 8 meters but the actual length is about 10 meters.  :-\
is it because my TV coax cable doesn't have a velocity factor of 0.66?
also, should the cable's 75ohm characteristic impedance matter too much? no right?




attachment contains EAGLE CAD schematic and board layout if anyone is interested in etching their own PCB
note: click the rats-nest tool to create the ground plane, EAGLE doesn't seem to save my ground plane  ???

Typical CATV coax is about 0.85 velocity factor.

The board is a nice clean layout. You should note however this is a case where a neat and clean layout might not give you the best results. This design puts several of the gates in parallel to increase the drive capability of the generator. Ideally, the inputs and outputs of each of the gates should be matched in length. The longer delay of the gate outputs on the left of the board compared to the ones on the right is probably contributing to the ringing you're seeing. For the TDR application though, it's not too big of a worry though.
YouTube channel: https://www.youtube.com/w2aew
FAE for Tektronix
Technical Coordinator for the ARRL Northern NJ Section
 

Offline jwm

  • Regular Contributor
  • *
  • Posts: 64
  • Country: us
    • NaN
Re: Super simple "fast-edge" pulse generator for scope-based basic TDR
« Reply #7 on: February 04, 2014, 04:45:55 am »
what was the rise time for your pulse generator?

I am not sure, faster than any equipment I had at the time (60mhz scope) could measure, is there a good way to measure rise time other than "find a really fast scope".

I have been meaning to build up the williams avalanche pulse generator too and see how it functions as a TDR. I'd like to be able to measure shorter lengths of cable than I was able to with my first version. though, I am probably still scope-limited. (have a 100mhz one now).

I'd be interested in any tricks to measure very short lengths of cable, like on the order of a light-nanosecond. Not for any particular reason other than since I built the 74ac14/oscilloscope version it has piqued my curiosity about how it might be done without crazy expensive equipment.
   
    John
John Meacham - http://notanumber.net/
 

Offline JoeyP

  • Frequent Contributor
  • **
  • Posts: 321
  • Country: us
Re: Super simple "fast-edge" pulse generator for scope-based basic TDR
« Reply #8 on: February 04, 2014, 04:46:54 pm »
The risetime is probably much faster than your 400MHz scope. I've built simple pulse generators using Fairchild NC7SZ14 (an ultrafast single element version of 74xx14. They cost $0.10 in quantity).  No need to parallel multiple units. Just the most basic resistive matching pad to produce a 50 ohm source. Example 80 ohms series with output to 100 ohms parallel. The 80 ohms accounts for ~20 ohm output impedance of the driver. The resulting output has a risetime <1ns and amplitude >1Vpp into 50 ohms, and surprisingly flat pulse top.
 

Offline dannyf

  • Super Contributor
  • ***
  • Posts: 8221
  • Country: 00
Re: Super simple "fast-edge" pulse generator for scope-based basic TDR
« Reply #9 on: February 04, 2014, 07:29:13 pm »
I thought the edges on an avalanche transistor is pretty fast - sub ns if I may recall.
================================
https://dannyelectronics.wordpress.com/
 

Offline free_electron

  • Super Contributor
  • ***
  • Posts: 8520
  • Country: us
    • SiliconValleyGarage
Re: Super simple "fast-edge" pulse generator for scope-based basic TDR
« Reply #10 on: February 04, 2014, 08:10:06 pm »
I thought the edges on an avalanche transistor is pretty fast - sub ns if I may recall.
My pulser reaches 180pS... (Jim williams based design but using a stripline as tank capacitor
Professional Electron Wrangler.
Any comments, or points of view expressed, are my own and not endorsed , induced or compensated by my employer(s).
 

Offline JBeale

  • Frequent Contributor
  • **
  • Posts: 302
Re: Super simple "fast-edge" pulse generator for scope-based basic TDR
« Reply #11 on: February 04, 2014, 08:21:08 pm »
If you only need 3.3V pulse with tr < 2 nsec you can use a $25 Raspberry Pi with no extra circuits at all. The pulse comes from GPIO 18 (pin #12 on P1 header), and this also gives you a programmable pulse width of any length from 4 nsec upward in 2 nsec increments.  I measured 1.6 nsec risetime with a 200 MHz scope & 200 MHz probe, so it is probably faster, but I just can't measure it.

Here are the details: http://www.raspberrypi.org/phpBB3/viewtopic.php?f=72&t=67741#p495574
 

Offline ice595

  • Regular Contributor
  • *
  • Posts: 98
  • Country: ca
Re: Super simple "fast-edge" pulse generator for scope-based basic TDR
« Reply #12 on: February 04, 2014, 09:12:52 pm »
sorry to be a little bit off topic here.
how can you tell the type of dielectric used in a coax cable by visual inspection?

I spliced open my CATV coax cable and the dielectric material is white and seems pretty solid.
since it has a velocity factor of 0.85, is the dielectric made of foam then?

and also, how would Solid Polyethylene (PE) look like?
 

Offline jwm

  • Regular Contributor
  • *
  • Posts: 64
  • Country: us
    • NaN
Re: Super simple "fast-edge" pulse generator for scope-based basic TDR
« Reply #13 on: February 05, 2014, 01:03:23 am »
If you only need 3.3V pulse with tr < 2 nsec you can use a $25 Raspberry Pi with no extra circuits at all. The pulse comes from GPIO 18 (pin #12 on P1 header), and this also gives you a programmable pulse width of any length from 4 nsec upward in 2 nsec increments.  I measured 1.6 nsec risetime with a 200 MHz scope & 200 MHz probe, so it is probably faster, but I just can't measure it.

Here are the details: http://www.raspberrypi.org/phpBB3/viewtopic.php?f=72&t=67741#p495574

Surely you kid. replacing a 12 cent chip with a combined 110ma drive output and schmitt triggering with a $40 computer requiring a boot up, has all the complexity of running a whole OS, uncertain timing due to not having a RTOS, a power consumption in the Watts vs mWatts, an unknown board layout that is most certainly not designed for fast rise times and coupling your GPIO signal with who knows what and a measely 16ma drive output? It needs to have that rise time while driving a 50 ohm load, I don't know the impedance of the path between the raspberry pi CPU and out through that header, but I doubt its 50 ohms.

Getting fast rise times is a pure analog thing, completely unrelated to what you use to trigger them, whether it be a computer like the raspi, a uC, or a plain RC delay. As in, you'd need a real driver even if you triggered from a raspi.
John Meacham - http://notanumber.net/
 

Offline JBeale

  • Frequent Contributor
  • **
  • Posts: 302
Re: Super simple "fast-edge" pulse generator for scope-based basic TDR
« Reply #14 on: February 05, 2014, 02:08:58 am »
Nothing wrong with your solution. The R-Pi happened to be convenient to me, and I know that a lot of folks have one already lying around, so it cost no money and very little time to connect up for a fast edge for a one-off measurement. For the kind of TDR I was doing, you trigger your scope off the pulse, it doesn't matter what latency the OS has.  Of course the R-Pi was not intended for this purpose... but it works.  The internal resistance of that particular GPIO pin driver seems to be about 44 ohms according to my measurement of the 10 nsec pulse.  It drove my laser directly with just a 25 ohm ballast resistor, and it can also drive a 50 ohm cable with enough signal level for the scope to detect easily- without any additional driver.
« Last Edit: February 05, 2014, 02:12:33 am by JBeale »
 

Offline ice595

  • Regular Contributor
  • *
  • Posts: 98
  • Country: ca
Re: Super simple "fast-edge" pulse generator for scope-based basic TDR
« Reply #15 on: February 05, 2014, 02:43:35 am »
but does the R-Pi have a fast enough ADC to detect the delay between the transmitted edge and reflected edge?
would be neat to make an inexpensive self-contained TDR.

what great idea that would be for my final year college technical project ;)
 

Offline jwm

  • Regular Contributor
  • *
  • Posts: 64
  • Country: us
    • NaN
Re: Super simple "fast-edge" pulse generator for scope-based basic TDR
« Reply #16 on: February 05, 2014, 03:10:11 am »
but does the R-Pi have a fast enough ADC to detect the delay between the transmitted edge and reflected edge?
would be neat to make an inexpensive self-contained TDR.

what great idea that would be for my final year college technical project ;)

I would think a fast comparator with one of its inputs driven by PWM to set its trigger level would be a better way to go. off the top of my head, Set your trigger level with the pwm, send the pulse and have the output of the comparator charge a cap, after it triggers use the slower ADC to read the voltage off the cap. sort of like an integrating ADC.... sort of.
John Meacham - http://notanumber.net/
 

Offline JBeale

  • Frequent Contributor
  • **
  • Posts: 302
Re: Super simple "fast-edge" pulse generator for scope-based basic TDR
« Reply #17 on: February 05, 2014, 03:37:09 am »
but does the R-Pi have a fast enough ADC to detect the delay between the transmitted edge and reflected edge?
would be neat to make an inexpensive self-contained TDR.
I was relying on the scope to do the measurement, the R-Pi has no ADC of any kind.  It might be possible to do something clever with a fairly cheap addon, like those DVT dongles that be programmed as general-purpose 1500 MHz+ receivers.  Note that the same R-Pi GPIO pin can also generate CW or FM around 100 MHz and has been used to generate FM broadcast-band signals.
http://www.icrobotics.co.uk/wiki/index.php/Turning_the_Raspberry_Pi_Into_an_FM_Transmitter
 

Online T3sl4co1l

  • Super Contributor
  • ***
  • Posts: 21974
  • Country: us
  • Expert, Analog Electronics, PCB Layout, EMC
    • Seven Transistor Labs
Re: Super simple "fast-edge" pulse generator for scope-based basic TDR
« Reply #18 on: February 05, 2014, 05:28:06 am »
I built this many years ago; it used a PH2369 (close enough).





Response on Tek 475, 10ns/div.

I also built a fancier one (100V generator, pulse line option, and trigger input included).



Note the avalanche element is a 2N3904 -- I have tried many transistors, most do not avalanche so dramatically or at all.  I have been able to avalanche all 2N3904s that I have tested, which apparently is unusual for a planar transistor type.  RF transistors seem to do it more commonly, which sort of explains 2N2369 (it's primarily a switching transistor rather than RF, but it's fast in any case).  RF transistors also have lower breakdown voltages, which suggests it might be reasonable to build one of these with a 9-12V supply and charge pump rather than a flyback converter.



Here's the response (350MHz TDS460), top being direct coax, bottom 10x probe on same node (obviously, the coax cable is longer than the probe cable, so it's time shifted).

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

Offline BravoV

  • Super Contributor
  • ***
  • Posts: 7547
  • Country: 00
  • +++ ATH1
Re: Super simple "fast-edge" pulse generator for scope-based basic TDR
« Reply #19 on: February 05, 2014, 05:46:30 am »
T3sl4co1l, suggesting you to read the Jim Williams Pulse Gen thread, if I'm not mistaken, you need to have tank capacitor to have a flat line after the pulse top to get accurate measurement.

Offline BravoV

  • Super Contributor
  • ***
  • Posts: 7547
  • Country: 00
  • +++ ATH1
Re: Super simple "fast-edge" pulse generator for scope-based basic TDR
« Reply #20 on: February 05, 2014, 06:16:53 am »
Alan, please, if you had the chance, really love to see the use of JW pulser at the 33GHz beast there to demonstrate the TDR, a centi meter resolution or use to profile the cable velocity factor maybe ?   :-X  ..<drooling>..

Guys, just fyi, this w2aew's screen shot of the beast when he tested at his JW pulser, a braggin 150 pico second.  :o

« Last Edit: February 05, 2014, 06:27:06 am by BravoV »
 

Offline JBeale

  • Frequent Contributor
  • **
  • Posts: 302
Re: Super simple "fast-edge" pulse generator for scope-based basic TDR
« Reply #21 on: February 05, 2014, 07:22:54 am »
I don't have any exotic scope, and with a different probe and cable setup I see only 4.6 nsec risetime. But just for reference I wanted to show that the Raspberry Pi can generate a pulse good enough for measuring cable length, without other circuits. In this case looks like I've got a 120 nsec round trip. You can see the pulse polarity switch when I change from end-open to end-shorted.  I'm connecting the R-Pi GPIO directly to the SMA connector, so the negative-going return pulse gets clamped by the ESD-protection diode making it look more square.

I also tried using a series resistor, a little 0603 part between the GPIO pin and the coax feed but it kept breaking off.  With no resistor the return pulse amplitude is clamped more by the output driver trying to hold the pin at ground, but you can still see it well enough.


« Last Edit: February 05, 2014, 07:29:35 am by JBeale »
 

Offline Redcat

  • Regular Contributor
  • *
  • Posts: 74
  • Country: de
  • Ask why not, not why.
    • redcatimaging
Re: Super simple "fast-edge" pulse generator for scope-based basic TDR
« Reply #22 on: November 29, 2014, 06:17:45 pm »
Hi Alan, I have seen your video some days before and decided to build up such a nice fast edge oscillator circuit based on a 74AC14N today on this cold saturday morning.
As always a great video and something interesting to play with :). Had only a 100nF cap, so it's RT is more like 5-7ns but thats great for the moment.

I would have a question: I have seen the so called "blue ring meter" (but i don't have one) on youtube, where you can measure an inductor/coil and see it's quality or condition based on the resulting ringing after a fast pulse was applied...
Hmmm...it came to my mind that your TDR might give a similar result - so ringing or more ringing based on the coils condition (?) - I would like to test some e-motor windings.
Can it be used for this application (would be great) and how would it be connected?

Thanks and nice greetings, Tom
Voltcraft 630-2,Tek 2215A,Tek 475,really handy DIY microcontroller component tester (R/C/D/Q...), ZD-915, ZD-931,Voltcraft 1062D - of course hacked :)
 

Offline marcusj

  • Newbie
  • Posts: 1
  • Country: es
    • My Blog
Re: Super simple "fast-edge" pulse generator for scope-based basic TDR
« Reply #23 on: April 28, 2015, 09:35:56 am »
For what it's worth, I also did a PCB build of Alan's circuit. I posted the full details on my blog, but here's a couple of snaps of the finished result. I had the PCB's made by Dirty Boards PCB's in China.
« Last Edit: April 28, 2015, 09:52:41 am by marcusj »
 

Offline dom0

  • Super Contributor
  • ***
  • Posts: 1483
  • Country: 00
Re: Super simple "fast-edge" pulse generator for scope-based basic TDR
« Reply #24 on: April 28, 2015, 11:25:04 am »
and no I don't have any 50ohm terminators at the moment so I can't just use a T connector to make the input 50ohms

Those things are crap anyway. For fast signals you need a feed-through terminator.
,
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf