I built the TDR circuit from QST years ago, and I used it at work with the better scopes to check coaxes on the reel, and the boss was so impressed, he went out and spent a ton of money on a box that had a microprocessor that did the math for you. But it never worked better, nor was it built under $5.00!
Hello Allen - on 0:21 you say you've made a video a couple of weeks ago which show the basics of TDR. The only video I can find about that sub. is your video nr. 37 from Mar 8, 2012 from A YEAR EARLIER. Which video # do you refer to at at 0:21 ?
Great video! Can this be used with a high speed 50 ohm input oscilloscope for TDR measurements? I believe the scope you used with this pulses is high-Z input. Does the source (pulser + scope) need to match the cable?
Why do you have to match the impedance of the output to the cable? Is it because the reflected wave would be shorted to the same potential than the output if you fed the cable directly from the output so that you could not measure it on the scope?
The output impedance of the source is ideally matched to the transmission line so that any reflections are perfectly terminate and don't reflect back down the line again.
w2aew Imagine that the signal at the output of the Schmitt-trigger is instantaneous, i.e. has zero rise time. When the transition occurs, the series resistor, which is there for impedance matching, and the capacitance and inductance of the cable, together form an RLC circuit, and the signal takes a non-zero amount of time to rise, no? So, isn't the rise time at the end of the cable necessarily longer than at the output of the Schmitt-trigger, i.e. on the other side of the series resistor?
One thing that you have to remember is that a 50 ohm coaxial cable appears like a 50 ohm resistor to RF when properly terminated, not an RLC circuit. So, the combination of the 50 ohm output impedance and the properly terminated transmission line looks simply like a resistor divider.
Makes my heart happy to see someone using an RPN calculator. I get the video is almost 10 years old, but had to comment on the lost art. :) Had to comment that this technology is also used in optical telecommunications networks, but is called OTDR (O for optical - of course). We use it every day to find optical fiber cuts/damage.
in the late 70s I couldn't wait to crack open my new Popular Electronics magazine just to build the next new project in what seemed to be a magical time for electronics hobbyist and home inventors. And I have to say everyone of your videos brings that magic back X10 in a way we could only have dreamed of in the 70s. Thanks a million for all the great videos.
Suddenly my 40-meter yagi had a high SWR. I used one of these little TDR circuits and determined the exact location of the problem. I'm sold. Thanks! 73 de K8CU
Great idea and excellent circuit! When I worked for the Navy I used Pulse Generators that cost hundreds of thousands of dollars, your circuit is outstanding, I have got to build one for myself. Thanks!
Great idea, thanks for posting! There should really be a hex inverter cookbook! Teachable suggestions: in addition to open-end, you could also show terminating the coax properly, and then shorting it. Many students find the result counterintuitive. Even cooler, show the far end voltage (with a low-C probe) on your second scope channel!
Hey there, do you have the files still for your pcb for this TDR, I need a 2ns square wave to externally calibrate my Tektronix 2430A as I'm going to have to change the NVSRAM and will probably lose the calibration data, I'm going to try and read it with an eeprom programmer but doubt the information is easliy transfered this way. Thank you
I think a good quality 100nF decoupling capacitor should be used. It will be from that capacitor that the current for a nice rising edge will come. This capacitor should be as close to the chip as possible, as shown in your circuit (newbies might try putting at the power supply). A good quality means low ESR, so look for COG/NPO (see wikipedia en.wikipedia.org/wiki/Types_of_capacitor) .
i wonder if adding more capacitors of an ordinary quality will help in case sourcing a high quality one is difficult. having two caps in parallel will make their resulting ESR r1*r2/(r1+r2)'ed, won't it?
Thanks for a great video! I built it to find a break in a Romex power cable. Got 1.2ns/foot. Used a 74LS14 as that was all I had available. Needed to change the 6.8 kOhm to 1 kOhm to get it to oscillate.
Man, you are a genius! I just build the circuit and it works great, I measured a 3m cable and it gave a result of 3.003m. I used the CD74AC14M96, SOIC-14 version from texas.
Kinda threw me a curve ball there. In my world the acronym TDR stood for time delay relay. Back in 1971 I set up and used this technique for testing coax in aircraft. If you ever had to take an airplane apart to test the cabling you would understand how valuable this is.
Great way to determine the impedance of an unknown bit of coax. (video #37) You can also calculate the VF for an unknown bit of coax. Take a measured length and back-calculate what the VF *must* be to get the measured delay from a known length. Thumbs up for bothe the #37 and #88 videos. (this comment posted on both videos)
+Peter W. Meek I did a video on that too (measuring velocity factor), exactly as you've stated. ru-vid.com/video/%D0%B2%D0%B8%D0%B4%D0%B5%D0%BE-TpIIftvQPFM.html
These are very informative videos and have answered a lot of age old questions in my head. Many thanks. Can I ask why you chose to utilize the remaining five schmitt triggers in parallel, rather than just add 50ohm resistor to a single schmitt output?
Great video Allan! I`m just re watching your video on this TDR pulse generator, which I first watched maybe 2 years ago, which inspired me to build it re purposed for something else. Two years ago I was looking for something that I could use as a fixture to calibrate a 60Mhz oscilloscope's step response. A "poor men's" approach for someone who didn't have a fast edge pulse generator (
The trick with AC14 is to use 2/6ths of the chip as a cable driver, and the 4/6ths to lower the impedance of the power delivery to the chip. How? Feed “0” to 2 inverters and connect their outputs directly to VCC. Feed “1” to another 2 inverters and connect their outputs directly to GND. You’re now driving the on-chip power distribution through two extra pins per rail. It improves things quite a bit. You need more chips then of course, and their inputs and outputs must be respectively equidistant. So the oscillator must connect to all the inputs via a wire/trace of identical delay. Same goes for the outputs connecting through the resistors to the BNC. With some extra care that chip can drive just under 2ns edges, extremely cleanly.
@@absurdengineering wow, what a trick, live and learn. i suppose it should also apply to other chips as well? for example the AC00 that I used in a similar little project.
@@s_s-g4d Yes. But always measure stuff and use an objective way to see the difference, like capturing a reference trace on the scope using a straightforward circuit, and then observing how the “improved” circuit’s output compares to the reference trace. It should be done using the intended load too, ie. 50 or 75 ohms or whatnot, properly terminated.
FANTASTIC content, congratulations!!! Based on the project in this video, it is possible to develop a UTP network cable tester that can detect the distance of each pair of cables and thus check whether the quality of the cable and connectors are good, perhaps using a microcontroller to transform the information timing in metric measurements???
Sure it would. You'd probably need a more precise way to measure the time delta between the steps in the waveform - maybe a gated integrator or something like that.
@@w2aew I'm good at programming microcontrollers, the digital part, but I'm a bit of a novice in advanced analogue, I don't have the slightest idea of how to make this closed integrator, it would be very interesting if you, when you can and have time, maybe make a video designing one prototype to talk to an Arduino for example, I think there will be many electronics, networking and electrical enthusiasts interested in this project !!!
Is there a simple way to perform TDR using this device on an unshielded cable? Will this work on a coil, as in a giant solenoid (#10 awg cable with unknown length). I am trying to find a way to determine if a short exists in the winding.
Just build one. Dead-bug style. Used 1206 for the buffer R. Works like a charm. One question though about measurement of the rise time in general. As this circuit is ringing, what value we take for the upper voltage? I was thinking maybe to use the 'end-value' (stable value) of the square wave, but that would probably not be right? I suppose we need to account the highest/max value? When accounting the max value I get ~2ns and if I take the stable value I get 1.76ns. Power supply was 5.0V
...and did some TDR measurements. The results that I got are: 200cm coax RG58 I get 185cm, and 20m cheap RG58 coax I get 16m. The conclusion is that the puls is coming back faster. In the case of 20m it's 25% difference, that I do not understand yet. Help 😀 If it would be slower I could understand, it's a cheap cable, but faster? I come with the propagation velocity factor of 85 [% of c] instead of 65.9% that I looked up for RG58, and what you also stated as typical coax speed. So something is off.. I get good clear reflection measurements with a puls of about 2ns. Scoop's input was terminated as 50 Ohm (but maybe was not needed). The TDR-PCB was direct mounted on the scoop input, and a T-split to the DUT/coax.
Love it! I’ve built a few 80:20 instruments like this to get at least 80% of the functionality at way less than 20% of the cost and complexity of a commercial instrument.
Search my channel for the super-simple TDR. If you have a digital or storage scope, you can do this with nothing more than a battery, a resistor and a couple of clip leads...
Looks like the "typical" rise time is less than half of that - cool! Parallel the outputs so each doesn't work so hard, and you can likely get a pretty speedy output!
I have this ringing with higher amplitude (about 2v p-p) in both edges at rising ede and in falling edge that made a negative voltage ringing I have decreased my ground lead inductance and did every grounding truthly but icant remove them what can I do????
You should have described also why the waveform looks like the way it does, it goes up half way and then after a bit goes up the remaining half distance.
Hi Alan - this is a very late comment to your TDR video (its now 2/01/2015); as a total newbee to this medium firstly let me congratulate you on your perfection, schematics and absolute clarity you present these details. As a semi-retired scientist (analytical chemist, geologist & gemmologist) who is now getting back into main stream electronics after a 45 year break - you are truly a breath of fresh air!!! I have read every reply since when you first posted this video and have not been able to get any details on your "decoupling capacitor" used - I have searched in vain on the internet the numbers that are visible on the yellow square plastic cap; viz '2073G, L1104 & E8515', and have drawn a total blank!!! Would it be possible for you to provide this cap details, its type and how you have connected it to your TDR apparatus as this is not at all clear in the video??? Also explaining what the 'decoupling cap' is actually doing to the voltage supply (a wild guess on my part, without knowing, is that it is removing or reducing to a very low level any AC ripple on the incoming DC voltage)??? Further, do you have a complete circuit diagram - including the decoupling cap - for your very cleaver piece of kit? I am wishing to build your TDR as I can see this will be very instrumental in teaching me how to use my very new 100MHz digital scope as well as providing a very practical application for it. If I may comment to the very few rude people who criticise your use of imperial units - I would say get a calculator and do the maths yourself if this is an issue for you; we all use what is or native tongue and what metrics we are totally familiar with! To the few who have complained about you saying "OK" every sentence - to them I say; "yes it can be distracting and break one's concentration when a viewer is hanging on your every word", but to them I say "suck it up" and move on if you can't really appreciate all that Alan is saying and presenting!!! Your adverse comments demean this guy and all the good he is doing for us very new and perhaps not so new 'Newbees'! Keep up the excellent work, mate! Cheers, Ian (from Gippsland in Victoria, Australia)
😅 you probably already know this by now, but a “decoupling capacitor” is in 99% of cases generally any capacitor with a value of between 0.001uF and 0.1uF. And you are correct in that it’s purpose it to remove any high frequency interference (you could call it ripple but generally that refers to lower frequency variations) on the DC power.
Alan, thanks for the great video. It prompted me to build my own 9 volt powered time base generator based on a 10 MHz clock oscillator. I wanted something a little more versatile so I I'm using 2 74HCT74 DFlops to give me a range from 10 MHz to 625 KHz. I tried it at the lowest frequency with an ~ 5' unterminated cable and lo and behold it worked! I got a 17 nanosec delay which after calculating the length equaled 66.1 inches which was pretty close to the length of the cable I made. Isn't math and physics great!
Great video, as always. Maybe I would have liked a bit more explanation about the Schmitt Trigger diagram. I understood that the first trigger was connected to VCC as well (we don't see that) and outputting VCC into the 6.8 k resistor as long as the 47 nF cap was discharged. Is that right? And why do you need the 5 other triggers?
+fabts4 There is more information on the schmitt trigger circuit in this video: ru-vid.com/video/%D0%B2%D0%B8%D0%B4%D0%B5%D0%BE-NuXitMK3HSA.html I used all the inverters in the package to increase the drive capability.
+w2aew My question was really about the schematics and wiring. You have to input a signal or connect the trigger to VCC, I didn't ssee that explained in #87 either. Is the current substantial, then?
+fabts4 The VDD pin is connected to the power supply, but there is no other signal injected. One gate is used as the oscillator, and the remaining are paralleled together as a buffer. The output drives into a 50-100 ohm load - which would be too heavy for just one gate, so that's why I paralleled them.
I saw a circuit which uses a single-transistor avalanche pulse generator. It needs a high-voltage power supply to work, but the advantage is that you can fit the entire pulse generator inside of a single BNC connector.
The ringing looks like a band limitation of the circuit. A band limited signal is a sum of weighted sines and since all real, physical signals are necessarily band limited (because of all the inductances, capacitances and QED in general), so is this one. In this case is looks exactly like a band-limited step response. The shortest constituent in your signal has a period of around 7ns (140Mhz), which sounds right compared to the advertised rise time. To remove the ringing you would need much better rise time (higher bandwidth). At some point the scope will low-pass filter the signal, removing the ringing from your view.
I went crazy enough to make a similar generator with a 16,384 kHz quartz oscillator followed by a 2^1...2^24 frequency divider made of three cascaded HEF4520B ICs with their outputs going into a dual-row pin header (2x25 pins: 24 outputs from the 4520's, plus one directly from the oscillator), from where the signal is routed by means of a jumper into a CD74AC00 quad NAND IC that produces as fast an edge as I could get from what was available for order in my country. ...this is how I found that I need a better oscilloscope :( I also added a virtual ground output at 1/2 the signal output peak voltage to be able to use it in both AC and pulsed DC modes. Btw I didn't care to put those resistors in series with the output, maybe I should, as well as add a USB power input.
Hey Alan, wanna hear a good one. The 74AC14 is discontinued according to Jameco Electronics. I used the HCT14 which has a rise time of 5 usecs. Just for shits and grins I tried measuring cables shorter than 4 ft and it didn't work. But... that's what they make tape measures for. :-D I use Jameco for most of my parts purchases since they give free shipping for orders over $25. Usually I'm lucky if I'm under $50.
GREAT video! Would a single 2032 coin cell (3V) have sufficient current to drive the Schmitt trigger? A better question...if I made a tiny project box to house all this, do you think the coin cell would be a good option for long term usage? Assuming a on/off switch of course.
Hello Sir, we tried constructing the same circuit, but we cannot get the pulse, we used general electrolytic capacitor rather than box decoupling capacitor, we are using DSO in range of mhz sampling frequency. do we need to use MSO?
Hi Allan , Could you give me your email address? I made the same TDR circuit like yours, But it's not working. Could you give me some suggestion ? Thank you in advance .
Hi, I only have a CD40106 Hex Schmitt trigger at hand. Would anything stand against building a square wave pulsing circuit around the 40106 for that purpose? thanks in advance!
Hi Alan, I have constructed your TDR model and incorporating an IC socket so I can swap various Schmitt Inverter chips rather than having a fixed IC. [I have only used SN74HC14N at this stage as I am waiting on delivery of my SN74AC14N). I seem to be having some issues because the TDR doesn't appear to be pulsing (when I set the scope to the same settings as you do) and have just got a continuous output on my scope of 5 V (I am using 5V input from my PS), I then get the same 5V reading when I have connected just the coax from the "tail" of the BNC connector on the TDR to the scope input as you have done. Note the TDR is not connected to the scope other than through the coax cable at this point. When I test the various circuit ground points - without having an IC in the socket - I get the following continuity readings, some of which seem stranged - SHORTED??? But I cannot see any shorted lines or blobs of solder where they shouldn't be??? With one probe of my MM connected to the GND point of the 37nF cap and the other lead touching: 1. Pin 1 on the IC socket - NO continuity as expected 2. Pin 2 " " " - YES have continuity to ground (where 6.8K resis. attaches to Pin 2) Is this normal??? 3. Pin 3,4,5,6 NO Continuity to ground but Yes on Pin 7 - which is IC GND pin. 4. Pin 8,9,11, 13 YES continuity; 5. Pin 10, 12, 14 NO continuity My TDR is put together exacatly as yours is and in addition I have attached a connecting wire from both Caps (common) GND to PIN 7, and then to the BNC connector GND as i have used perforated pcb board. Don't know if I can attach a picture to your comments or not?? But this would help if I could? Can you see any problems with the above connectivity??? Regards, Ian
***** Hi Alan - thanks for your reply however if you can re-read my last post you will see that I have said that all of the continuity readings were taken with the IC chip REMOVED from the socket - and this is what I felt was strange - since all of the points where the resistors are connected is what has given the "YES" results, other than the GND pin (where there is no resistor connected. The point I failed to convey in my last response was that the 47nF cap was a polyethylene type and not a ceramic - I don't think this should matter but I am just a newbee at this! The other point is that all of the resistors I have used are 1 watt rating since I didn't have 0.5 W on hand which yours seem to be. I felt that 1/4 watt were too small! For sake of clarity in my question - with the IC totally removed from the socket - when my MM probe is connected to the GND terminal of the 43nF cap and then in turn to each of the pins where the resistors connect to the socket pins I get the continuity "YES" results I said in my previous post - is this normal and what you would expect to get under these conditions? Sorry to be a pest but this one is really throwing me off the deep end! Do you have the facility for uploading pictures to your comments pages; if so how does one do this??? Cheers, Ian
***** Hi Alan - thanks to your help I have located the problem: Don't really understand why this was causing this problem but I had an earth wire running from the common GND of the caps to Pin 7 GND on the IC socket - no problem with this - but then I had connected a separate wire from the GND wire on IC socket 7 direct to the base of the BNC - which was pulling the resistor connections at the respective socket connections to ground(????) - simply removed this wire and run a totally separate GND direct from the common cap ground to the BNC base and all is good continuity wise! Now to add the chip and start playing with the TDR in earnest - I HOPE!!!!! Cheers mate and many thanks for putting up with my NEWBEE questions.
I want to build one that is a single hand held unit with a mini scope but no scope... let me explain, I want it to have an internal dedicated scope that measures and calculates the time delay reflection and doesn't output a scope trace but rather just a length, switchable between metric and us standard units that measures romex cable... a tool for electricians both to simply measure wire lengths and to find shorts and openings in circuits that may be buried in walls or in boxes for receptacles, switches, lighting, etc. if you know how far down the wire a short is and you have a toner you can find it quicker. I looked and it doesn't seem that they make them in a simple small handheld device, nor do they make them for romex... given that I'm no where near knowledgeable enough to make a small dedicated scope (by dedicated I mean it can be limited to only performing this sole function) and thus it may take me a few years to get started, anyone interested can steal my idea. I'm more interested in it's existence than the clout or money.
I just came back from the future to ask if this wouldn't be a good circuit for a "dead bug"-style setup. I guess that if it is layout sensitive, the answer would range somewhere between "challenging" and "nuts!".
Hi Alan. I've seen a similar design to yours, where there is a "RD" (resistor/diode in serie) in parallel with the feedback resistor, in the actual design, the 6.8k. What would be the purpose of such feedback "network" ?
Hi Alan, Great video's as ever. Do you think it is possible to build a TDR into a box to go portable. I doubt an Arduino is powerful enough to act as a sort of portable scope however a Spartan FPGA and ADC might and use the Arduino as an input for say VF to do the conversion and output to an LCD the length. Beats paying the megabucks for a portable TDR.
Thanks for your excellent video. I have been thinking about how to design a fast rise pulse generator for an experimental circuit I am working on. Your video helps immensley. I want to generate just one fast rise pulse so I have designed a debounced pulse generator that works with both TTL and 4000 CMOS - up to 18VDC supply. I am going to build your circuit.
We had a manufactured TDR which was invaluable for figuring the length of cable on a spool. Very good video showing a low cost way of doing the same thing. Not as fast, but you're not selling to a client either.
hey so we need to build TDR and we are having a 1KHZ SIGNAL GENERATOR SO IS IT POSSIBLE TO LOCATE THE FAULTS IN CABLES WHICH ARE BARELY 10Mtrs LONG WITH THAT FREQUENCY and how can we use this to locate faults in power cables such as electricity distribution line
Yet another great video, when wanting to make phasing lines at home I have used a signal generator, receiver and tee piece, trimming the cable for maximum signal "suck out" using the open circuit 1/4 wave... Your little circuit is a much more flexible solution.. should run for a while on 3 aa cells.
The two problems with using the compensation signal as a TDR source are that it generally won't have a fast enough risetime to give you good resolution, and it may not be able to easily drive a low impedance transmission line.
Once again well presented, very clever and cheap as chips to build, going to have lots of fun with my scope after watching your videos. keep up the good work 73 Phil
Hi I just tried your circuit here but couldn't get a low rise time then it dawned on me that my scopes all have 1 Mohm inputs. I was just using rg-58 between the pulse unit and the cro. I did also try my scopes probe set at 10x but get about 45ns. still no joy.
***** used the 74ac14 and the cro is a 100mhz digital. Its working as it should now. I'm getting about 4.7ns. Did the math on a 77 inch length of rg58 and it came to 77.8 inch
Hi for traveling a very fast rise time pulse signal from one board to another have i use coax??? If I use normal two wires can it decrease the rise time of pulse???
If the propagation delay (length or distance) is more than about 1/4th of the risetime (some say 1/10th of the risetime), then you should consider using a transmission line that is properly terminated. Using just a pair of wires *can* increase the risetime (slow it down), due to the distributed inductance and capacitance.
As always, great work! Do you think that equal path lengths on the inputs and outputs might reduce the ringing slightly? And surface mount resistors to reduce parasitic C?
That's really great to hear. I can't wait to see it if you ever make a video on it. Always a pleasure watching your videos and have built many of your test devices demonstrated in your videos. Thanks and keep up the sterling work. 73 Spence de M0STO