Perils of Probing: What's Inside That Scope Probe?

Did you know scope probes are not just wires?
In this 8 minute video Jack explains the real issues behind scope and logic analyzer probes, and why using one without understanding the issues can cause your embedded system to fail. Firmware people take note!
For far lots more about embedded systems (including over 1000 articles), visit www.ganssle.com."

[@TheNoBSZone]

As a fresh EE graduate I can't begin to thank you enough for sharing your knowledge. It's so clear you have invaluable experience and I'm so grateful to learn from you. I do have one request -- perhaps you could do a video on job interview advice particularly for those who are just beginning technical interviews? Thanks again.

[@deangreenhough3479]

Great explanation, I'm new to scopes and until now the whole probe issue has had me confounded. Thank you for sharing and providing a sure footing on which to proceed. You present the subject very well. I thank you kind sir

[@garthhowe297]

What an awesome video...I didn't realize this was such a problem.

[@isj-3227]

I get a very Fred Rogers vibe from your videos. Which is a positive thing, he was a very pleasant educator in his method of speaking.

[@daveb5041]

I always buy my probes at walmart. Also when I have a real technical question I ask the staff and they are very helpful. I learned how to do a differential equation that solved a high frequency harmonic problem with one of our boards that not even our my most senior EE could figure out, from a walmart associate in their electronics department. Donate your probes to good will? Poor people deserve better then a 50 pf probe, their life is bad enough as it is!

[@jafinch78]

Thanks a lot! Wow, that is really clearly explained and helps to re-enforce what I've not unified until just now, i.e. resistance, reactance, and impedance. I never connected so clearly for some reason. Amazing how the DUT really is part of a total system including the environmental systems conditions technically even, not just the internal and boundary systems conditions.

[@GlennPierce]

Thanks for the explanation. For the DIY probe how is the 0.5 pF calculated ?

[@wither8]

Best blog ever, Jack

[@CamiloMendivelso]

Simple and to the point, i like a lot your videos, greetings from Colombia!.

[@Neighbour_Al]

I was at Tektronix when they came out with their first 1GHz active probe. Amazing device with the active element embedded immediately behind the probe point. Yes, it was spendy!

[@HansBaier]

This was excellent! Where can I find a bit more details about that awesome DIY probe?

[@jeremyli2871]

Real good job Jack. Thank you for sharing.

[@daveb5041]

I use my walmart probe to fix crosley record players that are defective straight out of the box. Usually a lose connection from the hot glue. The $9.97 walmart probe was much better then the $99.95 radio shack probe that was just an alligator clip to a BNC that had a really fancy box. The radio shack probe was 50% off (normally $199.95) so I must have got a deal!

[@minimamo139]

Oh wow, Thank you. Very Helpful and at least I didnt had to listen to a very strange voice with creasy eyes, and a lot of cheap talk. :) Very Informing, Very interesting and very Direct. The Best on the internet so far :)

[@das250250]

Wonderful explanation jack ..

[@keen2461]

Excellent video! Subscribed.

[@SureshKumar-nk2ok]

thank u very much for this informative video.sir iam facing a problem with serial port cable ,how to check these cables

[@geezergeek1637]

These are great. Very helpful. Thank you!

[@BryanByTheSea]

Wow, very interesting. Thanks for the video

[@MathieuStephan]

Keep up the good work! Mathieu from Hackaday

[@PelDaddy]

Fantastic, educational video. Thanks.

[@jmsian3505]

I was checking the capacitance per foot of rg58u it states around 25 to 30 pf per foot. How did you come up with the .5pf figure?

[@frankbuss]

What is the disadvantage of your $2 0.5 pF probe compared to the k$ probes? If it is just inconvenient to solder it, I could live with it.

[@Germanwtb]

Maybe the precision, +-5% resistor? Things like that

[@vonnikon]

The primary disadvantage is that it puts a 1050ohm resistive load on the signal you are measuring. The 3200ohms mentioned in the video is incorrect. The advantage of this probe is that it has much higher bandwidth than regular passive probes. It is a good alternative to active probes when measuring high frequency digital signals where you typically have a good stong driver.

[@xgentecchannel2309]

Hello Jack, I was wondering if you would consider doing a video on microcontroller output jitter and phase shift. For instance I designed a transformer pwm controller adjustable duty time and dead time based around STM32 maple mini however it is getting like a ghost echo of the square wave on my oscilloscope where the frequency is jumping between 220khz and 180khz approx. I think it is period jitter. I have matched the output impedance eradicated all ringing. No one seems to address this issue with microcontrollers only point out what jitter is. This also affects atmega 8 bit also, One person pointed stm32 jitter corresponding to ppl however what causes this jitter and how do I go about eradicating it if possible. Can microcontrollers be used for PWM control reliably. I am changing to MBed using ARM processors soon and I am a completely self taught electrical engineering and system design and cad design hobbyist. Hope you can help. many thanks.

[@ChaosHusky]

This certainly explains why loudspeakers are being referred to as having an Impedance of 4/8 Ohms etc... Frequency/Reactance! :o

[@pesshau6508]

+ChaosHusky Yup, and also why most audio amplifier power output figures are pretty much useless on their own.

[@Alo762]

Actually, the DIY probe presented has at most 1kohms impedance as the reactance of the 0.5pF goes in parallel with the 1kohms resistor (actually 1kohms + 50ohms, but that is nitpicking). When the frequency goes up, it gets below 1kohms, according to the equation at 1:30.

[@vonnikon]

Indeed! The diagram is completely off. The DIY probe is a very good cheap high bandwidth solution for measuring digital signals. But not quite as good as that diagram would indicate...

[@JackZimmermann]

Great stuff! Thanks!

[@beehive808]

Thanks for the great video

[@garynorris4648]

Superb

[@wadehicks9270]

Good info thanks for sharing

[@KananDethin]

Thank you

[@bastonbalexendo9659]

tanks a lot

[@leonardmilcin7798]

I work to buy a house. I am not going to sell it to buy a set of probes:) For us mere mortals we need to accept that above 100MHz bandwidth there is no fly zone where we basically can't see what actually is going on and have to work intuitively supplemented by a piece of paper, a pencil and a calculator. The Catch-22 is that to build intuition you need to see what is going on, first. Or watch youtube videos from people who had chance to build it and want to share it. For which we are grateful.

[@alpedistas1]

awesome video, subscribing!

[@vidasvv]

Great video, tnx for uploading! 73 N8AUM

[@discoHR]

How good are 1M/10M scope probes? Is it safe to poke around a 1 MHz computer with a 1X (1M) 100 MHz scope probe? I've noticed that it doesn't mind multimeter probes as long as I don't touch one of the pins on the RAM chip, if I do then the memory contents get corrupted.

[@discoHR]

Thanks.

[@alanhillsplasmatweeter7318]

This is a very useful video. My one issue is declaring the specs "a Lie". The specs are correct: the resistance and the capacitance are both given. The user must understand the limitations from this data. You did a good job of teaching this, but the accusation was unnecessary IMHO. I am not a scope probe manufacturer BTW. Just another EE.

[@arthurboulton4511]

Very informative than you

[@bdnevins]

The 21 X probe should be 0.5pF//1kohm, the Z at 100Mhz cannot be greater than 1kohm!

[@umajunkcollector]

.Howz about WIRELESS probes? Go digital microwave RF, eliminate the cables. Make a microtransmitter probe, likened to bluetooth or wifi, that can be inserted into a circuit.

[@DanielReyes1]

+umajunkcollector The antenna for the bluetooth transmitter would essentially be a wire at the very least. You haven't changed anything. Just added noise from the bluetooth transmitter.

[@steverobbins4872]

Jack, you don't know what the hell you're talking about. Stop confusing and misleading people.

[@TheNoBSZone]

+Steve Robbins Lol, nobody takes you seriously... I'm an electrical engineer and off the bat can't see any bad information in this video. Either make a logical argument or shut your mouth.

[@steverobbins4872]

Fair enough. I should have pointed out a few specific problems, but there are so many that I didn't know where to begin. 1. At time 1:35 he shows the formula for impedance. It is wrong. Impedance (Z) has a real part (R) and an imaginary part (X). Thus Z=R+jX where j is the square root of -1. 2. At time 4:49 he shows part of a data sheet for a 2-input AND gate, with with his calculated currents added in red. Clearly, he doesn't know what he's doing. The data sheet shows a few points on the DC output characteristic of the gate. For example, if the gate is trying to pull its output down to ground, but there is 3mA load (from an LED for example) then it can only pull down to 1.1V, not all the way to ground. Now, what has this got to do with a scope probe? Nothing. Here is how to do the calc properly. Let's assume the probe impedance is say 2.2pF||10M, and we're probing the output of a 3.3V logic gate with a 3ns rise time. The correct equation to use is i=CdV/dt. (i=current, C=capacitance, dV/dt is the Voltage slew rate.) So the current into the scope probe capacitance is (2.2pF)(3.3V)/(3ns) = 2.42mA. (BTW, it's "mA" not "ma" because units that are named after people are always capitalized.) 3. At time 5:35 he says "you take a meter coax, you solder a 1k resistor to the end, and now suddenly we have a 0.5pF probe". Of course the scope input must be 50 Ohms, to match the coax. The first problem is that he neglects to mention that this is a x21 uncompensated probe. The impedance at the tip will be 1050 Ohms at dc, and the rise time will probably be crappy. But at time 6:30 he says the impedance is 3200 Ohms, which is silly. And where does he get "0.5pF" anyway? 4. He says you should get rid of any x1 probes in your lab. Gee, I guess the whole industry is a bunch of idiots for using x1 probes. He doesn't even know that the reason for having a x1 probe is that is is generally higher bandwidth than a x10 probe. 5. He doesn't mention one of the most important issues that should be including in any tutorial about probing: the ground lead inductance. The main advantage of his home-made probe is that he is connecting the shield of the coax directly to the ground plane of his breadboard, thereby avoiding the ringing that is normally caused by the inductance of the ground lead.

[@steverobbins4872]

Oh, and I forgot to mention one of the main things I hated about this video. He doesn't explain the real problem that he is trying to get at. Basically, very high speed signals on a circuit board are conveyed on special traces that have controlled impedances. In other words, these are transmission lines. If you probe one of these traces, it makes a reflection that distorts the signal, and can cause bit errors. The reflections have nothing to do with the driver (e.g. the logic gate he refers to). For example, the tip of the probe (being capacitive) will look like a short on the transmission line. (Assuming the probe ground is in direct contact with the ground plane, rather than using the inductive ground lead.) Depending on where this short is applied (trace length from the signal source) the source may see a low impedance, or a high impedance. Okay, transmission line theory is complicated (but worth learning!) and I can't go into it all here. But the bottom line is you can't probe controlled impedance traces with an ordinary scope probe.

[@vonnikon]

+Steve Robbins Believe it or not, but DIY coax probes like this actually have much better bandwidth than regular passive probes. Pretty much comparable to lower end active probes. Rise times are very very good. But the disadvantage of the DIY probe is that it puts a high resistive load on the DUT. The 3200ohm impedance mentioned in the video is of course incorrect. It is 1050ohms at DC, and dropping slightly at higher frequencies. I tend to use 450-470ohm resistors for DIY probes, since this creates an even 10X probe. But if the DUT can't handle the resulting 500ohm load, then I go for higher values such as 1k or 2k2. All in all these DIY probes are often a good fit for high frequency digital signals. But probably not for analog.

[@hugocoolens]

@@steverobbins4872 Dear Steve, I agree with most of your arguments. The only thing I don't agree with is "He doesn't even know that the reason for having a x1 probe is that is is generally higher bandwidth than a x10 probe", as a probe in its x10 position typically has about 10 times greater bandwidth than the same probe switched to x1.

[@FingerinUrDaughter]

wait wut. impedence is measured in OHMS? last time i checked it was measure in HENRIES.

[@Graham_Wideman]

Henry is the unit for inductance, not impedance.

[@FingerinUrDaughter]

@@Graham_Wideman indeed. i have no idea why i typed that.

[@itsmylifebitchodjodjjbbc1246]

hijack my name is Jay I was hoping maybe you can help me out with a couple things I'm trying to make a flash chip platform. And I have limited equipment and I was wondering if you can help me will be surprised and wonderful to hear from you thank you have a wonderful life