EEVblog #453 - Mysteries of x1 Oscilloscope Probes Revealed

[BiOzZ]

THANK YOU
my professor told us too do that once but never explained why so i forgot it and all the workbooks say use X10 so i just did it but at home i always used x1!
flipped my probe over ... 100mhz/6mhz!!!! i cant even think of how i might have fucked things up ... its like when i was 13 and always had my electrolytics upside down!

[xrunner]

I've been in electronics my whole career, but I've never thought about asking this question: Why was x10 decided on in the beginning? Why not x12 or x15 or x20?

I imagine it was so it would be easier to do the math in your head, but now, perhaps a better multiplier would be realistic and/or give better performance?

[BiOzZ]

I've been in electronics my whole career, but I've never thought about asking this question: Why was x10 decided on in the beginning? Why not x12 or x15 or x20?

I imagine it was so it would be easier to do the math in your head, but now, perhaps a better multiplier would be realistic and/or give better performance?

id say its easier to do it with a 1M input impedance, easier to do the math your self, easier for old scopes to calculate and now all scopes work on the OOM scale .. 1X 100X 1000X exc ... sort of why we still use QWERTY even tho it was designed to slow us down

and besides any more than 10x at like 5V noise is going to be more of a problem and scaling exc

[rf-loop]

Making passive 1x probe so that capasitance is low and resistance is high there is real problem. It is nearly impossible.
Ok, there is some old 34MHz model what have 32pF / 1Mohm (Tek P6101A01.  (I have not seen under 1m cable what still may be littlebit better) Perhaps also some others nearly this.) Also current Agilent N2870A is 1x and 1M, 39pF

But if it can be other than 1x but still useful with low levels, example 2x, situation is totally different.   Example:
Attenuation  2X Bandwidth  500MHz, Input Impedance at the Probe Tip 2 M?, 12.7 pF, Cable Length  1.3 m. (Tek. TPP0502)

This is best passive probe with low voltage divider ratio and "high Z" input what I know.
(but also it is not cheap. Good probes are not cheap. One probe and you can buy 2 entry level DSO) (also 12.7pF is some high for 500MHz, but this is 2x probe)

So, if break this 1x, 10x, 100x etc "rule", it may give some more.
But who want 1.27x probe or 1.57x probe. So, who want make these. No one.
(some scopes have fine adjust for probe factor)

[EEVblog]

I've been in electronics my whole career, but I've never thought about asking this question: Why was x10 decided on in the beginning? Why not x12 or x15 or x20?
I imagine it was so it would be easier to do the math in your head, but now, perhaps a better multiplier would be realistic and/or give better performance?

x10 was chosen in the days of analog scope when you had to multiply the scale factor yourself. Anything other than x10 or x100 would have been a pain.
x10 is also the classic "order of magnitude" difference that engineers love, and in this case it would have been the only logical compromise between attenuation and extended bandwidth while making the math easy.

[poida_pie]

I found these two links very informative on CRO probes and go down the rabbit hole of transmission line considerations.

http://www.dfad.com.au/links/THE%20SECRET%20WORLD%20OF%20PROBES%20OCt09.pdf
'n
http://www.sigcon.com/Pubs/straight/probes.htm

[ddavidebor]

So... If i destroy an old probe and connect the tip direct to a coax, can i get a decent bandwich?

[Pentium100]

Very interesting video. I always used the probes with x1 setting (unless the voltage was 50V or more), now I'll use the x10 mode unless the voltage is too low (I rarely measure very low voltages and when I do, it's usually audio frequency).

[landswipe]

I have a ds2000 and the 350mhz rigol probes that came with it.

I noticed that above the 1x 10x switch there is a rubber? plug, if you remove it, there is another adjustment, how does that fit into this?

Thanks

David.

[Fezder]

nice video, i thought 10x was only for higher voltages....newer knew about bandwith limits, i thought they are the same at 1x/10x ranges....thanks dave!

[cengland0]

i absolutely love the technical nature of these videos.  I learn something every time.  You rock Dave.

[Salas]

nice video, i thought 10x was only for higher voltages....newer knew about bandwith limits, i thought they are the same at 1x/10x ranges....thanks dave!

As the frequency increases the probe's impedance goes down and its voltage spec goes down too. When quoted in the hundreds of Volt at 10X mode, that is only true for few decades of kHz. It ends up to being a low voltage probe near its bandwidth limit. Another not much talked about catch to watch for.

[Fezder]

nice video, i thought 10x was only for higher voltages....newer knew about bandwith limits, i thought they are the same at 1x/10x ranges....thanks dave!

As the frequency increases the probe's impedance goes down and its voltage spec goes down too. When quoted in the hundreds of Volt at 10X mode, that is only true for few decades of kHz. It ends up to being a low voltage probe near its bandwidth limit. Another not much talked about catch to watch for.

hmm, interesting, and bit confusing, but makes sense thought .

[alm]

There is usually a curve of maximum input voltage vs. frequency in the datasheet. It's often flat for the 'DC' part of the spectrum, then goes down gradually until it reaches the lower limit which might be in the order of 20 V.

[KarlMonster]

So... If i destroy an old probe and connect the tip direct to a coax, can i get a decent bandwich?

Bandwidth yes, but ugly signal. As I think was mentioned in the blog, you need a 50 ohm termnator at the end of the coax before you attach a probe.

I believe (since Dave used bare coax - I didn't see any terminators) that the reason why unterminated coax works between the sig gen and the 'scope is because of the impedance of the sig gen output. The 'scope input is about 1Meg and the coax is rated at 50 ohms, but the sig gen output is something on the order of hundreds of Mega ohms. That's to prevent external signals from wandering upstream into the sig gen. As was also mentioned in the video, ohm specs that are orders of magnitude less can often be ignored; the sig gen dominates the circuit.

So bare coax needs to be terminated for circuits with less than a hojillion ohm impedance.

It's given me an idea for building an active 1x probe, but I'm new to terminating coax. This'll be fun!

Massive Edit: I did not understand what "termination" meant. I thought it was just a resistor in line with the conductor. Some coax has built-in termination, as Dave's cable probably does - though he didn't say so. Also; if you have a terminated coax cable, why attach a probe? A slightly longer conductor sticking out of the end should work fine, unless you need ground clip or other probe attachment.

[Computeruser]

LapTop006I wrote:  "wonder why FET probes aren't more common then"

JackOfVA  wrote:  "FET input gates are limited for maximum voltage to reasonably low values"

I have a Tektronix P6045 FET probe that I purchased used some years back. It still works. It is a 1X probe with a high impedance / low capacitance FET front end and an active amplifier/compensator at the 'scope end. It was very expensive new.

I foolishly attached it to a circuit that had high voltage on it and blew up the front end. Fortunately, when I did that, I could still get (and did get) a refurbish kit from Tektronix and fixed it. It still runs and I limit it to low voltages.

I use passive 10X probes most of the time, and if I need a 1X probe, I accept the limitations.

... C

[pfm]

Just came across this blog and video when searching for the answer to the question of why are 1x probes so low in bandwidth. I also read the pdf and noted the comment -
 "The effect of varying the cable resistance over the range 50 ~ 200ohm per meter can also be seen in Fig.10. Low resistances (50ohm/m) allow transmission-line reflections to build up, giving a peaking response. Larger resistances (200ohm/m) give an overdamped, sagging response. The optimum cable resistance was found to be around 115ohm/m. This gave a response which is substantially flat to nearly 600MHz!"


So based on this would it be practical to just use the shortest coax possible with ~100ohm resistor in series ? Perhaps a trimmer capacitor in parallel for compensation purposes. Wouldn't that be like the perfect 1x probe ?

Just another thought - why not just use a twisted pair cable ? would that be any better (or worse) than a coax probe ?

[zeaflal]

Just wondering if anyone knows of a source to buy scope probe cable.  We had purchased some years ago and are now looking (without success) for more.

Thanks.