TEK Scope Probe question

Apparently not.

I put a very fast step into my awful Tek DPO2024 (claimed 200 MHz) scope.

Direct Tr = 1.96 ns (it's not a 200 MHz scope!)

+500 MHz Probe Master Tr = 1.92

(All probes at 10:1, zero ground clip length)

Hi John, or anyone, can you explain what this all means?

Ok, so the scope has a Tr of 1.96 ns Does Tr mean transient risetime? if yes, Is it measured 10% to 90% ? How does that relate to BW, the reciprocal is 500MHz so that doesn't seem right.

What is the advertised BW of this probe?

What is the advertised BW of this probe?

I don't see a higher BW probe making the scope BW higher, but a low BW probe... Are You measuring 20ps differences, on a 200MHz scope? Thanks, Mikek

Tr is the risetime, as measured by the scope. It's 10/90.

That's not right. For a gaussian system, BW * Tr = 0.35.

Two of these probes do make the scope faster.

I'm reporting what the scope measured.

"Useable with scopes having Bandwidth up to 100 MHz "

"Bandwidth: 1X: DC-4MHZ, 10X: 100MHz "

Response at 100MHz not stated.

No stated risetime.

No derating curves.

Who can tell?

All the probes I use came with full specifications of 3dB bandwidth, risetime, and voltage derating.

I use PMK these days, less than half the price of the equivalent Tek, and more robust. Made in Germany.

In the days we used Tek probes, they came with full specification sheets.

One thing I've noticed, all manufacturers' probes come a bit over-peaked at the HF end. It's not my vertical amplifiers, I check those first.

How much overshoot, with and without probes?

As I've said before, probes tend to come over-peaked at HF from the factory.

I take it you were using your TDR pulser as signal source.

Ahh, yes, I should have remembered that from The EEVblog about Jim Williams Pulse Generator.

All my probes have spec sheets showing (minimum) 3dB bandwidth, and risetime.

System bandwidth is easy to measure, if you have a leveled generator. Usually Tek stuff comes in a bit better than spec, apart from some digital stuff (Yes, John L, I echo your comments about DPO2024).

Scope bandwidth is specified using a 50 ohm, terminated, source, direct, with no probe. At least reputable makers do it that way.

How probe bandwidth is measured, I don't know. What I am going to do is measure one of my PMK 500MHz probes with a calibrated source, and a vector voltmeter. I'll see if I can lay my hands on some lower-end probes to test, as well.

I'd not run across the effect on lower cutoff, since I rarely, if ever, use AC coupling. I did a quick and dirty Spice simulation, using a probe model I wrote ages ago. Yes it does lower the bottom 3dB point (actually, the -23dB point). I shall now have to do a proper analysis on paper. There goes the weekend :-(

Looks like JL's measurements support my contention that this is a system property, not the probe alone. I'd certain agree that any probe will add aberrations in the step response.

Take a look at an old manual with a full schematic of a scope's input. You should be able to draw a schematic, eliminating the attenuator switches (focus on x1, ignore the other attenuation settings which will all look the same (1Mohm || pF) and HF compensation, then reduce this to the LF transfer function with a few resistors and the single AC-coupling capacitor. You shouldn't need spice to understand its frequency characteristics! {confession: years ago I designed vertical amplifiers (and some other things) at Tek. But probes came from another department...}

Mathematical derivation posted to ABSE

No need.

The input looks like 1 megohm, in parallel with so many pF.

Putting, say, 0.1uF in series models the AC coupled condition.

Then it's a matter of modeling a basic probe, series resistance, and parallel capacitance.

I used spice 'cos it was a quick proof of concept. Took all of five minutes. A rigorous treatment comes later.

Mathematical derivation posted to ABSE.

A little, 5% maybe. Good scopes used to have gaussian step response. Modern scopes tend to be peaked, and have overshoot, so they can cheat and claim more bandwidth. On the DPO2024, Tek didn't quite cheat enough.

Yes. It's under 30 ps.

Speaking of picoseconds, Tek recently bought Picosecond Pulse Labs.

Mathematical derivation posted to ABSE.

Too many copies ;-(

Sorry, posting problem

Tektronix DPO2024 Review

Nice display. Most of the rest is terrible.

Boot time is absurd.

The software (Linux) occasionally ignores rotation of knobs

Force Trigger never works

AC coupled triggering is broken, probably a software bug

Average-N is almost useless; the dumb averaging code seems to add a new sample once a second.

Menus are annoying

It's NOT a 200 MHz scope!

And to add insult to injury, too many of the web sites that I visit show the stupid "Overachiever" tek ad for this very same scope, often multiple times on the same web page. I'd advise you to never google "dpo2024" or it will track/haunt you forever.

The LF thing is simple. AC coupling involves an input series cap and the scope's 1M input impedance. The probe adds 9 megs, so the time constant is 1/10 compared to no probe.

But scopes aren't first-order systems.

I haven't used Google for a couple of years, now.

Ixquick seems to deliver the same results, but without the tracking.

When you have 10:1 attenuation to work with, a little HF peaking is easy. Of course, that peaking will load the signal, but that's what passive probes do.

The scope measures risetime; I just read the number on the screen.

The probe tips were poked into a female BNC connector that had the fast step, and I used a screwdriver to short the probe ground (the one up near the tip) to the BNC shell, so no ground clip lead was involved. At 200 MHz bandwidth, a ground clip would no doubt change the numbers.