Fast?

I have one of those. It makes a pretty clean step at about the advertised speed. It's amazing for the price.

You can often get an old Tek 11801 or 11802 and an SD24 TDR head for under $1000 used... very used. It does 30 ps TDR. We use them in our test stands and have found some outfits that can repair and cal them for us. Tek never published schematics.

Here's some SD-24 stuff.

Well I thought this would up your street, John. Since you have one of these boards, what sort of resolution could you get with it, do you think? I mean, would you be able to resolve the difference between say one of your SMA savers being in or out of circuit? Or is that asking way too much?

It's just the step generator. You need a comparably fast oscilloscope.

With my 11802/SD24, 30 ps TDR, you can see PCB traces on various layers and the vias. Slide your finger along a microstrip and see the blip move. A good SMA hardline or connector saver is invisible from an impedance standpoint, but it does add measurable delay .

Something like a 50 ps TDR rise time is good for PCB work. Nanoseconds are OK for cables.

With, say, a 100 MHz scope, a fast CMOS gate is an adequate step generator.

Could you see this with a 350Mhz analogue scope or does it have to be a DSO of some kind (probably a dumb question but I'm feeling kind of exhausted right now).

It is a dumb question. 350MHz bandwidth translates into a roughly 1nsec rise time.

Digital sampling oscilloscopes can have more bandwidth, but they don't have to. It all depends on the nature and quality of the sampling head.

An analog scope is fine, except that it won't do the impedance math for you. 350 MHz is 1 ns rise time, so the faster step generator is worth using.

I'm not sure how one would connect the step generator and the scope input to make a clean 50 ohm TDR source. The Tek SD-24 uses a current source right at the 50 ohm sampler input.

A 50-ohm resistive splitter would work. BTW the $50 Leo Bodnar pulser board produces a 10-MHz square wave with ~30 ps edges, and has a separate trigger output.

Cheers

Phil Hobbs

Leo's thing is 50 ohms out. So a BNC tee into a hi-Z scope input would sort of work, but the scope capacitance will cost some speed.

Into a 50 ohm scope you'd need some splitter.

Leo uses a laser driver chip that has lots of output. We use similar chips as level shifters, to boost LVDS or PECL swings. (Hey, I just conceived a new product! Discussion can do that.)

Leo talks about TDR, so the OP could ask him for suggestions.

Mini Circuits sells resistive tees that are good up to at least 6 GHz. You'd lose some amplitude, of course.

The tee connector thing could work if the lengths of the cables are chosen such that the reflections miss each other.

Me mum's an Aussie, I guess I shouldn't comment.

Well, when dumb questions arise, you would be ideally suited to answer them, Bill. In fact I can't think of anyone better. However, on this occasion you were expressly requested to refrain from sticking your nose in:

"Constructive observations from non-Australians welcome." Too dumb to understand plain English are we?

OOOH !! Racist behavior, run for cover !

Cursitor Doom does claim to be able to think, but it's not a capacity that he has ever demonstrated.

As if a request from Cursitor Doom is ever going to be taken seriously.

Cursitor Doom is a little too dumb to appreciate that he's never going to be taken seriously, and much too dumb to realise that I'm not going to skip any opportunity to emphasise just how dumb he is.

There are people here who I treat with consideration and respect - not concepts that Cursitor Doom understands - but it is extremely unlikely that he will ever graduate into that not-all -that-select group.

Being Australian isn't being a member of a race - its a question of nationality. Cursitor Doom was merely being rude - as well as stupid - which is one more of his unattractive habits.