Home-brew 100:1 'scope probe.

[Skywave]

Originally Posted by mhennessy

We make assumptions all the time. Just as well, as otherwise we'd be explaining Ohm's law in every post. Luckily, there is always context.

Yes indeed: we do make assumptions all the time. But there are assumptions that are justified and there are those that are not. Then there is argument by analogy. As for your analogy with Ohm's Law, to me, that is an unjustified analogy. If I am explaining something to someone about resistance, voltage and current, I can state Ohm's Law without having to explain it. If the listener wants clarification about Ohm's Law, then he / she can ask and I will elucidate. If I say that to get zero reflections from a coax cable, I need to terminate that cable in its characteristic impedance, I don't expect to have to go into all the detailed mathematics to explain that. But again, the listener / reader can always then ask. For the topic under discussion here, I have done just that: asked for clarification. But I still feel that the author of the Z0 probe should have been a bit clearer in the probe's usage and not made what was - as far as I am concerned - an unjustified assumption.

Phew! In the end, you and I have a different 'take' on this. As I said, I think we'll simply have to agree to disagree. The validity - or otherwise - of the assumption the author has made is a matter of opinion, not scientific, provable fact.

Al.

[G0HZU_JMR]

Just to give my 2p.... I've read quite a few articles by Douglas Smith over the years and they usually contain useful info. I found some of his old articles about E and H field probes to be very informative.

However, his 'free' articles are often a bit sketchy in places and test results tend to be described verbally with minimal pictures or plots. I think he earns income via seminars and books so there will be a limit to how much free info you can expect to get on probes etc in the free online stuff.

In the case of the 40:1 probe article there isn't a proper circuit or system diagram or a typical response curve (showing under/over compensated curves as well?). Even the probe looks rushed and fragile. I'd expect the 1206 tip resistor to snap fairly often and the 200R resistors may crack and become erratic over time.

Even though I know it is supposed to be used with 50R test gear I do think this should be made clear in the introduction or at least in a decent system drawing including the (50R) scope or analyser. The introduction should ideally have given the ratio of the probe and given some indication of the impedance it presents. But I guess the people likely to build it will often know this stuff anyway so no big deal. But the article looks incomplete to me without these simple and basic bits of info at the beginning. You have to dig down to the construction steps to find out the ratio of the probe for example. This should ideally be stated in the introduction.

I'm not sure I'd recommend using it to measure 50V rms because the 1206 tip resistor will get quite toasty even in a 100:1 format. But I guess it depends on the application and how long you connect the probe for...

[Skywave]

Originally Posted by mhennessy

One thing I've learnt over the years is that it's impossible to proof-read my own work

I see. O.K., I am prepared to come out of retirement y'know!

Cheers,
Al.

[stuarth]

Making high impedance probes with good bandwidth and flat frequency response - necessary for accurate capture of fast digital signals, is very difficult. Hence the high price for good probes from the likes of Tektronix. Getting the signal into the probe with good signal integrity is also difficult - you can't just use the 6 inch earth lead that comes with cheap probes - good probes can be fitted with different length earth leads, an earth clip, or even be plugged into test sockets, depending of what it's being used for.

Low impedance probes can be made with good bandwidth much more easily, and are more tolerant of less than perfect earth connections. You don't need the 50 ohm termination at the tip end of the cable, but you do need one at the scope (it wasn't made clear in the article discussed here). For example, the Tektronix P6035 and P6034 probes are 100x and 10x probes with 5k and 500 ohm input impedance respectively, so don't seem to have the 50 ohm at the tip end of the cable. These probes are designed for high bandwidth and good signal fidelity, and sacrifice input impedance and the need to crank up the gain in the scope.

Without the 50 ohm at the tip end, low impedance probes are quite easy to make, all you really need is a resistor at the end of some 50 ohm coax - and a terminator at the scope end. How high the resistor can go depends on the amplitude of the signal you're trying to look at, the loading the circuit can tolerate, and how sensitive your scope is.

Stuart

[G0HZU_JMR]

Here at home and at work I've made several passive low impedance probes and these are mainly for use with a spectrum analyser. I just use a series resistor and a tiny compensating cap for the higher ratio probes and I use a coaxial DC block at the analyser to strip away any DC. The other gotcha with the 40:1 probe in the article is that it presents a 1000 ohm DC path to ground so could easily mess up the biasing of active circuits when they are probed. Again, the article doesn't warn the reader about this limitation.

The Pease active probe is quite elegant but I think it will generate a negative resistance up towards VHF so could give false readings depending on the impedance of the circuit under test. It could add energy and make VHF readings look bigger than they really are for example. It could probably also go completely unstable and oscillate up at VHF if the right passive load was presented to it. Also, how is it meant to be interfaced to any test gear and what is the input impedance of that test gear? It would also not be able to cope with 50V rms at the input but a passive divider could be fitted to the input I suppose.

[Craig Sawyers]

Originally Posted by Skywave

However, do I know that you can have a source whose internal resistance is not the same as the Zo of a coaxial cable and that provided that cable is terminated in a resistance equal to Zo, no reflections will take place at the load. In a past employment, working with high power VHF / F.M. transmission equipment, that principal was successfully used extensively.

That is absolutely correct, and is precisely the way that Tektronix Z0 probes work. And was the reason I said that the cable should be terminated at least at one end. (I should in hindsight have said at the oscilloscope end.)

Craig

[MrBungle]

Originally Posted by G0HZU_JMR

The Pease active probe is quite elegant but I think it will generate a negative resistance up towards VHF so could give false readings depending on the impedance of the circuit under test. It could add energy and make VHF readings look bigger than they really are for example. It could probably also go completely unstable and oscillate up at VHF if the right passive load was presented to it. Also, how is it meant to be interfaced to any test gear and what is the input impedance of that test gear? It would also not be able to cope with 50V rms at the input but a passive divider could be fitted to the input I suppose.

I've actually tried this probe up to 250MHz before my signal generator blew up on a Tek 475 and it doesn't oscillate measurably but the frequency response gets somewhat lumpy after about 100MHz and completely unusable around 120MHz. The killer is the interface with the test gear as you state. I tried terminating mine with an inline terminator at the scope BNC and it improved it a bit but not quite as much as I'd expect. That might be due to the follower not being able to drive the cable rather than the probe itself as it can only source current. Sucking electrons quickly out of a cable is quite difficult via a resistor! I'm using a 10 inch piece of RG-174 as the patch but it's relatively bad cable for this application.

I'm going to be honest though and I think poking around high frequency stuff is nasty with a scope. Too many side effects. Would rather approach the problem in the frequency domain if possible and sneak a known amount of power out somewhere.

[Craig Sawyers]

Originally Posted by stuarth

M
the Tektronix P6035 and P6034 probes are 100x and 10x probes with 5k and 500 ohm input impedance respectively, so don't seem to have the 50 ohm at the tip end of the cable. These probes are designed for high bandwidth and good signal fidelity, and sacrifice input impedance and the need to crank up the gain in the scope.

Without the 50 ohm at the tip end, low impedance probes are quite easy to make, all you really need is a resistor at the end of some 50 ohm coax - and a terminator at the scope end. How high the resistor can go depends on the amplitude of the signal you're trying to look at, the loading the circuit can tolerate, and how sensitive your scope is.

Stuart

There was a wonderful earlier Z0 probe made by Tek in the early 60's called the P6026, and really intended to be used with the Type N sampling plugin which has a 600MHz bandwidth and a 50 ohm input. So the probe itself had a bandwidth of around 1GHz.

The P6026 had:

7 attenuator heads 5x, 10x 20x 50x, 100x, 200x, 500x
2 loads to plug each of those into, both 50 ohms, one DC coupled and one AC coupled
1 adaptor from the load to a GR874 connector
1 10ns GR 874 cable
1 low inductance ground clip.

It was an eye-popping $140 in 1963 - about $3000 now.

The basic configuration of a series resistor feeding into a 50 ohm load, then a 50 ohm cable terminated in a 50 ohm oscilloscope load is exactly the same configuration as http://emcesd.com/1ghzprob.htm

It will probably come as no surprise that I have one of these pieces of Tek history.

Craig

[Sideband]

We use Hantek T3100 100:1 probes in our lab. These are available on eBay for not too much money (or our lab manager wouldn't have bought them....) and free postage. We use 100:1 probes for certain discharge measurements where minimal influence of the scope probe is required.

[Skywave]

I see that his finished probe - which is very short in length - has a BNC female attached. Presumably, a coax cable will be required to connect that probe to a spectrum analyzer. So what would be a suitable Zo for that cable? 50Ω ? And what about its length-dependent capacitance?

Al.

[stuarth]

Low impedance probes designed to drive a 50 ohm load at the scope or spectrum analyser end can have any length of 50 ohm coax between the probe and the scope, it's just a terminated transmission line. Of course, any coax is lossy, pasticularly at higher frequencies, but a couple of metres of good quality coax should not be a problem.

Stuart

[Skywave]

Thanks Stuart: your remarks confirm my thinking about this. Just wanted to check with other members.

Al.