Differential Scope Probe

Paul M

Does anyone have a simple(=inexpensive) circuit for a differential scope probe?
I have a need to scope the signals from a mains (240V) thyristor stack·firing circuit. The firing circuit derives its psu from the mains but is not isolated so everything sits at 240V+. I know I could use·the 2 channels on my scope differentially but I actually need 2 channels.
Commercial units start at about £300 and upwards. I know they are expensive because of the high CMRR required, the high·voltages, insulation etc.· I don't need super accuracy and·am only looking at frequencies up to about 40kHz

In anticipation,

Paul

kwinn

Two alternatives come to mind. The simplest is an isolation transformer between the scope and the mains. The other is an op amp as a differential input circuit between the probes and scope.

Phil Pilgrim (PhiPi)

Deinitely use the isolation transformer!!! Attaching scope leads directly to a mains-fed circuit may do real damage to your scope. The reason is that in many scopes, mains ground and logic ground are connected.

-Phil

Paul M

Phil, I am very well aware·that mains earth and scope input ground are connected. This is why I am after a differential scope probe.

kwinn, the op amp circuit is what I need but the common mode voltages are very high and I don't know how to design·a suitable circuit.

Here's an example and only £195 but still too much!

http://accessories.picotech.com/active-oscilloscope-probes.html

··

Phil Pilgrim (PhiPi)

Something like the following might work. It's a classic differential amplifier, but using optoisolators instead of discrete transistors. I have no idea what frequency range it would cover or how bad the offset voltage would be, but the latter might not matter that much.



-Phil

Paul M

Phil

Neat idea; I'll give it a try as I don't need very high bandwidth

Paul

Tracy Allen

A standard op-amp differential amplifier can have very high common mode rejection and handle high voltages. The attached circuit has a gain of 1/10000, so The output to your 'scope would be about 0.024 volts when the input is 240 volts.

-- input impedance of 100 MOhms. Could leave out capacitor for operation to DC.
-- 100 Mohm input resistor and 0.01 capacitor have to be rated for the voltage. Electronics Goldmine usually has a good selection of surplus high ohm resistors. Op-amp has to be unity gain stable.
-- Bandwidth from about 1 Hz up to roughly the unity gain bandwidth of the op-amp. Could operate down to DC by removing the capacitors C1 and C2.
-- Common mode range is over +/- 1000 volts, because of the voltage dividers.
-- At 240 volts, the current through the 100 MOhms is 2.4 microamps. Still, a zener clamp on the op-amp inputs might be a wise precaution.
To make this more professional, one would want to add a followup gain stage and an isolation amplifier (e.g. ISO122) with an isolated power supply. But it can work (with extreme caution) without those refinements. The probe ends are still Hot! Keep the "cold" end of the resistor and everything protected.

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Tracy Allen
www.emesystems.com

Post Edited (Tracy Allen) : 7/2/2009 4:07:15 AM GMT

Paul M

Thanks Tracey, that is very helpful. I am in the UK but RS has a good supply of high voltage resistors; I recall that in some high voltage circuits it is common to use say 2 50Mohm resistors in·series for the input resistors.·The·ISO122 looks like a good idea but pricey at over £20 each. However, could still end up with a probe for a fraction of the commercial ones.

One question: why such·high attenuation?

Paul

Tracy Allen

Hi Paul,

The solution with the ISO202+isolated power module and a few other parts could make a nice special purpose probe for a fraction of the cost of a commercial multi-capability differential probe.

You could probably reduce the input resistors to 10 MOhms if you are comfortable with that, or increase the resistors around the op-amp to 100k or 1M. Either approach would give a higher output voltage to the 'scope. If the signals are AC, the capacitors give an extra measure of isolation from ground.

I've been using a similar circuit this last week to probe the DC high voltage (~200 volts) generated for bias on a microphone capsule. The generator is a Senscomp coil in a switching circuit, and the source resistance is very high (nearly 100 M). So I used 500M (5 * 100M) in a divider with 1 M and then an amplifier. By measuring the voltage at 500M and 400M, I can extrapolate to the source voltage and also calculate the source resistance.

I have a portable 'scope (THS720) on which both input channels are isolated, so each channel can have its own ground point, and with a kV of isolation between channels. It is a very useful feature. But the differential capability is achieved by subtraction of the 8 bit digital readings, and that is seldom satisfactory.

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Tracy Allen
www.emesystems.com

Lawson

Phil Pilgrim (PhiPi) wrote: »

Something like the following might work. It's a classic differential amplifier, but using optoisolators instead of discrete transistors. I have no idea what frequency range it would cover or how bad the offset voltage would be, but the latter might not matter that much.

*snip*

-Phil

This quite a necro-post, but I just had to test a variation on Phil's circuit. As shown, Phil's circuit won't work well at all because the opto-couplers have no bias current. At best, it will have 2-3 volts of dead band, pretty bad distortion, and mandatory differential output.

The variation I tested adds input biasing, and uses the opto-coupler outputs in a push-pull configuration. I got the best performance form the opto-isolator with a 1.1ma bias current in each LED. (set via B1 and Rbias) Rin sets the input attenuation, while Rout sets the output gain and bandwidth. With Rin = 3.8Kohm and Rout = 2.8Kohm I measured a large signal bandwidth of 10KHz and a forward gain of 2x. linearity was excellent, with no visible distortion, and low noise. I expect that the bandwidth and gain/input-loading could both be considerably improved if a transimpedance amplifier was used to convert the current output of the opt-couplers to a voltage instead of a resistor bridge. (though I doubt it'll go over 1MHz bandwidth without opto-coupler changes)

Marty