Electronics > Projects, Designs, and Technical Stuff

Designing sort of a differential probe



I need to probe a differential signal (well actual 2) and I'd like not to use 4 channels of the scope just for this.
Ideally I would have two good differential probes but those are rather expensive so I'm looking for a DIY solution.

My signal is a IQ baseband signal, for ~ 10 mv up to 500 mV swing. It's mainly a 300 kHz signals but to get some details, having at least > 1MHz bandwidth would be nice. The signal is centered around 1.5 v, and since I'm already probing other single ended signals on that board, I don't care about isolation.

So my idea was to use an instrumentation amplifier with a gain from 1 to 10.

Unfortunately, finding one with a flat frequency response up to 1MHz isn't that easy. Currently the only one I found is the AD8250 and it looks OK. (albeit a bit expensive at ~15EUR).

The remaining question is how to power it ?

I'd like to have an input power range from 2.5v to 5.5v  and from there I need to get something like -6 +6V .

At first I was thinking of using a simple switcher like those in CCD bias generation (since in a single chip they have both boost and inverter), and follow that by two LDOs to 'filter it' further. But now I'm kind of worried that the inductors on those supply will generate EMI into my measurements. (The DUT would be connected to the 'probing' board with thin flying wires).

So I was looking into charge pumps but haven't found a fitting one yet .. they're either doubler or inverters but from 2.5 to -7V there is more than double or invert ...

Any suggestions ?



Yes, I know that it's not trivial but since I lift quite a few requirements of true differential probes (like complete isolation, wider bandwidth, higher voltage support), I tought off-the shelf part might be easier to find.

The LT1194 looks ok, but not in stock. However I didn't think to look in the "high speed" category for a "difference" amplifier. Not exactly instrumentation but I guess sufficient for my application.

For the power, I also think I could use the +- supply of the scope itself (via the "autoprobe" interface), but I'd rather not. I'd like to supply the power from the target board itself (which is powered via a 3.7v li-ion single cell)

Real differential probes often use a video line driver IC, and use only a limited part of the dynamic range to avoid slew rate/gain-bandwidth limitations. These should easily get you up to 100MHz. Matching the attenuators and other components to get a reasonable CMRR at that frequency is not so easy, though, but you might be OK at 1MHz with limited CMRR requirements (?).


[0] Message Index

There was an error while thanking
Go to full version