1:1 differential probe or isolation amplifier

[nctnico]

Does anyone know a 1:1 differential probe which can handle a few volts of DC offset, a signal of +/- 2V and has a bandwidth of >1MHz (preferably more)? This sounds simple but so far I have not found anything which matches. The 1:1 ratio is important because the signals can also be very small but the peaks can be very high.

[daqq]

If you don't need the DC value, you could get away with an (AC coupled) RF transformer, with an output sunk into a load resistor. This isolates the hell out of a signal and you get as much AC bandwidth as you need. You can even use a 2:1 for amplification.

The drawback is that this loads the circuit pretty heavily.

[Marco]

How about the Chipwhisperer differential probe? Easy enough to mod for unity gain.

[tautech]

Does anyone know a 1:1 differential probe which can handle a few volts of DC offset, a signal of +/- 2V and has a bandwidth of >1MHz (preferably more)? This sounds simple but so far I have not found anything which matches. The 1:1 ratio is important because the signals can also be very small but the peaks can be very high.

And a guy with your experience wouldn't float your scope for this? ^^^ 

[nctnico]

Does anyone know a 1:1 differential probe which can handle a few volts of DC offset, a signal of +/- 2V and has a bandwidth of >1MHz (preferably more)? This sounds simple but so far I have not found anything which matches. The 1:1 ratio is important because the signals can also be very small but the peaks can be very high.

And a guy with your experience wouldn't float your scope for this? ^^^ 

Well... I can't and I need DC coupling. I have to measure several other points in the same circuit. The 'thing' I want to float is a current shunt resistor which doesn't sit at the circuit's ground level.

Building something like the chipswhisperer has crossed my mind but if there is something out there to buy off the shelve life is easier.

[Marco]

And a guy with your experience wouldn't float your scope for this? ^^^ 

With some ferrite on the probe cable.

[tautech]

Does anyone know a 1:1 differential probe which can handle a few volts of DC offset, a signal of +/- 2V and has a bandwidth of >1MHz (preferably more)? This sounds simple but so far I have not found anything which matches. The 1:1 ratio is important because the signals can also be very small but the peaks can be very high.

And a guy with your experience wouldn't float your scope for this? ^^^ 

Well... I can't and I need DC coupling. I have to measure several other points in the same circuit. The 'thing' I want to float is a current shunt resistor which doesn't sit at the circuit's ground level.

OK, nevermind, seems like you don't know how to do it anyway and via the forum I'm not about to tell you. 

[daqq]

Maybe measure the DC offset by a normal probe, use AC coupling with high gain for the fine 1MHz+ signal?

[alsetalokin4017]

I suppose taking a differential measurement with two ordinary passive 1x or 10x probes is too obvious a solution.

[nctnico]

I suppose taking a differential measurement with two ordinary passive 1x or 10x probes is too obvious a solution.

Yes. I don't want to sacrifice 2 channels, the CMRR isn't that good, math updates aren't very fast and catching a tens of mV signal riding on a few V offset is likely going to kill accuracy.

[macboy]

Why not build an instrumentation amplifier? With a simple 3 op-amp circuit you get very high input impedance, very high CMRR, and you can even build in all the gain you want. At 1 MHz this is easy, you could even breadboard it with decent results.  This all assumes that the "few volts" of DC offset still keeps the signals within the rails of the op-amps.  Another option is a special purpose high side current sense amplifier IC.

[owiecc]

You can just buy an instrumentation amp IC.

[nctnico]

Why not build an instrumentation amplifier? With a simple 3 op-amp circuit you get very high input impedance, very high CMRR, and you can even build in all the gain you want. At 1 MHz this is easy, you could even breadboard it with decent results.  This all assumes that the "few volts" of DC offset still keeps the signals within the rails of the op-amps.  Another option is a special purpose high side current sense amplifier IC.

I'm thinking across these lines as well. But I thought I'd ask first before building something I can also buy off-the-shelve. Current sense amplifiers would have been an option too if they didn't had so much gain. I also need to be able to measure 1us spikes up to 1kA so an instrumentation amplifier with a gain of 1 is the best option due to the required slew rate. So far I have found the AD8421 (15Mhz / 35V/us) which can be powered from +/-15V.

[MarkL]

If the signal is small, how about a differential amp to increase the gain too?  The Tek AM502 starts at x1 and goes up to x100k.

[T3sl4co1l]

I suppose taking a differential measurement with two ordinary passive 1x or 10x probes is too obvious a solution.

Yes. I don't want to sacrifice 2 channels, the CMRR isn't that good, math updates aren't very fast and catching a tens of mV signal riding on a few V offset is likely going to kill accuracy.

Well, you opened with a spec of 2V.  Your SNR spec just went in the shitter by 40dB, so thanks for telling us about that in the first place...

I suppose your CMRR needs to be good well beyond 1MHz as well?? (Ex: current sense on noisy SMPS, let alone the switching node itself)

Tim

[nctnico]

Well, you opened with a spec of 2V.  Your SNR spec just went in the shitter by 40dB, so thanks for telling us about that in the first place...
I suppose your CMRR needs to be good well beyond 1MHz as well?? (Ex: current sense on noisy SMPS, let alone the switching node itself)

Fortunately this isn't for measuring in an SMPS. For a chance the circuit I'm working on is mostly DC which in this case is challenging enough. There are 2 things I want to measure: a rush-in current pulse up to 1kA and tens of Ampere during normal operation. The specs of 2V just is a realistic-ish target based on the differential probes I have found so far. The current shunts I have are 5m Ohm (5V @1kA) and 0.5m Ohm (0.5V @1kA). Measuring tens of amps would result in signals in the mV range but still with the occasional 1kA spike it should be able to survive.

[Marco]

Classical instrumentation amplifiers are a silly architecture for high side current measurement, the source impedance is essentially 0, symmetry in the input impedance is irrelevant.

[nctnico]

I'm planning to measure the low(est) side 
A generic instrumentation amplifier box may prove useful for other purposes as well though so going that route and making the input impedance 1M would allow to use it with a standard probe and a selectable gain of 1 or 10 could turn it into a neat pre-amplifier as well.

[tautech]

I'm planning to measure the low(est) side 
A generic instrumentation amplifier box may prove useful for other purposes as well though so going that route and making the input impedance 1M would allow to use it with a standard probe and a selectable gain of 1 or 10 could turn it into a neat pre-amplifier as well.

IMHO this is where TE equipment manufacturers have let the industry down. 

We now have all manner of cheap DSO's available with ranges of features that are astounding BUT none with Isolated Channel inputs for the arguably BASIC measurements that you need.

Massive marketing opportunity IMO.

[nctnico]

I agree. I have been thinking about an isolation amplifier design which doesn't use the classic differential amplifier approach. I hope I have some time to create the layout of a first prototype during the Christmas vacation. I keep running into needing isolated oscilloscope channels but there is nothing really good in the market.

[Someone]

I'm planning to measure the low(est) side 
A generic instrumentation amplifier box may prove useful for other purposes as well though so going that route and making the input impedance 1M would allow to use it with a standard probe and a selectable gain of 1 or 10 could turn it into a neat pre-amplifier as well.

IMHO this is where TE equipment manufacturers have let the industry down. 

We now have all manner of cheap DSO's available with ranges of features that are astounding BUT none with Isolated Channel inputs for the arguably BASIC measurements that you need.

Massive marketing opportunity IMO.

Tek has the specialist TPS2000 series which is the standard for isolated measurements, since it hasn't been updated in a long time you can get a hint as to how popular that market segment is.

Does anyone know a 1:1 differential probe which can handle a few volts of DC offset, a signal of +/- 2V and has a bandwidth of >1MHz (preferably more)? This sounds simple but so far I have not found anything which matches. The 1:1 ratio is important because the signals can also be very small but the peaks can be very high.

Start with a Tek ADA400, from there you could easily modify it to have lower noise and input impedance.

[tautech]

I'm planning to measure the low(est) side 
A generic instrumentation amplifier box may prove useful for other purposes as well though so going that route and making the input impedance 1M would allow to use it with a standard probe and a selectable gain of 1 or 10 could turn it into a neat pre-amplifier as well.

IMHO this is where TE equipment manufacturers have let the industry down. 

We now have all manner of cheap DSO's available with ranges of features that are astounding BUT none with Isolated Channel inputs for the arguably BASIC measurements that you need.

Massive marketing opportunity IMO.

Tek has the specialist TPS2000 series which is the standard for isolated measurements, since it hasn't been updated in a long time you can get a hint as to how popular that market segment is.

Price? 
That's the reason sales are low and compare their specs to any modern DSO and prepare to be 

Isolated measurements per se are not unordinarily specialised, and while Tek might have had this market almost to themselves for a while, is it not time for a change?

For one to set themselves up for isolated measurements there is significant cost involved, even 2 basic differential probes can cost more than a new 4 channel DSO. 
No wonder some people are looking for alternatives.

[Someone]

I'm planning to measure the low(est) side 
A generic instrumentation amplifier box may prove useful for other purposes as well though so going that route and making the input impedance 1M would allow to use it with a standard probe and a selectable gain of 1 or 10 could turn it into a neat pre-amplifier as well.

IMHO this is where TE equipment manufacturers have let the industry down. 

We now have all manner of cheap DSO's available with ranges of features that are astounding BUT none with Isolated Channel inputs for the arguably BASIC measurements that you need.

Massive marketing opportunity IMO.

Tek has the specialist TPS2000 series which is the standard for isolated measurements, since it hasn't been updated in a long time you can get a hint as to how popular that market segment is.

Price? 
That's the reason sales are low and compare their specs to any modern DSO and prepare to be 

Isolated measurements per se are not unordinarily specialised, and while Tek might have had this market almost to themselves for a while, is it not time for a change?

For one to set themselves up for isolated measurements there is significant cost involved, even 2 basic differential probes can cost more than a new 4 channel DSO. 
No wonder some people are looking for alternatives.

The TPS line is 10 years old and haven't been updated in specs or price, the market is so small they aren't interested but at the same time one of those scopes is comparable in price to a similarly specified differential probe so its what the market will bear. If its such a lucrative and competitive market perhaps you should setup shop and make your fortune.

[nctnico]

There are loads of handheld scopes with isolated inputs so people do buy them. The problem with Tektronix' TPS2000 series is their extremely poor value for money due to the 2.5kpts sample memory and lack of other options.

[rx8pilot]

I'm planning to measure the low(est) side 
A generic instrumentation amplifier box may prove useful for other purposes as well though so going that route and making the input impedance 1M would allow to use it with a standard probe and a selectable gain of 1 or 10 could turn it into a neat pre-amplifier as well.

IMHO this is where TE equipment manufacturers have let the industry down. 

We now have all manner of cheap DSO's available with ranges of features that are astounding BUT none with Isolated Channel inputs for the arguably BASIC measurements that you need.

Massive marketing opportunity IMO.

Tek has the specialist TPS2000 series which is the standard for isolated measurements, since it hasn't been updated in a long time you can get a hint as to how popular that market segment is.

Does anyone know a 1:1 differential probe which can handle a few volts of DC offset, a signal of +/- 2V and has a bandwidth of >1MHz (preferably more)? This sounds simple but so far I have not found anything which matches. The 1:1 ratio is important because the signals can also be very small but the peaks can be very high.

Start with a Tek ADA400, from there you could easily modify it to have lower noise and input impedance.

I got an ADA400A for around $100 used and it works quite well.

Sent from my SAMSUNG-SM-G890A using Tapatalk