DIY Cheap HV differential oscilloscope probe

[MarkKruetzmann]

Hi!

I am designing a cheap differential oscilloscope probe while trying out Altium Designer and so far I came up with the attached schematic. I figured a cheap differential probe would be beneficial to many labs doing switch mode power supplies, but the usual offerings start deep in the 3-figure-range. As this is out of the budget of many hobbyists, I came up with a low-spec design of such a device.

Specs are (or should be):
- about 30€ BOM Cost without case
- accepting Mains and a bit more
- at least 1MHz bandwidth, better 10MHz (as many SMPS work well under 100kHz, this should be fine)
- settable Attenuation

I ask you here what you think of my design and whether you see ways of improving it while keeping cost down.


The next post will be my thought-process regarding most of the elements here, but this is the (for now) finished schematic.

[MarkKruetzmann]

Mistake in the schematic! R5 and R7 are 22k for availability!

Somewhere in this forum I read about the AD8421, and I thought I could build a probe around this device without a pre-amplifying-stage with JFETs or so. Sure, performance wouldn't be phenomenal, but it would stay small and affordable. You may correct me regarding this early decision, as it may be a bad one and I based much of this design on this choice.

I started wit Daves teardown of the LeCroy differential probe and thought I'll copy the input impedance and main passive input architecture. I use the AD8421 and I want to power this device with two 9V batteries, which should be allowed to get as low as 7V. I decided for about 700V allowed input voltage (DC or AC-peak). The 7V batteries left me with a maximum of 4.5V at the input (see AD8421 datasheet, min input voltage is negative supply plus 2.5V) I would go for about 150 times attenuation. I decided for a bit more and chose 25k as attenuation resistor to ground. With 1% resistors, a 1k pot would suffice to calibrate for symmetry. As 25k is not as easy obtainable as I thought, I later switched to 22k which left me with an attenuation of 1/178,7.

Then there were long calculations regarding the capacitors, but in the end I chose to use 10pF ones as these are easier to obtain in 1% accuracy which really was necessary to be able to tune the HF path under worst circumstances. With the 3 10pF+-1% capacitors and an attenuation of 178,7 I ended up with a parallel resistance range of 587,4...599,26pF. The chosen capacitors net up to 588(+1%!)pF worst case, but I chose to include the 10pF here nonetheless because I did not account for stray capacitance loading the input yet and figured it would be better to have some headroom in the other direction. Also, usually it should work when the caps null out their tolerances a bit.

The effect of stray resistance, especially with shielding around the input, has yet to be discovered, but this is my first guess, which might or might not work (I hope and think it does though). In the rare occasion that the tolerances add up too much, I can just leave out C5 or C12 respectively.


The strange device S1 is a switch to set the Gain of the amplifier. I seem to have miscalculated the attenuation here as I calculated with 1/177,7 instead of 1/178,7, but oh well. This might still be more exact than the tolerances, which is why I deem it reasonable. The switch sets the Gain to 1,77; 17,7 and 177 respectively to yield a final attenuation of 1/1, 1/10 or 1/100 of the differential signal. This should in theory allow to measure a 1V drop over a shunt riding on 230V no problem.

On the output I added 47 Ohms before the signal feeds into an attached cable with BNC-plug to sorta match cable impedance. Overkill? Not enough?

Regarding construction:
Input cables shall be simple banana plugs to allow attachment of probes. This should be fine as bandwidth is quite low and source impedance should be low as well. (correct me if I am wrong, but it should work this way)

Output cable is directly attached BNC cable, say 1m length.

Case will be 3d printed with strain reliefs for the cables and a place for the batteries. Mentioning batteries, I should add a switch for these.

The complete front end shall be shielded. I think of doing that with a solderable can, but I do not know what impact they have on stray resistance. Input regarding this topic would be welcome! They are also quite expensive and small. A ground plane would shield the bottom side, adding yet more stray capacitance.

I have seen spark gaps in some designs, but I am hesitate to include them as the ground (which is provided by the BNC) may be unplugged and I think 3 2kV caps in series or 4 500V resistors should be safe enough - probably even safer than PCB spark gaps. Feel free to correct me here as well.

[daisizhou]

https://www.eevblog.com/forum/projects/oshw-diy-1kv-100mhz-differential-probe-(dilemma-vs-hope)/
Hello i am about to make.This is my reference

[David Hess]

For the simplest design, a high input impedance instrumentation amplifier like you selected will work but also consider using a current feedback "difference amplifier" like the AD8129 or AD8130.  They allow easier gain selection because the feedback network is grounded on one end and have much higher high frequency common mode rejection.  Their disadvantage is that some sort of high impedance buffer will be required for each input.

[floobydust]

Pintek DP-15K 35MHz, 15kV differential probe input stage uses two THS4631 with another op-amp and DC servo. Two 15MEG 10kV input resistors, +/-12V power, gain /100 or /1000.
You want to minimize drift due to the trimpots because of the high attenuation ratio in the voltage divider.  Best to use resistors and then the smallest trimpot value.

[tggzzz]

I ask you here what you think of my design and whether you see ways of improving it while keeping cost down.

With a circuit like this, the devil is in the details.

Any description should discuss how it works, and must include descriptions of how it might fail.

[MarkKruetzmann]

Okay, going there one by one:

@daisizhou:
Your design looks nice and seems to be in the prototype stage already, but is a bit too complex for this project.

@David Hess:
I looked at these devices. The AD8130 looks especially tempting, as it allows for low-ish gain to provide x100 attenuation.
It is also quite cheap. It does have some minor flaws though:
It needs to have a different (higher) input attenuation, but that is fine. That will lead to higher Gain though, which will result in less Bandwidth (still much higher than my design). It does have some input offset, which needs to be calibrated out (especially when considering G of about 200). Another trimmer to tweak reference voltage, some resistors to the rails, doable. I think input impedance would be fine with this one, as the source would be the divider, which is about 10-25kOhm. I might be wrong here, and an input buffer is indeed mandatory. The inputs would maybe need to be protected. Granted, a couple of diodes would do the trick.
Another bad thing is 10mA quiescent current compared to 2.3mA on the AD8421.

So, while the proposed 8130 would be probably very good to achieve solid 2-digit-MHz-bandwidth, maybe even 100MHz and above, I still prefer the AD8421 for my design. DC accuracy, low noise I didn't even talk about, built-in protection against overvoltage at the inputs, low quiescent current, higher allowed input voltage (which should reduce gain and thereby reducing noise while maintaining some bandwidth) and, last but not least, ease of design is nice here.

The 8130 and 8129 respectively are very good devices and I will consider them in a higher end design. Thank you very much for your input David, I appreciate it. A design based on that amplifier would probably be 10 times as fast or so, while only increasing parts count by say 50% and trading some noise for bandwidth. Maybe a future design, this one here shall stay simple.


@Floobydust:
Another nice part, I might have used it in a modern version of the LeCroy probe Dave disassembled. Maybe in conjunction with the part David Hess suggested. But for the reasons I stated to him, I will stick with the simple design. As tggzzz said, the devil is in the details and I have some details already, might as well sort those out before maing it more complex than it already is.


@tgzzz:
Yeah, devil is in the details and I try to convey those in my thought process writeup, but I should probably be more precise and direct. Using your guideline
"Any description should discuss how it works, and must include descriptions of how it might fail"
I will add a more detailed description now. I might not see all pitfalls though and under- or overestimate others, that is why I posted this design here in the first place. Description will follow next post.

[MarkKruetzmann]

Perfect, my PC just bluescreened. Anyway, again:

Okay, circuit description:

A differential signal comes in at P1 and P2. These are just short wires which terminate in a banana plug to allow clipping onto test probes. This might lead to RF capture, which would be bad, and bad HF response due to cable length. I assume a low source impedance though and the device has a low-ish frequency range, which is why I am okay with this sort of design. Some stray resistance may also shunt the highest unwanted RF, but I maybe should not trust on that alone. Or should I?
There follows the voltage divider comprising of R1 through to R11. This shall attenuate the input signal down to processable levels for the amplifier. Any imbalance caused by resistor mismatches can be calibrated out by supplying a high voltage at the shorted inputs, set the probe to max gain and adjust R6 until there is a negligible voltage at the output. If the resistors meet their spec, all mismatches can be calibrated out.
Parallel to the resistive dividers there are two capacitive voltage dividers. These allow for a HF path to the amplifier inputs. C4 and C8 allow for tuning these paths by connecting a square wave to one input while connecting the other to ground. Then the output waveform can be trimmed to best possible square response. The user should do that in the lowest possible gain setting (restricted by the scope used) to allow the highest possible bandwidth. These paths should be trimmed after R6 has been trimmed. Considering worst case, the tuning range of C4 and C8 might not be enough to achieve a good square response. This is under worst case only and when I dramatically underestimate the stray capacitance shunting the series capacitor string to ground. If that is the case though, just omit C5 and C12, respectively. If stray capacitance is way higher than expected, exchange C5 and C12 with something like 47pF.

Regarding safety: C1 to C3 and C9 to C11 are 2kV devices. If one fails with a short, there are still two more. Resistors R1 to R4 and R8 to R11 are 500V devices, and like the resistors 1206 SMDs. That should provide a reasonable safety margin (about 2.5 times) to the input and should suffice even when one or two resistors or capacitors fail. Definitely correct me if I am wrong here. I deem it safer than spark gaps at the input, as the devices' ground may be not connected.

The attenuated signal goes directly to the AD8421, which has three programmable Gain settings. I do not see possible failures in the gain setting circuitry besides a defective switch, and this should be non-destructive. However: I do see a possibility of unwanted (or unsupported) RF getting into the amplifier. Question to the forum: Is it okay to rely on shunt capacitance in the input path or should I implement a low pass filter somewhere? If so, I would implement it directly in the divider by putting a resistor between R4 and R5 or R7 and R11, respectively. I have not seen that yet on any probe designs, that's why I think I can omit it. It might be a nonexistant problem after all. I'm happy for your feedback regarding this.
Any voltage spikes higher than the rated input voltage will be (to a degree) shunted by the input protection of the AD8421, which withstands voltage spikesof 40V above the rails, so minor spikes (which are still attenuated by the divider) should be no problem.

After amplification the signal goes through a 47 ohms resistor directly in a soldered on BNC cable. As the input of most scopes are high impedance, this might cause reflections, but as the cable is short compared to the wavelength of the probe, this should be fine. Correct me if I am wrong here. Are the 47 ohm even necessary? The amplifier is short circuit resistant for an indefinite amount of time, so I should be fine here regarding safety of the amp.


I hope this clarifies where the questions are with this project.

[Mechatrommer]

https://www.eevblog.com/forum/projects/oshw-diy-1kv-100mhz-differential-probe-(dilemma-vs-hope)/
Hello i am about to make.This is my reference

although the basic theory is still the same, the detail implementation has obsoleted. trying to do manual tuning for 3 separated ins-amp opamps is super difficult.. ad8130 that came later gives much more hope with alot better cmrr and smaller footprint. I'm working on a whole new layout....

[David Hess]

@David Hess:
I looked at these devices. The AD8130 looks especially tempting, as it allows for low-ish gain to provide x100 attenuation.
It is also quite cheap.

...

So, while the proposed 8130 would be probably very good to achieve solid 2-digit-MHz-bandwidth, maybe even 100MHz and above, I still prefer the AD8421 for my design. DC accuracy, low noise I didn't even talk about, built-in protection against overvoltage at the inputs, low quiescent current, higher allowed input voltage (which should reduce gain and thereby reducing noise while maintaining some bandwidth) and, last but not least, ease of design is nice here.

It is about more than bandwidth and noise.

Noise with so much attenuation on the input is always going to be very high.  The only way to significantly lower noise is to use a bootstrapped amplifier which allows for a much higher input common mode range so the input attenuation can be much lower.  Image an amplifier with a +/- 20 volt input instead of +/-2 volts so the input attenuation can be reduced by 10 times lowering input noise to 1/10th.  This is a lot more complicated but some high voltage differential amplifiers have done it for lower noise.

The chief reason I would consider the AD8129/AD8130 for a lower bandwidth design is because of higher common mode rejection at higher frequencies and faster overload recovery time.  Voltage feedback (shunt feedback) increases overload recovery time.

The 8130 and 8129 respectively are very good devices and I will consider them in a higher end design. Thank you very much for your input David, I appreciate it. A design based on that amplifier would probably be 10 times as fast or so, while only increasing parts count by say 50% and trading some noise for bandwidth. Maybe a future design, this one here shall stay simple.

Like I said, the instrumentation amplifier you selected will work fine for your objectives.  I just presented an alternative.

[MarkKruetzmann]

Like I said, the instrumentation amplifier you selected will work fine for your objectives.  I just presented an alternative.

Apparently, my other mentioned thoughts seem to make sense as there was no response to that. Therefore I guess I just roll with the current design and figure out any problems from there on. I need to draw the PCB first though, so there is another possibility of getting things wrong, but we'll see.

The alternative is for the next, better probe or Mechatrommer will do something with it. Good luck with that, Mechatrommer!

I attach the new version of the schematic. R16, R17, C20 and C21 are optional if I run into RF issues. I hope I get away without them.

[joeqsmith]

I built up a few.  This one has the higher BW.  Not great performance.   I made a board for it in hopes of improving it's flatness but never got around to ordering them. 




This was the low frequency version. 

[Hydron]

To add to the comments about other amplifiers (not saying you should change the design, just that it may be of interest!), I seem to recall that the EEVBlog 70MHz diff probe uses both an AD8130 and an AD8129 - I assume with both always connected and the outputs muxed via the shutdown function.

This would allow for high bandwidth at both high gain (via the 8129) and low gain (via the 8130) - note that for gains over 10x you can use the 8129 for a higher GBW, but less than 10x the 8130 is the only option that is stable. (Note that I'm talking about the gain of the differential amplifier after the attenuator here, not the x ratio of the probe).

[oschonrock]

This is a really worthwhile endeavour.

As a benchmark would this commercial product be the sort of price/performance which you are shooting for?

https://www.siglent.eu/product/1142088/siglent-isfe-isolated-front-end-module
https://www.siglent.eu/_downloads/51338182

Just over £/$/€100  ?

Or for £150 there is this, which looks pretty fully featured.... can't vouch for quality.
https://www.telonic.co.uk/Micsig-DP10013-High-Voltage-Differential-Probe-p/mi-dp10013.htm

[MarkKruetzmann]

This is a really worthwhile endeavour.

As a benchmark would this commercial product be the sort of price/performance which you are shooting for?

https://www.siglent.eu/product/1142088/siglent-isfe-isolated-front-end-module
https://www.siglent.eu/_downloads/51338182

Actually a bit more performance (should be almost 10MHz in 10 and 100 times attenuation, at least 5MHz there), more functionality (switchable gain), and (for the DIYer) a lower price. BOM cost will be 20-30€ depending on quantity, leaving out the case and accessories. So probably closer to 40 with everything. Multiply that by pi and you get to a reasonable selling price and you are in the same ballpark as the mentioned probes.

Or for £150 there is this, which looks pretty fully featured.... can't vouch for quality.
https://www.telonic.co.uk/Micsig-DP10013-High-Voltage-Differential-Probe-p/mi-dp10013.htm

that seems to perform reasonably well, according to the internet
https://goughlui.com/2018/12/16/review-eevblog-hvp-70-70mhz-micsig-dp10013-100mhz-high-voltage-differential-probes/

To add to the comments about other amplifiers (not saying you should change the design, just that it may be of interest!), I seem to recall that the EEVBlog 70MHz diff probe uses both an AD8130 and an AD8129 - I assume with both always connected and the outputs muxed via the shutdown function.

This would allow for high bandwidth at both high gain (via the 8129) and low gain (via the 8130) - note that for gains over 10x you can use the 8129 for a higher GBW, but less than 10x the 8130 is the only option that is stable. (Note that I'm talking about the gain of the differential amplifier after the attenuator here, not the x ratio of the probe).

A schematic of said probe would be fantastic - a quick search didn't come up with one. That does confirm that David Hess' suggestion is probably a good one regarding these amplifiers. A look on how the probe is designed would be very neat, so if you recall where you have seen the schematic, I would be very happy.

I built up a few.  This one has the higher BW.  Not great performance.   I made a board for it in hopes of improving it's flatness but never got around to ordering them.

Design looks good, and given the fact that you just used what was lying around the results are very good! As this is a new design and I need to order parts anyway, I'll stick to my design as the performance may be similar while the parts count is much lower.

[oschonrock]

Actually a bit more performance (should be almost 10MHz in 10 and 100 times attenuation, at least 5MHz there), more functionality (switchable gain), and (for the DIYer) a lower price. BOM cost will be 20-30€ depending on quantity, leaving out the case and accessories. So probably closer to 40 with everything. Multiply that by pi and you get to a reasonable selling price and you are in the same ballpark as the mentioned probes.

Cool, that would be great. I don't think I can add anything technical to your design, it seems well thought through to me, but I am certainly not an expert in this area.

One thing that made my left eye twitch (usually a sign worth paying attention to) is your "20-30€ without the case" ... and without the probes....and "using banana plugs".... This is 600-1000V probe. Are you happy with 600+V on banana plugs? Does the case not deserve a bit more attention? Also (inbound) EMI stuff like metals cans around the input stage at least ...

Like most electronics engineers I hate thinking about this stuff. But it's kind of key for a project like this? Just a thought. 

that seems to perform reasonably well, according to the internet
https://goughlui.com/2018/12/16/review-eevblog-hvp-70-70mhz-micsig-dp10013-100mhz-high-voltage-differential-probes/

Good find. I might order one since I am trying to develop a small HV switch control breakout board board and don't fancy blowing up me or the scope. I can then report back.

[MarkKruetzmann]

@oschonrock

The other mentioned probes use banana plugs as well. Of course I will also use the safe variety of those, the cat rated insulated ones.

And an emi shield (one of these solderable cans) is actually part of the BOM cost. Quite a considerable amount of the BOM cost, I might add. These things aren't cheap!

But thank you for your concern! Those things often get looked over I guess. To me input protection is actually an interesting story, I find it interesting to which lengths the big manufacturers go to achieve that.

[oschonrock]

@oschonrock

The other mentioned probes use banana plugs as well. Of course I will also use the safe variety of those, the cat rated insulated ones.

And an emi shield (one of these solderable cans) is actually part of the BOM cost. Quite a considerable amount of the BOM cost, I might add. These things aren't cheap!

But thank you for your concern! Those things often get looked over I guess. To me input protection is actually an interesting story, I find it interesting to which lengths the big manufacturers go to achieve that.

 

[Hydron]

To add to the comments about other amplifiers (not saying you should change the design, just that it may be of interest!), I seem to recall that the EEVBlog 70MHz diff probe uses both an AD8130 and an AD8129 - I assume with both always connected and the outputs muxed via the shutdown function.

This would allow for high bandwidth at both high gain (via the 8129) and low gain (via the 8130) - note that for gains over 10x you can use the 8129 for a higher GBW, but less than 10x the 8130 is the only option that is stable. (Note that I'm talking about the gain of the differential amplifier after the attenuator here, not the x ratio of the probe).

A schematic of said probe would be fantastic - a quick search didn't come up with one. That does confirm that David Hess' suggestion is probably a good one regarding these amplifiers. A look on how the probe is designed would be very neat, so if you recall where you have seen the schematic, I would be very happy.

No schematic, just an explanation and some (unfortunately low resolution) photos here:
https://www.eevblog.com/forum/testgear/eevblog-hvp70-differential-probe/msg1394351/#msg1394351
Seems my recall was correct however

A few additions:
The Micsig is reasonable value for money, however be aware that the design has some significant limitations - rather than spend the money on a AD8129/8130 pair as the HVP70 does, it uses a single standard (albeit fast) op-amp as a differential amplifier, which means that it relies on good resistor matching to keep CMRR reasonable when switching ranges. If you're lucky this will work OK, but one of the two I have wasn't so great, and I ended up opening and tweaking it by stacking a resistor ontop of another. The probe also has some _serious_ peaking issues at ~70Mhz, which shortening the leads to ~half mostly solved, at the cost of a reduction of -3dB bandwidth (which had been artificially inflated by the peaking).

As for a DIY design, I looked at it once but the problem David Hess talked about regarding common mode input range is a serious one - makes it a compromise between input voltage range and attenuation/noise. Even a AD8130 with +/- 12V supplies only gives you +/-1kV input CM range with a 100x attenuator on the input, and whats more would only give +/- 250V of differential range (this is a bit different than a normal op-amp, which by design wouldn't have a differential voltage spec on the inputs).

[JohanH]

This is also an interesting variant. Nobody seems to have bought it and tried it yet (no feedback on ebay at least). https://mitstronic.com/

[NiHaoMike]

No schematic, just an explanation and some (unfortunately low resolution) photos here:
https://www.eevblog.com/forum/testgear/eevblog-hvp70-differential-probe/msg1394351/#msg1394351
Seems my recall was correct however

The 1080p width image being called "low resolution" is a perfect example of 1080p not always being enough. I attached the full resolution original I have.

[holla2040]

For those of still reading this post. This from my high-voltage measurement notes

AD8479
Very High Common-Mode Voltage Precision Difference Amplifier, 600V common mode
http://www.analog.com/media/en/technical-documentation/data-sheets/AD8479.PDF
    common mode halves when run from single supply

INA149
high voltage difference amplifier

[Terry Bites]

Watch out the AD8129/30 has input bias currents in the microamps and poor dc drift. Fine for AC though. I love 'em.

[jonpaul]

Bonjour MarkKruetzmann

SMPS testing rarely if ever needs  a diff HV probe.

We use the TEK AA501 diff amp,
TEK 7000 diff plugins and bench scopes TEK 2465B, in CH1 - CH2  mode, CMRR ~ 50 but OK.


1/ Off line SMPS have 360-600 V plus any transients or ringing, probes,  attenuators and amps must specified accordingly.

2/ As safety is involved, use rated components, connectors, test leads.

3/ You can  find examples of similar finished designs.

4/ Transient protect the PSU and input terminalss of the INA/OPA.

Bon chance


Jon