New MicSig/EEVblog DP10007 HV Differential Probe

[mrprecision]

I have shorten the inputs of the probe for the setting 10:1. I see some coupling from there DC/DC converter inside the diffprobe. I bought the probe at Bartonix, they could measure the same effect with my probe and with other probes of this type. They got some newer ones and there it is much better. I have two probes which are find now. I use this mostly for lower signals as a diffprobe, not mainly for higher voltages.

The noise of the 10:1 probe is lower than on the other Micsig probes. If you don't zoom in too much in the x-axis, you would think that the probe has higher noise...

[ExaLab]

I have shorten the inputs of the probe for the setting 10:1. I see some coupling from there DC/DC converter inside the diffprobe. I bought the probe at Bartonix, they could measure the same effect with my probe and with other probes of this type. They got some newer ones and there it is much better. I have two probes which are find now. I use this mostly for lower signals as a diffprobe, not mainly for higher voltages.

The noise of the 10:1 probe is lower than on the other Micsig probes. If you don't zoom in too much in the x-axis, you would think that the probe has higher noise...

Ok, now I remember the inconvenience you encountered ... Just for comparison, can you post the noise acquisition for the new probe (i.e. the replacement...) you received?

[EEVblog]

I have shorten the inputs of the probe for the setting 10:1. I see some coupling from there DC/DC converter inside the diffprobe. I bought the probe at Bartonix, they could measure the same effect with my probe and with other probes of this type. They got some newer ones and there it is much better. I have two probes which are find now. I use this mostly for lower signals as a diffprobe, not mainly for higher voltages.

The noise of the 10:1 probe is lower than on the other Micsig probes. If you don't zoom in too much in the x-axis, you would think that the probe has higher noise...

I don't recall seeing that on mine.
Unusual that you'd see that as a fault on one unit.

[mrprecision]

Noise measurement of two different DP10007 Probes (x10)

I measured the noise around 12.7mVrms of both probes (inputs shorten, attenuation x10, standard deviation). The datasheet shows a value of (max) 15mVrms. So the noise is within the specs. The noise of the DP10007 is much lower than on the other the two MicSig probes. The good thing on the DP10007 is the low attenuation (x10) which is the same like for the standard scope probes. For low voltage and regular power line voltages the DP10007 is out of my point of view the best choice.

[ExaLab]

Ok, but my question was related to the 90KHz noise effect you had with the old DP10007 probe for which you have requested replacement.
The new one you received as a replacement does it have a similar behavior? In other words, setting the timebase to 2uS are you able to trigger and to see the same 90KHz noise in the new one too?

If yes, can you post a screenshot of this 90KHz noise?

[ExaLab]

Yesterday i got my new probe (DP10007).

Input referred noise compliant with the manufacturer specifications (and with data measured by mrprecision):

13.7mV rms (x10) and 29.1mV rms (x100)

The "anomaly" a with frequency of about 90 khz observed by some people is not present...
Next steps, evaluation of accuracy, bandwidth and CMRR.
The BNC connector of the probe appears quite hard to fit into the scope socket. May be it is a bit out of specification. Probably I'll replace it to avoid unnecessary stress to the oscilloscope connectors.



To the guy who asked for the manufacturing date (and now he seems to have removed the related post):
my probe is dated 16-03-2021. The rms noise you measured in your probe seems a bit out of specification. To avoid the effects of any dc component I suggest to redo the acquisitions using the "Stdev" instead of the "rms". If still out of specification, measure the noise of your oscilloscope and try to subtract it from the acquisition made   ("probe noise"^2 = "measured noise"^2  - "oscilloscope noise"^2)

[ExaLab]

Some further measurements I've done on my probe


Static gain accuracy:

x10       -0.53%   (spec. +/- 2%)
x100     -1.12%   (spec. +/- 2%)


Input referred offset:

x10     + 9.8mV
x100   + 47.2mV


CMRR:

Freq.                 x10    x100    Micsig Spec.

DC (50Hz)         -89     -79      -80
20KHz               -75     -74      -60
100KHz             -72      -63      n.a.
1MHz                -57      -42     -50
10MHz              -39      -22     -40




Rise Time (Step response):

                 Rise Time   Fall Time     Bandwidth Estimation (from step response) (*)

x10            3.71 ns      3.71 ns      94 MHz
x100          3.37 ns      3.60 ns      100 MHz

(*) Single pole low pass filter model


Measured Bandwidth (-3dB):

x10     128 MHz
x100   118 Mhz


Bandwidth Flatness:

              +/- 1dB          +/- 3dB

x10        DC - 33MHz      33MHz - 100MHz
x100      DC - 33MHz      33MHz - 88MHz





Concluding:

Pros:

- good overall performances considering the price
- very nice and compact package design

Cons:

- CMRR below specification
- too long (and oversized) input cables can cause high frequency artifacts. I would have preferred them shorter (20-25cm) with a possible extension supplied (e.g. Tektronix approach...)
- BNC connector out of specification (it hardly enters the oscilloscope connectors, risking damaging them)

[EEVblog]

I've been informed by Micsig that a new version of the DP10007 is coming shortly, presumably to fix the CMRR issue I raised.

[voltsandjolts]

Thanks for the heads-up on that Dave.
Unfortunately I need to push-the-button on a DP10007 soon-ish, so I will probably get old stock.

Here's hoping the improvement will be retrofittable to older units 

[EEVblog]

Thanks for the heads-up on that Dave.
Unfortunately I need to push-the-button on a DP10007 soon-ish, so I will probably get old stock.
Here's hoping the improvement will be retrofittable to older units 

Yeah, it's not a big deal. It's just that they matched the specs of the Sapphire HVP70 because that's what I asked them to try and do and they said they were able to do that, but turns out it fell short on CMRR.

[RafaPolit]

This is a very useful thread, thank you!

I just bought a Keysight DSOX1204G, which should arrive in about a month or so (don't ask, living where I live has some serious disadvantages!).

After researching a bit, I would probably really want/need a differential probe.  Keysight's cheapest one goes for about $700+ (compatible with the 1024).  So, this really looks like a promising option without breaking the bank.  Should I just pull the trigger (Amazon US claims only 2 left in stock), or is the new version coming so soon that it would be better to wait?

Thanks again for any feedback, and since it's my first post... just can't praise EEVblog and Dave high enough, you are really an essential tool for us amateur / newcomers to the field.  Thanks a bunch!

Rafa.
EDIT: Amazon's seller is SainSmart Official, not sure if that is important info and if they are a trusted seller, price is really much lower than on MicSig's own site, so wondering if there is some weird issue going on?

[Ferrer]

Hi friends. First of all, I want to express my respect and gratitude for all the topics and comments that I have seen developed in this forum, of great value for those of us who enjoy electronics.

In relation to the Micsig DP10007, there is a question that confuses me and that I have not seen dealt with in this thread: What is really the maximum voltage that it admits as input: 700 V (DC + AC Peak), or 700 V (DC + AC Peak to Peak) ?.
- The specifications found on the Micsig website indicate 700 V (DC + AC Peak)
http://www.micsig.com/html/41.html#
- In the Product Review of Bob Kalpon's channel it seems that the limit is 700 V (DC + AC Peak to Peak)

- In some stores on the web it also seems that the limit indicated in the Micsig specifications is not correct, and they state that said limit is 700 V (DC + AC Peak to Peak)
https://www.sainsmart.com/products/micsig-dp10007-high-voltage-differential-probe-kit

I no longer know what to believe. Could someone tell me what is your opinion on this?

I think it is an important question if we are going to use the differential probes on the hot side of a switching power supply if the mains voltage is 240V (It could mean a Peak to Peak voltage of around 680V, and a small Variation of the mains voltage could cause the device to explode and, as a consequence, destroy the oscilloscope).

[bdunham7]

If they matched the specs of the ubiquitous Sapphire-OEM model (widely sold under various brand names like BK Precision, Probemaster, Tektronix, Cal-Test, etc) then it works like this:

That probe (the Sapphire) can withstand an input of 1400 volts peak +/- relative to the other lead, which if it was sinusoidal AC, would be about 1000V_RMS or 2800V_p-p.  It can read (produce a scaled output) of up to 700 volts peak +/- because it has a 100X scale factor and a 7 volt peak maximum output.  The original (or at least the ones I remember) Sapphire models had 20X/200X scaling, so they had the same 7 volt peak output, but could measure the full 1400 volts peak.  They later released a 10X/100X version like the MicSig, apparently because many scopes don't have 20X/200X attenuation settings and people don't like doing math in their heads.

The Micsig DP10007 may be based on the DP10013, in which case the peak number (for safety, not readings) may change to 1300 volts.  Or perhaps it is only the 700 volts that it is rated for.  In any case, it should be good for 1400V_p-p or 500V_RMS.  Note that they don't have a frequency derating curve for it.  I rather doubt it can take 500V_RMS @ 70MHz.

[EEVblog]

I think it is an important question if we are going to use the differential probes on the hot side of a switching power supply if the mains voltage is 240V (It could mean a Peak to Peak voltage of around 680V, and a small Variation of the mains voltage could cause the device to explode and, as a consequence, destroy the oscilloscope).

I believe that is it's operational range. i.e. the range for which it is guaranteed to give an output to within specification (subject to derating). It will not suddenly blow up above that.
It handles my 245V mains voltage waveforms no problem at all. I think I even took it to 270V RMS or so no problem, but I'll have to recheck this. I have a high voltage AC generator reference standard (up to a few KHz only) so I can take it up until the output waveform clips. I will have to do a short video demoing this sometime this week.

[Kean]

FWIW, on my DP10013 set to 50X (130V max differential voltage) the over-range indicator comes on at about 95V_RMS, i.e. about 270V_P-P
So I'm pretty sure it is referring to DC + AC Peak, not DC + AC P-P

[voltsandjolts]

The maximum measurement range: 700 V (DC + AC Peak)
The maximum rated safe range is more than that.

Note the input differential will never see the peak-peak voltage, so that doesn't apply.

[Ferrer]

Thank you very much for your kindness dear friends. Your verifications shed light on this very confusing issue considering the diversity of positions held on the matter. I will wait and watch carefully the video that will be published on the EEVblog channel to end any doubts.
And again, thank you very much for your extraordinary kindness.

[RafaPolit]

...
- In the Product Review of Bob Kalpon's channel it seems that the limit is 700 V (DC + AC Peak to Peak)
...

I looked at that video, seems like a good product! I purchased one and am waiting on the oscilloscope to arrive.

Regarding the video, there is a moment when he claims there is a slight phase shift at higher frequencies, and then, at even higher ones, that the wave is "ahead" of the original.  I was wondering if that is even possible or if he misreads it and it is so forward-phased that it almost reaches the next period and is delayed almost an entire period minus a small percentage?  Or can it really be presented "before" with a negative phase?

Thanks for any insight,
Rafa.

[EEVblog]

Teardown photos:
Micsig DP10007 High Volatge Differential Probe Teardown by Dave Jones, on Flickr

Micsig DP10007 High Volatge Differential Probe Teardown by Dave Jones, on Flickr

[EEVblog]

Video test of clipping. Decided to make it a main channel video with teardown.
TLDR; clips at 580Vrms, flashes overrange alert at 470Vrms, survives 1100Vrms just fine, output goes to +/-8V
Doesn't mean it'll meet the performance specs at those limits, but that's what they are

[Ferrer]

Superb verification of how far this differential probe can go. Clear, precise explanation and also fun and entertaining, like all your videos. I am amazed that the probe can withstand rms voltage above 1KV without exploding. I have enjoyed watching you dissect the probe like a CSI forensic surgeon. By the way, you are a magnificent detective at looking for bastard screws. Thank you so much Dave !!
A hug from Mallorca, Balearic Islands

[Kibabalu]

...
- In the Product Review of Bob Kalpon's channel it seems that the limit is 700 V (DC + AC Peak to Peak)
...

I looked at that video, seems like a good product! I purchased one and am waiting on the oscilloscope to arrive.

Regarding the video, there is a moment when he claims there is a slight phase shift at higher frequencies, and then, at even higher ones, that the wave is "ahead" of the original.  I was wondering if that is even possible or if he misreads it and it is so forward-phased that it almost reaches the next period and is delayed almost an entire period minus a small percentage?  Or can it really be presented "before" with a negative phase?

Thanks for any insight,
Rafa.

Dear Rafa,

your assumption is correct.

Due to the input voltage divider, with four parallel resistor-capacitor combinations in series, you get this high phase shifts.

[ExaLab]

Video test of clipping. Decided to make it a main channel video with teardown.
TLDR; clips at 580Vrms, flashes overrange alert at 470Vrms, survives 1100Vrms just fine, output goes to +/-8V
Doesn't mean it'll meet the performance specs at those limits, but that's what they are

A bit strange that the nodes of the four capacitors of the input HV divider are not connected with the related node of the symmetrical resistor network.
This approach does not seem to guarantee a balanced distribution of the input voltage between the various capacitors... I'm wrong?

[Hydron]

Immediately from Dave's teardown pics I can see a significant design improvement from the DP10013 - this new one has an extra trimmer to allow CMMR to be tweaked between ranges.
The older DP10013 only had a single trimmer but 2 gain ranges - it relied on resistor matching to maintain CMRR between ranges, and the single trimmer had to be adjusted to the best compromise between high/low range CMRR. When I opened mine up to shorten the leads (another story - it had some _nasty_ peaking at about 70MHz) I ended up tweaking one of the gain resistors (where VR2 is in the new design) to get it bang on in both ranges.

[mrprecision]

I have two MicSig DP10007 probes and did for both probes and for both measuring ranges some measurements. The CMRR is quite ok, only for the higher frequencies and for the 100x range the CMRR is quite out of the datasheet. What do you think?

Probe A, Bodeplot 10x


Probe B, Bodeplot 10x


Probe A, Bodeplot 100x


Probe B, Bodeplot 100x


Probe A, CMRR 10x


Probe B, CMRR 10x


Probe A, CMRR 100x


Probe B, CMRR 100x