I have a serial communications interface with high and fast common mode noise to test/measure/validate. Using regular 10x scope probes in ChA - ChB mode or even differential probes just doesn't give good enough common mode signal rejection in the MHz range. Optically isolated probes have great common mode rejection but boy are they expensive comparitively speaking. Looking over the market I see
Tektronix IsoVue series
https://www.tek.com/en/products/oscilloscopes/oscilloscope-probes/isovu-isolated-probes Lecroy HVFO108
https://www.teledynelecroy.com/probes/high-voltage-optically-isolated-probes/hvfo108 Saker XFVP, GEVP
https://saker-mv.com/products/ Micsig SigOFIT/MOIP
https://www.micsig.com/SigOFIT/Combining functionality, price, and availability I decided on the Micsig MOIP02P 200 MHz model. Below is a mini review/our first impressions.
Pros: - No Battery
- More Affordable than other options; significantly so compared to the "A" brands
- Quick push button calibrates offset, gain without removing from target
Note: Cal button sets to offset value set manually. Not to 0V.
Medium: - Uses relatively small MCX connectors to go between attenuation extensions and the target board but SMA for direct connection
Cons: - Male BNC/plug like on cables plugged right onto scope female BNC seems like asking for potential mechanical damage
We used BNC female to female adapter and a standard BNC extension cable to the scope
- Optical receiver has a fan. Not too loud but a tiny fan with a high pitch at 4 and 8 kHz per phone audio spectrum analyzer.
Potential Concerns: - Will SMA connector used to connect to the signal input/attenuators wear out quickly?
Have to remove the attenuator to put back into the carry case.
- Manual does not make it clear how high the input voltage can go before there is damage to the probe head.
Noise Seems Just Over the Spec::
Using MDO3054 at 10x 50 mV/div internal 50 Ohm terminator scope noise is 2 mVrms at full BW, full sample rate
MOIP02P measured with 10x atten installed and a 50 Ohm short in input
Mean p-p: 164 mVpp
Mean rms: 16.8 mVrms
Mean Mean: 0.024 mVdc
Assuming random noise and no correlation between scope, probe noise the added probe noise is
Sqrt(16.8^2 - 0.024^2 - 2^2)/10 equals 1.67 mVrms noise from the probe.
Probe spec is "System Noise < 1.41 mVrms"
There is a 310 kHz pattern in the noise that looks like coupling form a dc-dc converter. Checking the supply mains to USB power converter it seems to work at only 36.9 kHz. Active section is 8.5 uSec.
Perhaps lightly loaded and working with multiple cycle skipping? Putting a 100 MHz current probe around power source USB cable showed some noise in the 10 mA p-p region. 2.5 mA 0-p spike every 30.5 uSec on average or right about at the 36.9 kHz mains dc-dc switching frequency.
Seems to be some 300 kHz noise that is probably from some switching power supply.
Per FFT spectral peak is exactly at 310.9 kHz with amplitude -61.8 dBVrms.
Or, only 0.8 mVrms out of 30 mVrms total noise.
Main FFT Spectral Peaks:
310.9 kHz -61.8 dBVrms
621.5 kHz -63.4 dBVrms
99.5 MHz -62.2 dVBrms
Rest of spectral peaks below -70 dBVrms
Taking a standard 10x probe with the ground strap shorted to the probe tip to make a very crude magnetic EMI probe and putting next to the optical receiver yields a 99.5 MHz more or less sine wave. Does the probe use 99.5 Mhz inside it for something?
Square Wave Response:Square wave step response looks good but the fastest siggen square available only has 5 nSec rise time. The specified rise time is < 1.75 nSec and thus faster than my square wave source.
Frequency Response:Frequency Response seems to be exactly as listed. Right at 200 MHz for -3 dB.
Used an IFR 2025 RF generator with type N to BNC cable to a pass through 50 Ohm terminator at +4.4 dBM or nominally 1 Vpp into the MOIP02P 10x probe. Oscilloscope was Tektronix MDO3054 (500 MHz). See the frequency response plot.
Initially I used a Pomona short (6 inch) BNC to clip lead to measure the frequency response. Just that 4 inches of twisted wire greatly change the frequency response from 25 MHz and up. Peaked at 150 MHz at +6.2 dB and at 200 MHz was still +3 dB.
I then took a BNC chassis mount male and soldered an MCX connector to its leads to make a home-brew BNC to MCX adapter. The frequency response graph is using this home-brew adapter.
DC Drift With Time:Turned the probe on from cold/room temperature with the 10X attenuator on with 50 Ohm terminator and then recorded the dc level on an oscilloscope for an hour. Scope had already been on and settled for hours.
Time Vscope
0 +19.6 mV
90 sec + 9.6 mV
15 min + 3.6 mV
30 min + 1.6 mV
60 min 0 mV
Remembering that the actual input to the scope is 10x smaller than the above the dc drift from 90 sec after turn on to fully settled is only 1 mV and after 15 min to fully settled is only 1/3 mV. These results are pretty good. Even better when you can re-zero and calibrate with a simple button push and only 1 second. Certainly nicer/less annoying than Hall effect current probes.
Summary/Next Steps:So far so good. But, I haven't have the probe that long and haven't had a chance to look at signals with significant common mode. That is next on our list as we use it for its intended purpose rather than just general bench testing/checkout.