The procedure I am going to illustrate aims to harmonize the high frequency gain (from MHz upwards) of the two input channels. This will allow us to further optimize the CMRR compared to the results obtained from the Simplified Calibration (the improvement @10MHz could be over 10dB!!)
As already mentioned, this is achieved by altering the imbalance of the pair of zeros (referred to the TF of the input stage) of one channel relative to the other.
This step should be an integral part of the "Low Frequency Compensation" (generally called "Short Term" and "Long Term" Compensation) of each channel as it aims to align the zero of the Input Front End with the dominant pole of the following stage. For various reasons related to design shortcomings (which I will not address for the sake of brevity...), for our unit this is not possible.
However, this does not prevent us from using a "Preset and Evaluate" approach (simple, effective even if it is a little time consuming…). The aim is to acquire the trend of the CMRR as a function of the position of VC4 and then choose the best compromise. We will test a few different configurations for the HF frequency response of the positive channel and then choose the one that offers the best results in terms of HF CMRR.
In practice, we will recalibrate the probe for various positions of the VC4 trimmer while tabulating the CMRR at the frequencies of greatest interest (100KHz, 1MHz and 10MHz) (see picture below).
The procedure proposed involves angular steps of approximately 20 degrees counterclockwise. Naturally, if at the first step (+20 degrees) the average CMRR were to significantly worsen compared to the starting configuration (0 degrees), instead of continuing with positive angles (+40, +60 etc.) it will be necessary to switch to negative angles (-20, -40 etc.).
Before starting with the procedure, I remind you again that the angles I am referring to are relative to the initial position of VC4 and that this reference position changes from one unit to another.
I also remind you that the absolute zero of all the four capacitive trimmers coincides with the notch arranged horizontally and that the absolute rotation limits are worth +/- 90 degrees. Therefore, if during the Full Calibration Procedure VC4 (or VC3) reaches these absolute limits (i.e vertical notch of the trimmer) the evaluation angle of VC4 must stop here.
If the CMRR trend is positive but you have reached the above limit it will always be possible to continue with the analysis by unbalancing (with angles opposite to those adopted for VC4) the VC2 trimmer. This remote event requires a new LF compensation and we will analyze it separately (if necessary...).
The picture below ghraphs a typical behaviour of the CMRRs as function of the VC4 relative angle.
Full Calibration Procedurea) Simplified Calibration Procedure Perform the “Simplified Calibration Procedure”
b) Annotation of the VC4 Position Annotate the exact VC4 position and give it a ID name (e.g. 0˚)
Even if we will only act on VC3 and VC4, to be safe, take a photo of the initial position of all the four capacitive trimmers
c) CMRR Evaluation 1- Make sure that the x10 range is selected
2- Measure the CMRR @ 100kHz, 1MHz and 10MHz
3- Note these three values next to the actual position of VC4
4- If the actual position of VC4 is 80˚skip to point h)
d) Changing the position of VC4 1- Rotate VC4 counterclockwise approximately 20 degrees
2- Give a ID name to this new VC4 position (e.g. 20˚, 40˚ etc)
e) AC CMRR Adjustment 1- Maintain the x10 range
2- Twist the probe input leads and connect both (red and black) to a 20KHz, sine wave signal source (minumum amplitude: 20 Vpp / 200 – 300 Vpp recommended).
Make a direct ground connection between the signal source and the oscilloscope (this can simply be done with the probe used to monitor the signal of the source).
3- Adjust the oscilloscope to properly display the waveform (note that due to the strong channel decompensation, the signal amplitude in this step will be quite high)
4- Rotate VC3 until the amplitude of the displayed waveform is minimized (adjust the oscilloscope gain accordingly). Note that the direction of rotation of VC3 is always
opposite to that of VC4 at step d)!
f) HF CMRR Adjustment (for Advanced Upgrade Only) 1- Perform the HF CMRR Adjustment limited to the x10 range (VC5 adjustment only)
For details please refer to the step d) of the “Simplified Calibration Procedure”
g) Resume the procedure from point c)h) Choice and setting of the optimal position of VC4 1- Choose the optimal angle of VC4 based on the following criteria: aim for the best CMRR @ 1MHz and 10MHz values that do not excessively penalize the CMRR @ 100KHz.
As can be seen from the graph below (for which the optimal angle is about 40˚…), in this position the three acquired CMRR values are equally spaced by 8 -10 dB.
Note that the optimal angle chosen could also be an intermediate value between those tested (e.g. 30˚) and as already anticipated, if the trend of the CMRR for positive
angles is not favorable, the procedure will also have to be extended to the negative angles of VC4.
2- Set the optimal and final position of VC4
i) AC and HF CMRR adjustment
Perform the above step e) (AC CMRR adj) and step f) (HF CMRR adj) for the final position of VC4
l) x100 Range HF CMRR Adjustment (for Advanced Upgrade Only) Perform the HF CMRR Adjustment limited to the x100 range (VC6 adjustment only)
For details please refer to the step d) of the “Simplified Calibration Procedure”
OK, my discussion ends here!! I hope it has been useful to you in improving the CMRR of this differential probe. If yes, with reference to HF CMRR only, you will now have a performant differential probe!
Let me know!
Of course, there remain many other outstanding problems that for reasons of time (and professional ethics..) I cannot address.
Among these I remember only a few:
1) The actual bandwidth of the “pure probe” limited to approximately 45MHz
2) The obvious discrepancy between the advertised bandwidth (100MHz) and a pair of non-removable input cables over 45cm long
3) The abrupt degradation of CMRR (worse than -40dB) at frequencies close to 6700Hz
4) The presence of traces ( ̴ 0.20mVpp) of residual noise at 24KHz at the output
5) The absence in the user manual of any reference related to the derating curve for the input voltage (this lack speaks for itself!!)
For these and many other reasons, not least the terrible HF CMRR (in the absence of the changes I proposed…), at the moment, if you are interested in purchasing a differential probe I strongly suggest evaluating other manufacturers. Naturally, if in the near future Micsig were to propose some new product and after in-depth testing I should verify its effectiveness, I’ll have no problem enhancing its characteristics…
Don't be attracted by the low price of a product if it is not able to meet your expectations and above all, be wary of all those reviews characterized by superficiality and evident lack of professionalism of the reviewer.