[Test-Equipment] HV probe resistor selection

[David]

It hardly takes exotic high impedance circuits to cause a problem.
Oscilloscope CRTs for instance usually have at least a couple of
elements like focus, scan expansion, and grid driven with DC
impedances in the 5 to 20 megohm range which makes 75 megohm probe
measurements marginal.

On Sat, 23 Feb 2013 09:34:13 -0500 (EST), djed1 at aol.com wrote:

>Here's what I think are the design issues:
>
>1. You want the impedance of the probe to be say 100 times the impedance of the circuit under test. So unless you're measuring exotic very high impedance circuits, I would think 25 megs would be OK.  75 megs would be better.
>
>2. For a voltage division of 1000, the choice above defines the shunt resistor:  25meg/25K or 75meg/75K.
>
>3. You want the impedance of the meter to be constant and say 100X that of the shunt resistor.  This will give you accuracies of 1-2%- if you need better accuracy then put a trimmer in series with the shunt resistor.  You can also tweak the shunt value to compensate for a lower Z in the meter.  Of course, this assumes you have an accurate high voltage reference.
>Ed  W2EMN
>
>-----Original Message-----
>From: DaveC <davec2468 at yahoo.com>
>To: Discussion of Electronic Test Equipment <test-equipment at mailman.qth.net>
>Sent: Sat, Feb 23, 2013 5:12 am
>Subject: Re: [Test-Equipment] HV probe resistor selection
>
>Sent from my iPhone
>
>On Feb 23, 2013, at 1:57 AM, David <davidwhess at gmail.com> wrote:
>
>> Automatic ranging or even manual ranging on a correctly designed
>> digital meter will not allow an overrange situation to cause damage
>> unless it exceeds the meter's voltage rating.  For example if the
>> meter says it is rated at 1000 VDC and 750 VAC which is typical, then
>> that should apply on any voltage range without damage.
>> 
>> The input impedance should be constant, usually 10 Megohms shunted by
>> 30 to 100 picofarads, with the exception of some bench or labratory
>> meters which may have a higher input resistance on the lower DC volts
>> ranges.  My HP 3478A is greater than 10 Gigohms on its 30 millivolt,
>> 300 millivolt, and 3 volt ranges which I was not really aware of until
>> I used it with my Fluke 80K-6 and the readings did not match with any
>> of my 10 megohm input meters.  Some meters will have a lower input
>> impedance on AC voltage ranges (that pesky HP 3478A again is 1 megohm)
>> and they all have a fair amount of shunt capacitance which will affect
>> higher frequency measurements.
>> 
>> As Peter point out, the Fluke 80K-6 only has an input resistance of 75
>> megohms which will be too low to make accurate measurements of high
>> impedance circuits.  Sometimes with a little math you can figure out
>> what the unloaded circuit voltage really is.
>> 
>> Given the selection of resistors at Digikey that you linked, the
>> easiest replacement would be 3 x 25M units in series assuming the 75
>> kilohm shunt resistor in the probe is ok.
>
>Thanks for your comments.
>
>The 25 M resistor is 4 inches long. No room in the probe body for 3x.
>
>I'm leaning toward the 1 G resistor and selecting an appropriate R2 plus 10- 
>turn pot to calibrate.
>
>But still open to suggestions.
>
>Thanks 
>Dave