15-02-2021, 08:53 PM
I'm watching this carefully - I'm also aiming to be doing some trials this week, with IF amplifiers and transformers (as a matter of detail, I'm planning to use two RM5 ferrite core sets, which hopefully will have negligible external field, with coupling windings between them).
The 1nF that Amie reckoned on, for a x10 probe, alarmed me!
I'm in a different place from any capacitance measuring stuff right now. But I'm not away from my 'scope, a square-wave generator, and a box of resistors. So I had a play.
First thing: switch on 'scope (so any input circuitry is powered and behaving as normal) and use DMM to measure input resistance with my (non-switched) x10 probe fitted. Result: 10.07MΩ. In line with expectations.
Next, use two probes, two channels, to confirm that they show identical traces with my square-wave generator, and they are properly compensated. They do, and they are.
Next, add a 91kΩ resistor (came easily to hand!) in series with probe tip, and 'scope the square wave. This channel now shows exponentially slewing rise and fall, also a expected.
Such a rise, reaches 63% of final value, after one RC time-constant. Knowing R, and measuring time to 63%, I can calculate capacitance of the measuring setup, as seen at the probe tip.
The value of R in this, strictly is 91kΩ in parallel with 10MΩ, ie 83.4kΩ, and we also ought to allow that with 91kΩ added, the probe is no longer x10, but x10.1, however the error is only 1% and as can be seen, the square-wave rise isn't quite instantaneous, so there are other errors...
So, the time to reach 63%, using the amazing digitally-controlled measuring cursor* as shown is... 1.2μsec. Knowing R as 83.4kΩ, the value of C is thus 14.4pF.
It's a bit more than expected: I was reckoning on <10pF. But it is in line with figures discussed above, and even allowing for sources of error, I'd consider the result will be within 10% of the actual value.
* precision-controlled by digits of my right hand
The 1nF that Amie reckoned on, for a x10 probe, alarmed me!
I'm in a different place from any capacitance measuring stuff right now. But I'm not away from my 'scope, a square-wave generator, and a box of resistors. So I had a play.
First thing: switch on 'scope (so any input circuitry is powered and behaving as normal) and use DMM to measure input resistance with my (non-switched) x10 probe fitted. Result: 10.07MΩ. In line with expectations.
Next, use two probes, two channels, to confirm that they show identical traces with my square-wave generator, and they are properly compensated. They do, and they are.
Next, add a 91kΩ resistor (came easily to hand!) in series with probe tip, and 'scope the square wave. This channel now shows exponentially slewing rise and fall, also a expected.
Such a rise, reaches 63% of final value, after one RC time-constant. Knowing R, and measuring time to 63%, I can calculate capacitance of the measuring setup, as seen at the probe tip.
The value of R in this, strictly is 91kΩ in parallel with 10MΩ, ie 83.4kΩ, and we also ought to allow that with 91kΩ added, the probe is no longer x10, but x10.1, however the error is only 1% and as can be seen, the square-wave rise isn't quite instantaneous, so there are other errors...
So, the time to reach 63%, using the amazing digitally-controlled measuring cursor* as shown is... 1.2μsec. Knowing R as 83.4kΩ, the value of C is thus 14.4pF.
It's a bit more than expected: I was reckoning on <10pF. But it is in line with figures discussed above, and even allowing for sources of error, I'd consider the result will be within 10% of the actual value.
* precision-controlled by digits of my right hand
