From: "Terrell W. Fritz"
Date: Sat, 13 Jan 2001 20:03:42 -0700
Subject: Re: [TSSP] Top voltage testing
Hi Paul, I would like to disqualify the earlier P6015 probe voltage level measurements. The frequencies should still be ok. I found that the probe had some aged silver plated contacts in the ground path that added resistance and probably other odd things. Also, this probe has a nasty seven element compensation network that was far out of adjustment in a way that significantly affected the earlier measurements. After giving the probe a really good looking over, I feel it is much better suited to multi kV circuits rather than fine measurements like this. They have spring contact on aluminum surfaces and the compensation network is more just a trick to get to 75mHz rather than a help in accuracy. The only way I have to calibrate it is against the other 10X probe. I have got it working well now but I don't think it has any advantage over the 10X probe and I still have a bad feel about using it in this application. 100 volts is only 0.25% of its intended range and the network is not super accurate at that level. The network could very well foul the signals at very low levels. Great probe to stick into 15kV RF but sensitive resonant 50 volt levels are not it's thing. The modern 10X probe seems to work very well and I trust it rather than the big high voltage probe in this case. Cheers, Terry At 10:14 AM 1/10/2001 +0000, you wrote: >I wrote: >>> BTW, your relatively low Q coil would be ideal for a test of >>> pn2511, equ 9.18 which predicts the frequency difference >>> between that of max Ibase and that of max Vtop. Reducing Q >>> further by insertion of extra inner sonotube should increase >>> the frequency difference. > >Terry wrote: >> I can cut the thinner tube to fit in the terminated coil. >> I think the Q is humidity dependant so I would have to retake >> the Q measurements to insure they didn't drift over time. > >Yes. Probably adequate to use the bare coil for this experiment, >so no need to cut the inner tube. > >I wrote: >>> It'll be interesting to see if we can obtain a reliable match >>> with the third test result, ie the ordinary scope probe. Would >>> be handy if we could. Capacitance to the screen of the probe lead >>> is probably the biggest random factor. > >So far the ordinary scope probe seems to be the only one giving >the 'right' answers! > >Terry wrote: >> The 1000X probe has a very long cable unlike the 10X probe so I >> could come off the ceiling. I could probably mount the small >> light scope to the ceiling too if the top down approach is best... > >It's not so much the amount of capacitance involved - we can adjust >for that, but repeatability is something to aim for. > >Terry wrote: > >> So I guess the input Z is 361.46 ohms. > >Thanks. That agrees with the predicted 350 ohms. We've seen this >before when the loss factor is fiddled to match the measured Q, >the input resistance comes out right, which is reassuring. > >> The F3 overtone is at 319.226kHz > >OK, the model is in error by -3.1% at f3, thats within limits. > >Concerning the test #1 26% discrepancy in voltage gain. This is >very odd. The model needed a probe cap of 6.95 pF to reproduce >your f1. In test #3, which is giving a good 2.3% error, the >probe cap is a whopping 24.35 pF. I can't understand how just >a change of top cap alters vgain from 'lots less than Q' to >the expected 'just over Q'. Could there be a grounding >problem - when P6015 is in use is the scope ground floating >wrt the signal generator ground - that would just about account >for the low reading (which is just over half of the expected top >volts)? > >Perhaps you can setup with the P6015 and then add enough vacuum >variable extra ctop to pull the f1 down to 70.6 kHz to match >the test #3 f1 and see if P6015 still shows much less vgain than >the Q. > >One way or another, there's something very odd about the test #1 >top volts reading. Maybe the probe wasn't quite making contact >with the toroid so we only got a capacitive pickup? The only >situation we should get a Vgain less than Q is for short fat coils >having an Lfac > 1.0, and even then the gain is only a little less >than Q. > >Perhaps you can take another look at test #1 and poke around to >see if there's a problem with the Vtop measurement? > >Cheers, >-- >Paul Nicholson, >Manchester, UK. >--
Maintainer Paul Nicholson, paul@abelian.demon.co.uk.