From: "Malcolm Watts"
Date: Mon, 19 Feb 2001 09:23:12 +1300
Subject: Re: [TSSP] Top voltage testing
On 16 Feb 01, at 15:03, Paul wrote: > Hi Terry, > > The situation is looking a little better with the raised signal level, > is this once again the HP reading in error with small signals? > > > The reason the current is dropping with the probed coil is due to > > the Q of the coil dropping with the probe in place. > > Ah, I haven't been bothering to take account of the resistive part of > the probe, since other losses are not well quantified - I assumed it > wouldn't make much difference. In fact, I put in a 10 Meg load resist- > ance alongside the probe capacitance, and the probed coil Q fell to > 102 compared with 195 for the bare coil. Previously, without the > allowance for the 10 Meg, I got a Q around 170 for the probed coil. > Your base current readings have highlighted the fact that I must take > the probe resistance into account too. I did mention it a while ago didn't I? Again you have to consider the transmission line equations and the effect of Zft^2 on one end with a load of some sort at the other. A simple calc reveals this starkly. This problem reared its ugly head back in '94 when I was researching my article and it was from this that I decided Q measurements were best taken with the probe at a distance. I understand of course that the modelling will take this resistance into account for direct probe connection. Regards, malcolm > Adjusting the model's ctop to match your new f1, we now get: > > measured modeled error > f1: 310.9 kHz 310.9 kHz 0.0% Ctop adjusted to match: 3.145 > pF f3: 847.8 kHz 828.4 kHz -2.3% > > Zft: 34450 ohms 34877 ohms +1.2% > Les: 17.64 mH 17.85 mH +1.2% > > Lfac: 79.8% 80.8% > > You mention that the coil is not 'perfectly well known', and we have a > 2% discrepancy in the Ldc, with 2.3% in f3, so maybe more refined > dimensions would resolve the remaining error. > > Having got a reasonable match on Zft, perhaps we should have a quick > look at some of the Q-dependent measurements. Looking at your sweep > charts, the bandwidth of the probed coil f1 looks to be around the 3 > kHz region, so maybe the Q predicted at 102 is not too far from > reality. For the probed coil, the model predicts: > > Q: 102 > Rin: 293 ohms > Vgain: 119 > > So far it's still dry and sunny outside, so the prospects for tests on > a system with Lfac=110% are encouraging. It's also time that I > collated the various 'top voltage testing' measurements and put > together another web page. > > Cheers, > -- > Paul Nicholson, > Manchester, UK. > -- >
Maintainer Paul Nicholson, paul@abelian.demon.co.uk.