We bought two from Princeton Applied Research because we had to have .01dB amplitude resolution for a similar measurement.
Lower levels, interesting something going on there is. Explained by neither force even circuit theory there be.
Wenzel wrote a monthly column in an IEEE publication one was on Faraday's Law and non-conservative fields. He was great publishing really neat do it at home experiments in a professional journal unfortunately it is copyright and maybe he can't post it on his site?
If you were not using FFT and data converters -- you might get very different results at such low signal levels.
THx-RNMarsh
Last edited:
Any problem, with low power signals and sealed relays that does not oxidize ?
Energy penalty there be pulling signal from noise for. Mmmm, yes. Pay someone must, find the coins Luke should.
So we take about 1m of wire, bend it in half and then try to measure its impedance at audio frequencies by using a singularly inappropriate experimental setup. We do this twice, with thick and thin wire, and we believe we have found that a particular sample of thin wire has a lower impedance than a particular sample of thick wire. Is the world of science really expected to bow down to this nonsense?
If you are looking for a small signal, you don't deliberately design your apparatus so that it is riding on a much bigger signal which is almost identical in characteristics. Subtract one big signal from another (in hardware or software) and you don't have much left but noise and confusion.
If you are looking for an impedance change in bent wire, you need to be very careful to replicate the bending from one run to the next. The spacing (set partly by insulation thickness) is crucial - see the arccosh function for details.
If you are looking for a small signal, you don't deliberately design your apparatus so that it is riding on a much bigger signal which is almost identical in characteristics. Subtract one big signal from another (in hardware or software) and you don't have much left but noise and confusion.
If you are looking for an impedance change in bent wire, you need to be very careful to replicate the bending from one run to the next. The spacing (set partly by insulation thickness) is crucial - see the arccosh function for details.
Small signal switching is demanding of clean contacts, but useful if the impedances have to be quite low (I'm evaluating some reeds right now). Large signal switching has its own set of problems. The use of DMOS is problematic due to on resistance variations with current and rather high capacitance for low resistance.
What I prefer to do is JFET switching with T or 2/3 T networks in which entire feedback and feedback divider arrangements are switched to a summing node. This requires that the given amplifier be able to drive a plurality of networks, so it is demanding in terms of current if the impedances are low. However effects due to the variations in FET on resistances are rendered negligible.
Manual switches have their own set of problems, but with lots of wiping action and certain lubricants to protect from oxides and sufides can work well for a while. Little toggle switches are treacherous, especially beyond a couple of poles, but can be used to control relays.
I think Kamis has a valid point.
As far as I am aware, FETs are excellent switches, but also tend to create a spike when switching on. This of course can be avoided, I know this from my H/K 680 integrated amp, which uses FETs for input selection with abosute zero switching, but they did use extensive logic to control them which is way out of the reach of the DIY crowd.
As far as I am aware, FETs are excellent switches, but also tend to create a spike when switching on. This of course can be avoided, I know this from my H/K 680 integrated amp, which uses FETs for input selection with abosute zero switching, but they did use extensive logic to control them which is way out of the reach of the DIY crowd.
Amen to that! Quality toggle switches have incredible lifetimes, but I agree they are best used to switch on something else, a BJT, FET or even MOSFET.
- Status
- Not open for further replies.