Coil Signal Diff. Integrator Sampling Time

[bklein]

I haven't looked into this seriously yet but I haven't seen it discussed here. Everyone is talking about making fast coils but yet looking at typical sampling diff. integrator circuits, it looks to me that the sampling time is too long in duration. It seems to vary from 20-50uS or so. That is pretty major if you look at that first sample and what it contains. What do you think about modifying the typical diff. integrator circuit to utilize shorter duration samples?

[moodz]

I haven't looked into this seriously yet but I haven't seen it discussed here. Everyone is talking about making fast coils but yet looking at typical sampling diff. integrator circuits, it looks to me that the sampling time is too long in duration. It seems to vary from 20-50uS or so. That is pretty major if you look at that first sample and what it contains. What do you think about modifying the typical diff. integrator circuit to utilize shorter duration samples?

ppl should really be looking at ways to sample more effectively ... the diff integrator is old hat technology ...reliable but not cutting edge. You can get an ADC with a sample and hold built in for a few dollars that will sample in 80 nanoseconds ... the conversion time is 1 microsecond .... but thats a screaming sample time compared to your diff integrator samples. Even your 24 bit audio codec will sample at 5 microseconds ( 192 Khz ) and you get two channels in a chip for $4.50

Getting back to the integrator ... if you shorten the time you have got less time to integrate ( after all integration is done over time ) ...

Just some thoughts ...

moodz

[bklein]

I'm suggesting that the sample duration is too long for the first sample. As you suggest, the sample duration can be much shorter. The alternating integration between first and second sample can remain the same though... I guess it comes down to doing it cheap and dirty as it stands or going uC-based.

[Monolith]

I'm suggesting that the sample duration is too long for the first sample. As you suggest, the sample duration can be much shorter. The alternating integration between first and second sample can remain the same though... I guess it comes down to doing it cheap and dirty as it stands or going uC-based.

Integrating 10-20 us in the first sample reduces the amplitude of the signal for short TC targets. However, it also attenuates all high frequency noise.

So a short sample is a trade off between high, short TC signal amplitude and S/N.

Monolith