---+++ Filter design page for Cryoboard
On the simulation I will be comparing both the DacOut signal and the feedback voltage. Ideally the feedback voltage should be identical to DacOut input. When they are equal, both the demanded current, and measured current are the same (assuming C13 and C14 don't get in the way). From the schematic above you can see that C13 and C14 have both been set to insignificant values.
With a 1K load we get:
* DriveChan1kLoadOrigFilt.jpg:
With a 100K load we get
* DriveChan100kLoadOrigFilt.jpg:
There is considerable difference in output current levels between these two different loads. This will cause us calibration problems. To understand why there is a difference with the two loads, you have to think of the integrator being slew-rate limited (by its new RC time constant). In the case were the load resistance is higher, to get a specific current level, requires more voltage than that for the lower load resistance, and therefore take more time.
Updating the RC time constant of the integrator from 0.68ms to 0.05ms (make C10 & C18 500pF instead of 6.8nF) gives the following circuit:
(Note that I've now changed the caps on both sides of the load to 2.2nF)
* DriveChannelCircuit_ver2.jpg:
with a 1k load:
* DriveChan1kLoadNewFilt.jpg:
with a 100k load:
* DriveChan100kLoadNewFilt.jpg:
You can still see that there is a little difference between the 1k and 100k scenario in terms of their voltage waveforms. This means that our integrator could still be made faster but this looks unnecessary to me. We could even slow it down a bit more if a small difference isnt a problem.
This topic: CryoElectronics > WebHome > CryoFilterDesign Topic revision: r1 - 2010-12-02 - WinterlandUser
Drive Channel Circuit (Feedback part)
Since our sampling rate of the ADCs on the Cryoboard are lower than originally designed for, the anti-aliasing filters need to be redesigned. Originally I had assumed that the Drive Channel circuit, see next figure, had a much faster response time than any input. This is not the case. The circuit has been simulated in LTSpice (note that the THS4131 is modeled by an LT1994 - should make no difference) A 2V p-p 117Hz sinusoidal demand is present at the input. * DriveChannelCircuit.jpg:
On the simulation I will be comparing both the DacOut signal and the feedback voltage. Ideally the feedback voltage should be identical to DacOut input. When they are equal, both the demanded current, and measured current are the same (assuming C13 and C14 don't get in the way). From the schematic above you can see that C13 and C14 have both been set to insignificant values.
With a 1K load we get:
* DriveChan1kLoadOrigFilt.jpg:
With a 100K load we get
* DriveChan100kLoadOrigFilt.jpg:
There is considerable difference in output current levels between these two different loads. This will cause us calibration problems. To understand why there is a difference with the two loads, you have to think of the integrator being slew-rate limited (by its new RC time constant). In the case were the load resistance is higher, to get a specific current level, requires more voltage than that for the lower load resistance, and therefore take more time.
Updating the RC time constant of the integrator from 0.68ms to 0.05ms (make C10 & C18 500pF instead of 6.8nF) gives the following circuit:
(Note that I've now changed the caps on both sides of the load to 2.2nF)
* DriveChannelCircuit_ver2.jpg:
with a 1k load:
* DriveChan1kLoadNewFilt.jpg:
with a 100k load:
* DriveChan100kLoadNewFilt.jpg:
You can still see that there is a little difference between the 1k and 100k scenario in terms of their voltage waveforms. This means that our integrator could still be made faster but this looks unnecessary to me. We could even slow it down a bit more if a small difference isnt a problem.
| I | Attachment | Action | Size | Date | Who | Comment |
|---|---|---|---|---|---|---|
 jpg |
DriveChan100kLoadNewFilt.jpg | manage | 51.3 K | 2010-12-02 - 18:06 | WinterlandUser | |
| jpg |
DriveChan100kLoadOrigFilt.jpg | manage | 53.7 K | 2010-12-02 - 18:05 | WinterlandUser | |
| jpg |
DriveChan1kLoadNewFilt.jpg | manage | 52.2 K | 2010-12-02 - 18:06 | WinterlandUser | |
| jpg |
DriveChan1kLoadOrigFilt.jpg | manage | 53.7 K | 2010-12-02 - 18:05 | WinterlandUser | |
| jpg |
DriveChannelCircuit.jpg | manage | 25.5 K | 2010-12-02 - 18:04 | WinterlandUser | |
| jpg |
DriveChannelCircuit_ver2.jpg | manage | 30.2 K | 2010-12-02 - 18:05 | WinterlandUser |
This topic: CryoElectronics > WebHome > CryoFilterDesign Topic revision: r1 - 2010-12-02 - WinterlandUser
© 2020 Winterland Cosmology Lab, McGill University, Montréal, Québec, Canada