Designing an FSR 40*40 Matrix that resolves ghosting and other medical problems

[ahmedragia21]

Galgso
here is the update one with your apperciated comments

 

[Galgso]

Galgso
I need the lock in amplifier so that I excite the matrix with the clock so that to have average dc current of 0, because I don't want the current to flow to saliva of the patient. That matrix will be a force resistive sensor to be bite on. so I don't want current pass thru the saliva.

Yes but why do you need a lock-in amp? Your clock is already an AC source. If you capacitively couple the clock to the matrix, you will get AC only. Also note that you can't use the clock directly if you don't want AC on the output to the matrix because a 5v clock will have a 2.5v DC bias on the signal. If you filter if with an AC coupling capacitor you will get 2.5v AC peak

 

[ahmedragia21]

Galgso
yes but in that case how do I convert the current to voltage ? and sense it ? I'm exciting the rows, so I should lock in at column side

 

[Galgso]

Galgso
yes but in that case how do I convert the current to voltage ? and sense it ?

Is the output you need from the matrix digital or analogue? That is, do you only need to know if a particular cell is high impedance (megaohms) or if it's within the 10k to 100k range? Or do you also need to know the precise resistance of the cell?

 

[ahmedragia21]

Galgso
The output of the matrix that I need is analog, I will sense it with the ADC from the digital lock in op-amp. I need to know prcise resistance of the cell.

 

[Galgso]

A series resistor is sufficient for determining the resistance of the matrix connection.

The circuit below feeds AC into the matrix. The connection between the matrix and the 10k resistor can be through the mux. The negative mux supply will have to be your -5v rail from the charge pump because the mux is not rated for voltages below ground. The op-amp amplifies the signal to make it easier for the ADC to read precisely (may not be necessary depending on your ADC resolution). The op-amp has a diode in the feedback loop to ensure that the output voltage does not go below ground as this may damage your microcontroller. If you have a decent microcontroller like an STM32 you should be able to read the output fast enough to catch the positive part of the signal. Once you get a reading you can use the known resistance of the series resistor (I used 10k) to calculate the current and you know your supply voltage so you can determine the matrix resistance.

 

[ahmedragia21]

Galgso
Thanks so much for your input.

I shared the clock from rows to cols
For the Columns Mux, it's VCC will be 5V, and it's GND should be -5V from the charge pump?

 

[Galgso]

For the Columns Mux, it's VCC will be 5V, and it's GND should be -5V from the charge pump?

Yes to allow it to work with AC. Note that the maximum supply voltage for the 74HC4067 is 10v so make sure it you don't go above +5v and -5v. Also keep in mind that the mux must be in series with the matrix if you want to be able to switch between channels. The clock should go directly to the matrix through the capacitor and the other side of the matrix should then go to one of the mux channels.

 

[ahmedragia21]

Yes to allow it to work with AC. Note that the maximum supply voltage for the 74HC4067 is 10v so make sure it you don't go above +5v and -5v. Also keep in mind that the mux must be in series with the matrix if you want to be able to switch between channels. The clock should go directly to the matrix through the capacitor and the other side of the matrix should then go to one of the mux channels.

Sorry I didn't get the part of routing the clock.
I sketched it down, can you review the schematic please ? this is the final one, and final question and many thanks

Please note that I need the excitation voltage to be 1V for ROWS, that's why I used a different voltage divide configuration
Galgso

 

[Galgso]

The current schematic makes more sense, I was saying that the two terminals driving one part of the matrix should go between the output of the first mux on top and the input of the bottom mux which you seem to have done here in the final schematic. Also since you're putting in AC into U9's non-inverting input, it should also be powered by the charge pump or it won't be able to output an AC signal that's lower than ground into the mux. Everything on the schematic needs a -5v rail except the charge pump IC itself.

 

[ahmedragia21]

Galgso Many thanks, will test it <3 appreciated!

 

[ahmedragia21]

Galgso
final schematic