Well, good luck with your project.

does this mean you wont be helping me anymore with my project?

 

If your professor won't allow you to use opamps as buffers on the inputs of the 4051, see if you can use them on the outputs.

You need to convert the high impedance inputs to low impedance as quickly as possible, or it's just not going to work well.

 

If your professor won't allow you to use opamps as buffers on the inputs of the 4051, see if you can use them on the outputs.

You need to convert the high impedance inputs to low impedance as quickly as possible, or it's just not going to work well.

I have buffers at the output of the mux.

how should I convert the high impedance input to low impedance output?

I have another doubt, is my amplifier circuit slow enough that it produces its own delay???

i mean, before the amplifier can process the signal, rectify and produce an output, the input signal changes... Can this happen??

 

i mean, before the amplifier can process the signal, rectify and produce an output, the input signal changes... Can this happen??

It will do the processing, rectifying and outputting while inputting. It doesn't take a chunk, process it, output it, then grab another chunk. It happens on the fly.

 

Hello All,
I have fnally convinced my prof to use a buffer before multiplexer ckt. However, I am experimenting with a buffered mux AD8174.

Can anyone please tell me what does pin -Vin(pin 12) signify and how it should be used?

The following is the link to the datasheet.
http://www.analog.com/static/imported-files/data_sheets/AD8170_8174.pdf

 

That is the inverting input to the output amplifier. If you connected it to the output of the amplifier, it would be a buffer amp/voltage follower.

Do you see how much those IC's cost? Around $7 USD each.

They have a 250MHz bandwidth, which is nice if you're doing video processing, but overkill to the maximum for what you are doing. Your power usage would increase quite a bit. That is why I suggested the low power rail-to-rail opamps that I did.

Anyway, if you want to use something like that, look at a MAX4312. It's a buffered 8-channel MUX, also with ridiculously high bandwidth. At USD $11/ea, it would be cheaper than the one you're considering, as it has twice the input channels; $11 for 8 vs $14 for 8, and fewer components/connections.

 

That is the inverting input to the output amplifier. If you connected it to the output of the amplifier, it would be a buffer amp/voltage follower.

Do you see how much those IC's cost? Around $7 USD each.

Yes... I know, just wanted 2 show my prof, that using buffered mux is a costly affair

I did try connecting the inverting i/p to the output of the mux, but, i saw that the output was directly +5v .. so, i got confused ... Need to read more about it I guess...

They have a 250MHz bandwidth, which is nice if you're doing video processing, but overkill to the maximum for what you are doing. Your power usage would increase quite a bit. That is why I suggested the low power rail-to-rail opamps that I did.

Yes.... I changed the amps now...

 

So, does it work properly when each mux input is buffered like I suggested a number of posts back?

Did you try the "fallback" idea, which was putting some 10nF caps on the inputs of the mux, and buffer amps on the outputs of the mux? This would tend to reduce the signal from the foil sensors quite a bit, and if you had any cross-conduction in the mux during switching, it wouldn't be long before you had no signal again.

 

So, does it work properly when each mux input is buffered like I suggested a number of posts back?

Did you try the "fallback" idea, which was putting some 10nF caps on the inputs of the mux, and buffer amps on the outputs of the mux? This would tend to reduce the signal from the foil sensors quite a bit, and if you had any cross-conduction in the mux during switching, it wouldn't be long before you had no signal again.

Yea... will try tat out...

 

Hello All,
The project seems to be working well... with the buffers connected at the input to the mux.

However, the main task still remains... I have to built a carpet which would have these aluminum sensors below them. The carpet is extremely huge...
Covers a whole room... 12feet wide and should be approximately 25feet long.

There are wires running all over the places.

How should I make sure that I have extremely low noise and noise interference? Also, what care should I take while designing PCB?

 

Construct a large Faraday cage.

I'm tempted to suggest fill the room with saltwater, but that would likely cause damage to property, so I won't.

Actually, probably the cheapest thing to do is to keep the relative humidity around 60% to 70% If it's much higher than that, you will likely have mold problems. If it's a lot lower, you will have a lot of static electricity.

Unless you construct a Faraday cage (which would be very expensive) you won't be able to keep RFI out.

PCB design: Lots of info on this link: http://www.smps.us/pcb-design.html

 

Construct a large Faraday cage.

I'm tempted to suggest fill the room with saltwater, but that would likely cause damage to property, so I won't.

Actually, probably the cheapest thing to do is to keep the relative humidity around 60% to 70% If it's much higher than that, you will likely have mold problems. If it's a lot lower, you will have a lot of static electricity.

Unless you construct a Faraday cage (which would be very expensive) you won't be able to keep RFI out.

PCB design: Lots of info on this link: http://www.smps.us/pcb-design.html

Faraday Cage? .. but isnt it like pretty big? Moreover, I cannot have the room covered. It is meant to be used for old people... No other method? like having capacitors placed at particular distances? I have a ground plane below the aluminum foil sensors... Also, the sensor size is only 6"x6" and distance between two sensors is 1 feet... so, interference is somewhat less....

 

I don't know what to tell you.

You will need to experiment, find out what works, and what does not.

 

I think your sensors and wires will pickup all kinds of interference plus a lot of static electricity.
What do you expect to sense?

 

I think your sensors and wires will pickup all kinds of interference plus a lot of static electricity.
What do you expect to sense?

I am sensing if a person is stepping on the sensors or not... The system works pretty well for the previous 2 prototypes in which I had 21 sensors and 4 sensors respectively....

However, in the previous 2 prototypes, the maximum length of the wires was approximately 8 feet and these sensors were placed on a wooden block. Now, I dont have any wooden block, and also the maximum length of the wire would be approximately 20feet.

I know it sounds scary.... the wires would be acting like an antenna and producing loads of noise... Hence, I am pretty concerned about the noise factor.

 

I really don't know what your signal looks like.

If you wanted something that was more repeatable, you would need to use something like a resonant tank, using a couple of sheets of aluminum separated by a compressible dielectric material; perhaps something like a sponge or even carpet padding to make a capacitor.

If there was no weight on the capacitor, the plates of the cap would be further apart for lower capacitance, and the resonant frequency would be relatively high. If there were weight on it, the dielectric material would be compressed, the capacitance would increase, and the resonant frequency would change (decrease).

The exact resonant frequency would not be important; it would be the change of frequency that would be of interest, and how long it stayed at that frequency.

 

Hi SgtWookie,
I am using the aluminum sheets in the same fashion as you mentioned. I have a layer of aluminum on top, below which is a plastic sheet which acts like a di-electric medium. Below this plastic sheet is a ground plane (of aluminum foil)...
And the sensor(aluminum foil) does change the frequency and amplitude when it is activated.

 

OK, but you are just looking at static voltage levels.

When someone steps on a sensor, the charge would momentarily decrease as the dielectric is squeezed together.

You're not using it as a resonant tank; you're using it as a momentary change in voltage sensor.

 

OK, but you are just looking at static voltage levels.

When someone steps on a sensor, the charge would momentarily decrease as the dielectric is squeezed together.

You're not using it as a resonant tank; you're using it as a momentary change in voltage sensor.

Yes, exactly.... that's what my research is all about...

 

You are going to have some serious static problems with your aluminum laced carpet. You will most likely send some severe spikes down your line.

I have a question.

I have been following along for a while now, and there is something I have to know.

What is the requirement for the project?

Was it: Find a way to detect a persons placement in a room?

I had a very similar project a decade ago, and we built the room (12ft x 12ft) and had to show a persons movement through the room.

Is the point of the project to deal with frequency filtering and parasitics?