Trying to find a simple amplifier design/signal conditioner

[Jiggy-Ninja]

I have a sensor whose output voltage, in the conditions I plan to operate it in, will be between 4.5V (max) and 3.5V (min). I want to increase the resolution of the ADC measurement of this value, so I want to create some kind of amplifier or signal conditioner that can change this range into a 0-5V range.

None of the op amp circuits I've done in class have had a use for this kind of thing, and we didn't really do anything with signal conditioning at all, so I'm a bit lost here.

I've looked somewhere else and found a single supply difference amplifier circuit, but it doesn't look like what I need.

This is for a class project. I'm not asking to have the design done, just a pointer in the right direction

There are two main constraints on this design:

1) Supply voltages. I only have +9V (from a battery) and +5V (from a regulator) to work with. Any design will have to be single supply. I have a pair of 358 dual op amps (so 4 op amps total) designed for single supply operation available for use.

2) It all has to fit within 10 rows on a breadboard, so it can't be huge and elaborate. If I eliminate one of the other sensors (which isn't critical to this project) I can free up 3 more rows, but 13 is the absolute maximum I can use.

I don't think I have an instrumentation amplifier, which I suspect may be able to do this job, but I haven't used one of those yet either.

 

[berkeman]

I have a sensor whose output voltage, in the conditions I plan to operate it in, will be between 4.5V (max) and 3.5V (min). I want to increase the resolution of the ADC measurement of this value, so I want to create some kind of amplifier or signal conditioner that can change this range into a 0-5V range.

None of the op amp circuits I've done in class have had a use for this kind of thing, and we didn't really do anything with signal conditioning at all, so I'm a bit lost here.

I've looked somewhere else and found a single supply difference amplifier circuit, but it doesn't look like what I need.

This is for a class project. I'm not asking to have the design done, just a pointer in the right direction

There are two main constraints on this design:

1) Supply voltages. I only have +9V (from a battery) and +5V (from a regulator) to work with. Any design will have to be single supply. I have a pair of 358 dual op amps (so 4 op amps total) designed for single supply operation available for use.

2) It all has to fit within 10 rows on a breadboard, so it can't be huge and elaborate. If I eliminate one of the other sensors (which isn't critical to this project) I can free up 3 more rows, but 13 is the absolute maximum I can use.

I don't think I have an instrumentation amplifier, which I suspect may be able to do this job, but I haven't used one of those yet either.

That's going to be tricky with only the single supply. Any chance you can add a 2nd 9V battery to give you a -9V supply. That makes the opamp circuit much easier...

 

[yungman]

There are input, output rail to rail opamp that can do the job. You might loss a little at very close to 0V. But the rest of it is going to work. So you at least has close to 0 and to +5V. It really depend on the load, if you don't have to sink much current, it will drive to 0V. The way to do it is to use a +2.5V voltage reference to set up the mid point, then use the opamp as inverting amp and tie the +ve input to the +2.5V to set up the mid point. Then you can use summing to map the range of 3.5 to 4.5 and match it to 5 to 0V.

Notice the direction is inverted? This is because the opamp is in inverted configuration. Review summing junction in the opamp chapters, it should be easy. We have to do that all the time to map different output range to the ADC.

358 is not a rail to rail. Go to Analog Devices, Maxim, Linear Technology site. There are plenty plenty of input output rail to rail. In fact you don't even need input rail to rail as you range is only 3.5 to 4.5. It is the output rail to rail that is important.

If you have trouble designing the summing junction, I'll help you, but I am not going to give away right now, you need to hit the books.

 

[Jiggy-Ninja]

That's going to be tricky with only the single supply. Any chance you can add a 2nd 9V battery to give you a -9V supply. That makes the opamp circuit much easier...

Not likely. This is (hopefully) going to end up strapped to the bottom of a model airplane, so I'd like to avoid unnecessary parts.

 

[jim hardy]

have you looked at lm324?
four opamps in one package, made for single supply operation and real cheap.

see National application note 20 page 3 figs 4 and 5 for summing and difference amplifiers.
http://www.national.com/an/AN/AN-20.pdf

 

[Jiggy-Ninja]

have you looked at lm324?
four opamps in one package, made for single supply operation and real cheap.

see National application note 20 page 3 figs 4 and 5 for summing and difference amplifiers.
http://www.national.com/an/AN/AN-20.pdf

I just looked at the op amps I have, and I do have one. I don't know if they'd be better than 358s (they fit end to end anyway).

I also have an LM3900.

I haven't ignored your reply, yung. I'm still trying to digest it.

 

[yungman]

You need to buy a new opamp! Rail to rail output is a MUST, Looks like 324 should do the job.

You have a +5V regulator circuit, there are ways to float it and make it as -5V for the circuit also. That open up another can of worms about protecting the ADC input from swinging below -0.3V using schottky diode and it can be a pain! I suggest you stay with the rail to rail opamp, don't be too cheap, just buy one.

 

[Jiggy-Ninja]

BAH!

I just realized that I forgot another circuit that needs to go onto the board, which is probably barely going to be able to squeeze into the remaining space.

So this topic has now become basically meaningless.

 

[jim hardy]

lm324 is made for single supply and dirt cheap. .
about $4/dozen at digikey so ten bucks gets a lifetime supply

as you observe output will sink modest current to zero and input will sense zero.
but it needs 1.5v headroom at top end; with 9v supply and 5v signal he has it.

that lm358 looks good too.

i like havng 4 opamps in one package- you can buffer your input signal with a follower for high-z non intrusive monitoring.

edit just saw your post what's the other ckt?

 

[Jiggy-Ninja]

Frequency Shift Keying circuit to send the data using a cannibalized walkie talkie.

 

[cmb]

What exactly is the required resolution of the data from this sensor interface? I mean your question doesn't seem to clarify if, for example, you are only after 0 or 5V dependent on if your signal is below or above 4V?

Does it have to be a continuous analogue response, 0 to 5, and do you need it to be linear or log, or 5V to 0 (reverse volts) or what?

 

[Jiggy-Ninja]

Lol, with the FSK circuit in place, I have 1 free row, and 3 half rows. Who's up for the challenge?

What exactly is the required resolution of the data from this sensor interface? I mean your question doesn't seem to clarify if, for example, you are only after 0 or 5V dependent on if your signal is below or above 4V?

Does it have to be a continuous analogue response, 0 to 5, and do you need it to be linear or log, or 5V to 0 (reverse volts) or what?

I'll be more clear about exactly what I was trying to do.

It's an analog sensor measuring absolute air pressure. Output is between 4.7V (at 105 kPa) and 0.2V (at 0 kPa) and is directly proportional to pressure with a small DC offset. I am attempting to use it as an altimeter.

At the altitudes I'm likely to work it, it's probably never going to go below 3.5V. It probably won't even get close to to that (3.5V works out to around 1,000m above sea level), but If I can figure it out for that range I can easily adjust it if I feel like it.

What I wanted to do is extend that 1V range (4.5V-3.5V) into a 5V range (0V-5V) or a more realistic 3V range (1V-4V) to get better resolution of the microcontroller's ADC. (new calculations would need to be added to reading the sensor, but that would be trivial)

So, in short, a continuous linear analog response.

As was in my previous post though, this discussion is only academic now. There's no more room for it on my circuit board.

Using the summing amp idea yung gave, the design I calculated to turn 3.5V-4.5V to 5V-0V, arbitrarily choosing a R_F of 10kΩ, is to ground one input and use a 3.33kΩ resistor for that, and use a 2kΩ resistor for the input from the sensor. Is this the sort of design you were suggesting, yungman?

Op amps are cool.

 

[jim hardy]

altimiter , you say.
perked my ears - Dad was a weather man in aviation forecast office..

what is the nature of your sensor? Automotive MAP sensor?


the opamp part is very easy,,
...and you're in excellent hands with yungman and other folks here.

but i am confused by use of fsk

that implies your analog signal is digitized and you're sending the data bits with fsk?
good job for embedded micro with adc

or are you sending a tone with frequency proportional to signal?
good job for LM566 V to F

after you get this gizmo going, if the sensor isn't quite precise enough
i'd visit local weather bureau office and ask if they have a dead radiosonde that they'd let you have the aneroid element for a hobby project... it should be quite precise.
Radiosonde is the disposable electronic box on bottom of a weather balloon. ( I think they still use weather balloons ? )

 

[Jiggy-Ninja]

altimiter , you say.
perked my ears - Dad was a weather man in aviation forecast office..

what is the nature of your sensor? Automotive MAP sensor?

I have no idea what you just said, http://www.mouser.com/ProductDetail/Freescale-Semiconductor/MPX5100AP/?qs=sGAEpiMZZMvhQj7WZhFIALpLlyV3lh9LK4iCKYnc7ws%3d" .

the opamp part is very easy,,
...and you're in excellent hands with yungman and other folks here.

That part is just academic now.

but i am confused by use of fsk

that implies your analog signal is digitized and you're sending the data bits with fsk?
good job for embedded micro with adc

ATMega328. I could have probably gotten away with a much smaller uC than that (maybe even an ATTiny), but it's the only one I knew how to program since I already have an Arduino board.

after you get this gizmo going, if the sensor isn't quite precise enough
i'd visit local weather bureau office and ask if they have a dead radiosonde that they'd let you have the aneroid element for a hobby project... it should be quite precise.
Radiosonde is the disposable electronic box on bottom of a weather balloon. ( I think they still use weather balloons ? )

The main reason I'm doing this project is the "wireless communication" part, not the "accurate altimeter" part. The accuracy issue here was just an optional improvement, not something critical.

Thanks for the help though, everyone.