It will only feedback +0.01V. Do you think it already enough?
-
Categories
-
Platforms
-
Content
It will only feedback +0.01V. Do you think it already enough?
Actually if the OpAmp was perfect .. IE, swung from 0 to 5V, the feedback would be 20mV not 10mV. This because the OpAmp sinks when the output is low. That said you make a good point. There could easily be 20mV of noise at that node and drive the servo crazy. A 100K would be more effective and yield a 240mV hysteresis swing. Some good bypassing would still be in order though.
Chris
So are you saying that R16 needs to be a lot smaller?
I've done some calculations and worked out that the values I've given will give a hysteresis of 11psi which is about 5 times more than I want. I would have to increase R16 by a factor of 5 ie 5Mohms or divide R14 & R15 by a factor of 5 ie 2k2ohm to give the hysteresis I'm looking for.
Are you saying the opamp wouldn't be sensitive enough to cope with this, and if so would I be better off using a comparator?
Stew
I've done some calculations and worked out that the values I've given will give a hysteresis of 11psi which is about 5 times more than I want. I would have to increase R16 by a factor of 5 ie 5Mohms or divide R14 & R15 by a factor of 5 ie 2k2ohm to give the hysteresis I'm looking for.
Are you saying the opamp wouldn't be sensitive enough to cope with this, and if so would I be better off using a comparator?
Stew
Chris
A value of 100K for R16 would give me a swing of 104psi which might make the circuit stable but is as good as useless as far as holding a steady pressure is concerned. If you think that noise is going to be the problem how do we go about reducing noise?
Stew
A value of 100K for R16 would give me a swing of 104psi which might make the circuit stable but is as good as useless as far as holding a steady pressure is concerned. If you think that noise is going to be the problem how do we go about reducing noise?
Stew
- Joined
- Jan 21, 2010
- Messages
- 25,510
I'm coming in on this late, but if a tiny change in voltage from the sensor equates to such a large change in the pressure then you need to amplify the signal from the sensor so that (say) voltages between 2.1V and 2.2V are mapped to a voltage between 0V and 5V. This allows the feedback on the schmitt trigger to be larger (and it needs to be larger than the cumulative noise in the circuit and then some)
This is starting to get very complicated for what I thought would be a simple thermostat circuit.
Is it my choice of thermistor that is the problem, and if so would I be better off using something like this?
http://www.rapidonline.com/Electronic-Components/Gt-Thermistor-500k-3-61-0456
I was originally looking at 1 of these,
http://uk.rs-online.com/web/p/thermistors/2938402/
purely because it came with the wires attached and I've read it can be a problem soldering wires to a thermistor.
What I am after is a circuit that can be preset to a temp of between 130 and 165 degC and hold that temp to within approx 1 degC. I have half a dual opamp going spare and would like to make use of this if possible. I have attached a graph of pressure V temp.
If anyone knows of another way to do this I would love to know.
Stew
Is it my choice of thermistor that is the problem, and if so would I be better off using something like this?
http://www.rapidonline.com/Electronic-Components/Gt-Thermistor-500k-3-61-0456
I was originally looking at 1 of these,
http://uk.rs-online.com/web/p/thermistors/2938402/
purely because it came with the wires attached and I've read it can be a problem soldering wires to a thermistor.
What I am after is a circuit that can be preset to a temp of between 130 and 165 degC and hold that temp to within approx 1 degC. I have half a dual opamp going spare and would like to make use of this if possible. I have attached a graph of pressure V temp.
If anyone knows of another way to do this I would love to know.
Stew
- Joined
- Jan 21, 2010
- Messages
- 25,510
A simple thermostat circuit is unlikely to be accurate, stable, or maintain temperature within tight tolerances.
There is no real problem in soldering wires to the thermistor you have chosen.
The platinum RTD is a totally different beast.
A simple thermostat will be impractical to keep the water temperature to within 1 degree.
Firstly, it will regulate the temperature of the sensor, not the water. So you need the sensor to be placed so that the sensor temperature is representative of the water temperature. Note that there are several cases, the water will be at a consistent temperature, the water will have a temperature gradient, or the temperature profile will be complex and chaotic. Each will result in the temperature sensor seeing a different temperature profile.
It is almost impossible to prevent over and undershoot of temperature if you simply turn the heat on and off.
When the temperature decreases slightly to trip the sensor, it may continue to fall for a while even after the heat has been applied. Once the water is back to temperature, parts of the system will still be hot and will continue to heat the water even after the primary source of heat is removed, causing overshoot.
To be able to maintain a far closer control of temperature, you need a thermostat which is capable of reacting to the rate of change of temperature and modulating the heat source appropriately to cause minimal overshoot or undershoot.
For this you need to look at a PID controller.
There is no real problem in soldering wires to the thermistor you have chosen.
The platinum RTD is a totally different beast.
A simple thermostat will be impractical to keep the water temperature to within 1 degree.
Firstly, it will regulate the temperature of the sensor, not the water. So you need the sensor to be placed so that the sensor temperature is representative of the water temperature. Note that there are several cases, the water will be at a consistent temperature, the water will have a temperature gradient, or the temperature profile will be complex and chaotic. Each will result in the temperature sensor seeing a different temperature profile.
It is almost impossible to prevent over and undershoot of temperature if you simply turn the heat on and off.
When the temperature decreases slightly to trip the sensor, it may continue to fall for a while even after the heat has been applied. Once the water is back to temperature, parts of the system will still be hot and will continue to heat the water even after the primary source of heat is removed, causing overshoot.
To be able to maintain a far closer control of temperature, you need a thermostat which is capable of reacting to the rate of change of temperature and modulating the heat source appropriately to cause minimal overshoot or undershoot.
For this you need to look at a PID controller.
- Joined
- Jan 21, 2010
- Messages
- 25,510
Given your pressure vs temperature curve, you're probably better off regulating the pressure rather than the temperature.
This is partially because the pressure measured at a point will be representative, whereas temperature may not be,
This is partially because the pressure measured at a point will be representative, whereas temperature may not be,
Hi Steve I hear what you are saying, but monitoring the Pressure would involve using a pressure transducer which are both relatively large and expensive.
I looked at your link for PID controllers but that went straight over my head. Can it be done without any pic programming?
There's a circuit here where the writer claims hysteresis of 0.5C, I know he says at 25C but would it be much different at 150C?
http://www.craig.copperleife.com/tech/thermo/
I hope I'm not sounding too negative but sensors need to be very small and modellers tend to be people that are retired and don't have a lot of money to spend, nor are they able to program pics.
The latent heat of the water would tend to dampen down any sudden changes of pressure.
Stew
I looked at your link for PID controllers but that went straight over my head. Can it be done without any pic programming?
There's a circuit here where the writer claims hysteresis of 0.5C, I know he says at 25C but would it be much different at 150C?
http://www.craig.copperleife.com/tech/thermo/
I hope I'm not sounding too negative but sensors need to be very small and modellers tend to be people that are retired and don't have a lot of money to spend, nor are they able to program pics.
The latent heat of the water would tend to dampen down any sudden changes of pressure.
Stew
I know you dismissed PICs early in the thread and I'm equally certain that I mentioned Picaxe UCs.
http://www.picaxe.com/
ANYONE can learn to program and use them in a few hours! They will run rings around anything you design discretely, with logic blocks or otherwise. Download the free software and pdf manuals and you will become a believer in short order. UCs don't complicate things. Quite the opposite.
Chris
Hi Chris,
I appreciate what your saying about pics and picaxes and I'm pretty sure I could soon come to terms with programming them, I've been a computer engineer for the last twenty years. However the people who will be building these circuits are folk who spend all day in their garden sheds bashing lumps of metal and producing little works of art. Their hobby is not electronics or computer programming but model making, and as I said earlier many of them won't even own a computer.
There has been mention earlier about noise being the problem but no mention about how to get rid of it, do we have any ideas?
I think I'll probably use a 500K NTC thermistor instead of the 1K PTC one that I originally wanted to use.
Simple circuit anyone?
Stew
I appreciate what your saying about pics and picaxes and I'm pretty sure I could soon come to terms with programming them, I've been a computer engineer for the last twenty years. However the people who will be building these circuits are folk who spend all day in their garden sheds bashing lumps of metal and producing little works of art. Their hobby is not electronics or computer programming but model making, and as I said earlier many of them won't even own a computer.
There has been mention earlier about noise being the problem but no mention about how to get rid of it, do we have any ideas?
I think I'll probably use a 500K NTC thermistor instead of the 1K PTC one that I originally wanted to use.
Simple circuit anyone?
Stew
- Joined
- Jan 21, 2010
- Messages
- 25,510
Except for post 106
Sorry Steve I missed that one.
How would I go about amplifying the signal to give me 0v or 5v?
If I use the 500K thermistor would I need to do anything as the resistance changes a lot more with this one than it did with the original thermistor?
Could we use a capacitor to smooth out any noise? and if so where would be the best place to put it?
Stew
How would I go about amplifying the signal to give me 0v or 5v?
If I use the 500K thermistor would I need to do anything as the resistance changes a lot more with this one than it did with the original thermistor?
Could we use a capacitor to smooth out any noise? and if so where would be the best place to put it?
Stew
