I bought several pressure sensors from Freescale such as the MPX5700AP (pressure sensor up to 700kpa) to measure in real-time the air pressure within my pneumatic circuit of maximum 45psi.
Each sensors dies after about a day. I have no idea why andit drives me crazy.
Pin1 is Vout - directly plugged into the Arduino A0 pin
Pin2 is Ground
Pin3 is VCc - with my V supply being 5V max.
What am i doing wrong here? I can't understand why they all stop working. There can't ben any problem of pressure as it is 45PSI maximum. Same for voltage with 5V. Could it be that pneumatic air contains humidity? Are those sensors very sensible and fragile? Can't they be turned on of for hours in a row?
It seems to be the second time this month I have read through the datasheet for this device.
Some random suggestions.
Try and set up a basic test rig which you can monitor the performance for several days, without the possible complications ( over-pressure, humidity ) of your actual system.
Consider the possibility that you got a batch of dud devices, who did you buy them from ?
Double-check your interpretation of the diagram showing which are the connection pins of the device.
"There can't ben any problem of pressure as it is 45PSI maximum". Assertions like this, are often an indication of where the problem actually lies. Being sure it isn't a problem, is blocking your ability to consider that actually there really is a problem.
The terminology for pressure sensors is often vague and depends very much on the context. According to the first page of that datasheet, your device is an absolute pressure sensor. If your "45 psi" is pressure above atmospheric, that is actually about 60 psi absolute pressure. That should still be within the performance range of the device, though.
Those MPX sensors are not bad at all. They are in my opinion the first choice for a pressure sensor to be used with an Arduino (for a sensor with an analog output and a sensor that works at 5V).
MPX5700 : 0...700 kPa, with a maximum of 2800 kPa. It is a relative pressure sensor, mostly used to measure the pressure relative to the air.
700 kPa = 100 psi (according to Google Calculator)
2800 kPa = 406 psi
They should work without problem for many years. They don't mind a high air humidity, and they are not very fragile. With a pressure that is too high, they are not accurate anymore.
Therefor I completely agree with michinyon. Perhaps there are peaks in your air pressure system that exceed the 400 psi. For example when an outlet is turned off, the peak wave might travel through your pipes. Or the 5V is sometimes reversed or perhaps sometimes 12V.
Are the pressure values okay ? Can you read the actual pressure (while the sensor was not broken).
[ADDED] Oops, michinyon mentions in Reply #4 that this is an absolute pressure sensor. They would have already 100kPa from the air (baromic value), and the the range is up to 87 psi and they break at 391 psi.
I do confirm that it is a MPX5700AP (Absolute) one
Regarding the:
Pressure part: My pneumtic circuit is of max 45 PSI. I'm 100% sure of that as my compressor can't deliver air above that value. Those are made for 700kpa (100PSI). Even If I add the atmospheric pressure I'm still way above that limit which isn't even the burst limit. I'm definitly sure that it is not coming from the amount of air.
The Electrical part: V supply is a 5V 1A. So it can't be a 12V issue. My sensor is receiving 7 mA.
The pressure is okay yes. I could read the value without any issue for few hours it was working fine (that's why I'm getting crazy). This sensor is supposed to deliver 4.58 to 4.81V in Vout. The last MPX5700 I have that seemed "broken" as well is now delivering 1v to 1.5V ish. Although the others are totally dead (not delivering anything. This one has a much lower value in Vout). What could have done that?
I'm receiving two more MPX5700AP tomorrow. I don't want to do the same mistakes but because I don't know what my mistakes are I'm gonna have to do as Mychinyon said and try to monitor those sensors out of my circuit and see if they hold on. What process would you recommend?
I wonder if the compressor isn't expulsing oil with air that could, at the end of the circuit get into the sensor and break it?
In terms of schematic, there's nothing missing from my description right? Pin1 goes to Analogin pin, Pin2 to ground and Pin3 to Vcc 5V. Right? Pin 6 7 and 8 are not being used in my case.
Are those sensors okay to receive Vcc in full-time or do I need to unplug them between measures?
Like I said, they can be used many years (decades), they can have 5V continuously without any problem.
Yes, pin 1,2,3 are used like that. You may not connect the other pins. I had an MPX sensor, and I soldered an unused pin to GND, but that was wrong.
It is very unlikely that oil could travel into the sensor. There is no airflow into the sensor. And a little oil could not destroy it within a day.
It is however likely that you have bursts or peaks of pressure on your system. Even if you say that you are 100% sure that is not above 45 psi, you don't know what could be going on in your pressured pipes (like a peak of pressure that can travel through the pipes). It doesn't have to do with the compressor.
Or the power supply has nasty peaks.
Do you have another compressor ? and another power supply ? Could you build a full second system, with different components ?
Do air pressure silencers exist ? I don't know the English word. Perhaps with some porous stone. That could prevent damage to the sensor.
I don't have any other compressors but I can try to reduce the incomming pressure at 14.5psi and see if the sensor resist longer. I have another power supply yes. i'll try it and I'll add capacitor as well to make sure it doesn't go above 5V.
The peak of pressure in the pipe is an interesting thing. You mean that even with a 45PSI pressure at the entry of the circuit, it could be possible when the solenoid turn HIGH that the peak of pressure in the pipe (6mm) could reach a very high limit and therefore damage the sensor in a very very very short time?
Yes an air silencer reduces the sounds and you would suggest using it to reduce this potential peak you mean? That could do it yes. However I'd have trouble connecting my sensor to the silencer but I can find a solution. I also have an non return valve that may reduce the velocity.
Yes, that is what I mean. With metal pipes I would not be surprised of the peak is a lot higher than the normal pressure.
With silicon tubing, it will be a less problem
Even if that could be a possibility, I find it unlikely, as I said, my compressor delivers 45psi. Even with a peak of pressure, the sensor can go as high as 700kpa which is 100PSI and I think it can go as high as 2800 KPA (400 psi) before bursting.
That seems unlikely to find such an amount of pressure in my circuit but i'll try to run some more test tomorrow and will keep you guys updated
From the data sheet:
The MPX5700 series pressure sensor operating characteristics, and internal reliability and qualification tests are based on use of dry air as the pressure media. Media, other than dry air, may have adverse effects on sensor performance and long-term reliability.
You can isolate the sensor with a diaphragm arrangement, if your air is moist, oily or otherwise unfriendly.
That does not explain them failing after a day.
According to the datasheet, the sensor element is covered with a silicone gel.
That should stop moisture and oil for some time.
Leo..
Your supply is what exactly? Have you decoupled the sensor right at the sensor as
explained in the datasheet? Are long cables involved - if so they should be screened or
twisted pair or both.
Also, I would definately use a good filter in your pressure line. Particles of rust, dirt, moisture and especially water droplets and water buildup are likely causing your problems.
Yes, I mean by putting my finger at the end of the circuit, I do receive air that seems quite dry (even if, of course, they're always some humidity and/or oil that goes with it that shouldn't be enough to damage it that fast). I'll add something inbetween the sensor and the tube not to be directly at the end of the circuit. I have to think on how to do it.
My power supply is a 100-240V (50-60hz) 5V - 2A basic socket wall. I will try another one just in case.
No I haven't done the schematic on page 4. I'm currently building it for the next sensors I'm gonna try. I don't know if that was the cause of it but you're right it can only make it better (at least for the readings).
Long cables are not involved (20cm) and twisted pair.
I just need to think of a nice way to filter the air coming into the sensor. Will work on it today and will keep you guys updated.
Don't just try another power supply, test its quality! You want properly regulated 5V DC with
no AC component - this can be checked with a multimeter. Cheap power supplies can be
poorly regulated which will mean over-voltage.
noxxsound and MarkT, a power supply can be bad in a way that it can not be measured with a multimeter.
Sometimes a switching power supply has high frequency spikes (due to leaking capacitors or overall bad quality), other power supplies are capacitive coupled to the mains (you can feel that when touching one wire). It's no joke, I had a few of those from Ebay, but I have trown them away.
I'm gonna be extra careful with the power supply I use. Do you think I could use the arduino 5V pin as the power supply. At least that something of quality isn't it?
The Arduino 5V and 3.3V pins are to power sensors ! What did you think they are for ?
I assumed that you used an external power supply to power both the sensor and the Arduino.
When you used the external power supply for the sensor only, and the ground wire is not very good connected, that could cause all kind of trouble.
I say: more than 50% chance that the way the sensor is connected could be the trouble.
Maybe your silicon tubing and pressure system is okay after all