Mass flow sensor

[ZeroResistance]

To build a mass flow sensor for sensing air flow, I have heard the principle of measuring this with a hot wire, but

1. Would we need 1 hot wire or 2 hot wires  inline with the flow, and to measure the temperature  of  the hotwire would we need a thermistor each for the hot wire?
2. Is it possible to build this with a hot thermistor instead of a hot wire?

[Rerouter]

https://www.electroschematics.com/11900/measuring-air-flow/

1. Generally its 1 hot wire and 1 thermistor, due to this I don't think you could get away with a second thermistor, generally the wire is stupidly thin to keep the power draw low.

[mikerj]

The hot wire in a MAF is the sensor itself, it has a positive temperature coefficient so resistance increases as it gets hotter.  These typically work by having two wires, one to measure the temperature of the air and the second, heated one to measure flow.  They are wired in a Wheatstone bridge and the circuit maintains a fixed temperature difference between them by increasing or decreasing the hot wire current.  This cancels out the temperature of the air, and the heater drive current required to do this is the output of the sensor.

Platinum wires are used as they are fairly linear and provide very fast response.  Thermistors could be used but are horribly non-linear, so compensating ambient temperature would be more difficult.  A better solution for a DIY sensor would be to use the Vf of a PN junction in a diode or transistor which has a linear, predictable temperature coefficient.

[max_torque]

Why wouldn't you use an OTS MAF sensor? They cost buttons, and are very accurate and robust ?  For learning reasons sure, play around yourself, but if you actually want to measure mass air flow, just buy an existing device......

[Simon]

Why wouldn't you use an OTS MAF sensor? They cost buttons, and are very accurate and robust ?  For learning reasons sure, play around yourself, but if you actually want to measure mass air flow, just buy an existing device......

Ditto, it does not sound like something that is worth putting the efort into designing if and off the shelt unit will do the job. Many of the available sensors can be removed from the tube or you can buy just the sensing head.

[ZeroResistance]

Why wouldn't you use an OTS MAF sensor? They cost buttons, and are very accurate and robust ?  For learning reasons sure, play around yourself, but if you actually want to measure mass air flow, just buy an existing device......

ok, so an off the shelf MAF would have a display or something like that?, I needed to send the MAF data to the PC. I'm not sure in an el cheapo MAF sensor will have connectivity options

[ZeroResistance]

The hot wire in a MAF is the sensor itself, it has a positive temperature coefficient so resistance increases as it gets hotter.  These typically work by having two wires, one to measure the temperature of the air and the second, heated one to measure flow.  They are wired in a Wheatstone bridge and the circuit maintains a fixed temperature difference between them by increasing or decreasing the hot wire current.  This cancels out the temperature of the air, and the heater drive current required to do this is the output of the sensor.

Platinum wires are used as they are fairly linear and provide very fast response.  Thermistors could be used but are horribly non-linear, so compensating ambient temperature would be more difficult.  A better solution for a DIY sensor would be to use the Vf of a PN junction in a diode or transistor which has a linear, predictable temperature coefficient.

This is some pretty amazing info!
First you mentioned regarding Platinum wires,
Second you also mentioned that Thermistors also possible but highly nonlinear. In this case would the thermistor be used as a heater + sensor?
Third the diode / transistor version also seems to be a nifty thought. How does this work out?

[Simon]

Why wouldn't you use an OTS MAF sensor? They cost buttons, and are very accurate and robust ?  For learning reasons sure, play around yourself, but if you actually want to measure mass air flow, just buy an existing device......

ok, so an off the shelf MAF would have a display or something like that?, I needed to send the MAF data to the PC. I'm not sure in an el cheapo MAF sensor will have connectivity options

OK, so what are you artually tring to do? You can't buy a MAF with a display. A MAF is a Mass Air Sensor, it measures the mass of air passing, that does not imply a display. As for linearity, what makes you think the MAF is linear. MAF's are not linear. What you have to do is read the voltage into a µC, and linearise it's value based on the characteristic of the MAF and then you can send it to a diplay or a PC.

I work for a company that does a lot of flow measuring and we often use a car MAF and have a print out of the curve to hand to read the air flow from. Even our fancy flow grids that we paid a lot of money for just have a couple of studs on them that we connect a pressure sensor to and then with a printout of the flow grids characteristic curve we work out the flow.

[ZeroResistance]

https://www.electroschematics.com/11900/measuring-air-flow/

1. Generally its 1 hot wire and 1 thermistor, due to this I don't think you could get away with a second thermistor, generally the wire is stupidly thin to keep the power draw low.

So, how does this work out? In the article you linked they have a thermistor and a hotwire. Does the incoming air first flow past the thermistor and then the hot wire? Or both at the same time?
They say that the the thermistor measures the temperature of the incoming air.
However in a newer verison they have a seperate silicon temperature sensor along with the thermistor and the hot wire.

[james_s]

The AMM in my car has a fine platinum wire strung between supports within the tube that the air flows through. There is also a thermistor which is integrated into the base where the electronics are, air flows through the tube and past both of these things, the order does not matter. The thermistor measures the temperature of the air, and the cooling effect of the air flowing past the platinum wire is used in conjunction with the air temperature to calculate air mass flow. The result in the case of the Bosch sensors I've dealt with is a simple 0-5V analog signal which could be displayed or read by anything that can measure a voltage.

[Simon]

The thermistor measures incoming air temperature. The temperature of the wire will depend on the incomming air temperature and the speed of the air. So if you know the incoming air temperature when you measure the wire temperature you have the difference which allows you to extrapolate the air speed. now you know the air speed knowing the cross section area of the measuring duct you know the volume of air although techincally it is the mass as air pressure has to be taken into account. The latest MAF's have built in pressure sensors but previously these were seperate.

[Benta]

Go to a wrecking yard and rip out the MAF sensor from a crashed car. The car car just needs to be younger than ~25 years. Easy.

[james_s]

Older than that will work in many cases, Volvo started using air mass meters in 1983 with the first LH-Jetronic injection system. Turbo models followed in 1984 with the 760. By 1986 model year every car they sold in the USA was equipped with various versions of LH Jetronic across the range, many other European cars also used them.

American domestic cars were a bit slower to get fuel injection and quite a few older ones used flapper vane airflow sensors or MAP sensors that measure the absolute pressure in the intake manifold.

Salvage yard prices tend to be a bit on the high side though as the AMM is a relatively failure prone part so there is plenty of demand. I was surprised to find that some of the newer ones are actually cheaper, a friend of mine needed one for his Mercedes and it was only about $35 for a brand new one. A new one for my car is more than 5 times that price.

[radiolistener]

You can use ultrasonic transducers in order to measure gas flow.

[ZeroResistance]

The thermistor measures incoming air temperature. The temperature of the wire will depend on the incomming air temperature and the speed of the air. So if you know the incoming air temperature when you measure the wire temperature you have the difference which allows you to extrapolate the air speed. now you know the air speed knowing the cross section area of the measuring duct you know the volume of air although techincally it is the mass as air pressure has to be taken into account. The latest MAF's have built in pressure sensors but previously these were seperate.

I don't understand the following figure as attached.
Why is the thermistor arm also being corrected by the feedback?
Shouldn't the feedback only go to the hot wire platinum arm. and the thermistor arm be kept at a fixed voltage supply ?

[Simon]

i think they are doing somthing along the lines of removing the ambient air temperature "offset" with this configuration. Instead of putting the amplifier reference at at 0 and measuring the full input swing and thus loosing small voltage measurements with high resolution they are basing off the current ambient.

It's hard to explain and this is by no means a complete diagram. You would need more of the circuitry to see what they are doing.

[ZeroResistance]

It's hard to explain and this is by no means a complete diagram. You would need more of the circuitry to see what they are doing.

This is their whole schematic.
I don't see the thermistor being used to correct anything here except provide a temp output.

[Gyro]

The schematic says "wind sensor", not MAF?

[Simon]

Just get an off the shelf sensor. Your local auto parts place will sell you one for around £50 which is not a lot when you think of the complexity it resolves.

[ZeroResistance]

The schematic says "wind sensor", not MAF?

An article about it terms it as MAF https://www.electroschematics.com/11900/measuring-air-flow/

[Gyro]

You need to read the article again.... Reminds him of one, yes. Mentions and describes real MAF sensors, yes. Terms it as one, no. He says it is a wind sensor, possibly capable of room occupancy sensing experiments.

[ZeroResistance]

You need to read the article again.... Reminds him of one, yes. Mentions and describes real MAF sensors, yes. Terms it as one, no. He says it is a wind sensor, possibly capable of room occupancy sensing experiments.

They are using the constant temperature hot wire method, wouldn't the same principle be used for sensing Mass Air Flow ?

Having said that I have a schematic for their latest version attached.
I don't understand why they are taking an output from the bridge itself if they are maintaining the hot wire at a constant temp. that voltage would be constant wouldn't it? They should have taken the output from the feedback op-amp output.
I hope I'm reading the schematic correctly.

[mikerj]

I don't understand why they are taking an output from the bridge itself if they are maintaining the hot wire at a constant temp. that voltage would be constant wouldn't it? They should have taken the output from the feedback op-amp output.
I hope I'm reading the schematic correctly.

The higher the wind speed, the more power you have to dump into the sensor to maintain its temperature and therefore the higher the voltage across it.

[Simon]

The higher the wind speed, the more power you have to dump into the sensor to maintain its temperature and therefore the higher the voltage across it.

i think most comercial sensors use the current draw by the hot wire to maintain a stable temperature as the actual signal.

[radiolistener]

https://hackaday.com/2017/07/26/measuring-air-flow-with-ultrasonic-sensors/