Simple MAF Sensor I made today..

nput

ize

to

Nice,

Speaking of Mass Airflow Sensors... I was having some intermittent performance probs with my daughter's car (my car actually) and finally changed the MAF and it has been running smooth ever since (about 3 months without probs) but the prob was intermittent especially bad in wet cold weather. [Knock on wood]

The only MAF problem I ever had was on a '77 280Z... literally had a swinging flap driving a pot :-( ...Jim Thompson

nput

ize

to

Neat, Steel may not be the best choice... all that rust. But all metals are 'pretty much' equal in dR/dT. You might try an alloy. Phosphur bronze is nice, and availible in thin gages.

George H.

I was thinking of using some heat shrink fep type of tubing around it to protect it. The wire temperature only seems to get up around 300F.

We were using music wire today which seems to be an alloyed of some sort instead of high carbon steel .

Jamie

Simple in concept, but I'd be interested in how you solved the practical issues.

Smallest piano wire I have is .025". It measures 0.2 ohms/foot. If .005 wire has 25x the resistance, that's 5 ohms/foot. A foot of wire is gonna cost you 2 amps = 20Watts. And the shielded part of the wire is gonna get pretty hot. The tempco is not all that large. How much change in voltage do you experience at the tap?

You're gonna want a kelvin connection to the sensor. Voltage drop in connecting wires can be a substantial source of error if the power goes thru the sense wires.

Shielding one of the resistors from the airflow temperature compensates the system ONLY if the temperature of the air is the same as the temperature of the wall. That doesn't work well for sensing flow in a heating/cooling duct.

You can achieve the same device using two light bulbs in series. Bust the glass off one of them and stick them both in the air flow. You can get some measure of protection from physical damage by screening the sensor. But that doesn't do much for dust build up on the filament. Maybe if you keep it hot, the dust will burn off. Straight filament probably works better than a coiled one in this respect. I like #382 lamps because they're small and I have a lot of 'em. They have a filament coil that's tight compared to the size of dust particles. You can run them relatively hot without using a lot of power or burning down the house. The resistance is high enough that you might get by with the 4-wires you have to work with.

I think the classical solution uses a physically large and a relatively smaller thermistor. Servo the currents to make the temperatures equal. The differential power required to balance the bridge is proportional to the air flow. I haven't tried to figure out the math, but I think it has to do with the surface area being proportional to the square of the dimension and the volume being proportional to the cube.

I'd be interested in more details of your setup. I want a simple method to balance air flow in a Heat Recovery Ventilator. mike

You can clean some of them. I drove through a socal fire and managed to get the MAF coated with fine particulates. It was under warranty so that was the easier solution.

One of the reasons I'm not keen on hybrid cars is you get every headache of a gas engine (flaky sensors), plus whatever problem you get with an electric transport system. And electric car can be quite simple. No transmission required (unless you want Telsa speed) and even no differential if you use two motors. Of course the auto PC will be still be some overpriced computer.

I haven't looked into that Chevy Volt to see if they allow simpler emission rules for the standby engine.

input

size

t to

for

Well try some other metals. If you don't like copper there are nickel alloys too.

George H.

Nope, actually it's hard drawn carbon steel -- amazingly, the drawing process strengthens the wire so much, it has the strength of a high quality alloy steel, particularly in the thinner wires!

Tim

You can try phosphor bronze guitar strings.

Ed

Would coating the wire with h/shrink insulate the wire & defeat its purpose as a MAF?

stainless steel is pretty good for thermistors, ordinary steel not so much, and nichrome has a tc very close to zero.

I started this Wiedemann-Franz law project last year....(temporarily on hold.) I only then came to realize that the tempco of resistors is pretty amazing. Given that most metals go as 1/T (with T in Kelvin.)

George H.

(Soldering to stainless is also a bit of a pain.... nasty fluxes)

t.net ---

Not totally, we tried it today, it only slows the effect a bit, but it still works. The FEP tubing is very thin.. Good think too, it isn't the cheapest stuff around however, we just happen to manufacture that stuff at our location, not shrink tubing per say but we do extrude FEP and I had some steel wire pulled through a head and drawn down to about 5 mils wall thickness. :)

I suppose if you need instant response it would be an issue in which case, you could use platinum wire or something that is less corrosive.

Jamie

but copper does not have a TC like steel does.. so I don't think that is going to work..:)

Jamie

Today we connected it to a front end of a basic op-amp. Put a VD pot on the (-) input and the steel wire divider on (+) input. The same source as the VDivider Pot drives both.

The interesting part of this is, because both dividers are using the same source, which is current limited via a R. When current increases in the steel wire divider, it will of course drop the voltage at this source. The (-) input of the op-amp now gets dropped a little which gives you more increase of gain over all. I guess you could call this a cheap form of a bridge ;)

For basic room test, we get a dt of 50mv moving a 1 foot wire at ~ 20 FPM speed. We are using a 5 mil or close to that, wire. a 12 inch piece of wire is giving me ~ 2 ohms. and using 2 volts giving me 1 amp of current at room temp. I don't think 2 watts is much to worry about.

Of course, these calculations are under a load. I didn't measure the wire R with no current.

I should clarify that a bit, the 1 foot of wire had 6 inches of it covered, so only the other 6 was getting the actual cooling while we moved it that slow. And, the 50mv I was getting is at the input, not the output. The output was, well, full rail at very small air currents.

P.S. It seems that our stock of piano wire is inconsistent in quality but for what it was designed for, i guess it's fine.

Jamie

Jim Thompson wrote in news: snipped-for-privacy@4ax.com:

I took apart the failed MAF from my Nissan Sentra Spec V. it has a metal thin film sensor,and a hybrid ceramic circuit with a large bare-die IC gold wire bonded to it,with the usual thick-film resistors and chip caps. the ceramic PCB was covered in clear silicone rubber,it was still kinda sticky.the thin film sensor was wire bonded to the ceramic substrate,lots of gold wires. a microscope revealed a complex pattern etched on the thin film sensor.

nice thing was,I was able to find a replacement MAF sensor online for $78 without having to buy the whole duct/housing too($400)or mess with core exchange.

Does the ADC go into a fairly "intelligent" program that can determine if music can be detected and speech ignored?

That sounds like a derogatory remark?

If any one is interested, today I lengthen the wire "0.005"mil and was able to detect some one blowing air through their lips 10 feet away.

Actually, I had problems calibrating it at first due to the air currents in the lab that were virtually unnoticeable. I had to bring it in a closet area to calibrate it. I thought at first I was getting some RF demodulated at the input but scope readings revealed that was not the case.

The use of a JFET front end op-amp gave me this additional sensitivity instead of using a monolith front end, old stand by.

I really don't have a need for this range of sensitivity. Tomorrow we'll be receiving some phosphor bronze wire to experiment with, as suggested by some one here.

Jamie

Wonder if it helps to coil it up to increase the surface area?

Your source impedance is essentially zero. How did the JFET-ness help?