pcb-mount air flow sensor

In my case I got an output (once temperature compensated) very much like CMOS ID versus VDS and VGS curves.

Temperature compensation was the crux of my thesis... this was in Germanium days ;-) ...Jim Thompson

I've always like the use of DIODES pasted on the part of interest and attached to a sense circuit.

Use them for high wattage RF ceramic transistors, thermo sensing.. Jamie

I wasn't suggesting that John crams a expensive MAF into his design. Just to use the basic concepts on how they work on his air flow sensor.

I'm now at 3 successes out of 4 attempted MAF repairs on friends vehicles. The OBD2 box belches an error code that announces yet another expensive repair. I remove the MAF assembly, clean it out with the approved cleaner:

let it dry, ohmmeter check, put it back, reset the error message, and it usually works. Cleaning it while in the vehicle hasn't worked very well. There are several YouTube videos on how it's done. This one seems tolerable:

(milliohms???)

But the customer can specify "The crate controller shall include..." and disqualify our competition. And some potential customers may think these features are cool, or even useful.

We're planning to include some other unnecessary stuff, like mux-able SMB test connectors on the front panel to allow scoping stuff, and some registers that can control and measure bus transaction times. And some spiffy LEDs.

I once bid against LeCroy, who were shipping an expensive, flakey product. They discovered that they would have competition, so they dropped their bid price in half. My customer picked an obscure feature of our unit, made it mandatory, and disqualified them.

True. But only if you can keep track of budget (time & money). I've seen too many projects going way over budget because people had no clue how much work was involved.

You should allow the customer to configure the color of the leds :-)

I once put blinking leds on a device. Because of timing those leds could blink simultaneously or alternating. It took me quite a while to convince the customer that it was not a problem.

With uPs and FPGAs and cheap parts, featuritis is a serious problem, deciding what goodies to include, knowing when to quit. If the uP has an ADC, why not add a mux and digitize all the power supplies? Then why not compute noise? That's about where we are, trying to add sensible features that don't overload the engineering time.

We do have one nice orange LED called USR on our VME modules. A guy at Pratt&Whitney wanted his own led to control, and we don't tell that guy no. We call it Dean's LED.

In a crate full of modules, our "CPU" heartbeat LEDs tend to start together, and gradually drift out of sync.

We had customers specify that all the LEDs in the products they bought had to come from the same bin.

There'll be no skilled workers on your job!

sensor.

Don't forget testing time. If it's there, it has to be tested.

I had one project where there was a note added to Agile telling purchasing that Engineering had to be notified before a new reel of LEDs was used. Management forced me to use blue LEDs on a board where the highest rail was

heater.

sensor.

Designed, coded, tested, documented in the manual.

Maybe and maybe not. There are HMI standards that limit color choices = for indicators for various functions. Some of the standards conflict, and specific case / application / system functional requirements normally overrule general case standards. e.g. MV switchgear indication colors (somewhere in the ANSI/IEEE C37 series) versus the general case Mil-STD-1472.

These were those annoying HP 'Intelligent LED Displays' that could be scrolled.

Why are those things considered so important when they are likely to be defeated occasionally when somebody's foot is under the door but not inside enough to block the sensor? Even a car wouldn't block it if it has a high clearance.

I read through the suggestions. I think the chances of a false positive are too high. Nothing is worse than some protection circuit cutting in when there is nothing to protect. I'd scrap the air flow idea. Seems to me the air temp sensor is good enough, and that is a reliable circuit.

The mass air flow sensor on a car is just a wire that is heated by current flowing through it and then has it's resistance sensed. These MAFs are really flaky.

I like the old weather vane. Yeah, it would stick sometimes, but it was purely mechanical. All you had to do is clean it.

The hot wire is OK. I'm had one fail due to driving through a socal fire. The fine soot coated the wire. The car was under warranty, so it was easy just to have it replaced on the dealers dime. Otherwise you have to take it apart and clean the sensor with alcohol, or pay through the noise for a new unit.

I can't wait until electric cars can replace internal combustion vehicles. All this sensor crap will be gone.

Michael A. Terrell schrieb:

Hello,

even if there are LEDs in more than one colour?

Bye

You'd rather have blackouts like India?

You need to find out what bin means in optoelectronics. It has nothing to do with color.

LM71

heater.

sensor.

Engineering test takes a lot of time but so does manufacturing test. Though you don't have high quantities so it might not matter as much (but have less incentive to automate/BIST).