maf signal conditioning

ashford · Post by **ashford** » Mon Apr 04, 2016 10:00 am

i have some noise issues when using ford slot style sensors. i did some looking into of how ford handles this. on the sensor itself it has 12v power ground and signal ground. internally the signal and power ground are hooked together, so grounding the heater ground to the engine causes a ground loop. so far i have had the signal/heater ground hooked to a single wire to the ms, signal is better but still not as good as i would like. at idle the g/s reading fluctuates about 2-3 g/s and causes runability quality issues.

i grabbed a ford computer and started to check things out. the maf signal and maf return have their own inputs on the pcm, neither have continuity to ground, signal return( all other sensors), or case ground. so i take it that there is some sort of ground isolator or voltage comparator, i opened the case and both leads have small capacitors on them and go into an ic with a ford p/n on them so i cannot just go out and buy something.

does anyone know of a simple ic or circuit that can be made for this purpose?

subwoofer · Post by **subwoofer** » Mon Apr 04, 2016 10:04 am

Any instrumentation amplifier should do the trick.

ashford · Post by **ashford** » Mon Apr 04, 2016 1:16 pm

so something like an ad620? looks like it cant handle a load resistance of less than10k, that ok?
im looking for the least amount of other components

subwoofer · Post by **subwoofer** » Mon Apr 04, 2016 1:36 pm

Actually, you will need a rail-to-rail output amp, like the AD623. Program it for unity gain, connect the negative input to MAF local ground, positive to MAF output. You may want a 10k resistor between the two.

ashford · Post by **ashford** » Mon Apr 04, 2016 3:27 pm

is a rail to rail nessecary? most maf start to register at .2-.5 volt and if vcc is connected to s12c the upper end is not neccesary. i found an ina114 that is input protected and seems it would not require any other components to install.

don't mind me im just wrapping my head around this it has been 20years since my electrical education and i have forgotten 95% of what i did know, and appreciate someone elses input before i jump in

Post by **jsmcortina** » Mon Apr 04, 2016 3:37 pm

Non rail-to-rail can be a volt or more away from the rails, well worth getting the better part.

James

subwoofer · Post by **subwoofer** » Mon Apr 04, 2016 10:18 pm

For instrumentation amps, it is fairly typical for the dual supply amps to have an output range of V- +1.5V to V+ -1.5V. That would put most of the MAF operational range outside the amplifier output range. An INA156 would probaly work nicely as well, but that is surface mount only. I don't find that to be a problem, but soldering skills may vary.

On a side note: problems like this is why I suggest the next iteration of MS hardware should have differential analog inputs...

ashford · Post by **ashford** » Tue Apr 05, 2016 5:19 pm

well i ordered 10 of the ad623's, some various other items, a pile of heatshrink and acidentally ordered 503 .001uf instead of 50 from digikey.

anyone need capacitors?

vw_chuck · Post by **vw_chuck** » Wed Apr 06, 2016 4:24 am

Why do you think that the MAF fluctuating at idle is incorrect? Some MAF's are designed to actually read negative as the pulsations at idle are extreme in some engines.

subwoofer · Post by **subwoofer** » Wed Apr 06, 2016 5:02 am

Use the MAF high speed logger at idle to see whether you have a noise issue or if it is only seeing the pulsating intake air column. Tweaking the transfer function makes all the difference in the world.

ashford · Post by **ashford** » Wed Apr 06, 2016 9:59 am

pretty much if you look at the dot and follower in ts it is a jagged line everywhere. the engine is dead smooth on sd, so it is not the dead cylinder maf shake as seen on engines with a mis but looks like it. the settings are event average and maf lag factor of 20. my other install has some jiggle too but not nearly as bad, i'm gonna try this method on both installs with a twisted pair signal and signal return

subwoofer · Post by **subwoofer** » Wed Apr 06, 2016 10:59 am

Try going for a lag factor of 100 for MAF, with the new code you do not want it averaged.

ol boy · Post by **ol boy** » Wed Apr 06, 2016 12:09 pm

I am running a mid 90s style ford MAF and have noticed it's pretty noisy as well. My idle area however is about the smoothest area. It seems I have some harmonics at certain rpm points as the ADC value will jump about quite a bit. I'm also doing a 1k resistor to a 100uf cap to filter it's output but that only work so well.

ashford · Post by **ashford** » Wed Apr 06, 2016 12:22 pm

the code in the car is a 1.4 variant not sure which just that it is one of the first ones where you are forced to use the bias table for boost control. with maf lag set to 100 the signal is unusable. the wiring is probably not ideal either, but was planning to run a new pair from maf to ms through the body instead of the engine harness. the sensor is a pmas hpx in a 3.5 inch tube

ashford · Post by **ashford** » Wed Apr 06, 2016 12:23 pm

ol boy wrote:I am running a mid 90s style ford MAF and have noticed it's pretty noisy as well. My idle area however is about the smoothest area. It seems I have some harmonics at certain rpm points as the ADC value will jump about quite a bit. I'm also doing a 1k resistor to a 100uf cap to filter it's output but that only work so well.

i have a 99 mustang i dropped a older 5.0 in and used the 3.8 maf and factory harness, i have no complaints on that setup

subwoofer · Post by **subwoofer** » Wed Apr 06, 2016 11:57 pm

ashford wrote:the code in the car is a 1.4 variant not sure which just that it is one of the first ones where you are forced to use the bias table for boost control. with maf lag set to 100 the signal is unusable. the wiring is probably not ideal either, but was planning to run a new pair from maf to ms through the body instead of the engine harness. the sensor is a pmas hpx in a 3.5 inch tube

If the signal is noisy even when the engine is not running you definitely have a problem with noise. Needing a lag factor of 20 isn't a good sign either.

Let's cross our fingers that the differential connection cures the noise problem. It could also be that the MAF needs a lower load impedance, play with that if the instrumentation amp doesn't cure the problem.

Analog outputs from WBO2 controllers also have the same problem, but that signal isn't as sensitive.

ashford · Post by **ashford** » Fri Apr 08, 2016 8:13 pm

i did some testing with a bread board, signal generator and pcscope.
there is some bad news, if using 5v supply max output is 4.25 and it looses its linearity around 3.5v. but if using a 12v source it is perfectly linear from 0 to vcc-1. the bad is that the absolute input voltage is12v so another regulator@7v will be needed
i simulated some noise by using 2 capacitors one each lead to sig generator and one on each input. i had it up to about 2v p-p without any affect on the output signal, scoped on the inputs they were not equally biased past that point.
another oddity i found was that a small capacitor on the output created a small oscilation of about 100mv

ashford · Post by **ashford** » Sun Apr 10, 2016 8:12 am

this is what my brain figures to be a good circuit. any thoughts?

proposed conditioner.jpg

ol boy · Post by **ol boy** » Sun Apr 10, 2016 8:31 am

Nice, so -in gets tied to the sensor return ground from the maf ?

Sent from my SM-G920V using Tapatalk

ashford · Post by **ashford** » Sun Apr 10, 2016 9:06 am

ol boy wrote:Nice, so -in gets tied to the sensor return ground from the maf ?

Sent from my SM-G920V using Tapatalk

ahh i didnt catch that. sig return should be on the - on the chip
also the should be some sort of load on the maf i think

Megasquirt Support Forum (MSEXTRA)