DIYPNP Map Sensor Reading Off
 

Hi all,

Some of you guys have seen by car in the build threads section. Fast forward a few weeks ago, I did a shakedown run at Pacific Raceways where my car broke down.

I ended up towing the car back up to Canada and going over everything it seems like my map sensor is reading off. Thinking that the DIYPNP MPX4250AP died, I swapped it out for a GM 3bar. No luck either.

So I ended up pulling off the connector to the map sensor, and weird thing is, with the sensor unplugged, and the ecu powered ON, I am getting a weird back back feed voltage off the signal wire. It reads about 1-1.3v.

With everything completely soldered off the board (NO MAP SENSOR), I am seeing voltage on the map sensor signal pad with my meter. I go into Tunerstudio and it's telling me I'm at 20kPa but slowly rises to about 120kPa over a course of 5 minute. WTF. I shouldn't be reading anything here (or 20kPa, the "0 voltage" signal. This is with NOTHING plugged in at this point.

Any thoughts? Did my DIYPNP just died on me?

For reference, this is the video when it just went to shit on me:



Im obviously not doing anything to the DIYPNP when this happens on track.

It's hard to troubleshoot without schematics.

Everyone should tell DIY that they should publish the schematics.

No kidding. One interesting thing we found out today by probing around is that the ground for the map sensor goes through the microsquirt module. Seems like a lazy way of doing things versus finding an alternate way to route the trace.

This is normal behaviour. There is some voltage from the processor's ADC circuit, which rises as time goes by. You'll find that the same thing happens to the TPC is if you disconnect it.

Still does not say why, or give any clue as to why the DIYPNP is reading map incorrectly despite which map sensor is plugged in. Not sure if you saw the video but I would say that is pretty random behavior.

Quite frankly I'm lucky I didn't run super lean and totally kill my motor.

you say it was off, but how off was it?

Your video embed is teh brokenz.

Key on, engine off - when it should be reading atmospheric, it reads roughly 124-125kPa

Fixed now, my apologies! Skip to about 4:00 for broken diypnp

so it reads 124-125kPa at atmo, does it react in vacuum and boost as normal albiet 24-25kPa too high?

As indicated previously by Reverant, your voltage readings are normal and customary. The input is floating without input from an actual sensor. In fact, your readings show that the ADC circuit is likely to be working correctly.

The map sensor receives VREF and Ground directly from the same power supply as the processor in order to minimize any potential offsets. If you feel this is in error, please tell me a better way to route the traces between the map sensor and processor, and I will pass your suggestion on to engineering.

Did you calibrate the MPX4250? Did you calibrate the GM 3 bar? DId you take voltage readings from the sensor?

You seem pretty quick to point fingers, but so far I'm not seeing where our product is to blame.

You can say that there are almost 3 different reactions I get, depending on what mood the DIYPNP is in:

1) No reaction, stays at 125kPa at vacuum while revving

2) Some reaction if I do light revs, but the behavior is not consistent with engine behavior (vacuum does not increase)

3) Some reaction if I do light revs, but the behavior is consistent (vacuum does increase, goes to roughly 112kPa)

with both sensors?

I was simply asking the question if this is normal, and yes it seems that it is. So if we can write this out of the equation then lets look at other potential problems with this situation.

As for sensor calibration, yes I calibrated both. I confirmed that as well by looking at the data sheets. I did take voltage readings off both sensors and the sensors are fine.

Did you not see the video? Sudden power loss. I managed to tune on this fine on the dyno, by a respectable tuner at that. I'm quick to point fingers because my car suddenly died on track, and I was left with a towing bill due to an ECU fault. So far

Obviously no one was touching the DIYPNP at the time. I hope there is a better explanation than "did you calibrate the sensors".

Yes, with both sensors.

so it's suddenly reading high, and not reacting much at all to inputs. how did you install the 3bar?

I know the stock one can crap out, itll typically happen after a cold winter and condensation was allowed to form in the MAP line and it fubars the circuitry inside.

Take a couple of pictures of the DIYPNP, both sides (top and bottom) and post them here.

I just followed what they had here:3-Bar MAP sensor

As requested here are some shots of the DIYPNP front and back. Note, there is no mica filter on the transistor on the back because I pulled it off while probing around.

https://www.miataturbo.net/attachmen...ine=1371010213

https://www.miataturbo.net/attachmen...ine=1371010213

https://www.miataturbo.net/attachmen...ine=1371010213


Here is one of the logs that I did in the pit while going through two gears in the pit at Pacific Raceways. Note that at 6800rpm, it says Im at 137kPa (19.xpsi), there is no way my rotrex is boosting this much as much as Id like.

Just shows how all over the place the map signal was.

https://www.miataturbo.net/attachmen...ine=1371010213

137 kPa = 5.5 psi of boost at sea level.

It seems you are very laser focused at an alleged map sensor problem, and I think that's throwing you off from your real problem.

I don't see any other suggestions here for any other problems. The point being is that the car went to shit randomly. Not sure if you even saw the video, but I don't see a response to something just occuring out of the blue.

Maybe because the reason is that this product is terribly unreliable. I made it down to Washington twice, once to tune and once to track. This thing obviously died on me out of the blue but I see refusal to admit that nothing but the DIYPNP died on itself.

Also looking at the log you can see that MAP increases before RPM does. I got some response from DIY Autotune stating that all you guys see is "moderately slow throttle response" which is utter bullshit. I'll post the log showing TPS vs MAP later tonight.

With the ignition on, what's the voltage on pin 3 on the white PCB? Obviously with the MAP sensor soldered on the board.

I am glad to help if you can be reasonable. I will fall on my sword immediately, issue a public apology, and get you set up with whatever hardware will make it right if you can prove that there's a legitimate hardware problem. But right now, I'm struggling because you're attacking me and my company. You're not looking for solutions to troubleshoot your problem -- you're attempting to condemn the DIYPNP and DIYAutoTune.com for a problem that you have not shown to exist. Taking an antagonistic attitude towards the people you are asking for help is not productive.

The DIYPNP is only sold as a user assembled kit. The hardware is robust and well tested. It's one of our better selling kits, and of the various versions the N76 kit is the best seller. This is no doubt due to how popular they are with the Miata crowd.

I've been running a DIYPNP in my own car (turbo Miata, basically same application as you) for several years. I've built unkown dozens of them for various applications, including for enduro racers. I built many before I started working for DIYAutoTune.com, and I've built many after being hired on. I have never had a problem with one I've built. We have no QC issues with this product.

The weakest link in this product is the quality of assembly, and that's something we have no control over. This is no different than any other DIY oriented kit. Stray solder, incorrect connections, or poor solder joints can wreak havoc. Those are the typical issues, and they're all user error. A cold solder joint may only show up when the joint gets hot. If that happens, it's certainly not our fault in any way -- the fault is on the person who assembled it, as we are not responsible for the user's workmanship. You can buy materials to build a barn at Home Depot, but if the barn is poorly built, falls, and crushes your prize pig, is that Home Depot's fault?

I'm sorry that something happened while you were on track, but that "something" could have been any number of things. From the evidence you have supplied, that something does not appear to be a problem with a map sensor. The only problems I can find with the map sensor is your incorrect understanding of how the sensor works -- your interpretation of the reading was incorrect, your interpretation of troubleshooting the ADC circuit was incorrect, and your understanding of how to wire an analog device to the ECU is incorrect.

The datalog screenshot you provided shows the sensor working and showing reasonable values. Posting an actual datalog .msl file is much more useful than a screenshot, but the map sensor is clearly working in the image you provided. It is clearly showing 137 kPa absolute. That is 5.5 psi of boost, NOT 19 psi of boost (that would be 235 kPa).

The map sensor itself is generally extremely reliable. It is more likely to damage or dislodge the tube that feeds the map sensor its pressure signal than for the sensor itself to fail. Did you check the tube for obstructions?

If I had a turbo Miata that suddenly fell on its face when hot, under high load, and at high RPM, the first place I would look is at the ignition system. As it gets hot, it has a harder time firing the spark (high resistance due to higher heat). As ignition components wear over time, the typical failure mode is to first break up under high heat, high load, high RPM. Over time, it will start breaking up under lighter heat and load conditions.

It's fair to say that something failed on you at the track. I don't think it's the map sensor that came with your DIYPNP or the other sensor that you replaced it with. I think it's something else. We'll help you, but you'll have to help us to help you. Part of that is not throwing mud.

What Ben said. As the founder/owner here, I've got to say he's right on, that said, our policy is pretty much 'the reasonable customer is always right'. We want to help you, but right from the get go- there was blame and attack straight from the video. Let's figure out what went wrong, and help you make it right? Maybe there's a problem with the ECU, maybe it was something else, either way, we're your friend, not your enemy, and we'll help you figure it out.

I had very similar symptoms years ago with another MS install while on the track. Minor difference, mine only showed up when I made a hard left turn with a passenger in the car. Root cause, the vacuum line to the MAP sensor was getting pinched when my passenger braced and pushed against the firewall, that was locking my map reading in one place by pinching off the line, and i fell on my face. Easy fix, reroute the map sensor's vacuum line somewhere it won't get stepped on ;).

If the map sensor IS reading off, the first place to look is the calibration. We can support any MAP sensor in the world with a 0-5v output. That means it's up to the person who built the ECU to calibrate it. It defaults to use the MPX4250 which is a 2.5 bar map sensor, anything else will require calibration to match the voltage output curve of the sensor to the ECU so the ECU can correctly interpret the output from the sensor.

We want to help-- please help us help you.

Ben, you are right on some points you have made, I have been shooting off words without organizing my thoughts. But you seem to have missed some important points that I have also made, and only pointed out the flaws in my arguement.

I haven't responded yet because I haven't been able to do the testing that DIYAutotune has requested (via email) or what Reverant has suggested due to Father's Day events and working a busy week. Hopefully I will have some time tomorrow to put together a coherant response and try to address the mistakes I have made in some previous posts.

Having a hard time locating Pin 3, maybe my eyes are just not working right, mind circling the general location?

Pin 3 is right next to pin 1 on the 50-way connector.

I get 1.8V at this point.


With the sensor hooked up to 5V and SG GND, I get 1.8V as well. Seems like the map sensor itself is fine.

Does the voltage change if you apply vacuum to the sensor?

Yes, while blowing into the sensor I can see the voltage fluctuate.

Get a syringe, remove the needle and replace it with a small piece of vacuum tubing. Apply vacuum/boost to the sensor, and make sure the voltage goes from approx 0.2V to 4.8V.

Yep, works for both the DIYPNP map sensor and GM 3bar. Also borrowed a friend's GM 3bar out of his running car and they all seem to be working fine.

Do you see the MAP reading in the tuning software moving up and down during these tests?

Tried to plug in the laptop today to verify the map and fuel loads as per DIY's email response today and I was unable to connect to the laptop.

But to answer your question, based on past testing the logs show that there is map activity but it is very non-linear.

https://www.miataturbo.net/attachmen...ine=1372708191

Notice a few things here: Map signal is moving but looks way off compared to the log below. Someone mentioned how my map signal moving before rpm does is related to slow throttle response. Bullshit. Look at the log below.

https://www.miataturbo.net/attachmen...ine=1372708191

There was a comment earlier about how my "map sensor issue" was related to slow throttle response. This is just not the case. Look at how my map slowly increases as rpm increases. Everything here is coming on smooth, no random spikes like in the log above.

I understand these screenshots are hard to interpret due to scaling, etc. I'm more than happy to post the log for those still not convinced that there is something seriously wrong with the DIYPNP.

Could it be that you have a blockage/kink in the hose to the MAP sensor? That would explain things.

Hose to the map sensor is fine and clear. I blew through them to make sure they were clean at track. No kinks as the hose lengths are pretty short, I tried to reduce the amount of hose as much as possible, hence why the sensors are mounted in the engine bay, and not on the diypnp board. I pulled 6 wires through the firewall.

Then I'd disconnect the sensor and use a potentiometer instead. See if that speeds up the response. Just turn it left-right and see if the MAP-reading in TS changes instantaneously.
If it looks OK with the potentiometer, your MAP sensor is shot. Replace it.
If not, attach the potentiometer directly at the MS and not via the wires going into the engine compartment. Disconnect everything else to eliminate interference from other areas. Make sure to use sensor ground as reference and not some random ground found on the chassis.
If the potentiometer attached directly to MS gives you slow responses, then I'd say something is wrong with your MS. If the responses are immediate, your problem is likely in the wiring.

BTW: the potentiometer should be rated at similar resistance as your MAP sensor. Doesn't need to be the same- just similar.

BTW2: If you attach screenshots of logs, trying to troubleshoot MAP issues, TP-data would help. For all I know, the slow ramping-up of MAP could be related to throttle-movement. And a description of conditions. Were you free-revving? Accelerating hard in low gear?

I'm still trying to figure out why the MSQ and log files still haven't been posted.

QFT.

There is a lot of talk of logs and looking at them, but a screen dump does not = an actual log.

In for PEBKAC.

Because no one asked for them, although I have said numerous times that they are readily available. This goes for the tune file as well. With my dealings with the tech support @ DIYAutotune, they were not asked for as well.

Post #21 begs to differ

Mention of log:

Really?

My opinion, both professional and personal: You're chasing the wrong thing because you initially mis-interpreted a datalog. If you had any map sensor issue, it was probably a kinked hose between your intake manifold the sensor's pressure inlet. A kinked hose will cause slow reporting of pressure changes. I hate to use absolutes (that's a pressure joke, get it?) but there's really no hardware failure possible that would cause "sometimes" slow pressure readings.

You have continued to not demonstrate your reported failure. Meanwhile you have not removed your video that condemns our product for the failure you can't demonstrate--while at the same time you continue to ask for our assistance, but generally not follow advise.

I'm trying to help you here, our whole company is. As are people on this forum. But I can't continue to hold your hand as you go further down the wrong rabbit hole.

I have feeling the map sensor is working properly but you have a mechanical failure that is showing up on the logs and you're misdiagnosing the failure.

I've been dealing with Matt via email and he has been helpful. I dont see anywhere within this thread that I haven't been following advice or suggestions people have made.

As for your comments, I don't see any helpful advice except for asking me to take down a video. I have tried 3 map sensors. Would you like me to take a photograph of my map sensor setup? There is no kinked hose, its probably as straight as you can get with a hose. Much better than pulling hose through the firewall like the original DIYPNP design is made for.

Yes, I made an error in reading a pressure signal. I deal with gas in pipes all day in my work and forgive me because we use gauge pressure for all our readings. Even if it is reading a map signal, I honestly don't see a suggestion from you for what else to test. On the other hand, Reverant has at least suggested for me to probe the board and see whats going on.

Of course I won't remove my video. There is nothing in this thread that suggests map sensor failure, or any mechanical failure. I have done my compression tests and leak downs, mechanically everything is firing, I have tested my coilpacks and spark plugs, everything checks out?


How do I know this? I had to borrow a Vipec V88 and order a boomslang harness for temporary use. Let me say this, the car runs fine and it is using the same hardware and sensors. Want video proof? This whole thread seems to be about proving ME wrong, which I have expected from the beginning seeing as this board has a high population with Megasquirt users.

If I'm wrong, then please put it plain and simple for me: What should I test and what should I be expecting to see from these tests? Funny thing is that Matt has actually been trying to help me diagnose the board, vs. you seem to be hung up on mechanical,wiring, etc issues.

EDIT: If its not a map sensor failure (seeing as we have tested all 3, and through email DIYAutotune has confirmed what I'm seeing is correct, then what else could it be?

Braineack, I have been running the car successfully these past few weeks using a Vipec. There is no mechanical failure.

As for those being so hung up on the log files and tune file not being uploaded, well they have been attached so please enlighten me. I have nothing to hide with this issue and I think I have showed pretty good cooperation with trying to reply in a timely response. In the attached logs there is a log from going to close to top of 6th, and then the other two are from the trackday that I was logging in the pits.

Although this thread has been useful, I have also been doing my own due diligence and trying to sort this issue out by testing all the individual components.

If there is one bone to pick with DIYAutotune its that there is no phone number to call. If this was my daily driver I'd be pretty frustrated seeing as the car was down for a couple of weeks.

I am no BIG expert by any means - but I a had a look at your tune and compared it with my old DIYPNP tune.

2 things I noticed:

- your Map sample setting at 35 is veeeery low - it will kep you from a very instant reaction. (I think you should be able to set it at ~60 - 70 without any problems

- weird thing is your injector timing (mine is set at -350°) yours at -175 - I am not sure if this could cause strangeness in the map reading?

Another thing when I re-read this thread - looking at your picture:
https://www.miataturbo.net/attachmen...ine=1371010213

I would clean the solder spots of the removed map sensor to be sure there is no short of any sort between the spots...

I'd really like to see a log of the car running poorly. The logs you posted don't show much in the way of MAP sensor issues. Here's my take on the data logs you posted in detail. I've sent them by email, but wanted to post them here for the benefit of everyone reading this thread.

2013-05-14_23.08.12.msl: There's a small amount of MAP jitter recorded. There is much more at full throttle than at low throttle amounts. This is causing the TPS driven decel fuel cut to activate. My recommendation is to disable TPSdot fuel cut by changing the Decel Fuel Amount in Acceleration Wizard to 100%.

While the MAP noise can be caused by electrical noise, my experience has been that MAP noise that only manifests itself at full throttle is usually caused by turbulence in the intake. Moving the point where the MAP sensor is plumbed up often can reduce this noise.

2013-5-29_15.43.56.msl: The TPS appears to be broken or calibrated wrong; it's reading -17% through the whole log. There is a point where the MAP bounces up at record 657, but with the TPS not reading, that could have been a stab of the throttle.

2013-5-29_17.55.11.msl: No anomalous MAP readings that I can see. I did spot two points where the RPM input lost sync at idle - possibly a flakey CAS - and several points where a bit of noise on the TPS triggered acceleration enrichment, including one at record 700 that caused a dip in idle.

2013-05-29_17.58.33.msl: MAP reading is not manifesting any noise that I see. I did see a couple spots where the overrun fuel cut kicked in for a very short instant that may have made the engine hiccup.

If you are seeing MAP behavior that looks odd to you, please let me know what record number and file, and I'll have a closer look. But I'm not seeing it.

Based on what I saw in the logs, it's possible that I might have seen the start of a CAS problem. However, this is kind of like trying to diagnose a broken bone with X-rays taken before the accident.

Just wanted to repeat this for emphasis. We've spent a lot of time going over the MAP sensor readings, and other than one point that appears to have been a calibration error, I'm not seeing a MAP sensor reading problem. This thing needs to be diagnosed as a whole.

Since we are posting emails here, this is my response to your last email Matt. I have yet to hear back.

I have to emphasize again: Nothing has changed since being able to run the car on the VIPEC, no change in CAS, coils, injectors, map location, map sensor, TPS.

As it is in fact a DIY unit I would look over loose solder joints, solder blobs shorting or something like this - this might correlate with your running "other" ecu...

We were closed July 4th, and I can't seem to find your email. I'm emailing back now, but I can't really answer this without a data log of the car as things are now.