Playing with OBD data

Status
Not open for further replies.
Joined
Jun 24, 2004
Messages
14,505
Location
Top of Virginia
I got my ELM327 adapter in the mail yesterday, so I set out with Torque on the tablet mounted on the CR-V's dash to observe. Very interesting stuff. Some thoughts.

TPS

It's widely postulated that the throttle "ramp" on DBW vehicles is not linear, and often aggressive, perhaps giving 90% throttle at 50% of pedal travel. At least with our Honda, this is very much not the case. Idle TPS is about 14-15%, and in most normal driving, even spirited driving, TPS never goes above about 40-50%. If you mat the throttle, it will go to wide open, which registers as 78.8%. 50% of pedal travel equals about 40% TPS. As engine technology marches on, with things like variable valve timing and even camshaft-less engines, the function of the throttle butterfly has changed forever from what it used to be...it has been supplanted from being the only means by which airflow into the engine is controlled.

Ignition advance

This one is interesting as well. Idle is around 4-8 deg advanced (which I assume means BTDC). Normal cruise with the torque converter unlocked gives about 20-30 deg of advance. If the torque converter is locked, I can easily push the advance down to near zero with the throttle...as the engine is more prone to knock in those conditions, it ratchets back the spark timing. What's interesting is that the advance dial also acts as my fuel cut indicator. Whenever the gas pedal goes to idle, and you feel engine braking as it cuts fuel, spark advance goes to -17 deg. Slow down from 25 mph, say, down to about 20 mph, you can feel the ECM give fuel back to the engine with a reduction in engine braking. At this exact moment, spark advance moves from -17 deg to around 0, then to the idle range in the positive single digits.

Fuel trim

It's not labeled either way, but I believe this is short term fuel trim (vs. long term trim). The ECM keeps it to a maximum of about 7-8% either pos or neg. It's most often in the 2-3% (pos or neg) range. When fuel is cut, fuel trim immediately goes to 0.00% (as it's not trimming anything at the moment). Once fuel is applied again, it takes a second or two before trim reacts, so while it's a good indicator that fuel is cut, this display alone will not indicate exactly when it's been re-applied (but the timing advance shows that). In doing some reading, it looks like fuel trim that is kept close to 0.00% is very good, and anything within 10% pos or neg indicates good ECM control of the fuel.

Engine load

This one surprised me. It acts very much like a "boost gauge" on a turbocharged engine. It's very surprising to see how often the engine is at 100% calculated engine load. It can be at 100% load at 1,800 rpm with the torque converter locked in 5th gear and not downshift. This gauge reacts immediately to throttle opening in a very non-linear way. I understand this is calculated based on current airflow vs. maximum airflow at that time. I need to research this PID more. I'm sure I can learn something from it, but I'm not sure what. The load never goes below about 20%, even under heavy engine braking.

Other misc data

Coolant temperature runs at 185-190 deg F. Intake air temperature is generally 10-20 deg F above ambient, though if I sit at idle, it will slowly rise. GPS and OBD vehicle speed are extremely close, with the speedometer pretty much displaying the OBD speed on the nose without much of an intentional error as I've seen on other vehicles I've owned. I've played with MAF (grams/sec) and it's interesting, too. It idles around 3 g/s and moderate throttle will give 30-50 g/s, even though the actual throttle butterfly has only moved from 15% to 25%. Again, there seems to be a lot more to airflow in a modern engine than throttle position alone.

Altogether, this has sparked a lot of interest in me to learn more about these values and what they mean. Will also play with the Acura this weekend.
 
Torque is a very cool app. It's the only reason I keep an Android device around!

If you Google around you will see some very good tutorials on what all the values mean.

I do think some of the things I've seen (I've seen similar behaviors as you have on our Fords) may also be accounted for by the refresh rate of the devices. I'm sure they are "dropping packets" somewhere along the lines but it's still cool to watch.

One interesting thing I've noted on my Taurus is that the throttle position never really goes above 50-60%. No matter what - you can nail the gas and it will just go to 50-60% and the boost will go up but not the throttle position.

Don't forget to setup the CSV logging and you can get that data sent to a CSV which you can then crunch in your spreadsheet program. That's where you can have fun with charts and such!
 
very cool writeup, I am still debating whether to

a) buy a droid tablet, obd2 connector, and torque -or-

b) buy an ultragauge
 
Last edited:
I found a write-up on the engine load PID, and it said that it's intended to show how much reserve torque is available from the engine. So when I'm running at 55 mph with the torque converter locked and rpm at 1,800 rpm, it doesn't take very much throttle to deliver as much torque that's available at that moment. If I give more throttle so that it unlocks the torque converter, engine load drops (even though I'm giving "more throttle"). If I give even more throttle so that it downshifts, the engine load drops further.

The opposite is true also, during light-to-moderate acceleration. Load might be 50% through 1st gear, then increase to 60-70% in 2nd gear, then increase to 80-90% in 3rd gear, all with a completely steady pedal position. And that makes sense; it generates more of its potential torque with each stepped loss of mechanical advantage as it upshifts.

Fascinating.
 
I would wonder whether the single fuel trim readout you are given is the sum of short and long term fuel trim, since knowing either singly is pointless. However, the ECU isn't "limiting" your fuel trim to 7 or 8%, it just happens to be where your engine is operating. It has the ability to trim 25% if it wants. I would actually suggest that your engine running at 7 or 8% trim is not all that great. I get concerned when mine breaches 3% under steady state. I know they give condemnation points of "if it's over 10%, blah blah blah" but at 7 or 8% I would be concerned something might not be quite right. With that said, it would just be something to look into when you are bored or have spare time. There are plenty of sites out there that give at least hints on possible causes of long term enrichment or leaning.

You need to evaluate the source of the TPS reading. You are saying that it appears to be linear, but oftentimes that data is filtered by the ECM. There is sometimes a raw TPS and a filtered
TPS available through the ECM. Even if the ramp-up appears to be linear, it may not actually be linear. The fact that you say it does not go beyond 78.8% leads me to believe it may be a filtered value.

As an aside, does the CR-V have an electronic throttle or is it a cable set-up?

With regards to engine load, this number is of varying accuracy. For engines without a MAF, the calculated engine load is probably not all that accurate as its having to approximate engine load through the MAP sensor (if I remember right). Some scanners will report a MAF reading but it's calculated, not from a MAF sensor. Your saying that the engine load never goes below 20% leads me to believe that it is not a reliable number for this vehicle. You should be sitting at something like 2-4% load at idle.
 
I wanted a permanent and unobtrusive setup as opposed to a tablet on the dash, so I installed a ScanGauge and find it very useful. It is fascinating monitoring spark advance and other PIDs.
 
nice write up Hokie - thank you! Anything on EGR? y 1.5L has a big controlled egr valve on the intake/head and I was wondering if where knock is detected that the egr flow ramps up. The high static cop motor seems to use egr and timing to limit power with marginal fuel. Wondering whats more influential.
 
Last edited:
Regarding the STFT, I may not have used the best wording. I know the ECU doesn't "limit" the STFT to 7-8%...that's simply as far as it's ever had to adjust. And as I said, those are the maximum values I've seen in either direction, and it only happens with very abrupt changes in throttle...in other words, when conditions rapidly change within the engine. During steady state driving, STFT is in the range of about 1-3%, and usually in the positive direction (meaning that it is adding fuel). Like you, I would find it hard to find fault with steady state STFTs below 3%. It doesn't appear that I can access LTFT, at least not with the free version of Torque.

Regarding the TPS, are there multiple PIDs in the OBD-II specification for that, one for a filtered value and one for a raw value? I would assume that a filtered value would be useless for any type of diagnosis, and thus not allowed in the specification. Or at least, the specification would require that a raw TPS PID also be available.

Geroge, the 7" form factor tablet sits in a mount I made for it in the dash. It works great, and also serves as my GPS if I go out of town with the car. No suction cups and no irrepairable damage to the dash. The mount is screwed to a small panel in the top of the dash that is easily removable. I bought a second one from the Honda dealer for 15 bucks, so I can either have my tablet mount on the dash, or I can swap it for a stock panel and it's back to stock. It's almost like it was designed that way.

Anyone who has used both the paid and the free version of Torque, what additional PIDs did the paid version unlock? The Torque website is not exactly clear on that. Thanks.
 
Originally Posted By: Barkleymut
very cool writeup, I am still debating whether to

a) buy a droid tablet, obd2 connector, and torque -or-

b) buy an ultragauge


I have the Scanguage, which I think tucked into my truck decently well. A tablet with datalogging would be the only reason I'd change, and I'd have to buy a tablet to go with it. AFAIK Torque doesn't work on my iPad, which is a shame, it might be an app I'd be willing to pay for.

So far the SG works well, except when I'm wearing my sunglasses--angle of the display, dark glass tint, something--find it rather hard to read quickly. But usually I don't need to read it quickly anyhow.
 
Originally Posted By: cchase
As an aside, does the CR-V have an electronic throttle or is it a cable set-up?


Our 2008 has electronic throttle.

I did find that there are two throttle position PIDs in the OBD-II specification. PID 11 is "throttle position" and PID 45 is "relative throttle position". I need to do some more research on those, see what they mean, and see if I can figure out which one Torque is reporting.

Based on this short explanation...

http://www.mp3car.com/engine-management-...sition-pid.html

...it appears that a minimum above 0% and a maximum below 100% is normal.
 
LTFT is a good parameter to check, since it will show any kind of operating deficiencies in the engine. STFT responds quickly to throttle input but LTFT will show small air leaks, lazy O2 sensors, poor injector spray pattern.

Typically STFT is just fine tuning the LTFT, so it will always hover around 0%. For example, if STFT stays +5% for more than x period of time (usually some number of seconds), it shifts +5% onto the LTFT and returns the STFT to 0%. That's why only knowing STFT without LTFT is kind of "dangerous knowledge". If you don't know the LTFT, it could be sitting at +10% and the STFT will look dandy at 0%.

It's probably better to consider the throttle position values as "actual" vs "commanded" rather than "raw" and "filtered", but still it is important to determine what you are looking at. If the vehicle does not have electronic throttle, it may not have 2 parameters. "Commanded" will be the position of the gas pedal, "actual" will be what the throttle servo is actually outputting.

I've never played around with Torque so I don't know how specific the labels are in determining what you're seeing.
 
Me too; never could get good explanation about the "load". It is *reported* by ECM i.e. it is not computed by scanner. Interestingly, ECM then uses this load value to tweak the other parameters.
 
Quick update. I bought the paid version tonight, and there are TONS more PIDs in that, including one called "relative throttle position". That seems to be the true throttle position, varying from 8% at "closed" to 92% at "open". Progression is still very linear, at least with the engine off/IGN on. Will have to see how it goes during driving.

Also on my Torque Pro dashboard are EGR %, commanded and measured AFR, STFT, LTFT, and fuel flow rate in gal/min (though it also will show in gal/hr).

How cool. I will be geeking out on this this weekend...
 
Sounds like the paid version is more than worth it.

When I got my Ultragauge I spent months trying to convince other people to drive just so I could experiment on the road.
 
I have to correct a few things I posted earlier. First, the minimum reported TP with the engine running is actually around 2-3%. Maximum is about 97-98%. So I think I have found the right PID here. It's still very linear. I think the camshaft phasing might have a lot to do with how "peppy" modern engines feel compared with those without that feature. At least in this vehicle.

My fuel trims are interesting. My route to work this morning was all rural roads with speeds of 45-60 mph. I never spent any time idling. But this evening, I was in a lot of traffic, and STFT had much more wild swings. It would swing through about 12% on either side of 0. LTFT varied from a minimum of about -0.5% to a maximum of about 1.5%. So it seems that the system is happy overall, but the swings in STFT surprised me.

Stopped at idle in drive, the STFT would slowly creep UP. Not immediately, but progressively. I would think a true vacuum leak would show a consistently elevated FT, rather than a slowly rising one. I didn't notice when the STFT went so negative, but that generally happens when the throttle quickly closes.

AFR was also consistently maintained very close to 14.7. At idle, the commanded AFR was always 14.81, and the measured AFR would jump around that measure, like between 14.7 and 14.9. The AFR is a great meter showing when the ECM cuts fuel. When it does, the commanded AFR immediately pegs at the gauge maximum (30 as I have it set up), and the measured AFR quickly responds. The split instant I touch the gas pedal, the commanded AFR jumps back down to around 14.7, and the measured quickly follows.

I was hoping to see evidence of torque management during shifting, but I haven't quite seen it yet. I was looking for either the throttle to scale back quickly and return, or for the AFR to quickly jump then return, etc. I don't see evidence of it, but that doesn't mean that it doesn't happen.
 
I think some vehicles pull timing during shifts, which could be something to look for.

That aside, some vehicles have strange throttle mapping. There's no way around it. Every BMW I've driven has very soft throttle response in first gear and immediately sharpens in 2nd and above, for example.
 
My OBD data-

Commanded AFR runs .990-.997, but when cold runs 1.0.
STFT runs -5% to +2%
LTFT runs -2% to -10% depending on air intake temp
Air intake temps from ambient to 150f
Coolant temps 180-190f mostly
Timing 0 to 50 degrees
Closed loop after CAT1 hits 600f (runs 1200f-1400f mostly)
.65 volts O2 average reading (.08 to .9 have been seen)
MAP of 13.7 has been read a few times, (14.07 is the normal reading before starting the engine).
Average gph over a tank ranges from .85 to 1.4
Highest gph rate was 3rd gear full throttle at 6k rpm which was 11.0 gph.

Can list a few more since the Ultragauge on my car will show 70+ parameters. Mainly used for learning how to drive more efficiently and getting feedback every trip.
 
I have also looked for timing changes, but I haven't seen those either. I will play with the MDX some; I perceive the torque management more from that car than I do from the CR-V. It's got the V-6 with the V-6 transmission, and though it has the updated V-6 transmission, Honda may have put some more aggressive torque management in that car vs. in a 4-cylinder powertrain like our CR-V has.

I was surprised at how linear the throttle ramp is in the CR-V. I do believe that there's more to modern engine "response" than throttle position alone. It may feel like the throttle is open more than it really is, but it may be the result of camshaft phasing or something else. The MDX doesn't have camshaft phasing, so I will be looking closely at the throttle ramp in that car.

As an aside, does anyone know how closely ignition timing would correlate with valve timing? I can't see the valve timing in this app (or at least I haven't found that PID yet), but can obviously see the ignition timing. Would it make sense that the valve timing would advance and retard with the ignition timing, or would the two operate relatively independently of each other?
 
Originally Posted By: Hokiefyd
Quick update. I bought the paid version tonight, and there are TONS more PIDs in that, including one called "relative throttle position". That seems to be the true throttle position, varying from 8% at "closed" to 92% at "open". Progression is still very linear, at least with the engine off/IGN on. Will have to see how it goes during driving.


If that "relative throttle position" is truly the pedal sensor, then I would expect reasonably linearity. Which the ECU could then remap to be as non-linear as it cares for. Since it works together as a system, the pedal sensor could be non-linear, and the ECU still mapped to it linear, non-linear or whatever the programmer wanted.

Pretty cool stuff. I'd be curious how many of the sensors it can see; that would be cool if you could interrogate any and all systems in the car.
 
Status
Not open for further replies.
Back
Top