DIY Fuel rate meter/injector duty cycle meter/GPH

[theholycow]

This technique was presented to me (http://www.gassavers.org/showpost.php?p=99613&postcount=18) by user monroe74 on the realtime monitoring tools for pre-obd2 cars (http://www.gassavers.org/showthread.php?t=7345) thread. Thank you, monroe74! See also fumesucker's work measuring the same thing with his laptop computer sound card (http://www.gassavers.org/showthread.php?t=7608), which allows him to graph and log the data.

More with the sound card, using programs to calculate MPG:
http://www.gassavers.org/showthread.php?t=5233
http://www.gassavers.org/showthread.php?t=4250
http://opengauge.org/diympggauge

If you want to combine fuel rate with vehicle speed to get MPG, see these threads:
http://www.gassavers.org/showthread.php?p=64732#post6473232
http://www.gassavers.org/showthread.php?t=4250
They talk about using laptop computers, and the second one also includes some circuit diagrams.

Extremely cool project with really great information about OBD protocols, connectors, FE calculations, and how to put a MPG gauge in for cheap:
http://www.circuitcellar.com/avr2004/first.html

Another idea is to use an OEM MPG display or DIC (Driver Information Center) made for your model or a similar model, or one that might be adaptable to dissimilar models...
http://www.gassavers.org/showthread.php?t=4258

If this is pasted to another forum or otherwise sent elsewhere, here's the link to the original (and please leave the link in):
http://www.gassavers.org/showthread.php?t=7714

I've mentioned this in many threads, but I think it deserves its own thread. The result is a meter/gauge that shows you in realtime just how much fuel is being used. It's different from an instant MPG gauge, which shows you fuel rate compared to distance. This is a better way to directly and immediately compare one strategy to another in some circumstances -- for example, determining when you're in DFCO (Deceleration Fuel Cut Off) or if more fuel is used at a given speed in one gear than another. It also gives you fuel usage information when the vehicle isn't moving, such as at idle, which you won't get from an instant MPG reading.

1. Get a duty cycle meter, or a digital dwell meter. I have not seen a cheap/easy/appropriate duty cycle meter, but Harbor Freight (http://www.harborfreight.com/cpi/ctaf/displayitem.taf?Itemnumber=95670) sells a digital multimeter (DMM) with Dwell function for $33 (plus $8 shipping if you can't get to one of their stores). That one has a very large display and is built well, though if you leave it under your windshield on a hot sunny day the label will peel.

You need a digital meter because you will need to easily see very small, precise numbers as well as larger numbers. It's tough to see such different scales on an analog display.

2. Tap one of your fuel injector wires and run the wire to wherever you'll want the meter (I ran mine near the center console). It doesn't matter which injector, but you must use the injector's positive wire. If you tap it with something like a 3M Scotch-Lok QuickSplice, be careful not to damage the wire. To avoid that, you can just disconnect the injector's electrical connector (either from the injector itself or the ECU, whichever is easier to reach), jam the stripped end of your wire into the female end of that connector, and re reconnect.

3. Connect the injector extension wire to the positive lead of the DMM. Connect the DMM's negative lead to ground (I used one of my car's power outlets).

4. Set the DMM on Dwell for 4 cylinder engines, regardless of how many cylinders you have. We're not actually measuring dwell (which is adjustment of old-fashioned points and condensor ignition systems); we're using it as a duty cycle meter. The 4 cylinder provides a 0 to 90 scale, which is almost exactly the 0 to 100% scale required.

5. Start the engine and look at the meter. You are looking at fuel injector duty cycle information. It is off by a relative 10% -- so if it reads 1.0, the actual measurement is 1.1; if it reads 20, the actual measurement is 22; if it reads 50, the actual measurement is 55. This is the percentage of time your injector is injecting; the balance of 100% is the time that it's closed.

If you want to translate that into an absolute fuel rate measurement, as in Gallons Per Hour, you've got some math to do. You need to find your fuel rail pressure and your injector size/flow rate, I think. I'm not entirely sure how to do this, as it hasn't mattered to me. I think it might be possible to get a decently accurate estimate by knowing the idle fuel usage, which you may be able to find online, and figuring that compared to your reading at idle.

For reference, warm idle on my 2002 GMC 5.3l V8 reads 1.0 to 1.5, and on my 2008 VW 2.5l I5 it reads something like 1.3 to 1.8. If you're idling and reading .1 or 10, you may have done something wrong.

[theholycow]

I've already beat my previous best MPG in both of my main vehicles due to this meter. Things I've managed to learn about my vehicles with it:

2002 GMC 5.3l:

- DFCO (Deceleration Fuel Cut Off) is only useful if you can engine-brake for 30 seconds or more. It takes 6 to 10 seconds to get into DFCO but does hold it down to ~1100 rpm.

- Due to that DFCO issue, using N is far better than coasting in D at any time. Not only does coasting in D slow the vehicle down, but even with my foot off the gas the computer must inject more fuel at higher RPM to keep the air/fuel ratio correct.

- While stopped (at traffic lights and such), idling in N uses a very tiny bit more fuel than idling in D, but is worthwhile.


2008 Volkswagen 2.5l I5:

- DFCO is very aggressive whenever I'm above 40mph and will continue down to 1000rpm. If I'm going below 35mph when I take my foot off the gas, it will almost surely NOT go into DFCO.

- Like the GMC, when not in DFCO and foot off the gas, it still has to use more gas than if it's in neutral, to keep the air/fuel ratio correct...and of course still slows the car more.

- WOT (Wide Open Throttle) actually uses less fuel than 90% of the gas pedal travel while still making more power. It is actually more efficient with my right foot all the way to the floor than if I back off a little. This is useful knowledge for someone who uses higher gears and wide throttle openings to reduce pumping losses. (Note: The car is Drive By Wire and the gas pedal isn't physically connected to the throttle, but the data is as observed.)

- When I combined this meter with Ross Tech Vag-Com OBDII data collection, I found that the engine does not enter open loop under any condition, even at WOT. It only uses open loop for about 30 seconds after a cold start, and when in DFCO.

[MarvHein]

I've already beat my previous best MPG in both of my main vehicles due to this meter. Things I've managed to learn about my vehicles with it:

2002 GMC 5.3l:

- DFCO (Deceleration Fuel Cut Off) is only useful if you can engine-brake for 30 seconds or more. It takes 6 to 10 seconds to get into DFCO but does hold it down to ~1100 rpm.

- Due to that DFCO issue, using N is far better than coasting in D at any time. Not only does coasting in D slow the vehicle down, but even with my foot off the gas the computer must inject more fuel at higher RPM to keep the air/fuel ratio correct.

- Idling in N uses a very tiny bit more fuel than idling in D, but is worthwhile.



I too became a big fan of N coasting vs. D coasting with my 2004 5.3L Yukon XL when experimenting down a 2-mile hill. The onboard computer showed 99 mpg in neutral vs. 55 mpg in drive. I'm sure the actual mpg is significantly better than 99 as coasting in N as much as possible improved my city driving by 2 mpg, a huge percentage for these vehicles.

(Caveat: My front diff blew out recently and I just don't know if there's any way it might be related to this technique. I never noticed any obviously stressful situations when re-engaging drive and the tranny seems fine.)

I took your comment about idling consumption to be a rolling idle comment. At a stop, have you found idling in N uses more or less than idling in D?

[theholycow]

As I mentioned in the other thread, it's impossible for your diff to have gotten messed up by neutral coasting with the possible exception of if you were in 4wd when doing it.

The comment about idling was while stopped. I will edit it, as I really should have said that. While rolling, there's nothing "tiny" about the difference between N and D.

[theholycow]

As requested in another thread, photos. Click on the photo to see it at PicasaWeb where you can go to view it in a larger size.

A fuel injector wire in the VW:
http://lh5.ggpht.com/ronanian/SEiFZ-l9BHI/AAAAAAAACV0/s6JDaMGuYIA/s576/IMG_1039.JPG (http://picasaweb.google.com/lh/viewPhoto?uname=ronanian&aid=5192607384817217601&iid=5208559650399388786)

The meter hanging on my dash, left-center in this photo:
http://lh4.ggpht.com/ronanian/SEiFoyucbgI/AAAAAAAACWA/Dcvb06xy17Q/s576/IMG_1088.JPG (http://picasaweb.google.com/lh/viewPhoto?uname=ronanian&aid=5192607384817217601&iid=5208559904911814146)

In my truck (look at the red wire):
http://lh6.ggpht.com/ronanian/SDqyo9SAhoI/AAAAAAAACSI/jk_RiHXg9SY/s576/IMG_0932.JPG (http://picasaweb.google.com/lh/viewPhoto?uname=ronanian&aid=5199448830687159425&iid=5204668736094635650)

The red wire where it meets the injector:
http://lh6.ggpht.com/ronanian/SEiE4CzhyVI/AAAAAAAACVE/G2yRl_gY1MQ/s576/IMG_0929.JPG (http://picasaweb.google.com/lh/viewPhoto?uname=ronanian&aid=5199448830687159425&iid=5208559067414514002)

Close-up of the injector wire:
http://lh5.ggpht.com/ronanian/SEiFHC3Ak7I/AAAAAAAACVY/jzFZ4v4-kNo/s576/IMG_0930.JPG (http://picasaweb.google.com/lh/viewPhoto?uname=ronanian&aid=5199448830687159425&iid=5208559325127152562)

Where the wire goes in to my dash. It comes out at the steering wheel, though I could also have it come out at the center console.
http://lh5.ggpht.com/ronanian/SEiFSAcEp0I/AAAAAAAACVs/de7QyBMLtEc/s576/IMG_0933.JPG (http://picasaweb.google.com/lh/viewPhoto?uname=ronanian&aid=5199448830687159425&iid=5208559513455863618)

[monroe74]

Oops, I didn't notice this until just now. Cow, you did a really nice job of putting a lot of good info in one place. Thanks for doing that, and thanks for mentioning me! I was just sharing bits and pieces I've learned in various forums like this one.

You're coming up with lots of interesting observations, and I look forward to hearing more about what you're discovering.

[rlkwnc01]

Great idea. Any thoughts on using this set-up with a throttle body injector?
('92 Tracker) I haven't investigated the system thoroghly yet. I have the same meter. Just thought I'd ask.
Thanks

[theholycow]

Either monroe74 or fumesucker could probably answer that question better than I can. It ought to work with TBI fine, I think. Since you already have the meter, all you have to invest is a few minutes of your time...I'd love to hear if it works for you.

[monroe74]

Thanks for the compliment, but I don't think I know any better than you. I figure it's worth a try.

[rlkwnc01]

I assume I need to try to find a signal for an individual cylinder, just as you did for the injector? It has a Camshaft Position Sensor which pulses and breaks for each cylinder. I guess I'll just need to play with it.
Thanks for the help.

[theholycow]

You definitely need to hook it up to the injector wire. I'm pretty sure a TBI only has one injector, right? The point is that it reports duty cycle to you -- it reports time on vs. time off in percentage form. You will get meaningful numbers even though it's reporting for all cylinders vs. one cylinder.

[rlkwnc01]

Thanks for the feedback. To your point, it's just a number for comparison. I just need to get into it. I'll report back but it may take alittle time till I can. Too bad real life gets in the way of real stuff.

[rlkwnc01]

Just wanted to let you all know I did find the injector wires with alittle probing. The Haynes manual made no direct mention of it. On 4 cylinder scale at high idle it read 4.5 and at warm idle 1.5. Looks to be in the ballpark. I'll install a permanent in car set-up when I return from travel this week.
I've got several experiments lined-up to try including vapor/steam injection and hho. I'll keep you informed.
I think this will be a great evaluation tool. THANKS!!

[monroe74]

On 4 cylinder scale at high idle it read 4.5 and at warm idle 1.5. Looks to be in the ballpark.

Yes, that sounds about right.

I'll install a permanent in car set-up when I return from travel this week.

In my own situation, I found it was very helpful to use a temporary setup for a couple of weeks. I feel like that taught me a lot about how my engine was operating. Then I felt there was no need for a permanent installation. Although I may end up with some kind of permanent device later on, as some point.

I'm just suggesting you live with a temporary installation for a while before you make a big commitment to something more elaborate. The temporary installation will help you formulate what your real long-term needs are.

[sonyhome]

Hollycow,

So you wanna polute your thread with what we discussed in the other thread, or make a new one?

To recap what we talked about:

- Collect all that on a PC via sound card
- The injector tap provides
* Duty cycle = gas consumed
* RPMs
- A VSS tap woul provide
* Car Speed MPH
- Computer provides
* time

With all of the above we can make a program that detects

* MPG
* MPH
* GPH
* Which Gear is in use
* Gliding/Neutral

I saw another thread on collecting data with a PC sound card, which might not be fumesucker's (If I find it again I'll update the credits). That one thread had audio samples and a java applet to parse the audio.

Since I've decided I need to upgrade my skills from C to Java, that would be a good mini project for me.
I have some ideas on how to present the data to help analyze hypermiling techniques efficiency.

[theholycow]

So you wanna polute your thread with what we discussed in the other thread, or make a new one?

Might as well pollute here, I can breathe that crap. ;) I brought in your message from the other thread, too.

I'm really slow at projects like that.
- This WE I stayed home, and got around to fix my CRV speakers
- ...and cut foam grill blockers for the CRV. Dunno if the foam worked, but got 30MPG @65MPH one way, and 26MPG @74MPH the other way (Short distances, flatish), without overheating apparently. Still warry to use that on long trips.

I'm slow too. I didn't do anywhere near as much as you did this weekend. :)

I'm going to rig up a light that tells me when my cooling fan is on so I can see if it runs more after blocking the grille. If so, maybe it's worth blocking a little less, so the energy to run the fan doesn't have to be used.

[theholycow]

In my own situation, I found it was very helpful to use a temporary setup for a couple of weeks. I feel like that taught me a lot about how my engine was operating. Then I felt there was no need for a permanent installation.

I too found that a temporary setup is good enough. A couple weeks usage and I don't need it again until I try some modificatons, though I use it sometimes anyway. I wouldn't mind permanently having a small gauge, maybe an LED bar graph or something.

[suspendedhatch]

Wouldn't you need to take RPM into account? And fuel pressure?

[theholycow]

You definitely don't need to worry about RPM. You're measuring fuel per time (GPH), not fuel per revolution when you measure fuel injector duty cycle. If the fuel injector is open for 30 seconds out of every minute, that's 50% duty cycle, and it's feeding the same amount of fuel whether you're at 500 or 5000 rpm.

It's my understanding that in most systems, fuel rail pressure is the same under all conditions. If not then it could explain the observations I've made about WOT in my VW better than my guesses.

[monroe74]

I basically agree with what you said, but I want to add some details.

You definitely don't need to worry about RPM. You're measuring fuel per time (GPH), not fuel per revolution when you measure fuel injector duty cycle. If the fuel injector is open for 30 seconds out of every minute, that's 50% duty cycle, and it's feeding the same amount of fuel whether you're at 500 or 5000 rpm.

There's another subtle issue, which has to do with injector latency. That means the injector doesn't open the exact instant the ECU asks it to. This is on account of inertia (magnetic and mechanical). Injector latency is the reason that injector specs cite a dynamic flow rate as well as a static flow rate. The latter means what would happen if the injector were open permanently. The former means what happens in the real world, with the injector opening and closing all the time, and therefore injector latency has an effect on the flow.

This ends up also having a relationship with RPM. Because of injector latency, a series of short pulses will emit less gasoline, as compared with one long pulse, even if the overall pulse width adds up to the same total. Theoretically what you really want to do is subtract a small amount of time from each separate pulse. And this is something the SuperMID does, for example.

Having said all that, I think it's probably a minor factor, in the context of what's being discussed here. But it might be good to realize that a 10% (let's say) duty cycle in top gear probably represents a little more gas than the same duty cycle in a lower gear. Because in top gear, that duty cycle is composed of a relatively small number of relatively long pulses. Therefore injector latency has less of an effect. Injector latency has more of an effect when each pulse is relatively short.

It's my understanding that in most systems, fuel rail pressure is the same under all conditions.

I just want to add another clarification to this. My understanding is that most or all EFI systems use a fuel pressure regulator, to hold fuel pressure constant. But it might be helpful to understand that fuel pressure is held constant relative to manifold pressure, not relative to atmospheric pressure. This simply means that the pressure differential across the injector is held constant. The result is that a given pulse duration will always pass a constant amount of fuel. (This is also a bit of an oversimplification, since fuel density changes relative to temperature. But we can probably assume that engine operating temperature is relatively constant, which means we can ignore this factor.) And that's exactly what's wanted, both from the perspective of an engine designer/tuner, and from the perspective of someone who wants to measure fuel flow.

[sonyhome]

Hey,

Toss the duty cycle meter!

I found the thread about the PC-based MPG gauge. It was posted by skewbe... With Java applet and DIY injector connector.

http://www.gassavers.org/showthread.php?p=64732#post6473232

http://www.gassavers.org/showthread.php?t=4250

[theholycow]

Thanks! Interesting stuff there.

The first thread is essentially the same thing that fumesucker did in this thread (http://www.gassavers.org/showthread.php?t=7608), which I link in the first paragraph of my original post. It's a great idea if you want to log data and review it after (and maybe graph it against OBDII data you've collected at the same time).

The second one starts that way, then turns into using a DIY duty cycle meter.

Both also talk about figuring in the VSS so you can have a MPG gauge, instead of just a fuel rate (GPH) gauge. Neither one of them provides a $33, five minute, large font display on your dashboard like hanging the dwell meter does.

I think I'll add them to the original post.

[mkiitmr2]

Could this be used as lower profile duty cycle monitor?

http://www.skycraftsurplus.com/index.asp?PageAction=VIEWPROD&ProdID=1001

Very cheap, but looks like it maxes out at 50% duty cycle. But would that even be an issue? I may just pick one up and throw a ground and injector wire to it to see what I get from it. Im just looking for something lower profile that looks more integrated with my cars interior vs mounting a large DVM :)

[theholycow]

That's cheaper and looks smaller than the meter I used. The display is described as "5mm" (which I assume is it's height), and that's tiny for this purpose, but then it's less than half the best sale price of the Harbor Freight dwell meter. I don't know much about duty cycle meters but it looks to me like it ought to do the job.

The size of the whole unit means it won't take up the whole dash like mine does; mine hangs over most of my stereo in the truck, and in the car it covers a whole vent and part of the stereo.

I've only ever gotten to 50% under WOT near redline.

[mkiitmr2]

I've only ever gotten to 50% under WOT near redline.

Nice to know. I figured as much. My only concern is that my cars alternator keeps the car at 16.2volts while it is running which is out of the range for this duty cycle meter. Is that how much is actually going to my injectors or is the voltage going to my injectors limited so that it is kept constant?

[theholycow]

16.2 volts is awfully high. Are you running a race car DC-DC converter like in this thread (http://www.gassavers.org/showthread.php?t=9256)?

Anyway, that's a good question. It might be difficult to measure with a volt meter because of the low duty cycle. My dwell meter reads 100% with the key turned to "On" and the engine not running; perhaps it would be possible to get a good voltage reading in that condition.

[mkiitmr2]

I have an Toyota MR2 Turbo. That is considered normal. I think 16.5volts is the highest acceptable good range for my car. So Im in the normal range. I was just looking at the wiring diagram for my car and I remembered it has an "injector resistor pack". I could measure the resistance of the pack and then calculate the voltage drop with input voltage of 16.2 to see what it puts out. But Im pretty sure that will put me at 15volts or less which is good enough for that meter based on its absolute specs. I may just try what you said also with key in on position. My injectors are grounded by the ECU and have constant power with key in on or run. Is that good? Do I tap into the power side or ground side in my case?

ERIK

[theholycow]

I'd try to add a resistor and lower the voltage going to the meter to help protect it, since the spec is 12v nominal max, 15v abs.

"constant power with key in on or run" sounds like the same behavior as my 100% duty cycle in that condition. You should be able to measure voltage like that.

You'll still tap the power side of a single injector, I'm pretty sure.