Ok, quick question or 2...........

If you get VE in perfect harmony and you show 128 BLM's across the boards all the way till you get into PE, does this mean that the 128 BLM is actually 14.7 afr across the boards when in VE ?

And if so, then you can tune PE to hit 12:1 or so from point of PE threshold to redline? Would that be the way to go about it if tuning for closed loop?

Or am I asking the wrong questions?

 

Ok, quick question or 2...........

If you get VE in perfect harmony and you show 128 BLM's across the boards all the way till you get into PE, does this mean that the 128 BLM is actually 14.7 afr across the boards when in VE ?

And if so, then you can tune PE to hit 12:1 or so from point of PE threshold to redline? Would that be the way to go about it if tuning for closed loop?

Or am I asking the wrong questions?


Anyone ???

 

Well in Theory if you can get Upper KPA range of VE to 14.7:1 or 128 then from there you can use the PE F/A adder to obtain 12/13.x whatever. The problem is tuning under such a load @ 14.7 is a little risky.
I prefer not to use the PE adder, I just use VE for most WOT fueling. If you did get upper VE to 14.7 in WOT areas then you could just add 10% or what you prefer to obtain your 12/13.x :1
to those portions of VE as well.

 

Gees folks...sorry if I sounded like a jerk Here's the deal.... I am willing to share the extensive data I took on some injectors if I can get bruce to agree to share his extensive data as well. This would include the basic injector flow curve at different operating voltages, and data showing the effects of injector duty cycle on flow. Not to mention a brief description of the setup it takes to take the data, and the problems with calibration. Heck, I'll even go through some math equations that characterize the injector. I think it would be a nice compliment if bruce can share his data, setup, and calibration along with mine. How about it?

For the moderators, if this is an inappropriate place to do this or maybe you think I suck and don't want me to post, then let me know. Thanks for listening!

Dave

 

Dave, I have a suggestion. Why dont you and bruce talk via email. It seems that you two are cut from the same cloth. Besides, I know everyone would love to hear what you have to say. I think that you came across as somebody that cares about what you have learned. It didn't seem rude of you at all. Why should you share stuff IF the board is not going to "welcome" the
data? It seem that the mods are not upset with you. I think they are on the edge of their seats waiting for you to drop some info. It use to seem(to me) that guys like grumpy was rude or something. But bruce is a really nice guy. I know that you are cool. So why not get together with bruce so we can learn something new

 

I have questions about how the Pe a/f ratio wot vs rpm table functions. Cause on the dyno that table doesn't make the cars respond for me. I always have to change the upper ve to get a response with the wideband. That pe adder just didn't show any response. Maybe it was because the upper ve table was off. OR could the pe adder only apply when the blms are above 128? That might sound crazy, just a shot in the dark. Cause I can't understand why the 730 (any bin) has a stock pe f/a vs rpm table that looks like that. I was told that the f/a vs rpm is the % change from 14.7:1 to net aprox 12.5:1 for wot. Well the stock chip has f/a vs rpm entries that are negitive % from the ve tables. I have just always needed to hit the upper ve tables for Wot tuning, the f/a vs rpm doesn't cut it....

 

keep in mind that (at least for $8d), the PE percent modifier comes from at least two tables:
P.E. % change to AFR vs coolant temp
AND
P.E. % change to AFR vs RPM

in AUJP, the vs.CTS table is 22.7% for all the "normal operating temp" values (> 56C).

The vs.RPM table does indeed fluctuate from -5.5 at 2000rpm to 7.8 at 4000rpm and back to -2.3 at 4800 and higher. but when you factor in the CTS table, you can see that you are still getting an enrichening of the AFR because the magnitude of the values in the vs.CTS table are much greater than those in the vs.RPM table.

As a side note, in order to tune the accuracy of the 100kpa VE cell spots, I set up both of the PE tables to give me a flat AFR across the entire RPM band of about 12:1, and then tuned the VE cells using a WB to make sure I hit 12:1 at WOT at 100kpa across the appropriate RPM range. If I know I'm supposed to be 12:1 based on the PE etc., then I know that any discrepency from 12:1 must be the VE. After I get the VE pinned down, I can then manipulate the PE tables as I see fit, knowing that if I want to increase or decrease the PE AFR by some arbitrary value, I can get quite close if not dead on, because I know my VE table is right.

So, I've had success using the PE tables to control AFR, they seem to work as expected to me. You can use a program called AFRtuner (thanks tim/traxion!) to make it easier to manipulate these two tables in a logical manner.

 

I'm not exactly sure how this would work but let's just be hypathetical for a minute and say your car breaks down. Nothing serious, just a sensor failed and defaults to something relatively close (as is usually does) but still it's not a pretty AFR. What if you were in closed loop when this happened. Any chance that closed loop would save you some damage? This is the only reason I've kept closed loop. I've had some flukes happen on long road trips that make me wonder what if.

 

11s....I will start posting some stuff either tonight or tomorrow. I plan on scanning one of the digital drive boards and posting a picture of it to get a feel of how this BB works. I will also give some explaination of how it works and why I chose to go this route.

I'm sure bruce will be chiming in pretty soon with some of his setup. For some strange reason, I picture bruce walking over to a very large filing cabinet, rumaging through it, and pulling out this gigantic folder with his vast notes and test data. This is probably gonna take him some time for him to do, so I'll just start plugging away with what I've got.

And as always, feedback is welcomed or even desired. If I get too technical or not technical enough, or maybe someone feels they need to use this.... then speak up please!

Dave

 

NICE :hail:

 

Here are three pictures and a brief decsription
of the setup needed to properly drive a
high Z injector. Driver.jpg is a digital one-shot,
a load compensated voltage supply, and a transistor
injector driver. This board is controlled by a laptop
to produce the pulses that drives the injector. Control1.jpg
is some electronics that controls two solenoids for fluid
pressure control, and some circuits to read a pressure
transducer. Another laptop was used with this board to
control fluid pressure. Dead time.jpg are two boards I had to
add in order to do injector duty cycle tests. One board
is used to set the off time of the injector, the other board
is used to set a precise length of injector operating time. Here
is a link to the pressure vessel I used to supply the fluid.

http://www.harborfreight.com/cpi/cta...emnumber=34202

This vessel was modified to accept the pressure transducer, and the two
solenoids. One solenoid lets air into the vessel, the other lets the
air out.

Other items used for the setup are a digital scale, graduated cylinder,
and a radio/cd player.

Picture quality is probably not to good because of the 100K limit. Better
pictures are avaiable upon request.

More to come later.

 

control.jpg

 

dead time.jpg

 

Here's a little explaination on why I had to build a digital drive
board instead of two other methods and it's gonna be brief. My first
choice to drive an injector
was to use a simple one-shot to generate the pulse. This kind of works
but was very touchy to adjust precisely due to having a pot control
the PW. Scratch one method. The other way was to use my ECM bench to
control the PW. This also proved to be very touchy because of the pots
used to control the operation. Bad idea...scratch method #2. I decided that
a dedicated digital control board was the hot ticket.....and it was. With
this dedicated driver, I was able to select any PW from .001 ms to 99.999 ms
with +-.5 microsecond resolution just by turning a few ***** (actually
switches). The same can be said for the two extra
control boards I had to build in order to do the duty cycle tests. I could prcisely
set the duty cycle any where from .01% to 99.99% just by setting some more switches.
The duty cycle test results
are very interesting to say the least. Fuel pressure was controlled by a laptop by
reading a calibrated pressure
transducer and then operating two solenoids. The pressure was held to X psi +-.02
psi. The voltage supply used to drive the injectors is load compensated and was
adjusted to X volts +-.02 volts. The digital scale was checked for accuracy over
it's operating range by employing calibrated weights of different sizes and
observing the results. The scale can measure up to 500 grams with .01 gram accuracy.
A calibrated burret was used to establish the weight of the working fluid, in this
case mineral spirits was used.

Now, all of this is presented to you so as to prove that I actually performed the
injector testing myself and with a high degree of accuracy and that I didn't get
this info off of somebody's web page, or from some crappy automotive magazine.

Next up: the graphs. But I'll wait until bruce checks in with his setup. I hope
he doesn't embaress me to much with his setup. Knowing bruce, he probably used
a CRAY computer to control his injectors.

Dave

 

OK, Dave, <raises hand in back of room> I have a Q if you don't mind!

Whilst waiting for your fellow scientists to collect their notes, there is a test-construction point I'm curious about:

On the old DIY-EFI site there was a design for an injector tester - it relied on a PC running a DOS(?) timing program for its pulse train, IIRC. Without diging up the article, I think it used the parallel(?) port output to drive a big Darlington, that then fired the injector (from a car battery). Fluid flow was measured in a graduated cyl.

I always meant to build that (one of these days) but now I wonder - what is the downside to a test design like that, compared to the one you created? Other than the obvious lack of a variable V+ supply? I'm just curious - if it turns out that the old experimental design was seriously flawed, I'll cross it off my list of rainy-day projects (that I may never get around to building anyway). No sense running inaccurate tests... even for a hobby. I'm mostly interested in the pw timing here; fluid measurement is whole 'nother topic.

Thanks for any insight you can lend.

 

I have that article and just re-read it so I could give you an answer. I take it that you have the articles in question? The article is well written but there are some things done and said that personally I would not do or say, at least not without further explaination. There are enough things that are not done quite right that it makes me wonder if the author actually built the thing. For instance, if you look at the injector drive electronics in the article, and how the transistors are driven, then look at the spec for the transistors, it's entirely possible they won't even turn on. And if they do turn on, you'll probably end up shorting out the transistor. Also, there is one glaring omission...there is no data presented! That omission is one great clue that it was not built. But to answer your question about how good the timing is, I really don't know because I don't have the software that was used. I chose to do it in hardware because I know that it will work any time, all the time, just all by itself, and don't have to worry about crappy printer ports, or a finiky laptop.

Dave

 

Dave, thanks for the information.

Sorry to stray so far off-topic, but since you have built such an apparatus I thought I'd pick your brain for a minute.

I have the article, but hadn't looked at it in a while. You raise a good point about whether it was actually run for data collection. I think I have DeArmond's e-mail address around here somewhere - I'll send him a note and ask.

I also have a file with the software in it - be glad to send you a copy if you want. I haven't tried to run it, but now you have me curious.

Re the output device, let me make sure I understand your concern. Assuming the output from an old desktop PC is a buffered 5v signal, you think it will need another stage of current gain to turn on the TIP-120? Or that the grounds between the PC output and the UUT need to be isolated? Here is a datasheet I dredged up from Digikey:

 

Not to sound out of line,
This thread has gotten way off topic...Dave you have some great info but injector flow rates had little to do with BLM correction while in PE mode or usage of BLM cell matrix.
I think it would be helpful to many to start a new thread with what you have put in here..
Especially for search purposes.

 

One problem is that I have never seen a printer port actually put out 5v. It's usually much less. Also, darlington transistors are notorius for having much higher turn-on voltage than just a regular transistor. This particular one has a max Vebo of 5v. So it starts getting pretty iffy. 87_TA does make a valid point here about being off topic, but what is off topic is the diy flowbench. If you want to continue, contact me off list. But what does an injector flow curve have to do with BLM's....everything! Something to ponder....what is the main reason for BLm's? Why is the matrix for BLM's dedicated to such low rpm/low load? Heck, my ECM can access most if not all BLM cells just backing down the driveway! If you can answer these questions without seeing an injector flow cure, then you are a smarter person than I, and I will leave with my head hung low. I didn't know the answer either until I produced the curve. I'm tryng to help, but I'm also trying to make you people think. And of course bruce popped up out of nowhere looking for a fight, so I took up the challange and if he doesn't produce the data, you can draw your own conclusions. We are all waiting for him to produce his vast personal data on fuel injectors. I've got mine ready to go. It's starting to look like there is no interest in this thread anyway, so it just may die here, right now. As a closing note, the injector flow curve was the most enlighting thing I've seen to help me understand and appreciate ECM operation. I could go on and on but this is it! Sorry if I ticked anyone off

Dave

 

there is interest in your post. We should start a fresh thread though. I wouldn't say because it's off topic but it's worthy of a dedicated thread. It seems that grumpy has been too busy to ring in. Like I said, you shoud extend your hand via email. You'll find that he is a really cool guy.....

Tim

 

11s..email me privately.

Dave

 

I agree in making a new post for this, I would be interested. I'll send you an email Dave.

 

Just for the record, a search on 'BLM Cells' returns 158 threads in this section
BEFORE this one. Whoever started this topic apparently didn't find the answer
he wanted in the other 158 threads, so how do we know he won't find it in
here???

Dave's major point seems to be this:

How injectors physically modulate fuel flow at low PW's has much
to do with the why and wherefor of GM's BLM code. It's an
interesting thought, which I don't think I've seen articulated before.
(The thought wouldn't exist without flow bench testing. Therefore,
testing is related to BLMs.)

With that out of the way:
Considering the implications of Dave's thought, it could in fact give insight into
why the BLM cells are arranged as they are, and what they are there to do.

A quick google yields a few sites that speak to injector flow as it relates to PW
and injector electrical properties, particularly inductance. Of the ones that aren't
engineering research documents-

There is a Wells site here, that yielded the illustration below (which I
hypothesize may be part of Dave's epiphany):

http://www.wellsmfgcorp.com/counterp...v5_i1_2001.pdf (small chart of Flow vs Amps)

A diagram and some typical specs of a Bosch injector are here:

http://frwilk.com/944dme/injector.htm (spec & diagram of Bosch inj)

And I discovered that Bowling & Grippo have a design to cover the same kind of testing
gear (though they too fell victim to using the dreaded TIP-120 Darlington):

http://www.bgsoflex.com/FI_tester_doc.pdf (B&G Injector Tester)


As for how low PW flow data relates directly to BLMs in an $8D mask-

I'm not sure that having graphs of Buick injector flows will mean much to my L98 SB.
If someone posts measured flow vs. PW data for 24 - 30 pound/hr SVO saturated injectors
@ 3 bar fuel pressure, it would indeed be useful, and worth another thread. Lacking that
specific data, I think I'm free to infer (in a general sense) the gist of what Dave means
via an illustration such as the one here:

That's my take on Dave's idea, anyway. YMMV.

 

Interesting stuff. Can you help me out and explain about the max Vebo and what it means?

Thanks,

RBob.

 

The information in that Wells article is exactly what driveability techs have been going to for testing fuel injectors vs. a static resistance test. If you hook an injector up to a lab scope with a low amp clamp, you can actually see in the pattern where the pintle moves. There is actually a "blip" in the current ramp when the pintle moves. An injector without much of a "pintle hump" in it's pattern is not moving well. If you pressure clean them, or replace them, they can be re-tested and the pintle hump will be more pronounced.

I suppose using this info gives an idea of the lag time between the voltage being turned on, the the injector actually opening?

 

Over 24 hours, Mr Jones, are you there? I could really use an explanation on this parameter. . .

Thanks again,

RBob.

 

Thank you Mr Jones for not responding...this is waaaay off topic. I'm sure a google search will turn up something. Please you it.

Nick

 

I think we can all agree that GN jones is on a higher level than 99.9% of the members here. So I would suggest having patience. He might have a demanding job and little time to relpy. Last thing this list needs is to scare him off. I have much respect for him . Just a suggestion.