Jeep with old Howell kit - it's WAY rich at lower RPM...

[DanInTX]

Hi All - First real post here - asking for suggestions, advice, guidance, support, help, etc. for tuning up my Jeep. I've included as much information as I can think of.

Background in a nutshell:

The Jeep is a 1990 Wrangler with 4.2L straight six with three speed automatic. I put a Howell TBI kit (emissions version; 1227747 ECM) on it in 1990. Starting and driving improved tremendously but some problems persist. Most noticeable: the thing STINKS terribly, especially at idle, like it is way too rich. Upon startup a lot of white smoke comes out the exhaust. It isn't oil and it isn't coolant. I used to think it was condensation in the exhaust, but it lasts too long. I now get the impression that it is "smog" from unburned fuel. Recently I immersed myself in learning about TBI, and I bought a Moates AutoProm setup with TunerPro RT. Datalogging revealed low BLM (108) at all RPMs at lower loads, although they get close to normal (a few points on either side of 128) from 50 to 80 MAP (and go low again at 90).

Shortly after initial installation I contacted Howell and got a new chip. It didn't make a difference that I could tell. With my new AutoProm I read both chips - initial and replacement - and found that the new chip had somewhat lower values in the VE1 table, and the BPW was reduced from 185 to 181.

Interestingly, I found that the throttle body is a small bore one from a 2.8 or 3.whatever - I acquired one from a real 4.3 and the bores / butterflies are bigger. I imagine it has real 4.3 injectors, though, since the BPW was set at 185, which seems right for 45# injectors for a 4.3L.

Other relevant information: I have an unrestricted cold air intake connected to the small throttle body and stock, low flow Jeep intake and head. I have a 2.25" exhaust with a high flow (clean) catalytic converter, and a header. The engine has been rebuilt (overbore is 0.060"); compression is around 8.8-8.9; camshaft is a Comp Cams design intended for lower RPM use; and the distributor is an HEI conversion - originally a DUI, but more recently a Summit Racing version (not controlled by the computer).

So...


I have a theory about my RICH problem. I am guessing the small throttle body represents some restriction, lowering my VE, especially considering the improved exhaust flow and (partly) the camshaft. This isn't the only problem, but I guess it's part of it.

To fix this thing, I have a few different directions to go. Ultimately, I will put on the 4.3 throttle body and tune it from there. More immediately, however, I need to do a little bit to the existing setup. My main questions at this time are:

(1) Should I further lower the BPW a little bit? This sounds like a band-aid fix, but I see that it's done sometimes.

(2) Should I adjust the VE1 table based on my BLMs? I am guessing this won't be a one-time deal, since my minimum BLM is 108 and I have no idea how low it would really go.

(3) Would I see a benefit in lowering​ my fuel pressure?

(4) What chip and buffer addresses do I use when I write the 27SF512 chip? (I installed the G2 adapater in my computer.)


A few points to make:
- Fuel pressure is 12.0 and super steady at all RPMs; return pressure is zero.
- Vacuum at idle is about 17 inches and very steady.
- Timing is 8 degrees BTDC at idle, and the mechanical curve (NOT computer controlled) is moderately aggressive; vacuum advance is set to the point just before detonation.
- The spark plugs are black.
- The EGR is vacuum controlled and seems to work right.
- The sensor readings (MAP, O2, TPS, Coolant) and voltages make sense and move the right direction when viewed on my scan tool.
- The Jeep actually runs pretty smoothly - it just stinks bad, drinks gas, and has less power than it should.
- TunerPro RT displays my BPW as 34.9 and calculates my injector duty cycle to be ~95% at 800 RPM, all the way up to 314% at 2800. I have no idea where these number come from or what they indicate.


Thanks VERY much for any help anyone may offer. I will be monitoring this thread closely and updating it when I make progress.

Dan.

[Six_Shooter]

1) I would not, since you say the BLMs look pretty good at mid throttle, I'd leave BPW alone unless you're running out of VE. Keep in mind that the VE1 and VE2 tables get added to together (in the '7747) and if the total is above 100% (99.6 actual) then the VE numbers get truncated to 100%. Going on to #2...

2) I would start with the VE1 table in the trouble areas and get that dialed in or get it closer anyway. 108 might be the lowest that the BLM limit is set to. I usually tune for AFR using a WBO in open loop, but if you don't have a WBO2 sensor, you can tune by BLM in closed loop, but I find it takes longer since the BLM updates slowly.

3) No, that will cause you other issues and would not recommend using that type of fuel delivery "tuning".

4) Reference this: https://www.moates.net/documentation...entation_id=75

I will also recommend adding spark control into the mix. YOu can gain a lot of drivability power and economy by having the ECM control spark as well as fuel.

[DanInTX]

Thanks Six_Shooter!

I like what you said - those are ideas I can understand. Weirdly I found the chip and buffer addresses right after I made that post, despite searching for them for days - sometimes the right thread just doesn't show up when you want it.

I hadn't planned to do this right now, but at this moment I am changing over to the 4.3 throttle body (which I had gone through a few months ago). I should have it running again shortly, then I can let it tune itself a little, and I can datalog a little more. I've never won the lottery before, so I'm sure I won't be lucky enough to have my problems go away with this simple swap, but by doing it now I can concentrate on getting the right combination dialed in.

Although I have studied this stuff for months, I still find each new step to be a challenge. Reading the chip was harder than I thought (had the wrong addresses), datalogging was a chore since I had never plugged the AutoProm in, even installing the G2 was a chore since I had never worked on a computer board. I figure things like getting the emulator to work, modifying the bin, burning the chip, and finally making the thing run right will all have plenty of surprises built in. But I'm deep in the middle of it now.

Back to the garage to finish the throttle body. I'll keep posting updates as things happen - hopefully my experience can add a little bit to all the wonderful information on this site. More later!

Dan.

[DanInTX]

OK - some progress has been made.

I put the real 4.3 throttle body on yesterday in place of the small one, and it actually improved my BLMs noticeably. While a few cells at low Map and low RPM are still 108 or 109, most of my cells at lower MAP have improved significantly - a lot of them to near normal, and some even flopped over into the lean zone. This makes me believe that the weeny little throttle bores that I had with the original Howell-supplied 2 or 3.something TB (what size would it be?) constituted a measurable restriction which contributed to poor VE and subsequently caused the lower MAP areas to run rich. There is still a lot of adjustment that needs to be made, but I think it can be salvaged.

However, while datalogging last night after the swap, TunerPro RT gave me some continued disturbing news:
(1) The whole run occurred in 'asynchronous' mode, and
(2) My displayed BPW was 34.9 and my [presumably calculated] injector duty cycle ran from ~80% at idle to over 300% around 3000 RPM.


Now, I have a couple theories about this. I know my BPW (or really BPC, I suppose) is set at 185 on my chip, so I am guessing that the displayed BPW is a calculated number by TunerPro, and MAY be due to injectors that are too small. I know that the original TB was the small one, but I DON'T know if the injectors that came in it were the little 33# ones for a smaller engine, or if they had been swapped to the 45# ones from a 4.3. There are no markings (numbers or stripes) that tell me. My theory is that I have smaller ones, and the ECM has to overdrive them with highly elevated duty cycles in order to supply the correct amount of fuel. Does this sound logical? I DO have a set of new NAPA injectors for a 4.3 - if this sounds like it makes sense, I will swap those in. Feedback here would be most appreciated!

My other theory is that, due to my weird BPW (34.9) and my illogically high injector duty cycles, asynchronous mode was triggered to run about 99% of my on-time. Now, I am only vaguely familiar with 'asynch' mode, so I did a lot of reading about it last night on this forum. I gather that it is used only for certain operating modes (maybe PE? WOT?) and is not intended to be used for regular operation. It seems as though it gets triggered when things aren't going right, say, like when your injector duty cycle is too high? Or something similar, anyway. I also got the impression that many BINs (like the V8 ones) aren't saddled with this mode, and that mainly the V6 ones got it. It was also stated that async was looked on with disdain in many cases, causing problems getting idle tuning right, being especially contrary (per one guy who tuned some straight-six Jeeps) when trying to get those Jeeps with their very long intake runners to idle properly. My question here is: is there a BIN available for my 1227747 that doesn't use async mode? I saw one post in which the person mentioned using a non-async V8 BIN by changing it to run a six cylinder (I guess by just changing the 'cylinder' constant to '6'?). Now, solving problem one may solve problem number two, but any information (or pointers to appropriate threads) would be greatly appreciated as well.

Please feel free to chime in and offer your experience, especially anyone who used the Howell kit with the small TB and actually got it to run right.

Dan.

[DanInTX]

Another comment:

Six_Shooter recommended enabling electronic spark control on my Jeep engine. I like this idea and have been studying about it some. I get how to modify the distributor, and I think I know which parts I need to acquire (feel free to refresh my mind). But I wonder which knock sensor to use. It seems they are tuned to detect knock in specific applications. Obviously GM didn't make one for my convenience so that I could put their GM TBI on my AMC/Jeep motor, so I wonder if anyone who has done this has a suggestion about what sensor seems to work best in this application.

While I'm full of questions, here's another one. My EGR is a 'negative pressure' one and it is connected to ported vacuum via a CTO (coolant temperature override) valve. I intend to keep my EGR since it doesn't hurt performance and may be beneficial in some areas. Would it be worthwhile to convert my EGR so that it is controlled by the ECM?

Again, thanks to all who would like to post any advice on my setup!

Dan.

[Six_Shooter]

The old TB would be from a 2.8, or 3.1. Most likely a 2.8 though since the only 3.1 TBI application was '90 to '92 IIRC Pontiac Transport. The 2.8 TBI was used from 1986 to 1994 in the S10 and there were a LOT more S10s produced than the limited Transport.

Ignore the injector duty cycle, IIRC it uses a bit in the datastream that requires a modified bin to be accurate. Basically it was re-purposed from the stock data to sending out injector duty cycle.

BPW is not the same as BPC. IIRC BPW is a calculated value that will not change during engine operation, since it's the base pulse width (hence "BPW") that the rest of the fueling calculations use before getting to the final PW.

As far as Async and sync modes, I've never worried too much about what mode is actually being used, as long as the engine runs well. However, as I understand it Async is only used when pulse width are too short to actually control an injector and goes to a timed period between injector firings instead of in sync with the DRP. It is kinda strange that you never saw this change to sync at any point. However it could be just that the datastream is so slow on the '7747 that this bit didn't get updated in the time you were observing it before going back in to async mode.

As far as KS goes. There's different thoughts on how to select a proper KS for a non-stock application. Most people go by bore size, finding a stock GM application (that is also OBD1 and similar ECM) with a similar bore size. Some go by engine displacement, and some have other theories. You may have to swap this sensor a few times if you want to get real picky about how it picks up knock, you may also need to use some mechanical tricks to desensitize it, again depending on how you want to tune the KS.

Either way you control the EGR will likely be fine, however, since the ECM has the ability to control the EGR, I would likely have the ECM control it. :)

[DanInTX]

Hi Again:

I have been doing some tuning by changing my VE1 table based on BLM history from datalogging with TunerPro RT. The changes have made more of a difference than I expected, so this has been a pleasing exercise. At the outset my engine was WAY rich at lower RPMs and especially at idle, but strangely got lean at higher RPMs. It was very smelly while idling, and it was nearly impossible to drive with the windows out or top down. (It's a 1990 Jeep with a 4.2 straight six that has been converted from a carburetor with a Howell kit.) It would produce lots of white smoke at idle that looked like water but never stopped.

After some effort the BLMs have improved at all ranges, and my white smoke has decreased tremendously. It starts better and idles more smoothly. It doesn't seem to have any noticeable power increase, but it seems happier driving, with less stumbling and less roughness.

There is more work to do, mainly at the lower RPMs, but I am running out of room in the VE1 table. Initially I used the formula (BLM/128) X the value in the VE1 cell I wanted to change. This didn't make big enough changes at once - my BLMs were settling at 108 (a fixed low value) and my integrator would be as low as 70 at the same time.

I made myself a spreadsheet and used the following formula:

(BLM/128 X VE1) - (VE2 - (BLM/128 X VE2)

This formula calculates the change percentage (BLM/128) needed for both tables, and subtracts that total amount from the value in the VE1 cell. It seems that doing the percentage change only to the VE1 table doesn't create the complete amount of change needed when considering the fuel used is based on VE1 plus VE2.

The new way brought my BLMs in line faster but has created a problem, which leads to my questions: How important is the VE2 table? Are the values there used for other things besides just adding to the VE1 table? I have seen people say to zero that table out and put the whole combination in VE1. I have seen other people say leave a value of, say, 5 or so, in VE2, and put the rest in VE1. Others say leave it alone. This has become important to me because now I have VE1 values at low RPM/high vacuum of 5 to 6, leaving me no room for change. I figure that if I transfer the VE2 amounts (or at least percentages of them) over to VE1, I will have some space for more changes, but I don't know what this will do to other parts of the programming, since I don't know what else VE2 does. Any ideas?

A related but possibly more important question is this: what can I do to depress ALL the values in both of the tables? At the upper end I have some cells at 105 or more. I know that anything over 100 gets capped at 100, so my changes aren't going to be effective. I also know that these values don't resemble reality, because after all it's a Jeep 4.2L. The head isn't even the newer 4.0L version, so there's no way it can get close to 100% volumetric efficiency. So, is there a way to skew these numbers down? Two ideas that come to mind are altering the BPW constant AND/OR adjusting the fuel pressure up a pinch. My fuel pressure is at 12psi, and my current BPW is 185, which seems right based on my engine and stock 4.3 injectors. (A replacement chip I got from Howell years ago actually did decrease that value from 185 to 181 - it helped some but not enough.)

Any guidance or suggestions are very much appreciated!
Thanks again!

Dan.

[DanInTX]

Hi again:

(This seems to be turning into a log of my progress in learning to tune my Jeep)


After some time and lots of effort and studying, I have been successful at improving my Jeep conversion. In the beginning my BLMs were way off: super rich at the low end, and very lean up high. I ended up combing VE1 with VE2 to get one table to deal with. I left a value of '5' in all the RPM levels from 400 up, but left the zero RPM line at '25'. I don't know what leaving values in the table will do (not sure if it's used for anything else) and I left '25' at zero because there was no similar row in the VE1 table. The thing runs smoother now than it ever did, BUT it is still way smelly like it's very rich, especially at idle. Don't quite know what this means, and don't quite know how to fix it.

My next move will be to increase my BPC (wrongly called a BPW in my file) a few points. I'm doing this because, after getting my BLMs all good, my new table has values over 100 near the bottom right (and it mysteriously has some 118's on the lower left side). I gather that if I run the BPC up a pinch, it should shove my VEs down across the board - then I can log/retune and hopefully get my whole table between probably 30 and 95.

However, anyone know why it's still smelly, especially at idle? Since my BLMs are mostly good - and getting better - I would like to think that it would eventually quit smelling so rich. It still stays in 'async' the whole time it's running. Does this matter?


-----


I have a backup plan. I have read in various threads that a bin from a 454 is actually a better place to start tuning a Jeep 4.2L than the 4.3 Chevy engine - something to do with chamber design and things. And it doesn't have an 'async' mode. And it DOES have a spark table included, which my chip doesn't have - I have fuel only - no timing, no EGR, no nothing else. So I have to ask some very specific questions:

(1) I plan to use the 454 bin on this site - the one for an automatic transmission (some things included in the auto file seem like they would improve my Jeep's behavior, like the throttle follower, etc.). HOW EXACTLY do I change the thing to work on my six cylinder - is it just changing the cylinder count constant, or is there more to it?

(2) I need small words and lots of details for this: I have never been able to get my VSS working. My particular Jeep has been modified so that there is no mechanical speedometer provision. I have a transfer case that has a hole for a 2-wire VSS, and it has a 40 tooth reluctor wheel. The one time I think it worked, it caused my engine to quit at over 5mph. Obviousy this is because my ECM wants a lot less than 40 pulses per revolution and it thought I was going a Warp 2. I also tried a Dakota Digital signal converter without success. I never could get the ECM to tell me a speed greater than zero, so I guess it wasn't getting the right signal. Can someone PLEASE tell me exactly what I need to do to make it work? Like how many pulses per driveshaft revolution? (I hear things like 2 pulses per revolution, 4 pulses per revolution, 2000 pulses per mile...) I have a mind to take out my reluctor wheel and machine off most of the teeth, maybe leaving only two or four. Any suggestions?

Once again, thanks for any information you may offer me regarding both my tune AND my new bin and VSS!

Dan.

[1project2many]

The 454 suggestion is from me... after tuning a fuel only Howell setup years ago. You're welcome to try the chip I've posted. The engine I tuned had some incredible torque so it might not be a good match but it is a place to start. All the 7747 calibrations use asynchronous mode fueling, but the 4.3 seems to be "locked" into asynch unlike the V8's. The chip you have should have spark tables. Even if they're all zeros. The reason the 7747 doesn't control timing is because the ecm is not connected to the ignition module. Make sure you connect the ecm and use a GM module if you're connecting spark.

I will not tell you "exactly" what you need to do to make the 454 calibration work with your six. No, not because I'm a jerk. But because it's a nearly impossible task. You do need to change the cylinder count, yes. After that you will need to adjust the volumetric efficiency tables, the power and acceleration enrichment tables, the spark tables, and any other variables necessary to achieve best performance. A job like this is not small. I made 25-30 stabs at making that Howell setup go the way I wanted. These days it would probably take less time but it would still be no small feat.

[DanInTX]

OK. I've made exactly three new "chips" for mine so far. It has helped, but there's still more to do.

I do have tables, but like you said the values are all '0'. I would very much like to let the ECM do the ignition, but it isn't hooked up. Currently I have a large cap HEI conversion with 4-pin module. I pretty much have a handle on the process, but I don't know what a reasonable 4.2 spark table looks like (yet).

I will dig up your 454 file - thanks! - and compare it to what makes mine run the best so far. I do think my best performance will show up only after I get the 'automatic transmission' stuff working (whole new bin? figure out what to do in mine?) AND get the VSS functional. Some things are easy to do, and some things are mysteries. With time I'm sure it'll get done - a little more each night. The VSS has been especially troublesome - by eliminating my access to a cable driven speedometer, I can't use one of those converters - I have to get the real sensor to work right. I believe I'll have to revisit the Dakota Digital thing, maybe a DRAC if the numbers are close, or just cut most of the teeth off my 40-tooth wheel.

Does "nearly impossible task" mean it can't be done on my computer without changing code? Or does it mean "doable if I change half a million different things"?

Thanks again!
Dan.

[Six_Shooter]

"nearly impossible task" meaning telling you what needs to be changed.

From my experience I've always found with the '7747 start with a bin that has the same cylinder count and similar displacement. I know others have had sucess starting with a bin of different cylinder count, but it's not seemed as straight forward as one would have hope.

[1project2many]

What he said ^^^

It is rare for two different engines to work well with the same calibration. The process of revising a calibration over and over to make it work well is trial and error. Knowing exactly what to change before taking readings and driving the vehicle takes some type of magic, psychic ability, or extremely complex and expensive computer simulations. Most of us drive, look at the logs, and make changes based on the results. You will probably have to change a "lot" of things, some more than once.

[newellshk]

Your stinky idle might also be improved significantly once you get the ECM controlling spark. I think that some vacuum/centrifugal advance distributors tend to run less than ideal advance at idle to help with starting, a compromise that doesn't have to be made with electronic spark control. You can also (temporarily) disable closed loop idle, if it's even a thing with that OS and calibration, and lean your idle out to see if that helps. I don't think you can really hurt anything running leaner than stoich at low load and rpm.

[DanInTX]

Thanks to all who are replying to my posts!

I have piddled around a LOT with this stuff. For the most part I have played with my VE tables (I combined most of the stuff into VE1, and left a value of '5' in VE2, so now when I do my modifications according to BLM history, I only change VE1). After much trial and error and logging and reading and burning, I have gotten all of my BLMs to within just a few points of 128. This took changing my BPW (really the BPC - it's just called 'BPW' in $42 - right?) from 181 (from Howell original value) to 195. I went only a few points at a time, but by doing this I got all my VE numbers to be less than 100, although some of them are right up there near 100, and the lowest ones are in the 20's.

I also followed the notes on my definition to get out of asynch mode (most of the time - it's still available for a few certain things).

And I used TunerProRT 's "Smooth" function on my final (or most recent, anyway) fuel settings - my 'happy place' seemed to come at washing three times with a value of 0.8. This REALLY smoothed my graph but left a little valley through the middle, in the same place that the BLMs always put a BIG valley (with really rocky edges everywhere as well).

On top of that, I altered a few trivial things to fit my particular needs, such as:
(1) I added a pinch to my target idle speed - this helped smooth out the shakes that occurred because of my automatic transmission and torque converter setup - at idle the Jeep would vibrate - a tad bit of throttle smoothed it out, so I added it to my tune.
(2) I caused the thing to raise the idle speed 300 RPM with the A/C on. I don't have A/C but I do have a winch. When I activate the power to my winch and controller, I also cause the computer to think the A/C is on and elevate my idle speed.


HOWEVER:

I am NOT having a good time figuring out how to deal with idle. Like I said earlier - it smells really bad at idle. And this means not only at idle, but when driving at low speeds where the throttle spends a lot of time closed. Since I don't have a VSS working yet (working on it - trying to make the "Rostra" pickup and the Dakota Digital SG-5C work together - any suggestions?), I gather that my computer thinks that I am always sitting in my garage playing with the gas pedal - it never knows that I am actually moving, so I never get any benefit of DFCO, throttle follower, and other little goodies.

The suggestion all of you guys have made about using the ECM to control spark is great - I want to do that - but I can't get it done exactly right now. It's coming, but not here yet.

The suggestion was made above to use open loop for idle and adjust the AFR down until it quits smelling so bad. I am wanting to play with that, and I have read where some people use open loop idle, then go back to closed loop just off idle. I have John Wilson's PDF open, and I have perused multiple forums and chats looking for how to do it. Sorry - it just isn't coming together in my brain very fast. Any suggestions on how to make this happen would be GREAT!

My main reason for investigating this avenue: while datalogging, I happened to to glance over at the "dashboard" and noticed a curious thing: when driving at about 30-40 mph, the BLM and INT needles would slam down to minimum levels when I take my foot off the pedal, and they would instantly jump back up to the relatively normal values that they were running at before I took my foot off the gas. It happened every time I tried this. I switched the page over to BLM history and could find no evidence of this occurrence. I guess if I was on the 'immediate' value setting instead of history I would have learned more, but I didn't think of it in time, and it's hard to do that stuff when driving anyway.

That behavior at closed throttle while moving has had me looking closely at AFRs versus RPMs versus temperatures in closed and open loops, but so far nothing has jumped out at me. I think I might be fishing in the right pond, but no luck finding anything to solve the issue yet. I REALLY need to lose the stinky smell which I assume comes from running too rich down there - I use my Jeep as a Jeep a lot, and a lot of trail running occurs at low RPM - it's hard to drive this thing at low speed due to the stench.

My other important reason for fixing this soon (before I can piddle with the timing control) is that I have a Jeep trip at the end of the month - Jeep Jamboree in Bradford, PA. I would like to have it a little more bearable by then.

I am continuing to ponder this situation, but any suggestions on how to make it not smell so bad (i.e. run so rich) at idle / closed throttle / maybe (?) low throttle would be very much appreciated.

Thanks again!
Dan.

(I read over this post and it is VERY wordy - sorry! - but hopefully there is a nugget of information someone can find that will trigger a flood of information to help me!)

[DanInTX]

Update on VSS:

I figured out part of why I can't get my VSS to work: I have a wire labeled "Vehicle Speed Sensor" but it doesn't seem to go anywhere. There is one end that I guess is supposed to hook to the sensor, but the other end sort of gets lost in the loom somewhere. The ECM connector that is supposed to have the pin for A10 is empty in that spot.

Now I read the wiring diagram for my computer - 1227747 - and it says the VSS is supposed to go to A10. Is that right? I pirated a wire from another harness with that little hook thing on the end and I will place that into the A10 hole.

With the wire in the ECM connector, I will attempt to hook up the thing: one wire from the sensor to my Dakota Digital box - this will convert my sensor signal from sine wave (AC) to square wave (DC). I have a choice of signals - a 40 pulse from my transfer case, or a 2 or 4 pulse from my Rostra sensor. The other wire will go to ground - actually to the ground pin on the signal converter.

I am assuming this will work better than it did in the past, now that I have a wire going to the computer.

Dan.