F14CRAZY

Supercharging (this time, it's real)

Recommended Posts

I don't wish to frustrate or [censored] anyone off in my troubles. However, I feel that this is what internet message boards are for.

Why are we here? A turnkey Series II L67 and 4T60e swap by Ryan was going to cost upwards of $6k. We're well off, yes, but we can't afford that.

My L67/LN3 hybrid project, with current expenses and estimates, will run a little over $1k, which is far, far away from $6k. Is it the same as a Series II L67 and 4T60e swap? No. But figure the cost versus benefit.

I discussed the full blown swap with Ryan for a while, but in the end we couldn't afford it. However, with the memcal programming, I'm still using Ryan's services. I understand what long-distance engine diagnostics is like though. But before too long, I'm going to drive the car down to him for finer-tuning. Hell, I'll take him out to dinner and buy him a few beers (or whatever he's into). If you guys come to the Reatta Homecoming in Lansing, I'll buy all you contributers beers too. These messages are going to be of great help to anyone that wishes to do the same to their Reatta. If you held a poll on the most significant complaint of Reatta owners, it would probably be "not enough power", at least by today's standards. This may be the easiest and most affordable solution.

We have nearly 6 pages of photos and information on this project. A project that has never been documented to have been done before. Before I started, nobody could give a definite answer on whether it would all come together or not. Previously, this was unknown territory. This will all be freely and readily available to anyone that wishes to do the same. Hell, when we're done, I'll compile a laminated book at cost for anyone that would like it.

Sometimes, one needs to "screw it and do it" and go ahead and press through with something. Think of all the good ideas that have been thought up by people, but were shot down because others didn't think it would work?

My point: this is real research, experimentation, and innovation. The payoff is soon going to come.

I wish to add that if you guys would like me to try another regulator, I have access to a used '92 Lesabre TPI regulator. I don't know if they're compatible, but those regulators are expensive enough to make you not want to buy them if you don't need them (about $70).

Share this post


Link to post
Share on other sites

I don't think the balancer moving is what is suspected, I think it is the phasing of the windows to the keyway. I don't know why GM would have changed it from the N/A engine, since the operating system is the same, but it is simply an avenue to look into. Fuel or timing could cause the problems you are experiencing, so we are looking for a way to narrow the search. We don't have conventional timing marks, and with a stock system they aren't really needed, but you aren't stock anymore. The cam sensor isn't used for timing, only the reference from the crank sensor. If it is off, nothing downstream will be right.

If the regulator is the one I think it is, it can be taken apart. There is a snap ring on the top. Remove it. I grabbed the top of the vacuum can with a channel lock pliers and rocked it back and forth while pulling straight up. It is sealed internally with 0-rings and it didn't want to come out. Once out, the inside of the can can be viewed for contamination. The fuel enters the side of the can, travels upwards through the screen, which is removable, into the area below the vacuum diaphragm. The operating mechanism is pretty well protected above the screen, but there could be old deposits from stale fuel out of sight. The fuel regulator pintle is out of sight but you can spray carb. cleaner up the central return passage and through the outer passages the screen normally covers.The can where the regulator sits can be cleaned out, pretty much a manual operation. The central passage in the can connects to the return line to the fuel tank. Disconnect the fuel line and spray carb cleaner through that central passage, which should flush out anything that may be blocking it. As for what it should be doing when the vacuum is disconnected is exactly nothing. It should give you a static fuel pressure reading, neither rising or falling when the engine is running at any speed. Personally I don't like the small diameter, construction and long length of that fuel rail, but apparently it works okay. It may cause a small change in fuel pressure just from change in demand, but that should only be momentary and should return to the base reading, that same as what it would be with the key on and engine off.

If the injector is a Bosch, the part number should start with 280, likely followed by 150 followed by another three digits. It might be a Rochester or possibly Motec injector. Any numbers would be helpful.

Share this post


Link to post
Share on other sites

The harmonic balancer idea is an intelligent suggestion.

sc23.jpg

L67 on the left, LN3 on the right. Under examination, the "teeth" are identical between the two. Though an excellent suggestion, the only way this would happen was if the balancer was not properly set in relation to the crankshaft. Since it's keyed, this is unlikely, unless I suppose the key happened to fall out or something. When I was putting it on, I found where the key and the keyway matched up. The pulley only went on a very slight bit. You could still wiggle it off, but I spun it and it didn't slip, and turned the motor. This pretty much proves that the key and keyway are matched with each other.

But hang on...as we can see, the design of the rings is the same, but would a different diameter make a difference? Recall that when I fitted the L67 balancer, I had to swich to an L67 CPS because the rings were not going to fit in with the sensor. I can't remember which had the rings with the greater diameter, and it's difficult to tell in the pix, but there's surely a difference. Could this be throwing the signal off? You can switch the rings, but the LN3 had propriety fasterners, so I gave up and just went with the L67 CPS. Oldstrofeo confirmed that the signals from the different CPS would be the same. I'm not sure if I'm properly visualizing the effects of the different ring diameters, but I don't think it would be an issue, but still something that I think I should bring up.

I'll get those injector numbers tomorrow and clean out the regulator. Thank you 2seater for the instructions. And I'll put my cat back. And I'll see about the spark advance for a steady 2k rpm.

Padgett: if it helps, at 1500-1650 steady rpm, my ED08 spark advance reading was 18 degrees (oxygen sensor disconnected, engine at operating temperature).

Share this post


Link to post
Share on other sites

Guys. If the computer is commanding 4-6 degrees of advance, the MAP is wrong. Period. Has nothing to do with the shutters in the wheel, they just give the zero references - where to start. This is why a lost EST signal just loses the ECM spark adjustment and the engine will run on a "guess" from the IM.

Hokay. I don't care where you are in the spark map you should *never* be below 18-20 degrees advance even at idle and I would expect 30+ (and a lot of plus) above 2000. Is a matter of flame propagation time.

And according to Willie the Barber if the pipes are glowing and you have no sign of extra lean, look for a retarded spark and that is exactly what we are seeing.

Ryan: please chime in here: what is the main spark map of the PROM and can anything other than knock (which does not seem to be happening)explain the under-ten-degrees advance we are seeing at very light load (neutral).

If the balancer (reference) was off we would see the right spark numbers. We isn't.

ps at 1600 rpm/minimum load my spark table is showing 29.3 degrees (BTDC) and if you are tuned for premium gas I would expect 4-5 degrees more.

pps - remember at light load/no boost the spark numbers should look just like an N/A engine.

Share this post


Link to post
Share on other sites

<div class="ubbcode-block"><div class="ubbcode-header">Quote:</div><div class="ubbcode-body">Guys. If the computer is commanding 4-6 degrees of advance, the MAP is wrong. Period. Has nothing to do with the shutters in the wheel, they just give the zero references - where to start. This is why a lost EST signal just loses the ECM spark adjustment and the engine will run on a "guess" from the IM.</div></div>

That is very true. At idle, this engine will want way more than 4-6 degrees.

<div class="ubbcode-block"><div class="ubbcode-header">Quote:</div><div class="ubbcode-body">Hokay. I don't care where you are in the spark map you should *never* be below 18-20 degrees advance even at idle and I would expect 30+ (and a lot of plus) above 2000. Is a matter of flame propagation time.</div></div>

On my S/C engine, I run as low as 9* under high boost (~12psi). At idle, however, it's in the mid-20's (this is looking at my Series1 tables).

<div class="ubbcode-block"><div class="ubbcode-header">Quote:</div><div class="ubbcode-body">If the balancer (reference) was off we would see the right spark numbers. We isn't.</div></div>

This is very true. I would still check the crank sensor phasing with the timing light for a guess, but that is probably not the problem.

<div class="ubbcode-block"><div class="ubbcode-header">Quote:</div><div class="ubbcode-body">pps - remember at light load/no boost the spark numbers should look just like an N/A engine. </div></div>

Just thought I'd throw this in for those who have never seen a spark table. This is viewed in an editor program, not the raw hex data. These ECM's can be programmed either way.

post-41748-143137886948_thumb.jpg

Share this post


Link to post
Share on other sites

Sorry its taking me so long to post replies to this thread, guys. I am really busy in the shop right now so I am getting to this thread as often as I can...

As far as the timing goes, I have set up Philip's main timing table as close to the stock 95 3800 SC timing table as I felt was "safe". Given this, it is very possible to see only 6 deg of timing at lower RPMs...

The stock 94-95 3800 SC timing table:

BHDR_timing_table.jpg

Philip's timing table on the chip I made for him:

philip_timing_table.jpg

This is my first attempt at using an LN3 ECM to control a supercharged engine, so please bear with me. As far as the reluctor wheels are concerned, I don't think that is the problem here. However, the LN3 and L67 balancers do have different counter-balance factors, as the 3800's are externally balanced. While I don't think this is causing any of the tuning problems Philip is experiencing, this is something that must be considered when trying to use the L67 parts on an LN3 engine.

What we need to determine next is if the LN3 computer is traveling into higher load fuel and spark maps too soon in relation to actual engine load. My tuning software does not show me every table that exists in the LN3 and L67 computers, it is very possible there is another table that I cannot see in the L67 programming that has an impact on engine load based on supercharger logic. Obviously the LN3 computer won't have this table. If this is the case, I can rescale the MAF tables to work around this problem.

At this point I am inclined to believe the MAF scaling is a problem between the two systems. Based on what Philip has provided as far as data, it sounds like what may be happening is the LN3 computer is thinking the MAF is flowing a lot more air than it actually is; at least above a certain flow rate. I agree it should be very difficult for one to get into the higher LV8 portion of the timing table while the vehicle is in neutral. With this happening, the computer can be dumping way too much fuel causing the bog. Philip, sounds like we need to try another chip with different MAF values.

-ryan

Share this post


Link to post
Share on other sites

I am also using the Tuner Cat software and agree it is difficult to see exactly where he is pulling the data from (need to go into the editor and pull up every value, then see what locations are missing. For example the 5B1 (1988) version I had did not have DFCO so I had to add it.

Suppose I have been a bit lax about developing the software but

1) is not my rice bowl

2) my 88 runs well for me

3) Corvair is taking most of my time

Think I have about all of the ECM programming just in bits and pieces and have forgotten a bit since I programmed out the automatic and EGR elements for Greg's.

Point is that at light load (holding 1600 in neutral) you should have close to 30 degrees of advance and *something* is commanding the rong value, either reading in the rong "LV8" or dialing back the advance for some reason, or just plan reading the rong value (if the tdf - tuner definition file does not match the PROM program, at at least four different tdfs for 3800s that I know of, may be more, you will have the rong locations).

A good description of how to develop a TDF is found here.

Frankly, this is not rocket science, rather basic embedded programming and is lots of documentation on the Motorola 68HC11. Shouldn't be anything too difficult for a PC whiz. Do need about $200 worth of equipment and some software to make things easy.

Big point is that the ECM is not commanding enough advance and need to find out why.

Share this post


Link to post
Share on other sites

I don't know what Philip's MAF scale might be, but it looks like a dead ringer for the L27 one I have. If it is the same, my testing indicates it is almost exactly the same as the stock LN3 flow rate vs frequency. It looks like you guys have narrowed the problem area to a timimg issue. Low timing without boost to speed the flame front would certainly cause what seems to be happening. The one thing lacking in the Reatta diagnostics is the LV8 reading. It would be nice to have a scantool connected to see what cell the ECM is reading.

I would guess GM spaced the rings on the balancer differently to prevent them being switched, but the reason baffles me? The internal construction of the engine indicates the only difference is that some of the s/c engines used a floating piston pin. Maybe something else?

Share this post


Link to post
Share on other sites

<div class="ubbcode-block"><div class="ubbcode-header">Quote:</div><div class="ubbcode-body">

I would guess GM spaced the rings on the balancer differently to prevent them being switched, but the reason baffles me? The internal construction of the engine indicates the only difference is that some of the s/c engines used a floating piston pin. Maybe something else? </div></div>

Very possible. I looked up some GM part numbers and have the following conclusions...

88 LN3 crank is different than 91-95 L27 and L67 cranks; I already knew this because all LN3's I have seen had a 6-bolt flywheel flange and the L27/L67's had 8 bolts. Crankshaft part numbers are the same for both the L27 and L67.

Connecting rods: LN3 and L27 have same p/n. L67 has different part number. Of the engines I have seen apart, the L67 connecting rods are much beefier than the n/a applications; IE: heavier.

Pistons and pins: Again, LN3 and L27 uses identical part number. L67 is different and is beefier than n/a counterparts.

Since the 3800 engines are externally balanced, most of the weight change to account for the differences between the SC and N/A differences must be made at the balancer and flexplate/flywheel. That is probably why the L67 dampener Philip got wouldn't work with his crank sensor.

As far as the tuning is concerned, it would be nice to have some LV8/load data to see if the ECM is going into "high load" too soon. Not only would this give the engine too little timing, but it would probably give it too much fuel. I use Tunercat also and there is no defined table in the $5B1 (88 LN3) programming that assigns LV8 values based on MAF and RPM. In fact, in all of the 3800 platforms I support, such a table does not exist.

With that being said, there is another thing we might want to consider. Concerning MAF calibration tables, I noticed they were nearly identical between Philip's stock LN3 tables and the 1995 L67 tables I pulled from a stock .bin file. However, when I look at 92-93 L67 tables (use a different computer than the 94-95 L67), the MAF values are much different. See below...

Stock 1988 LN3 MAF calibration tables:

LN3_MAF.jpg

Stock 1994-95 L67 MAF calibration tables:

$5b_L67_MAF.jpg

Now here is the one that is really different... 92-93 L67:

$5B4_L67_MAF.jpg

Share this post


Link to post
Share on other sites

The tables would seem to be back asswards smirk.gif I don't remember exactly where the break points were for the L67 engines but I think the initial ones were like 205hp, then they went to 225hp before the later Series II with 240hp.

Two things come to mind. GM is treating the output signal differently from one year to the next, meaning the gm/sec reading is a reference, not a real number, or the signal is modified from the MAF (I found this can be manipulated easily), again making the signal a reference number only? It doesn't make sense since they have the ability to change everything to make the signal reasonably accurate and the computing power to use it? The later model L67 should definitely run beyond that limited MAF table, no matter what conversion I use for the mass of air. Once I corrected some errors in my flow bench and Excel formulas, I find the four 3800 MAF sensors I have (including the TPI L27), max out (10.4kHz) between 300 - 318 cfm, converted to 171-183 gm/sec (34.4 grams/cu.ft) at my test conditions). My numbers are close enough to seem to indicate the MAF readings "should" be real numbers. There is probably some manufacturing tolerance, and the MAF signal isn't entirely stable, it floats around a bit, no matter what the inlet tract looks like, so I am relying on averaging several sensors to get decent numbers.

Did I just luck out getting mine to run reasonably well using essentially the same platform? The only difference between Philip's car and mine are the balancer/CPS and the MAF, aside from the way the boost is made. Greg and Don have running S/C's, although they were complete engine swaps. Strange behavior.

Share this post


Link to post
Share on other sites

Sorry I was away from a computer for a day.

Uh, yeah. So Ryan, do you wish that I send my memcal back to you? I'm pretty tired at the moment, so I'm not fully reading on this MAF stuff.

Was able to get the injector numbers. I found 927, 467, and 0-280-150-934. It had a symbol on it too. Not sure if it was Bosch

Share this post


Link to post
Share on other sites

Big change was that GM went to a 2 byte MAF value for the L-67s I have seen. There are ways to play with the one byte value though. Tables look like I remember, think of it as six gears.

From the values and pipe colors mentioned it does not sound like it is running over rich which it would be if the MAF readings were high, just at least 10 degrees retarded and that is what the ECM is commanding. Next is to figure out why. One problem at a time.

Perhaps if you could hold it at 1200-1500 rpm and record all of the ECM values, something might jump out.

Share this post


Link to post
Share on other sites

<div class="ubbcode-block"><div class="ubbcode-header">Quote:</div><div class="ubbcode-body">Sorry I was away from a computer for a day.

Uh, yeah. So Ryan, do you wish that I send my memcal back to you? I'm pretty tired at the moment, so I'm not fully reading on this MAF stuff.

Was able to get the injector numbers. I found 927, 467, and 0-280-150-934. It had a symbol on it too. Not sure if it was Bosch </div></div>

Actually what I think I am going to do is lend you an adapter that will plug in between your mem-cal and the ECM. This adapter has a socket on it which will allow you to swap out individual EPROM chips. That way I can send you a couple of different chips to try out.

Give me a couple of days to get the new chips done, but I should get them sent out to you in time to be there by next weekend.

-ryan

Share this post


Link to post
Share on other sites

Better yet any chance of recording the advance at 900, 1100, 1300, 1500, and 1700 rpm ? That should show us wher in the LV8 map it is operating.

Share this post


Link to post
Share on other sites

That part number is a Bosch injector. The information I have indicates that is a 28#/hr injector and comes from a Pontiac 3800. The one thing I could not find is the pressure where the flow rate was determined. Usually the default is 43.5 psi if not listed. That should be about right, a little more than 50% larger than stock. No problem there.

Share this post


Link to post
Share on other sites

<div class="ubbcode-block"><div class="ubbcode-header">Quote:</div><div class="ubbcode-body">

As far as the tuning is concerned, it would be nice to have some LV8/load data to see if the ECM is going into "high load" too soon. Not only would this give the engine too little timing, but it would probably give it too much fuel. I use Tunercat also and there is no defined table in the $5B1 (88 LN3) programming that assigns LV8 values based on MAF and RPM. In fact, in all of the 3800 platforms I support, such a table does not exist.</div></div>

LV8 really isn't a table, it's a calculation based on MAF and RPM. Adjusting the MAF tables adjusts LV8.

<div class="ubbcode-block"><div class="ubbcode-header">Quote:</div><div class="ubbcode-body">With that being said, there is another thing we might want to consider. Concerning MAF calibration tables, I noticed they were nearly identical between Philip's stock LN3 tables and the 1995 L67 tables I pulled from a stock .bin file. However, when I look at 92-93 L67 tables (use a different computer than the 94-95 L67), the MAF values are much different. See below...</div></div>

I question if John has the scalar right in the 94-95 .tdf file. I've seen MAF scan values in the 200+gm/sec on these cars. The Tunercat program has it's own scalar that is separate from the scalar used in the .bin file that's actually loaded in the car. I don't have a .bin here for the 94-95 PCM, but if you could take a look at location $8E78 and let me know what's loaded in there I could tell you what is actually used.

Here is the MAF table from my Series 1 S/C engine running the 8253:

MAF.jpg

Share this post


Link to post
Share on other sites

Location location $8E78 is going to be a constant at offset $E78 in the 32k prom contents. The ECM just loads the prom in the upper half of 64k (16 bit) memory. This is because the processor always starts by executing the instruction (usually a jmp) found at the top of memory.

BTW the ECM produced MAF values along with the Advance readings would also help.

AFAIK the early (92-93) L-67 was the only one that used the 26C256 PROM, later versions used a 27C512 (double the memory so offsets are different). Does someone have the stock bin I can look at ?

BTW for the math impaired a leading dollar ($E78) is one form of expressing a hexadecimal value. My original larning was to use a lower case h suffix (E78h)

Share this post


Link to post
Share on other sites

Been away for 10 days, this is interesting Sh**

Looking at the image of the two balancers, a question? did you straighten out the bent tang before installing the L67 Pulley?

Because there's some skew in the angle of the image from one pulley to the other it does appear the keyway aligns Ok.

Share this post


Link to post
Share on other sites

This is indeed interesting.

Yes, I did notice the bent tine or whatever you call it and bent it back before putting it on. CPS sensors are not a good thing to kill.

This afternoon I'm going to look into fixing my cut exhaust and then getting those ED08 readings

Share this post


Link to post
Share on other sites

My 2 cents.........put on the high flow cat from Summit or Jegs as suggested instead of using the Fernco couplings. Also, if you haven't done it yet, remove the exhaust restriction in the manifold. I think you will burn up the rubber in the Fernco coupling. (PEE-Yew!) cool.gif

Share this post


Link to post
Share on other sites

I actually used the metal part that goes over the old school Fernco bands. Just got it done.

I did the above and used the old throttle body "plate" to attached the transmission cable that goes to the throttle body. Needs a little modification to be just right though, but won't pose any problems now.

New problem: I fired it up to get those ED08 readings, but found I've got a substantial (trickle) coolant leak between the intake manifold and the supercharger. [4 letter explicits]. It was a pain in the butt putting the supercharger on, and I suppose those two yellow O rings got misaligned or something. A setback, but I'll order another gasket set tomorrow. Another couple hours to deal with that mess.

Share this post


Link to post
Share on other sites

While you have it off, you could paint it (and the valve covers) RED!!!!

They should be dry by tomorrow, and you're not busy getting ED08 readings now.

Once you have it together, y'know it'll never get COLORIZED!

Share this post


Link to post
Share on other sites

<div class="ubbcode-block"><div class="ubbcode-header">Quote:</div><div class="ubbcode-body">

LV8 really isn't a table, it's a calculation based on MAF and RPM. Adjusting the MAF tables adjusts LV8.</div></div>

I was under the impression that there is "something" in the programming that tells the ECM how to assign LV8 readings. If it isn't a table, is there a constant? (my information on $5B1 and $5B2's is quite limited, so any info you can provide would be quite helpful).

<div class="ubbcode-block"><div class="ubbcode-header">Quote:</div><div class="ubbcode-body">I question if John has the scalar right in the 94-95 .tdf file. I've seen MAF scan values in the 200+gm/sec on these cars. The Tunercat program has it's own scalar that is separate from the scalar used in the .bin file that's actually loaded in the car. I don't have a .bin here for the 94-95 PCM, but if you could take a look at location $8E78 and let me know what's loaded in there I could tell you what is actually used.

Here is the MAF table from my Series 1 S/C engine running the 8253:

MAF.jpg </div></div>

As I mentioned earlier, my information on the earlier 3800 computers is quite limited. TrofeoSC, I noticed you have reset the scalar on your $5B2 MAF tables so you could attain 256 gm/sec flow rates (stock maxes out at about 171 gm/sec), do you have a "good" commented disassembly of the $5B2 or $5B1 codes you would be willing to share? Please understand I don't want to take your hard work and use it without giving you due credit and compensation. However, any light you can shed on this issue will help the Reatta community.

Share this post


Link to post
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now