Balancing S38 ITBs

Specific conversations and info for the BMW E28 M5 and M535i.
HealeyBN7
Posts: 251
Joined: Feb 06, 2009 5:22 PM
Location: Thousand Oaks, CA

Balancing S38 ITBs

Post by HealeyBN7 »

Thought I would throw this out there...

After completing a dyno tune on my mildly stroked S38, I am still working to dial in the ITBs on the M5. The ECU was hunting just off idle and depending on where it was in its cycle it would "jerk" slow speed transitions. We didn't spend a lot of time on the dyno working the idle and low RPM range as I was not confident my ITBs were properly synched before we started.

I am getting close after chasing a couple dead ends, so thought it might be beneficial to share my experience and pick up any extra tips. Because the new engine uses a MAF (hot wire) instead of a more restrictive AFM (spring loaded flap), the engine is much more sensitive to idle air leakage through the butterflies and bypass screws. I am sure the higher compression and increased stroke also plays a part.

The factory set the butterflies with a resting position of .010. Zero is set where the butterfly is fully closed into the housing. A dial indicator is positioned at the leading edge of the butterfly and a stop screw is cranked until you get a .010 reading. All three pairs of ITB are set in the same way. Unfortunately this is too much free air flow for the ICV and bypass system, so the small needle valves on each ITB are ineffective as tuning individual contributions.

I have been reducing the opening of the ITBs far enough in order for the needle valves to have an effect. When I started at .010 you could move the needle valves and there was no change in engine tune.

I have been setting the mechanical adjustment by removing the manifold/trumpets and using the dial indicator. Not difficult but fiddly with all the little nuts and tight spaces. I noted a significant improvement by tightening the clearance to .005, and then made a further adjustment by "feel" to .003. The final setting has the potential to change a bit when the linkage is reinstalled and all three ITB pairs are linked. Instead of setting up three dial indicators I ordered stack of .002 and .003 feeler gauges from McMaster Carr so I can ensure that the linkage does not change the static setting of each ITB. For the final mechanical adjustment, I'll set the stops, install the linkage and ensure all six shims have equal drag. I plan on cutting the shims to ensure I am not introducing issues with the curve of the barrel. They are mechanically close now, but I need one more final pass.

I tried three methods to balance the ITBs with the bypass screws.

1. The manual instructs you to install a vacuum gauge to the ported ITB and record the value. Move the gauge to each ITB and repeat and average the result. Complete repeating this process until all values are close... I found this impossible as the idle speed changes as you make adjustments and invalidates the previously recorded values. A complete waste of time. I thought I could improve in this by buying 6 vacuum gauges. Each required a dampener needle valve to keep the gauges from bouncing. Then I found that they were not calibrated the same (a friend warned me about this), so I disassembled the gauges and hooked them to a common manifold and pulled a vacuum from a small HF pump I used for my AC, and reset the needles. They were then calibrated at a specific vacuum, but were our of spec outside of a narrow range. All this really determined is what we all know - Chinese gauges are junk. On to method two.

2. An S38 guru, suggested that I use exhaust temperatures. I tried using an IR gun, but short of burning my hand to ensure I was pointing at the right pipe, this was more inconsistent then the individual vacuum gauge approach. Not learning from my experience with the vacuum gauges, I found an instrument that can measure four individual readings at one time. I bought two (good for 8 readings) and made clamps to hold the k-type sensors to the outside of the exhaust tubes. Temperatures varied by 50-80 degrees and large adjustments to the bypass screws, so big that you can hear/feel cylinders working or shutting off, didn't seem to make any differences to the temperatures. I gave this a really good try, repositioned the clamps several times (ouch) and finally gave up when I realized that all I was measuring was the thickness of my ceramic coating and perhaps the distance from the exhaust valves. Given my euro headers have no provisions for EGT ports (the US exhaust manifold does), and removing them to weld bungs would be almost impossible without dropping the engine this was a hill too hard to climb. Besides I started to wonder how effective exhaust temperatures are at idle with the engine completely unloaded. So, besides determining that ceramic coating works, all I can conclude is this was great way to burn your hands. On to method three.
Image
Image

3. I had a discussion with a friend that manages large ship engine installations. He said they use manometers, or mercury u-tubes to measure absolute vacuum. I spent a few hours searching for mercury and although I found a few sources, as I learned more it seemed that the EPA would be knocking at the door. I found a motorcycle tool that measures relative vacuum. This eliminated the mercury issue and resulting jail time. I found a DIY site that I extrapolated that to a six cylinder ITB application. 100' of tubing, a little Marvel Mystery Oil and a few staples in a board resulted on a tool that seems to have real value. I was able to quickly turn the bypass screws and watch the fluid rise and fall. I ordered some fancy parts from McMaster Carr to make it a better tool with a common aluminum manifold and provisions for filling, draining and adding a vacuum gauge, but as it is it seemed to perform well. Finally!

Image
Image

A short test drive to the snake and pretending to creep along in traffic convinced me that it is just about perfect.

So the M5 will need the mechanical throttles adjusted one more time with the shims and I should then be able to perfect the balancing with the manometer. And capping off two of the ports should allow me to use it on the Lotus.

I am anxious to return the LM-2 back to it's case or move it to the Lotus.

What have I missed?

Dean
Edited to fix the pictures...
Last edited by HealeyBN7 on Mar 31, 2015 4:04 PM, edited 1 time in total.
tn535i
Posts: 5585
Joined: Jul 14, 2006 1:30 PM
Location: Middle Tennessee

Re: Balancing S38 ITBs

Post by tn535i »

Good work and detail... I think I agree all the way but not sure if you missed something so I'll elaborate.

First the Jerk, I think this happens when after all the balancing it still results in a bypass flow that requires too much transition every time the TPS open or closes. See if that Jerk goes away by just unplugging the TPS, assuming it will still idle OK. That would prove if it's the TPS/ICV On/Off idle that's causing it. The trick I've found to getting the transition as smooth as possible is to get it to idle just slightly above the point the TPS and ICV will control it even if TPS unplugged. If unplugging the TPS causes the idle to drop too much or jump very high you'll have a rough transition. It might still do a bit of this cold but warm you want the speed plugged in or unplugged to be pretty close.

So the above completely drives (IMO) the best position for the ITB throttles to be set. I can't say what my settings were since we have probably measured differently but my home position is the one that gives a warm idle with the TPS/ICV out of the equation (next paragraph) and has the individual bypass screws about 1-1/2 turns and the big bypass screw about the same. I think (just a guess) that your base throttle setting might be a little tight and if you unplugged the TPS or closed down all your bypass screws your idle would be too low. Also the finer or tighter the throttle opening the harder it would be to get them all the same I think. A little more open should also be a little easier to match up but you might be just fine using the adjustment screws at this point.

Setting the whole system with the TPS in the jumpered position is the only way to go as I recall without looking at my notes. I'm not sure you mention that but I think jump BOTH the idle and WOT to the center pin on the harness side and it sets the ICV to default home position and also ignores the O2 sensor. I think that's an S38 DME thing? You don't want it either trying to control idle speed or adjusting mixture and dithering based on O2 while you balance ITB's or it will drive you mad. This may stop some of the hunting and give you the steadiest possible idle to do your work.

I see you think the cheap vacuum gauges are junk and I agree. Many buy the setups with multiple gauges and like them but I've been trained not to trust them like you found out until you verify. Even then the S38 intake vacuum level is so dynamic the size of the snubber and everything else has some impact and getting everything identical is nearly impossible even with quality parts so why bother trying. It probably works OK but not really best.

I agree that a U-tube manometer (maybe no snubber at all) that allows you to reference one cylinder to another is the best way. It's maybe possible that equal length tubing from each side of the manometer is best and longer tube will act like a small snubber. Basically the same set up you would use on a two cylinder motorbike. You don't want Hg because it's so dense and less responsive but oil works great because it's only 0.8 and even less dense than water so more responsive than either. The only thing is you balance each cylinder to #1 now for example but then need to work back through another time or two till it's no longer changing. And take care not to unplug one end and leave the other to suck the fluid out. You need to clamp it off every time you switch it around. Then basically you are looking for NO reading no matter which two cylinders you hook it up to.

Sounds to me like you have really good methodology and maybe just need another run through.
HealeyBN7
Posts: 251
Joined: Feb 06, 2009 5:22 PM
Location: Thousand Oaks, CA

Re: Balancing S38 ITBs

Post by HealeyBN7 »

Thank you for the detailed reply. This is just what I needed!

I have the TPS jumper in my valve adjustment kit and forgot all about it. Ugh. Yes, essentially all three contacts are grounded together. Good tips on checking to see if the TPS is set properly. I was using the multi meter/click approach, but no real way to see if I have it right on. thanks. I'll shoot for 1.5 turns out. You are correct that I am restricting more air. I am closer to .75 now but one or two are just about closed. The big plug is around 1.5.

Absolutely right that all six of my manometer tubes need to be plugged in or it will drain the marvel mystery oil before you can shut off the car. First time it was close as one of the fittings was a bit loose.

My AFR is pretty steady at 14.7 ~.2 at PT cruise. I an still dithering between 13.5 and 15 at idle. Is it correct to assume the ITB balance will get me to 14.7 or is that just too much to ask?

If my shims come today, I'll give it another go tonight and report back. Thanks again.

Dean
wkohler
Posts: 50921
Joined: Oct 05, 2006 11:04 PM
Location: Phönix, Arizona, USA
Contact:

Re: Balancing S38 ITBs

Post by wkohler »

Interesting read. Kinda glad you hit this stumbling block so you could share it with us!
ahab
Posts: 6174
Joined: Jun 11, 2006 9:12 AM
Location: Chalfont, PA

Re: Balancing S38 ITBs

Post by ahab »

Wow. Excellent info here. I will be setting off down this same road in another month or two and this is invaluable. I bought the 4 bank CarbTune vacuum gauge but would be interested in the site that has the instructions for the 6 bank manometer. I also have a Miller MAF and cannot get the car to run in any sense of the word driveable when it's installed. Connecting the CarbTune shows I am pulling 28-30 cm/Hg on each cylinder when the spec calls out 22. Driveablilty with the AFM is actually pretty good despite this but the AFM shows evidence of adjustment which I believe is somehow compensating (or masking) for the increased vacuum.
Karl Grau
Posts: 9708
Joined: Mar 10, 2006 7:34 PM
Location: Sandy Eggo

Re: Balancing S38 ITBs

Post by Karl Grau »

Wow, I'm impressed! The Roundel is a nice touch too. :)
tn535i
Posts: 5585
Joined: Jul 14, 2006 1:30 PM
Location: Middle Tennessee

Re: Balancing S38 ITBs

Post by tn535i »

I studied your manometer a little more closely and got to thinking about it. Since you are basically referencing each cylinder against every other I guess while you balance them all at once everything is dynamic but should get you there. You might just test it by driving one way out and making sure it rises above all the others or drops. In any case the oil at low density is a good choice to get more resolution.

Please let us know how it works out.
HealeyBN7
Posts: 251
Joined: Feb 06, 2009 5:22 PM
Location: Thousand Oaks, CA

Re: Balancing S38 ITBs

Post by HealeyBN7 »

Well, what I thought would take just a few minutes before dinner, had me glued to the garage for the most part of the evening, but I am here to report success.

I swept back through the mechanical adjustments. The .002 shim idea did not work as the throttle plates are far tighter even at the upper end of the .010 adjustment, so the whole shim idea didn't work. Good news I was only out $1.25 per shim or $15.00 if you include McMaster Carr shipping... We don't right?

So, back to the traditional dial indicator method. I was able to carefully set each ITB pair to .004.

Image

When I hooked up the linkage I found that I was introducing a .002 error that compounded from the rear to the front. So, I remounted the dial indicator on the middle ITBs and set it to zero. Then I hooked up the linkage. I can now measure the error and adjust the linkage back to "0". Perfecto. I did the same for the front pair. This is a far better method then the manual's approach of measuring the linkage length with calipers and hoping for the best. All the mechanical components are now spot on. Done.

Image

Next I remounted the intake and connected the manometer. You can see in the photos that I swapped out the series of "tees" and installed an eight port aluminum manifold. This allows the installation of a vacuum gauge and fill port. Both good improvements and it makes it look like a real shop tool, well if shop tools were made of scrap wood.

Image
Image

Using the TPS bridge this time, thank you TN535i, I was able to work to a non fluctuating baseline. duh!

Each of the vacuum tubes are connected to the ITB ports. I have them numbered with yellow tape as I am easily confused.

Image

In this version of the manometer I removed a micro dampening orifice from each of the tubes. I don't know if you could see them before, but they were short lengths of copper pipe that I soldered closed and drilled a very small bypass hole. With the orifices restricting the volume as it dampened the intake pulses. I don't think they were necessary. Now full vacuum volume pulls on the tubes. The fluid jumps around a little more, but you can easily eyeball averages and work to align the liquid just as easily as before.

Here are a few photos in process. Once I figure out how to link/upload the video, that may help explain things. Slight bypass adjustments make a very big impact on the visual manometer gauge. Less then a quarter turn will swing the fluid up or down 4 inches. It didn't take long to get them all within an inch of each other. I didn't seem worth the effort to make it any closer, as I was just breathing on the adjustment screws and the fluid would rise and fall a half an inch. Here I caught it mid movement on the video screen snap. On average it was much closer.

Image
Image

I ended up with the main bypass out 1.5 turns. The ITB individual bypasses are around .75 to 1.25.

The test drive was WONDERFUL. It is new car smooth on and off the throttle even at the slowest of speeds. I should be able to cruise through parking lots and traffic just like the Range Rover. The LM-2 shows a near perfect 14.7 at just off throttle, where it would have been hunting slightly before.

Hoping for a fun drive into the office tomorrow!

Thanks for all your help and encouragement. YAHOO.

Dean
wkohler
Posts: 50921
Joined: Oct 05, 2006 11:04 PM
Location: Phönix, Arizona, USA
Contact:

Re: Balancing S38 ITBs

Post by wkohler »

This is great.
tn535i
Posts: 5585
Joined: Jul 14, 2006 1:30 PM
Location: Middle Tennessee

Re: Balancing S38 ITBs

Post by tn535i »

Nice Win and thanks for trouble of pics and sharing.

You need to give the manometer a name now like the arthropodometer (for the six legs) or since a preying mantis has long legs the mantismanometer. I need to build one for next time rather than referencing to one cylinder and working through them one at a time like I did.

Did you still use a snubber on the vac gauge or did it read pretty steady since it was downstream of the fluid? I think the fluid would have some impact on it's reading since you have to lift the weight of the fluid? It looks like you get only 10" Hg which is a bit low but as I recall mine reads somewhat low also like maybe 15" while an M20/30 should be 20 ish. I think due to the ITB's you won't get a nice high and even vacuum like other engines and each trumpet is very dynamic at idle. So another question... Did you come up off idle a bit to see what the balance looked like? Another worthy measurement/diagnostic attribute of the mantismanometer :laugh: might be that you can see if the linkages all tip in the same going from idle to maybe 1500 rpm or so.
Last edited by tn535i on Apr 01, 2015 10:34 AM, edited 2 times in total.
jdb
Posts: 140
Joined: Jan 01, 2007 6:56 PM
Location: Humboldt County, Ca.

Re: Balancing S38 ITBs

Post by jdb »

Very well done!
This is a real journeyman's job.

Jeff
ahab
Posts: 6174
Joined: Jun 11, 2006 9:12 AM
Location: Chalfont, PA

Re: Balancing S38 ITBs

Post by ahab »

Bravo! Nicely done.
HealeyBN7
Posts: 251
Joined: Feb 06, 2009 5:22 PM
Location: Thousand Oaks, CA

Re: Balancing S38 ITBs

Post by HealeyBN7 »

tn535i wrote:You need to give the manometer a name now like the arthropodometer (for the six legs) or since a preying mantis has long legs the mantismanometer.
Mantismanometer it is! I love it.

The vacuum gauge has a very small orifice/snubber built into the connection, and yes it read pretty steady. It didn't when I connected it to a single ITB. The combined pull across all six ITBs and the fluid does seem to have a dampening effect. It didn't move around much at all.

The spec calls for an average of 292 m/bar.
http://home.insightbb.com/~todd.kenyon/throttle9.JPG

I don't know what an "m/bar" is. I searched. I found conversion tools for everything but m/bar. If I convert 292 bar into inHG, I get 8623 inHG , which is so far off the vacuum scale I think it approaches the creation of a black hole. Moving the decimal place to .293, I get a reasonable 8.6 inHG, which is close to what the gauge is showing. A little education here would be helpful, hint hint:)

Yes. I came off idle to see what happens. I did this earlier, before and after adjusting the linkage. That is where the part throttle issues throw off the O2, and what convinced me to readjust the linkage. What I saw was steady idle, then with the slightest throttle input the ITBs started to vary by several inches. This told me that the bypass screws were compensating for minor (.002) linkage issues. I do think that was helpful insight.

The ride into the office this morning was terrific. Freeway cruise is smooth and the AFR is steady.

On the "Mantismanometer", with the new vacuum gauge hooked up, it carries fluid at the same level as the service legs which attach to the ITBs. As the vacuum pulls on the ITBs, it drops the fluid level in the vacuum gauge tube. This has the potential for bringing air into the service side of the lines. I think I am going to install a larger tube for the vacuum gauge, so it can act as a tank for the service legs.

Fun fun.

Dean
wkohler
Posts: 50921
Joined: Oct 05, 2006 11:04 PM
Location: Phönix, Arizona, USA
Contact:

Re: Balancing S38 ITBs

Post by wkohler »

Mbar is a millibar, which is equal to 100 pascals or .1 kpa.
HealeyBN7
Posts: 251
Joined: Feb 06, 2009 5:22 PM
Location: Thousand Oaks, CA

Re: Balancing S38 ITBs

Post by HealeyBN7 »

wkohler wrote:Mbar is a millibar, which is equal to 100 pascals or .1 kpa.
Chris. Thanks that helps. So, 292 mbar is .292 bar, which is indeed 8.6 inHG. It is a bit lower then what I was registering. I was showing 10.25 inHG or 347 mbar, so I am pulling excessive vacuum over spec.

Hmm. It runs so nice. The OCD part of me is saying another round of testing to see if I can lower the vacuum by (347-292) 55 mbar.

Dean
ahab
Posts: 6174
Joined: Jun 11, 2006 9:12 AM
Location: Chalfont, PA

Re: Balancing S38 ITBs

Post by ahab »

Doing the conversion against the 28-30 cm/hg that I'm reading equates to 373-399mbar. Also excessive vacuum. As I said earlier, the car runs well enough that driveability is actually very good but I'd like to get my car closer to spec as well.
HealeyBN7
Posts: 251
Joined: Feb 06, 2009 5:22 PM
Location: Thousand Oaks, CA

Re: Balancing S38 ITBs

Post by HealeyBN7 »

I reset the OBC MPG tracker before heading out to work this morning. 80 mile round trip. Mostly 80 MPH freeway on the way into work. Traffic on the way home. I haven't see this average before, except in the wife's '84 528e.

Image

Now that's a measurable improvement even on crappy California summer gas blend.

Dean
Mark 88/M5 Houston
Posts: 8548
Joined: Feb 12, 2006 12:00 PM
Location: Far North Houston

Re: Balancing S38 ITBs

Post by Mark 88/M5 Houston »

Great research and documentation! Thanks for sharing with the forum.
tn535i
Posts: 5585
Joined: Jul 14, 2006 1:30 PM
Location: Middle Tennessee

Re: Balancing S38 ITBs

Post by tn535i »

OK wait a minute both of you (ahab and healey)... Just look at your gauge and it shows a healthy engine should be 20 ish. I used to tune carbureted engines at idle by maximizing vacuum with both mixture and advance to get as high vacuum as possible. I really don't think there is anything wrong with higher vacuum on the S38 and the spec might be more of a minimum. If properly tuned and the vacuum is high you are better converting fuel and air to motion or power. That's good and I'm not even sure how you would adjust it down other than to change the valve timing, mixture or something you can't easy do like the spark advance to make it run worse. The S38 is already lower than an M20/30 because of the valve timing and ITB set up so higher numbers for vacuum are good numbers so long as they are even on all 6. Then there is the real possibilities of differences in gauges and such. To get a true reading you probably need to build another u-tube manometer that is filled with something dense like Hg (call the EPA) open to atmosphere. A water manometer at 10" Hg would need to read about 11 feet H2O which is impractical. I would be glad the reading are high if the O2 says you are not lean.

Go here for lots of conversions

http://www.unitconversion.org/unit_conv ... ssure.html
buzby
Posts: 466
Joined: Feb 10, 2009 8:21 PM
Location: Auburn, CA

Re: Balancing S38 ITBs

Post by buzby »

Great gearhead information Dean! Thanks to all the other contributors as well.

Dave
ahab
Posts: 6174
Joined: Jun 11, 2006 9:12 AM
Location: Chalfont, PA

Re: Balancing S38 ITBs

Post by ahab »

tn, thanks for the input. I was under the impression (perhaps a misguided one) that the higher vacuum translated to a leaner base mixture, and my AFM had been re-calibrated to compensate. I suppose the O2 sensor should be forcing the ECU to maintain stoich and maybe I'm off base here (O2 sensor is new and injectors recently cleaned). I'm certain that the AFM has been adjusted but I'm not sure to what extent. I don't have the LC-1 installed yet to know by how much, but I do have a rich condition which is evidenced by the soot at the back of the car and the black smoke at high rpm. I ran seafoam through and it went away temporarily but it has started to reappear as the carbon builds up again. I also have the issue with the Miller MAF. The car will run and I can nurse it around the block but it is in no way driveable. Even the base tunes from Miller shouldn't be that far out of spec. I was attributing this, perhaps incorrectly, to unbalanced ITBs.

Not trying to hijack you Dean. I was only attempting to round out the discussion with some additional info about ITB balance.
bobby in Derby City
Posts: 980
Joined: Feb 12, 2006 12:00 PM
Location: Hidin' in a bowl of rice ready to put a cap in an a$$

Re: Balancing S38 ITBs

Post by bobby in Derby City »

Kudos to you Healey for the in-depth write up and pictures. :up: I made the same type of manometer about 7 years ago to balance the throttles on a Suzuki Katana 600 I had very briefly (brief motorcycle itch) after I rebuilt the carbs since I did not want to buy the Carbtune devices (cheap motorcycle, cheapo tools :laugh: ). I still have my "Mantismanometer" somewhere albiet it has only 4 tubes. I had wanted to use it on my S38 years ago, but was just lazy about retro-fitting it for six throttles. Where did you get the little "7-port block thingy" you substituted for original tees you had?? Maybe I need to make another one and keep my original one for the S14 (if I ever rebuild the ventilated engine ... LOL)TIA!
HealeyBN7
Posts: 251
Joined: Feb 06, 2009 5:22 PM
Location: Thousand Oaks, CA

Re: Balancing S38 ITBs

Post by HealeyBN7 »

bobby in Derby City wrote:Where did you get the little "7-port block thingy" you substituted for original tees you had??
I found the manifold in McMaster-Carr's vast selection of all things impatient tinker's need.

http://www.mcmaster.com/#standard-manifolds/=wlswo3

Here is a new picture of the manifold. You will see that I added a few feet of larger 3/8"ID tubing for the vacuum gauge port. It now acts as a reservoir to remove the risk of introducing air into the legs of the "Mantismanometer".

Image

We need another S38 to try it on. Any takers?

Dean
ahab
Posts: 6174
Joined: Jun 11, 2006 9:12 AM
Location: Chalfont, PA

Re: Balancing S38 ITBs

Post by ahab »

HealeyBN7 wrote:We need another S38 to try it on. Any takers?

Dean
If East Coasters are included in the offer, let me know how much to send for deposit/shipping to 18914.
bdalessio
Posts: 39
Joined: Jun 01, 2013 9:42 PM
Location: San Francisco, CA

Re: Balancing S38 ITBs

Post by bdalessio »

PM sent.
Post Reply