ITB/EFI Vacuum Manifolds- let's see em'!

[mikesarge]

Hi all-

I'm having a helluva time getting a reliable vacuum signal out of my ITBs. I've done smoke tests, replaced hoses blindly, tested all ancillary vacuum components, and no luck! Al has been as helpful as he can be over email (and remarkably patient with me!), and I am going to bring him my car in a little while here anyway because I don't want to trust myself with tuning this thing on my own, but since I've got some time until he's available, I thought I'd start a thread on vacuum manifolds, and vacuum signal.

Plenty of threads out there for pics of ITBs, but very few show clear shots of the vacuum system I'm finding. Also plenty of threads on msextra and other forums about the difficulty in obtaining a reliable vacuum signal with ITBs.

So, here's the request:

Post up a picture or two of your vacuum manifold setup, and along with it, how about-

Cam grind
Displacement
Vacuum signal at idle (according to your ECU)

[winders]

I think it is a mistake to try to get a usable MAP reading with ITBs on a high performance setup. At least at anything above 20 to 30% open throttle. The vacuum goes away so you need to use throttle position.

If you have mild cams, you can use a MAP for partial throttle mapping below 30%. But you need to switch to throttle position once the throttle opens up. This would be a blended approach that not all ECUs support.

My ITBs have vacuum ports...but they are sealed.

[OSC911]

Exactly what Scott said. I have a Motec and use TPS only. Have the map for barometric only and ports sealed on the throttle bodies.

[jpnovak]

I use an 8 port machined aluminum manifold block. 6 ports to each individual throat. 1 to ECU MAP sensor and one to FPR. It is the lower manifold attached to the rear shock cross member. I drill and mount a hose barb at the bottom of each cylinder throat below the ITBs. I then fill them with epoxy and drill a very small hole to help dampen the pulses. The manifold itself has a section of open-cell foam to help dampen pulses from cam reversion.




So, you didn't say what issues you are having with the vacuum MAP signal. A few things can happe.

- low MAP readings due to cam choice. the more overlap (narrow lobe center) the less vacuum , especially at low speeds.
- sawtooth patterns caused by cam overlap. This creates lots of signal noise and makes fuel metering difficult.
- sawtooth patterns caused by incorrect linkage setup. This happens when one ITB opens more than the other (or individual throats).

The biggest issue I see with most EFI installs is using a linear scaling for either MAP speed density or TPS alpha-N scaling. Both of these processes are highly non-linear and so don't even consider even spacing on your load rows in fuel or spark tables. If you do, you will have problems. I 100% do not agree that a car can not be tuned with speed density and ITBs and that you have to have TPS signals. That just means the tuner doesn't know how it all works. In some cases you can get better results with one or the other. And has been mentioned a blend often is the best compromise depending on the entirety of the engine build. It all has to work as a system.

[thetorch]

Originally Posted by jpnovak

...
The biggest issue I see with most EFI installs is using a linear scaling for either MAP speed density or TPS alpha-N scaling. Both of these processes are highly non-linear and so don't even consider even spacing on your load rows in fuel or spark tables. If you do, you will have problems....

Can you elaborate here? I don't understand why even spacing of load rows matters if you can tune the map values (fuel or spark) to be non-linear and therefore create an optimal curve that matches your preferences. Trying to learn here. Thx

[Showdown]

This is copied from my thread but I'll post it here for posterity too. This has resulted in a steady 45-48 kpa at idle and a solid signal throughout the RMP range- there's no pulsing or noise. All told, I think it was about $100 or so but I have the tools and whatnot to make it work.

After my little issue with getting a good vacuum signal I started reading a lot about EFI/ITB systems and the vacuum signals on which they're dependent. I'm still not sure if I'll use speed density or MAP but before I can evaluate which is best for my application I need to have a good signal.

Lots of the writing I came across was about the importance of isolating the vacuum signal from noise or pulsing. The best way to do this seems to be a central manifold and uniform travel distance from each TB to said manifold. Apparently hoses of different lengths can create a pattern of pulsing that can distort the signal.

I looked at EFI/ITB systems from lots of companies who charge lots of money for their product and almost universally I kept seeing a second vacuum rail near the fuel rails. Not a single one used hoses from the TBs to the manifold. Hmm.

When I plumbed my system I didn't know this so I just ran hoses from each TB to the manifold. Those were the instructions and I followed them to a T. Today, instead of taking a lunch break I decided to rectify that.

I started with some untapped fuel rail and cut it down to 12 inches. I drilled and tapped three 1/8" NPS holes.



After that I tapped each end of the fuel rail for an AN-8 plug.



Then before installing those plugs I drilled and tapped one per rail again for a 1/8" NPT fitting.



I now had two air tight manifolds; one for the left bank, one for the right.

Needing to fit them to the TBs I drilled and tapped more holes for an M6 bolt and fabricated a few aluminum brackets that would allow me to utilize the fuel rail mounting points on the RHD fuel rails.



I figured two per rail was fine but if I need another I can add it as the fuel rails have three mounting points.



Once the fabrication was complete all that was left to do was say a prayer and install them and hope that my measuring was accurate.



theres enough space for the injector wiring plugs to fit in between each rail and they don't interfere with any other function or access.



I did have to plug 4 of the 6 holes in my main manifold. I may fabricate a new one at some point but I had the plugs handy and this was the path of least resistance.



I still have to run the short hoses from the TB barbs to the vacuum rail barbs but that's a piece of cake- I did actually have to eat lunch otherwise I would have completed that part too.

So not only does it look a LOT cleaner without all those hoses running everywhere, said hoses are now out of the way of the linkages and can't get tangled up or interfere with any moving part, but the distance from each TB to the first manifold, is identical and the distance from each of those manifolds to the main manifold is also identical which should do as much as possible to eliminate pulsing and create as even of a vacuum signal as possible.

[jpnovak]

Originally Posted by thetorch

Can you elaborate here? I don't understand why even spacing of load rows matters if you can tune the map values (fuel or spark) to be non-linear and therefore create an optimal curve that matches your preferences. Trying to learn here. Thx

When setting up parameters in your tune file you must maximize dynamic range and resolution within sections of that range. Here is a MAP vs TPS graph. This is a twin plugged 3.0 with 46mm ITBs and Mod_S-102 cam. As you can see it is non-linear.



to maximize the range of MAP you need a minimum about 40 and max at 100kPa. Anything lower is a waste of a row. The idea would be relatively large steps up to 80kPa where the MAP rise rate is linear. Then you go to smaller steps ending up with 2 (or even 1) kPa steps above 92 where the curve really flattens out.

The opposite is true of TPS alpha-N tunes. You want 1-2% TPS down low and up to about 30% TPS and then much larger steps to 100.

Yes, you can tune with even bins but you loose resolution. ALL ECUs interpret a value between actual bins. the linear weighting algorithms often don't do well enough job for the accuracy I target.

[al lkosmal]

exactly.....

[mikesarge]

This is hugely helpful- thanks to everyone who has contributed so far,

Today I had a minor breakthrough, as I re-cut new vacuum lines to be closer to the same length, then pulled a bunch of fuel out of my initial map, because it was so rich the wideband was showing a false-lean. Just pulling fuel from the map improved my MAP signal a ton.

Now I’m seeing a vacuum signal that’s quite a bit stronger, though still pretty bad, around 65kpa at idle. It was drivable, so I drove it a bit to give my closest neighbors a break from the smell/noise, and get some temp into it. Lots of popping and banging, but it’s a big improvement nonetheless. Still quite rich.

Feels like there’s a vacuum leak, but it’s strange- let off the pedal, you feel the vacuum load into it with deceleration, but then it’s like the engine finds more air and the revs hang, stopping the deceleration and just coasting.

Anyway, I may try again with something like Julian/showdown’s setup. It’s clean, and he has great vacuum numbers. I’d like to eventually use ITB mode to dial in the tune, with both alpha-n and MAP maps, but to do that I have to get the vacuum situation taken care of first. This car is going to be driven on street >90% of the time, so lots of time spent at part throttle where a MAP signal should be an accurate way to fuel it.

In the meantime- keep posting your vacuum manifold setups! Let’s see em!

[Showdown]

Mike,

Any improvement is great and it’s always a combination of a lot of little things that eventually gets you there! My money is still on getting. Those ITBs perfectly balanced.

[mikesarge]

Ah, I should have mentioned, I also made another attempt at re-balancing today. I think they’re close. I also found that the trumpets were leaking above the throttle plate, which would present as an air leak after the flow meters and make balancing impossible. I think that’s sorted now, which likely also contributed to my gains here

I went from pulling around 10 kg/hr down to about 6, idling at 1150rpm or so. I still think that’s a bit high, and it was hard to stop it from low-frequency surging, making it a challenging and long process. I’ll try again next time I can get out to the car, unfortunately that’s a couple weeks out.

Thanks again!

[chrismorse]

Originally Posted by Showdown

Mike,

Any improvement is great and it’s always a combination of a lot of little things that eventually gets you there! My money is still on getting. Those ITBs perfectly balanced.

Hi Julian and Mike,

I haven't even started tuning my 2.8 hot rod build yet, (PMO itbs & Al's MS2 kit), but my sense is that it is important to get the cylinder balance right before dealing with the Map/tps input.
When I was rebuilding the 40 DCNFs on another car, i ran across a clever utilization of manometers used to balance 4 cylinder motorcycle engines. The vacuum ports under the throttle plates were used to individually feed a row of manometers. These could be carefully adjusted to get an accurate signal, then used to adjust the air bypass and throttles, for smooth idle and transition. I duct taped two units together on a board for the eight cylinder engine. All cylinders would register at the same time.
The car still has a vicious tip in hesitation, but i think this is due to uneven accelerator pump shot due to Shi**y alcohol gas deterioration of the diaphragm.

When i can upload a photo of the manometer set up, i will.
chris

[Showdown]

I’m no expert at all, but as far as I understand, balancing the ITBs is absolutely critical to success- without this step, one will simply chase their tail.

I’m super interested to see what you worked up with manometers, I’d imagine the precision of data they you had was leaps and bounds beyond a carb synchrometer.

Fwiw, I do think (and know based on other success stories and my one experience) that balancing the ITBs and getting a good baseline map signal is totally possible.

I was out driving and gathering tune data today and had a consistent 45 kpa vacuum at idle, so it’s totally possible.

[jpnovak]

MIke,

you make no mention of any engine parameters or EFI system. Some engines depending on cam and compression may only pull 65kPa at idle. I would not let absolute numbers bother me. Can you tune a steady idle or not? Can you tune the light to mid throttle street driving without any surging or weirdness? Just about anyone can tune WOT.

Some EFI systems allow you to choose the crank angle at which the MAP reading is taken. With reversion in an ITB if you are at the wrong crank angle then you will not see the lowest value. There are often settings in most to report mininum, average, time weighted average MAP values, etc. If your configuration is wrong you will have less than expected results.

I am just saying that modern EFI is highly tolerant of a low MAP signal. You just have to know what software knobs to turn to get the most out of it.

A few other tricks. You can pinch off your vacuum lines so that only 1 cylinder is driving the ECU reading. Then use a datalog to look for absolute minimums per cylinder. Helps identify hardware problems and air leaks.

[mikesarge]

Originally Posted by jpnovak

MIke,

you make no mention of any engine parameters or EFI system. Some engines depending on cam and compression may only pull 65kPa at idle. I would not let absolute numbers bother me. Can you tune a steady idle or not? Can you tune the light to mid throttle street driving without any surging or weirdness? Just about anyone can tune WOT.

Some EFI systems allow you to choose the crank angle at which the MAP reading is taken. With reversion in an ITB if you are at the wrong crank angle then you will not see the lowest value. There are often settings in most to report mininum, average, time weighted average MAP values, etc. If your configuration is wrong you will have less than expected results.

I am just saying that modern EFI is highly tolerant of a low MAP signal. You just have to know what software knobs to turn to get the most out of it.

A few other tricks. You can pinch off your vacuum lines so that only 1 cylinder is driving the ECU reading. Then use a datalog to look for absolute minimums per cylinder. Helps identify hardware problems and air leaks.

Hey Jamie-

This is all very helpful. My engine may as well be a stock SC: 3.1 CR 9.5:1, 964 cam. I suspect I’ve got a bit more to gain in terms of precision balancing, and continuing to improve vacuum signal. I know I’m having a hard time getting the throttles on 1 and 4 to completely zero out.

I’ll try your trick of pinching off 5/6 lines once I’m nominally ‘balanced,’ I like that as another method of checking.

Thanks for all of your input on this thread. Finally making slow, steady progress.

[jpnovak]

with a 964 cam (112 lobe center cam) you should be able to idle in the 48kPa range. Either your throttles are not synced, you have an air leak or you have some tuning to do. Target about 13.3AFR for idle. This should be smooth and give you the MAP minimum value.