LM-1 Sensor Position
 

Hi

I'm intending to install an LM-1 AFR in my MFI 911, Where is the best position for the sensor, Headers or muffler?

If you put it in the header you would only be getting left or right cylinders, muffler is mixed but further downstream

suggestions?

Thanks

Neven
73 911E, 70 914/6

What is the english version of your post?

Would it help if you looked at

http://www.pelicanparts.com/catalog/shopcart/MISC/POR_MISC_INNOVT-detail.htm

Were the TLA's (Three Letter Acronyms) to much?

Neven

Neven,

I would choose mounting two bosses on the two muffler inlets, just behind the muffler flange and inwards and angled slightly up. This allows the sensors reasonable protection, you can monitor each bank of cylinders independently, and average to get the overall mixture.

Another location is on the heat exchangers, just ahead of the muffler flange. This puts the sensors in the path of debris from the tires. OK, if you protect the sensors.

There is not an easy way to install a sensor that combines both sides of the engine. If you put it at the tail pipe, there is unreasonable delay in the measurements and accumulated contamination of the muffler will dramatically affect the readings.

A MFI 911 should always use an OEM muffler.

Best,
Grady

Thanks Grady

I've got a Danske SS OEM Style muffler and SSI Heat Ex, so I was thinking that the muffler inlet pipes would be good (They cast, not SS)

PS Ive got my megasquirt board and bought a spare MFI pump on Ebay so the next step will be my eMFI system

Neven

Neven,
You're right. You can only use one WB sensor with the LM-1. So it's either the left bank or the right bank, or one bank at a time. I suggest welding the bung(s) in the collector, right after the individual pipes (3) merge together.

Another company uses two WB sensors and a single controller which is more convenient for multi-bank engines.

Sherwood

Sherwood

The single bank thing is not so much an issue with MFI as I can't
adjust left to right anyway (assuming throttles are balanced). Maybe I'd be better off putting in dual MAP sensors :-)

I've always found it amusing that the vaccum advance is connected to both throttle bodies. Very german, zere is zee
slight chance zee left bank has a different manifold vacuum at
idle zan zee udder bank so we connected zee distrubutor to both!

Neven

Well, there's an idea. Weld the sensor bung in the muffler instead of in the heat exchanger. A lot easier to weld in that vs. stainless steel.

Neven wrote:
"The single bank thing is not so much an issue with MFI as I can't adjust left to right anyway (assuming throttles are balanced). Maybe I'd be better off putting in dual MAP sensors :-)

I've always found it amusing that the vaccum advance is connected to both throttle bodies. Very german, zere is zee slight chance zee left bank has a different manifold vacuum at idle zan zee udder bank so we connected zee distrubutor to both! "

With or without MFI, a vacuum leak in one bank of cylinders (or one cylinder) is not going to be read by a single sensor installed on the opposite bank. That's why Porsche uses the total vacuum signal from both banks (or at least from a few opposite cylinders).

John wrote:
"Weld the sensor bung in the muffler instead of in the heat exchanger. A lot easier to weld in that vs. stainless steel."

Not sure about the response time from a wide band sensor installed in the muffler as this is a distance from the cylinders. The reading would be an average of 6 cylinders. The best setup would be to monitor each cylinder. Next best is each bank of three. Assuming all cylinders have the same A/F, it may not matter for steady state throttle positions. But that's an assumption, and you know how that sometimes shakes out. Anybody tried this?

Sherwood

Sherwood,

I was considering having the bung welded in the muffler inlet, just aft of the SSI's output flange. That's only a couple inches downstream and before the body of the muffler begins. In fact, I could weld a second bung in the other side, to be able to switch the sensor back and forth to verify that the crossover was set correctly.

I once tried to drill through Stainless Steel and destroyed a couple bits. My thought is that the steel muffler is easier to breach and simpler for the average muffler shop employee to weld on.

Having said that I'm going to get in touch with James Achard, maybe he can help, being a certified welder and all . . .

Hi

I have danske polished SS muffler and ended up putting the bung in the muffler inlet (which is the only non SS bit!)

Neven

Neven,

Could I trouble you for a photo of the installation?

John

If you can wait, unfortunately the engine has been reunited with the car down country and I won't be seeing it again for 2 weeks

Actually its not unfortunate, this is close to the end of a 10 year project, I'm getting a great xmas present!

Neven

John, a high quality Cobalt drill bit makes quick work of stainless steel. Not quite like a hot knife through butter, but still pretty easy.

Jeff

I disagree completely. I have used this tool on many cars (incl Ferrari 308, Nissan Pathfinder, BMW 3-Series, and 911s), and the muffler location gives instant readings that correspond with readings from O2 sensors located upstream. You can monitor this by comparing the output of the LM-1 to the output by the OBD-II computers on post-1995 cars (like my BMW 5-Series).

Bottomline, for measuring and tuning, tailpipe is fine. Since you have two separate banks on your 911 engine, installing the O2 sensor on one or the other will only monitor 1/2 of your MFI system. It won't detect throttle-body imbalances, or clogged injectors on the opposite side. The side-to-side measuring difficulties are more pronounced with carbs.

-Wayne

Wayne,

“I have used this tool on many cars .…”

By “this tool” do you mean the LM-1 reading a WB O2 sensor? These are usually best closer to the exhaust port. Yes, you can buy the LM-1 tail pipe clamp for non-cat cars. I will yield to your experience that it reads the same as an upstream O2 sensor. Have you tried this comparison with a muffler contaminated with unburned fuel, oil, and carbon?

If you mean an Exhaust Gas Analyzer (CO & HC), I agree.
Many portable Exhaust Gas Analyzers don’t like exceedingly high temperatures for extended periods, such as on a dyno or road testing up a long hill. I made an ice water-to-exhaust gas heat exchanger to keep the temperatures under control.
I would love to have an EGA and a pair of LM-1s on a MFI 911.

John, Neven, Remember, an O2 sensor must be hooked up and running or it will get damaged. If you use two sensors, one for each bank, you either need two LM-1s or some method of powering the unused sensor.
Another issue is the sensor must be mounted somewhere in the 10-2 o’clock position, best is 12 o’clock, never in the 9-3 position, worst is 6 o’clock position.

Here is the manual for the LM-1:
http://www.innovatemotorsports.com/support/manual/LM-1_Manual_3.0.pdf

Sherwood, who manufactures the dual WB O2 sensor unit? Perhaps that is something Wayne might consider selling.

There is always subtle differences left-to-right on a 911. The CR can be slightly different, the cams can have slightly different timing, the cooling (or accumulated dirt) can be different, and of course the carburetors or fuel injection. I suppose it would be best to monitor each cylinder but side-to-side will yield an immense amount of useful information.

Best,
Grady

Wayne, Grady, thanks for the additional info.

I think I'll put two bungs in, one for each bank, and keep one plugged when not in use. This way I can switch the sensor back and forth to monitor each bank.

Grady, you must be tuned into my wavelength, as I was just reading that very passage from the manual this morning, where it says, don't mount the sensor in the bottom of the pipe. Given my tendency toward off-track-excursions I'm going to mount the sensor bungs as close to 12 o'clock as possible. I can just imagine going off and tearing the sensor wire out. . .

John,

Let’s do a little simple engineering here and design a switch for left-right. The sensors are relatively inexpensive. How is an O2 sensor powered in Factory configuration? The only issue I see is the LM-1 calibration cycle.

The advantage is tuning at “just off idle” and “very light throttle” at 3500 RPM.

How is your exhaust configured? Is there room to put a 3rd bung on the top of the tail pipe, inside the valance? That might be the best of both worlds.

Best,
Grady

one thing to look at would be a dual channel wb meter and run 2 sensors one for each bank, as far as the tail pipe goes with the short distances of the 911 exhaust your lag time will be minamal the lag time also has to do with what sensor you are using , ntk,lsu,ect they all have different response times.

"...... who manufactures the dual WB O2 sensor unit?"

Grady,
I thought I had a dual WB sensor vendor archived. I saw such a system in a recent issue of Race Tech magazine (new products section). I'll try to locate.

The following vendor offers a "dual, WB sensor system", but I don't know if this is standalone or used in conjunction with their fuel/engine management system.
http://www.3sxperformance.com/electronics.asp

Innovate Motorsports (LM-1) recommends no switching between multiple WB sensors in their single sensor system. The unswitched (and thus unheated sensor) needs power to prevent combustion byproducts from rendering it inaccurate or ineffectual. Don't know if there's a workaround this (e.g. a constant source voltage applied to the unswitched sensor).

BTW, it appears the term, "dual-channel" shouldn't be confused with two sensors.

Here's a list of single WB sensor vendors I've archived:
http://home.earthlink.net/~jcgebhart/msindex.html
http://www.widebandcommander.com/
http://www.zeitronix.com/Products/Products.htm
http://www.cyberauto.com/index.php/fjo_products.htm
http://www.techedge.com.au/vehicle/wbo2/default.htm
http://www.plxdevices.com/
http://www.innovatemotorsports.com/lm1.php

Sherwood

Hi

If you putting the LM-1 in an MFI setup, I can't see any great advantage in 2 online sensors, You can balance left to right by other methods (syncrometer) or if you wanted log dual MAP sensors in the intakes, There would be an argument for dual bungs in the muffler (I haven't bothered tho as the LM-1 only shipped with 1 bung) so that you could check the
LR balance. If they are out for any particular reason (cam timing, intake leak, faulty injector) then you are stuffed until you fix the problem. as youb can't correct the side individually

Neven

I was under the impression that the MFI systems that use vacuum advance use two runners (#1 and #4 IIRC) symply to get strong enough a vacuum. I run an MFI on a 2.2, without the vacuum, so I could be wrong, but I believe the vaccum is not drawn from both "banks" as such, only from two cylinders that happen to be one on each bank. Staggering the intake events may also have been a factor in choosing the cylinders to use, I'm sure.

FWIW, I've got the LM-1 and I installed the bungs at the very end of the headers, just ahead of the muffler flange. One bung per side. Initiallly, the plan is to use the LM-1 only as a diagnostic and tuning tool, one bank at a time. Later there is a plan to couple the LM-1 to an electronic box and a servo motor that would adjust the position of the main rack on the MFI pump. Neven is doing the heavy lifting in that regard.

Michel Richard
914 2.2E MFI

Hi Grady,

That's not possible. The WB sensors are not simply 'powered', but have a heater that needs to be controlled. The operating temp of the sensor MUST be regulated to 800 degC. The controller (LM-1) regulates the temperature to that. Otherwise the sensor will live only a very short time.
We (Innovate) are coming out shortly with what we call a 'Lambda cable' or LC-1 (prototype was shown at SEMA last week). This product has the WB controller of the LM-1 molded waterproof into the sensor connector itself. It has the same programmable analog outs as the LM-1 and also has 2 serial ports for logging. With the serial ports it can be daisy-chained with the LM-1 (serial out of LM-1 to serial in of LC-1, serial out of LC-1 to laptop or digital display(s) ). Up to 12 LC-1's can be daisy-chained.

Regards,
Klaus

Forgive my inexperience on this topic, but if your carbs/throttle bodies are balanced and your fuel supply is accurately accounted for (proper psi, having fuel injectors test for flow) and your timing is accurate, what benefit will you gain from measuring each bank?

If you have aftermarket engine management aren't the changes you can make with fuel supply and possibley timing is usually limited to the whole system and not each bank? With batch injection you can't control fuel to each indvidual cylinder.

It seems if the muffler location can give an accurate reading, as Wayne stated, it looks as if it is the way to go.

Klaus,

Cool. Is this up on your site yet?

Yes, I understand the issue of controlling the temperature. I should have used the word “controlled” rather than “powered.”

As I understand what you posted, you can now have up to 12 sensors (each controled by a LC-1) and the data and control daisy-chained back through the LM-1. I assume you go through the calibration process (heater & free air) for each sensor via the LM-1.

Good for you guys! Take a great product and make it better.

Is there a reasonable method of having an additional calibration point in addition to “free air”? Are there suitable calibration test gasses available? What would be the procedure?

Best,
Grady

Hi Grady,

It's not on the web-site yet. We just introduced it, but it will take a month or two before we ship.

No, each LC-1 is stand-alone. It can run completely independent. Calibration can be performed individually on each. In it's simplest form you can just use a LC-1 and an analog gauge. Or 2 LC-1's and a double gauge like Westach makes with 2 needles. You can still real-time log with LogWorks then. The LC-1 has a special interface wire for calibration. Just ground that wire through a pushbutton switch to initiate free air cal.


The LM-1 (and LC-1) have 2 calibration points already. Free air is one. The other is stochiometric. As the pump current is zero for all sensors at stoich, independent of sensor tolerances, there is no need for a separate cal point. Different to other widebands, the LM-1's measurement principle yields inherently a linear output. For that only 2 points are needed.
Calibr. gases are of course available. We are using lab quality cal. gases to verify and test LM-1's. A simple way to test Lambda 1.0 is to use MIG welding gas (Argon/CO2 mix).

Regards,
Klaus

Good question, I guess my answer is that's a lotta ifs.

Since you can't practically install a bung in each cylinder, if you measure each bank it allows you to verify your "Check Measure Adjust." One element of that regimen is to adjust the throttle rods side-to-side, and you are supposed to set them the same. But no measurement of the rods is perfect, and if you came up with an obvious lean-running bank it would allow you to isolate the problem, either to that bank or to an individual cylinder within the bank.

Beats guessing. With the sensor in the tailpipe you could have a rich bank and a lean bank and not know it.

Would I be correct in assuming the process of setting up two sensors would be more applicable for MFI, due to the complexity of the system as opposed to carbs or throttle bodies (efi), or it would apply to all because they all have throttle rods that need to be in proper adjustment?

Mike,

For MFI you have the issue that each throttle body has a throttle plate adjustment. The injection quantity, if your MFI pump is calibrated correctly, is the same for all cylinders. So the issue becomes one of adjusting the throttle stop and the air bleed screw to match the injection quantity on each cylinder.

EFI would have this issue, insofar as multiple throttle bodies also need to be synchronized, and then you have the issue of injector pulse width for your fuel quantity. So there are two variables.

Carbs still have the left-to-right issue, because of the cross link.

To add to John’s comments about carburetors, the “just off idle” movement of carb butterflies not only adds more air&fuel but uncovers various air-fuel transfer ports. This is the critical transfer progression from running on the idle circuit to eventually on the main circuit. Being able to isolate one bank (carburetor) or even a single cylinder is the “art” of carb tuning.

Best,
Grady

I appreciate all the info guys. I am starting my webers to throttle bodies project this weekend and now I have a little more to digest. I will be following this thread with interest.