EFI and exhaust leaks + sensor placement

Old forum posts ending on Oct 21 '09

Moderator: TechMOGogy

Locked
EvanH
Posts: 175
Joined: Wed Apr 14, 2004 4:14 pm
Location: Northern Utah
Contact:

EFI and exhaust leaks + sensor placement

Post by EvanH »

I spent some time last weekend talking to Jim LaGuardia about EFI and installation. One thing came to mind that others may be take for granted.

Jim's (and I presume Jake's) EFI systems are intended to be "closed loop" systems. This means an oxygen sensor is added to the exhaust, and the data from the oxygen sensor is used by the controller to fine tune the mixture.

The problem is, if you have an exhaust leak it will very likely impact the accuracy of the readings from the O2 sensor and therefore cause problems with the mixture.

I bring this up beause both of my Pinzgauers have exhaust leaks, and I would be surprised if all of the trucks slated for an EFI install are leak-free.

On a related note - are any of the major EFI players willing to discuss their findings for optimal sensor placement?

-Evan
Jim LaGuardia
United States of America
Posts: 1704
Joined: Wed Apr 14, 2004 3:42 pm
Location: San Bernardino Ca
Contact:

Post by Jim LaGuardia »

I put mine in the muffler inlet. My new muffler has a tap in each mufler inlet. I can run dual sensors or use the second tap for a PLM (portable Lambda Meter), the PLM is a stand alone O2 sensor and display that can be used as a secondary tuning aid.

If you are not using a heat exchanger and want slightly faster sensor response, then locate the sensor near the front of the pipe.
Cheers, Jim LaGuardia
http://smg.photobucket.com/albums/v510/Goatwerks/
"Arch Magus of Machines."
pinztrek
Barbados
Posts: 131
Joined: Mon Aug 01, 2005 9:27 pm
Location: Atlanta Area
Contact:

Re: EFI and exhaust leaks + sensor placement

Post by pinztrek »

EvanH wrote: The problem is, if you have an exhaust leak it will very likely impact the accuracy of the readings from the O2 sensor and therefore cause problems with the mixture.
This is absolutely the case, and will make the mixture appear artificially lean. But this should not be an issue if the sensor is properly placed in the header pipe.
I bring this up beause both of my Pinzgauers have exhaust leaks, and I would be surprised if all of the trucks slated for an EFI install are leak-free.
Placement is key. And if you have a header to exhaust port leak, you need to fix it. It's not that hard. :-)
On a related note - are any of the major EFI players willing to discuss their findings for optimal sensor placement?
o2 sensors are designed to be hot and in the exhaust flow near the head. Nearly all OEM sensors are in the header 12-24" from the exhaust port. The exception to this placement is post catalytic converter sensors on OBD-II 4 sensor systems which are not applicable to aftermarket efi. And even then they are heated sensors as it's not hot enough for them to operate on their own. And their front sensors are 12-24" from the port. :-)

running the sensor further back than that can occasionally work, but you set yourself up for the exhaust leak, slow response, & other issues. If you do run them further back you must use a heated sensor, otherwise at best you will occasionally get slow response, and many times highly inaccurate response.

Even at medium to high rpm's a sensor back around the heat exchanger or further can drop below operating temperature. The pinz has pretty effective cooling in that regard. The sensor is also lower (water issues), and less protected there.

The pinz header is easy to work with, and there is no good reason not to use the sensor the way it was originally designed. After trying 3+ different locations we have one we like that is well protected, easy to access, and works very well.

Any concerns over imbalance between cylinders needs to be addressed during plenum design, as mixing the exhaust streams will not compensate for that. (It actually makes things worse.) So once you have your system designed right, 1 sensor in one header pipe is all you need. (Which is also how OEM's did it since the inception of fuel injection)

Have fun,

Alan
pinztrek
Barbados
Posts: 131
Joined: Mon Aug 01, 2005 9:27 pm
Location: Atlanta Area
Contact:

Re: EFI and exhaust leaks + sensor placement

Post by pinztrek »

Dupe post due to forum crashes deleted
Last edited by pinztrek on Wed Dec 31, 2008 6:30 am, edited 1 time in total.
pinztrek
Barbados
Posts: 131
Joined: Mon Aug 01, 2005 9:27 pm
Location: Atlanta Area
Contact:

Re: EFI and exhaust leaks + sensor placement

Post by pinztrek »

Dupe post due to forum crashes deleted
Last edited by pinztrek on Wed Dec 31, 2008 6:29 am, edited 1 time in total.
User avatar
Jimm391730
United States of America
Posts: 1456
Joined: Wed Apr 14, 2004 10:58 pm
Location: Idyllwild, CA

Post by Jimm391730 »

The problem is, if you have an exhaust leak it will very likely impact the accuracy of the readings from the O2 sensor and therefore cause problems with the mixture.
OK, someone tell me why this should be true -- the exhaust system should always have a slightly higher air pressure than the atmosphere (the exhaust isn't being SUCKED out, it has to be pressurized to FLOW out). So any leaks would reduce the amount of exhaust (very slightly) but (unless I'm missing something) should not change the makeup of the exhaust gasses that the O2 sensor is measuring, right? I'm probably wrong in this, but I can't figure out why my thinking is incorrect.

Lastly, I have Jim L's system with a wide band heated O2 sensor mounted in the front of the muffler (bad location, I know, but was installed by the old SoCal people - 'nuf said). However, I am using Jim L's flex coupling, so I'm confident there is no leaks in the system. There is a very definite )2 sensor time delay being that far away (something like 1-3 seconds, depending on condtions) so I have my system running open loop: the computer does nothing with the O2 sensor except to give me a display on the laptop. I've optimized the mixture map as well as possible to keep a reasonable lambda across load and RPM and this seems to work OK. It obviously does not tweak the mixture in realtime for maximum efficiency (mpg), and in fact I know (hope) that I'm running slightly rich which is why I'm only getting 12mpg with a 7500 lb, 9' tall 712 and W box on the highway. But frankly, I'm pleased to be getting that!

The EMS system that Jim L uses does have a dedicated barometric sensor, so it compensates for altitude very accurately; my mixture does not change from sea level to 8,000 feet, even open loop.

My biggest improvement is in starting at any temperature (from 15F to 110F) and no distributor to get wet! (plug wires straight from coilpack to plugs are easy to waterproof in comparison to the stock dizzy).

Jim M.
712W and 710M
pinztrek
Barbados
Posts: 131
Joined: Mon Aug 01, 2005 9:27 pm
Location: Atlanta Area
Contact:

Post by pinztrek »

Jimm391730 wrote: OK, someone tell me why this should be true -- the exhaust system should always have a slightly higher air pressure than the atmosphere (the exhaust isn't being SUCKED out, it has to be pressurized to FLOW out). So any leaks would reduce the amount of exhaust (very slightly) but (unless I'm missing something) should not change the makeup of the exhaust gasses that the O2 sensor is measuring, right? I'm probably wrong in this, but I can't figure out why my thinking is incorrect.
Venturi effect http://en.wikipedia.org/wiki/Venturi_effect
results in most leaks allowing air to be sucked in. Like a compressed air paint sprayer. Or garden hose fertilizer sprayer. It's physics. :-)

Very common in emissions inspections, and detected as excess O2 in the exhaust. Since it dilutes the exhaust, it contaminates the readings.
There is a very definite )2 sensor time delay being that far away (something like 1-3 seconds, depending on condtions) so I have my system running open loop:
Even for being mounted that far away, the delay appears to be higher than I'd expect. The delay could be the result of a degraded/failing wide band sensor due to water contamination or excessive rich running. That's one of the early indications of failure. Or it's heater is not working, and you only see proper operation if the exhaust flow is enough?

I'd give these a scan for WB placement & troubleshooting: http://wbo2.com/lsu/position.htm and http://www.performancetrends.com/wide_band_uego.htm

Selective closed loop operation is one of the big EFI wins, it'd be a shame not to be able to see that advantage. You may want to try closed loop with a 3 wire (heated) NB sensor, they are only about $45-50.

Have fun,

Alan
MASSIVE PINZ
Posts: 234
Joined: Tue Sep 13, 2005 11:30 pm
Location: Cleveland Georgia
Contact:

Post by MASSIVE PINZ »

Any exhaust leak will cause O2 sensors to do odd things. I have even had systems that had no audible exhaust leak pick up air and introduce it into the exhaust, causing tuning issues.

I made the 02 sensor an optional item on my system, because it made zero impact on driveability and tune-ability in my direct experience. The truck also seemed to get the same MPG with or without it enabled, and all it did was add complexity to the install.

The Pinz likes to run at ricjher AFR settings than most other engines, even richer than the VW and Porsche aircooled engines that I deal with primarily. Thee engines generally make their best power and efficiency at 13-13.2:1 but the Pinz doesn't like tthat, it likes to run no richer than 12.5:1 in my experience.

In "closed loop" my system reaches to work at stoichometric AFR of 14.7:1, which other engines love (even my aircooled daily driver Porsche) but the Pinz didn't like it at all. I put 1,000 miles on my system in closed loop and gathered data, then did the same thing in open loop. The results were nil.

As people install and learn my system some may elect to go closed loop with the O2 sensor and that will be just fine, but for now the added complexity at install time and the added user interface is something we are avoiding. Simplicity is key. Streamlining the installation is mandatory, the more crap that has to be installed the faster the installer will lose their patience. When that happens its all over.

(I also had a hard time keeping exhaust leaks totally eliminated when using the o2 in my truck, even with all the hardware/tubing/flanges in very good shape, there was always an issue with some false indications)
Jake Raby
Owner, Raby's Aircooled Technology
www.aircooledtechnology.com
Jim LaGuardia
United States of America
Posts: 1704
Joined: Wed Apr 14, 2004 3:42 pm
Location: San Bernardino Ca
Contact:

Post by Jim LaGuardia »

Jake, good to hear there these guys will finally be enjoying EFI like myself and a few others.
I disagree with a few things you have stated on the subject due to my 30+ years as an automotive technician and training, but when in life does everybody agree on anything :lol: . I moved from wasted spark for more fuel efficiency and better accelleration, a move for me that resulted in improvement.
My stock worn out 2.5L runs very well tuned at 13.3.
Ok guys, get busy installing :!:
Cheers, Jim LaGuardia
http://smg.photobucket.com/albums/v510/Goatwerks/
"Arch Magus of Machines."
MASSIVE PINZ
Posts: 234
Joined: Tue Sep 13, 2005 11:30 pm
Location: Cleveland Georgia
Contact:

Post by MASSIVE PINZ »

Trust me, I generally loke 13-13.5:1 engines but what my data proved was this particular engine didn't like it, then again that could be my plenum which has about 35% more volume than the unit you are using.

In a couple of years, when people have the system understood fully and perfectly tuned advanced users can move to the O2 arrangement. I deal with users around the world and the one thing I have learned over my 20 years of engine development is that features can be less than attractive in some instances.

Wasted spark has always worked well for me and I prefer it over sequential in some instances. I believe the reason you saw a benefit with sequential is due to the position of your injectors and the turbulence that exists inside the plenum at some speeds that will be aggrevated by batch fire. My injectors are positioned in the runners and thats where I saw the biggest benefits and the least EGT differential between all 4 cylinders.

Zif I hadn't ever exposed my arrangement to the lab envoronment here I would have never learned most of the things that were included in this kit, the differences before and after the lab work in the same truck were phenomenal.

We are all different and our backgrounds, theories and experiences are all illustrated in what we create and thats why there is no right or wrong way to do any of this. I just went with what I thought was best, based on my direct interface (and data collected) with the engine and then wrapped that into a package.

What I am proud of is the fact that EVERY piece of that plenum and it's design was done under my roof :-)
Jake Raby
Owner, Raby's Aircooled Technology
www.aircooledtechnology.com
Jim LaGuardia
United States of America
Posts: 1704
Joined: Wed Apr 14, 2004 3:42 pm
Location: San Bernardino Ca
Contact:

Post by Jim LaGuardia »

Jake, just to be "Chrystal clear" my injectors do not mix in the plenum in any way shape or form, they, like your system go directly to each port.
It is nothing like the Draggus manifold you originally tested(#10) :shock:
Cheers, Jim LaGuardia
http://smg.photobucket.com/albums/v510/Goatwerks/
"Arch Magus of Machines."
MASSIVE PINZ
Posts: 234
Joined: Tue Sep 13, 2005 11:30 pm
Location: Cleveland Georgia
Contact:

Post by MASSIVE PINZ »

I saw those mods that you did a while back but was unclear how far they protruded.

BTW- The draggus plenum only has 40% of it's volume thats actually usable, but I tested two more as well, another of which I built from scratch with the same hands that are typing this message :-)
Jake Raby
Owner, Raby's Aircooled Technology
www.aircooledtechnology.com
User avatar
Jimm391730
United States of America
Posts: 1456
Joined: Wed Apr 14, 2004 10:58 pm
Location: Idyllwild, CA

Post by Jimm391730 »

Thanks for the detailed explaination of exhaust leaks and O2 operation. That does make sense now that you've helped me see past the very simplistic idea I had.
Even for being mounted that far away, the delay appears to be higher than I'd expect. The delay could be the result of a degraded/failing wide band sensor due to water contamination or excessive rich running. That's one of the early indications of failure. Or it's heater is not working, and you only see proper operation if the exhaust flow is enough?
Well, the delay may not be quite that bad, but it is noticeable to me on the computer screen as I adjust the mixture up and down under different conditions. Of course some delay could be in the ECU reporting it to the laptop, but that should be minimal. My engine seems happiest around 13.5:1, but as Jake mentioned I'm not worried about running closed loop; trying to tune the ratio a few tenths either way doesn't seem to make much difference in my power or efficiency (perhaps it could improve emissions somewhat more, but I can't tell).

End result: the EFI conversion is perhaps the second most impressive thing I've done to my truck, lead only by Jim L's 2.7l rebuild!

Jim M.
MASSIVE PINZ
Posts: 234
Joined: Tue Sep 13, 2005 11:30 pm
Location: Cleveland Georgia
Contact:

Post by MASSIVE PINZ »

4 cylinder engines have less cycle time in the exhaust system to maintain the reqired heat to make O2 sensors work at their best and live their longest lives. Other engines with more cylinders have problems with sensors getting too hot and some of those pull the sensor partially from the exhaust with a specially prepared tube.

This is one issue I had, as the Pinz primaries are HUGE for such a small bored engine and I could not get exhaust temps hot enough while being able to sample all 4 cylinders due to the merge points of the exhaust system.

The delay thats mentioned above is a serious reality and I can see it plain as day with my data logger and 5 gas arrangement when compared to the standard wideband O2 on some engines.

The Pinz runs very low exhaust gas temps, the lowest of any engine I have ever worked with and its obvious this was designed into the engine as the characteristcs mimic those of a steady state industrial engine more than that of any automotive engine- its perfect for a Pinz.
Jake Raby
Owner, Raby's Aircooled Technology
www.aircooledtechnology.com
Locked