I'm doing some remote tuning through PMB for my injected 2056. The tune is close and I've been driving it but there are improvements that could be made now that I have had about a year and change since my last tuning session.

Eric suggested I move my O2 bung from the muffler (where I knew the location was not ideal but it got to sample all 4 cylinders) to one of the exchanger pipes (much better/hotter location, but only get to sample one cylinder.)

Tapping into the Cylinder 1/2 bank would be easiest, and I was thinking of going into Cyl 1 since the pipe is shorter and maybe a little hotter.

So now the question is what cylinder is the one I want to TIG weld the bung into. I don't track EGT so I don't really know if one pipe is hotter then another. And I want to eventually get heat back in my car so I can't go to headers which would allow me to sample all the cylinders before the muffler.

Thanks! Zach
@Eric_Shea

Zach, can you comment on what Eric told you or why you don't think your location is ideal?

I've got a Microsquirt setup with Chris Foley's EVO II stainless steel muffler. He pre-welds a bung right into the muffler. I seemingly had no issues with it when I was running D-Jet but for some reason when I switched to Microsquirt (and a different O2 sensor) my idle AFR readings were completely useless for tuning purposes. Worse than useless. Kept pointing me to super lean when I was super rich. At throttle it seems ok, and I know tracking type 4 idle AFR can be challenging (so says Jake).

Sorry, don't have an answer to your question, but I'm curious.

zach, what muffler are you running? i had mine put in the Triad at the back center pipe that goes from the lower to upper section so its sensing gases from both sides where they come together.

Phil

I had a Triad, with a bung in the exact same spot. I wanted something a bit quieter (and as it turns out a TON lighter) that did not give up on power.



I am using one of Ben's "Quiet" mufflers. The exhaust bung is in the base of the tail pipe. As long as there is open throttle, it works but is not ideal. The exhaust gas has cooled quite a bit by then. And there is reversion when I'm shifting or at idle which leads to false lean conditions. Latency was almost a full second. It was not a huge issue for the early rounds of gross tuning but it needs to be moved to someplace earlier in the exhaust system for more accurate fine tuning.

When I had Ben build the muffler for me, he did warn me that sampling all 4 cylinders with one bung would be a challenge with this muffler. IMO its worth it over the Triad I was using before. The noise is tamed and long drives are not nearly as fatiguing and leave my head throbbing even with ear plugs. And the Triad weighs about 2X what Ben's Quiet muffler weighs.

If push comes to shove I know I can monitor two O2 bungs with Microsquirt. I'd prefer not to do that, as I REALLY don't want to have to take apart the AMP Seal plug again.

Zach
@mb911

Don't put it in the muffler it's not hot enough for the lambda.

I'd put it in #2 as it's usually the most nominal operating P&C.

A better idea is to weld in a chunk of 1-1/4" to 1-1/2" tube between #1 and #2 pipe, about halfway between the muffler flange and the start of the "exchanger" outer shell. Weld your bung to the top of this cross pipe.

You don't have to cut the SSI pipes to match, just drill say a 1/2" hole into each of the SSI pipes.

Agreed. As I said I knew going in it was not an ideal location, but a compromise to get all 4. Not all of my ideas are great. Just ask my ex-wife. ;-)



That's in interesting idea. Would enough gas make that hard 90 bend to give an accurate reading though? You want the sensor in the main gas flow as much as possible.

Zach

As stated, reading from a single pipe is still not an ideal situation either as you want it as close as possible after all 4 pipes merge. So while you will get better response in the exchanger pipe you are only getting 1/4 of the exhaust.

As you really only need the wideband output while you are tuning provided we disable closed loop operation when done, I recommend just running a different muffler temporarily while tuning.

I have recently learned from discussions with a "techy" at INNOVATE MOTORSPORTS, that the Bosch O2 sensors should be mounted in 1" long bungs (which they supply in their kits. I will be modifying the 1/2" bungs in my twin headers (911 engine). This will place the tip of the sensors just inside the header pipes - minimizing protruding into the gas flow.

O2 sensor placement is a lot less of an issue when just outputting to a gauge. Only becomes a noticeable issue when you start increasing the precision of a system. I have found idle tuning is best done by aiming for best vacuum rather than a specific mixture and that ultimately the best mixture is specific to the motor. Though being able to see the mixture is still helpful as a reference.

@JamesM

James! How did I miss that you were the JamesM on 914World!
I'm very glad you chimed in as I clearly missed the thrust of your email to me on this.

Should I just put the Triad back on for your tuning? Is the bung in the center pipe sufficient? I don't have access to a bursch with a bung for testing. I have to think the gas in the Triad will also be too cold - but it won't have the reversion issues.

Zach

IMO it won't matter as there will be a flurry of crossflow. Make the drilled hole a bit bigger if you like.

I'd use square tube, easier to weld the bung to a flat surface.

Tuning with the muffler that will be used every day is ideal. Its funny there was another thread recently about how all of mu mufflers should come with o2 bungs in them.. I felt it unnecessary as it is not an accurate way to measure.. It really should be in the heat exchangers.. The alternative would be in the inlet of the muffler however that would impede flow without doubt.

I cant say for sure as I haven't ran that setup yet, but in theory the Triad should produce better results. The glass packs used on the Triad are pretty wide open though and the pipes are pretty large so the chance still exists that fresh air might be getting pulled far enough back into the system at idle to screw up the readings. Wont know until someone trys it.

If anything I am hoping it will at least decrease the response time for the O2 reading. right now im seeing over 1 second between a change in injector pulse width before the corresponding change in O2 reading shows up in your datalogs. I actually disabled closed loop operation in the last update to your tune that i sent this weekend for that very reason, to much lag in the sensor reading to be of any value. Dont want the engine attempting to adjust mixture real time off a reading that is already over a second old.

No muffler swap on 914 heat exchangers makes enough of a difference to the motors fuel requirements to matter. If it did anyone running d-jet with aftermarket mufflers would be screwed.

In theory putting the 02 bung in an exhaust manifold would be ideal but 914s don't have single manifolds that join all the cylinders so there is no option but to compromise. Options are to put it in the heat exchanger pipe and only measure 1/4 of the gases or put it later in the system.. Wideband O2 sensors have integrated heaters so distance from the head from a heat perspective is not a problem, only becomes an issue when its to near the exhaust exit that fresh air gets back in and messes with the readings. Generally you want to be as close as you can get it after the pipes merge and before the muffler. The Bursch mufflers are probably as good as you can get with stock heat exchangers for O2 placement. the Triad looks promising, its just unknown atm how much fresh air gets back in with those.

I've tried different locations for the Lambda Sensor with a muffler which is quiet simmilar to Ben's one I think (Mittelmotor).
I don't have pics, but I have mounted one sensor in the center of the Muffler and one sensor at the right sidewall of the muffler. They read exactly the same under all load conditions and might be 0.1 Lambda off at idle conditions.
Center of Muffler is also not to cold...

@Rob . i think the rootcause for your lean reading is your cam.

Will post some pics of the Sensor location tomorrow...

@mb911 - I never said that all your mufflers should include an O2 bung. I simply asked if your mufflers came with one. That's all.

I have a Triad also and I'm about ready to do FI. Like you I realize that the muffler is not an ideal location for an O2 sensor. And I also would like a quieter muffler than the Triad at this point in my life. Keeping an eye out for an alternative.

Subscribed to this thread.

Sorry @Gint . No worries was just pointing out how it is not ideal on a type 4 muffler..

NP at all. I don't know of a decent solution at this time though, short of using two O2 sensors. I'm very interested in this thread at this point in time.

I had two wide band meters on my /6 with carbs, I'm done tuning and one is off the car now. If anything I found it confusing and ended up only looking at the one meter at a time. Basically they read the same and if your engine is sound they should all be the same.
Since I have carbs and I don't really like the look I'll likely remove the other meter as well. If you're carbs, L-jet or D-jet there's no real reason to leave it on after tuning. Aftermarket EFI leaving the WB installed depends on if you run open or closed loop.

That's a fair point Mark. That had occurred to me that unless something was really wrong, one side should be an adequate representation of all. But still...

Why?
I would agree to the statement for a Narrowband Lambda Sensor as they are not temperature controlled.

Is there really somebody still using Narrowbands?

From my data there is almost no difference between the sensor location at the inlet of the muffler to the center where it is much cooler.
Also the Widebandcontroller is indicating a good operation temp. range.

In my case the issue isn’t temps so much as reversion. Fresh air going into the muffler indicating a false lean condition. The other issue is latency. The farther the probe is from the heads, the longer it takes to report back to the ECU. In my case that latency is almost a second which makes using the wideband to fine tune almost pointless.

Zach

14point7 (Alan To) disagrees with you, but what does he know he only builds wideband controllers.

Where is that said? Not trying to be argumentative, but curious. I'm interesting in reading his reasoning...

I'm not seeing a thermal issue either, assuming the sensor is hitting correct temperature setpoint. Sure, you're likely to see a little less free O2 further out (conversion from CO to CO2 takes some time), but other than that I can't really see an issue. Obviously, reversion is an issue as you're introducing free air into the muffler, but this isn't something that always happens.

In the manual.



He talked/wrote about this in greater detail but I don't have time to look for it.

Kind of interesting that the LM-1 use to have a tail pipe clamp.. Not sure if they still do? Anyway good discussions for sure.. Maybe a real set of 4cyl heat exchangers should be made with a collector that would allow a 911 style muffler be fitted to it and o2 sensors in the collector's

I've often thought that something like a Eurorace but with exchanger jackets would just the ticket. Chris is the only one I know that experimented with jackets for his headers. Not sure if he still offers that or not, but IIRC the boxes were small and only allowed limited heating. He was advertising them as good for keeping the windshield defogged and crisp fall/spring drives. Full jackets would probably require the exhaust to be removed for valve adjustments though. Would be tough to keep them equal length without blocking the oil change bung or valve adjusters.

That said... If you build it they will come...

Zach

The 14point7 product idicates a good operating temp. in the middle of the exhaust.
I was just sharing what I have meassured, there is not nuch to disagree on that data. But if his point is Sensorlifetime the Heatexchanger is the goto location as you don't have to deal with to much condensate there wich could cause sensor problems due to thermal shock...

Hi Zach, I think your cam is killing the meassurement at low load, not the exhaust.
If delay is your problem, whay not using the lambda delay table? It's made to overcome the time delay problem...

Not at all far enough up that learning curve. Which is why Eric and James are doing the remote tune.

Zach

Good points, Frank.

In my case I had a 2056 and a 2270, both with a Raby 9550 cam. On both I used the Tangerine EVO II muffler with the bung welded right in the middle.

On the 2056 with D-Jet my AEM AFR gauge seemed to provide valid idle AFR's.

On the 2270 with Microsquirt using the PLX wideband sensor/gauge I get crazy lean idle numbers. It took me a while (and a lot of black smoke / smell and piss poor mileage) to realize this wasn't the case. I know it was calibrated correctly, so that wasn't it. Because of these readings I really can't go closed loop for tuning purposes.

Weird. It could be aggravated by the cam, but on the 2056 I got totally different results.

Data point: I use a collector-mounted heated wideband for my Microsquirt installation.

I don't plan to tune to the nth degree, nor do I suspect my cylinders are running radically different. Plus, I'm batch processing and wastred spark so even if I find the cylinders are different ain't nothing I can do about it. So collector-mounted is fine for my needs.

Experience: I've already replaced one LM-2 wideband sensor due to suspected overheating (!) and I have purchased the 14Point7 sensor/controller to replace it.

With the 2270 you should still be able for closed loop EGO correction, as you can chose when it should kick in. Above a certain engine load point it should be fine.

Here with delay the pics I had promissed:
Click to view attachment

Click to view attachment

Bung here is right in the middle of the muffler.

We're getting excellent reading in the Triad center pipe in Greg's car.

@JamesM is doing the remote tuning. Eric is sipping cocktails watching. @Frank S I believe his bung is out by the tip. Maybe Zach can post a pic.

This and Marks point about over-reading data for most T4 models and applications is why (bringing it back to the original post) I recommend the warmer #3 cylinder HE tube.

When I have a moment I'm going to put the Tirad back on. Maybe Monday since I think I have off work. Really hard time of the year for me since I'm in the woods so much.

Zach

Under some conditions I was seeing a delay up to 1.5 seconds between a pulse width change in the logs and the associated change in AFR at the O2 sensor. A lot changes with engine operating conditions occur in 1.5 seconds. When that is combined with the false O2 readings under lower RPM and overrun conditions you wind up with a lot of pretty meaningless data, specifically when trying to tune anything sub 2500ish rpm in this case given its rare to record steady load/throttle position in datalogs that low in the RPM range.

Tuning in person or real time it would be less of an issue as there other ways to determine what is going on with the engine in addition to make quick trial and error adjustments, but when attempting to tune entirely from logs you need the data to be able to know what is going on.

If reading only #3, will #3 be richer or leaner than the other cylinders? My guess is richer since it is the hottest cylinder

I never tune with data logs, I only tune in real time, but I have SDS. I can totally fuch up my fuel values, look at my WB, turn my rich/lean knob to the correct value, look at the knob value and correct the fuel map. I can do this in about a minute for each value point, 30-45 minutes I can tune SDS to about 90-95% done, just driving down country back roads.

I've never had or needed to tune a SDS system on a dyno. I broke in my T4 2600cc nickies engine on a totally untuned SDS system, I just started it up, set my baseline timing, ran it up to 3k rpm, turned the rich/lean knob to 12.7-ish AFR and let it sing.