tuning with a wideband for running with a nb.

kanenas · Post by **kanenas** » Wed Feb 22, 2006 2:23 am

My son lent me his wideband sensor (Zeitronix) and I have been tuning the old benz very nicely with it, using VExME and the MT logs. WB tuning is fantastic. But as the fine tuning progresses, I need to remember that in actual use it will be a narrow band O2 sensor that will be controlling the MS. So the question is weather it is better to fill the cruise part of the VE table with 14.7's or should the afr be graduated, say from 15+ down 14- as we move from 30kpa to the 80's range? Can the nb sample fast enough with adequate discrimination and can ms calculate ratios fast enough to respond to graduated tables, or is it a waste of time? or, should i just confiscate the wb sensor?
thanks,
d.

mops · Post by **mops** » Wed Feb 22, 2006 11:03 am

hm.... thats a good question,

I'd definitely like to see the educated reponse to it

i've been tuning my BMW and I think that once you got it all tuned, you could remove o2 sensor altogether. but i'd leave a narrowband there i suppose.

wes kiser · Post by **wes kiser** » Wed Feb 22, 2006 1:40 pm

That will work just fine, but you must be aware of the narrowband's limitations. One, when using a narrowband, you can't use an afr table. The only information it can convey is whether it is rich or lean of the setpoint (which has to hover pretty close to stoic). So at whatever point you want to close loop to a narrowband, the engine needs to be tuned to run close to 14.7:1 , because that is the value the megasquirt is going to try to bring it to.

whittlebeast · Post by **whittlebeast** » Wed Feb 22, 2006 2:27 pm

This question is going to be a little difficult to answer.....

A few points

1) At one point I had 4 different O2's connected to the CRX including two narrow bands 10" from the valves, a heated narrow band at the header collector and a wideband at the tail pipe. Long story short the vast majority of the tuning was done with one single wire narrow band high on the header driving the MS and the two stock leads going to the stock computer.

2) The ski (800cc, 7000 RPM, 145 HP beast) was tuned entirly with seat of the pant and no O2 at all

3) More than one O2 will give more than one answer.... 4 O2's will give 4 answers

4) Whenever I see strange results on anyone trying to tune a MS install the first question I ask is Are you SHURE of the calbration and output function on your wideband. Often the problem will be somewhere in that question. The inline Innovate Motorsports causes the most greaf.

5) Narrow bands tend to be very consistant and very fast and are just fine for most any car in the cruse ranges. I like to keep a motor at stoch at anything below 70 MPH and steady state. I watched the narrow band (we had a ricer round display on the A piller) for any hint of stoch when at full throttle.

6) A good $19 Bosch single wire that is up to temp is always better than a wide band with the wrong transfer function 100 times out of 33

7) When running anything with boost GET THE WIDE BAND AND READ THE DIRECTIONS 6 TIMES BEFORE YOU START IT

8 ) I try to get all noobees to start on narrow even if they have a wideband

(EDIT_2/22/06)
9) Autotune on a mis-calibrated wideband or on a mis-behaving narrow band 02 will take a motor right into lean detonation and destruction or go the other way and wash the rings right out. By design it was made difficult to setup to make people think before they continue. As with all parts of MS and all related software the guarentee is "If you ever melt down the motor or squirt we guarentee that you get to keep every part you can find when things come apart. If you dont like the guarentee, buy a new Civic and change the radio station all you want

Hope this helps and Have fun tuning.

AW

mops · Post by **mops** » Wed Feb 22, 2006 2:46 pm

yes, few good points here.

i never tuned on narrowband, but definitely I think tuning with narrowband would be simpler. I'd just get the motor to stoich at low loads/low rom and then just guesstimate/calculate the other fields. and if you want lean at cruise you can calculate ve's as you want form stoich.

it took me a while to get my wideband organized and setup to read correctly, but now it's all good

for boosted application i'd strongly suggesst EGT meter.

Mike_Robert · Post by **Mike_Robert** » Wed Feb 22, 2006 3:44 pm

GIGO theory applies here - Garbage In, Garbage Out. Andy's right - you gotta make sure your data is valid before committing expensive engine parts to the results of your data! Acquisition of accurate AFR *and* EGT data are pretty much an excellent idea with any boosted or high CR engine, IMO.

-Mike

whittlebeast · Post by **whittlebeast** » Wed Feb 22, 2006 4:50 pm

Sorry

I forgot to answer the original question. If the wide band is correctly calibrated, your MSQ and INI filed are correctly set and your callibrate o2 was done on the wideband and then all reset when changed back to narrow band ...... then rock and roll

Most of the OEM's have done it that way for years good enough for Ford and EPA...

AW

kanenas · Post by **kanenas** » Thu Feb 23, 2006 10:52 am

Thanks for the great comments, experts please keep them coming. Regarding wb calibration, a complete kit like Zeitronix "should" come calibrated, correct? Other than that, the only thing I can go by is a pair of logs, the "last" nb log and the first wb log, both were recorded using the same VE table. I did not make exhaustive comparisons, but a sanity check verified that the recorded ratios were not too far off under similar speed and pressure conditions, so it is more likely that both sensors are reasonably accurate rather than both being defective in a similar way.
It is very interesting that the nb seems very capable under "normal" conditions and that there is little to be gained using a wb on a mild mannered daily driver. That probably means that son will get his wb back, i guess he will be happy about that:)
So, as soon as i am finished with the brakes, i will punch in a bunch of 14.7's and redo the cruise portion of the VE table. But I am still curious about other real world experiences and theoretical comments, so please share your thoughts!
thanks,
d.

ChevelleFan · Post by **ChevelleFan** » Thu Feb 23, 2006 1:31 pm

kanenas wrote: It is very interesting that the nb seems very capable under "normal" conditions and that there is little to be gained using a wb on a mild mannered daily driver.

I would tend to agree. When I first got my car running on MS last summer, I used a NB to get the cruise stuff dialed in. Once I figured out MSTweak, I was able to get my car running nicely in a very short amount of time.

-Dave

T3Bunny · Post by **T3Bunny** » Thu Feb 23, 2006 11:13 pm

kanenas wrote:Regarding wb calibration, a complete kit like Zeitronix "should" come calibrated, correct?

DEFINATLY, DEFINATLY NOT! First with the widebands I have played (others could be differant...) you have to calibrate every sensor to the unit first thing. If you change the sensor later on you have to reclaibrate it again. Calibration is usally a few simple steps though. Thats part one...

Part two, and I think this is what Whittlebeast is refering to, is the calibration for your sensor through your MS unit and MegaTune. There are a lot of settings to be configured and its not uncommon for default settings to be close, but slightly off due to minor voltage shifts. AKA bad grounds... Even good grounds are rarely perfect though. So after you have your MS amd MT setup correctly, its still possible your going to need to make minor tweaks to the outputs from your wideband. On Innovate products, this is a configuration program that allows you to fine tune the output voltages.

DISCLAIMER: I run the supposed "worst of the bunch" for this issue. I run the Innovate LM-1. I haven't had a lot of direct experiance with the other brands of widebands. In fact I have removed other widebands when setting up cars to pop the LM-1 in... I tell myself the reason Innovate is so picky about this, is its "extra sensitive" and accurate...

Also, I always start my tuning with a NB sensor. Because of the way it works and ease of setup, if the car is a PIG to drive at first the NB is definatly the way to go. I have delibertly scrambled VE tables so bad a car would barely run. Give it an hour of dataloggs and focusing on just driving the car, then run those datalogs through MST3k, upload the new VE and BAM! On VW's, this first POORLY tuned VE table usally runs better than the stock Digifant EFI EVER did! And now we have AutoTune!

But to dial it in, I then swap over to a wideband. I like to lean out cruise VE's a bit for economy. This goes without saying, but if your running a turbo setup you will need to lock the turbo out of boost until your base maps are pretty well dialed in. Personally I will only tune boost with a properly setup wideband.

And I need to send props out to Andy Whittle for his patience with me when I first started! I have learned a LOT from him! Thanks Andy!