Misc: Everything Else Air Fuel Gauge 02 Sensor Lambda Sensor

Monkeypants

Well Known and Active Forum User
Non-VOC Member
I haven't read thru this entire thread yet, but I did see mention of John McDougall's bike with its O2 sensors up close to the exhaust ports.
Those are narrow band sensors which rely on exhaust heat. They have no heating element of their own so must be up fairly close to the action.
When fabricating the pipes for my 1360 I included an O2 bung for a wide band sensor just after the 2 into 1 collector. John's info was that with all of the O2 related tuning he never found the need to use different jets on one cylinder vs the other.
So the single hidden bung should do the trick.
Now an admission.
Ive never gotten around to checking the AFR and I'm not sure that l will.
The jetting was as suggested by Terry Prince for that motor fitted with Dellorto 41 pumper carbs. It seems spot on, which is interesting because Terry also mentioned that at that time no one had assembled an engine at that spec.
I think he based the info on Vincent 1200s with the same Manx style combustion chambres which he or his customers had built and fitted with Dellortos.
Making the engine bigger actually meant reducing the mainjets a little from a 1200cc engine size, which I did.This makes sense if you wade through Bernoulli's work.
Somewhere in there he explains that the flow in a Venturi increases at the square of the flow thru the main body.


I thought that if anything the engine might still be a tad rich, however on an easy cruise it managed 60 mpg Imperial, just a shade better than my standard 998 Rapide.
That was a pleasant surprise. I had made an assumption that with the large displacement high performance carbeuretted engine, fuel mileage might not be great, maybe something similar to to Buick Wildcat numbers.

It's feels good to go Green!

Glen
 
Last edited:

BigEd

Well Known and Active Forum User
VOC Member
VOC Forum Moderator
Don't shoot me, as this post is not about the topic! This a great post by Glen. Honest, informative, a bit technical but not too technical for the average punter and nice humorous comment at the end. Thank you Glen.:)
 

Magnetoman

Well Known and Active Forum User
VOC Member
Warning: Vincents use Amals, and this post deals with jetting an Amal aided by an air/fuel meter, but other than that it has nothing whatever to do with Vincents

I've been tied up with other projects the last couple of weeks but yesterday returned to jetting. The background is I'm trying to determine the minimum modifications needed to make a 2-stroke 1000-Series Amal Concentric work with a 500 cc Gold Star.

Yesterday I only had time to make a quick jetting run from which I determined the AFR was 10:1 at full throttle with its 290 main jet. From that I calculated to decrease it to get to 12:1 would require dropping the main jet to 10/12 x 290 = 241.7. Or, that using a 240 would give 290/240 x 10 = 12.08:1. However, after looking at today's results, which I'll discuss momentarily, I looked again at yesterday's, shown in the first graph. The fat green line shows where an AFR of 9.4 would be. Note that I was at full throttle yesterday for only ~2 sec. so it's possible the thin green line would have reached 9.4 in another second. Make a mental note of that.

OK, turning to today's results, and remembering that maximum engine power should be achieved when the AFR is in the range ~12-13, the second graph shows that with it held at ~1/8 throttle the AFR was 12.3:1. So far, so good.

The third graph shows that at ~1/3 throttle, where basically every variable in the carburetor is contributing, the AFR drops a bit to 12.0:1. Still, so far, so good. However, the graph also shows that at full throttle it drops to 11.0:1 despite my above calculation that it should have been 12.08:1. I'll return to this in a moment making use of the mental note you made two paragraphs ago.

The fourth graph shows that the second full-throttle run today resulted in a slightly higher 11.4:1 than the first run. However, note that it took 2-3 seconds for the ratio to drop to that value.

I used 10:1 from yesterday's run to calculate that I should use a 240 main jet today. Referring back to the first graph, if instead I had been able to hold full throttle for 2 sec. longer, and if the AFR had dropped to 9.4 in that time, then today with the 240 main jet the AFR should have been 11.36:1 at full throttle, which it was. That is, within experimental error the calculation I did yesterday to arrive at 240 gave the "correct" answer. However the "experiment" I did upon which I based the calculation wasn't run for long enough to achieve the steady state AFR value. The important point from this is that not only does the air/fuel meter tell you the AFR with the current configuration of the carburetor, if you run the tests for more than ~2 seconds it quantitatively tells you what changes to make to achieve the desired final result.

In light of today's longer full throttle runs, and using 11.0-11.4:1 as the values they gave, changing to a 220 for the next run should give 12.0-12.4:1. The smaller main jet also will increase the AFR somewhat at partial throttle resulting in values at all throttle settings falling in the sweet spot of 12-13:1 that gives maximum power.

As an aside, the jetting I used yesterday and today felt perfect at all throttle settings and rapid changes in throttle positions. At least, to within my ability to run full throttle tests on a Gold Star in an urban setting. Without the Innovate A/F meter I would have been happy with yesterday's 290 main jet, or with today's 240, even though it looks like 220 will be the one to use. Also, all of these settings are rich compared with stoichiometric 14.7:1. If I were tuning for economy rather than performance I'm not sure I could achieve it across the range with the present 2-stroke spray tube and air compensating passage that are part of all 1000-Series Concentrics (although, both can be changed). Certainly not with a .106 needle jet.
Innovate24May2019_01.jpg Innovate25May2019_01.jpg Innovate24May2019_01.jpg Innovate24May2019_01.jpg
 

Attachments

  • Innovate25May2019_02.jpg
    Innovate25May2019_02.jpg
    449.7 KB · Views: 7
  • Innovate25May2019_03.jpg
    Innovate25May2019_03.jpg
    379 KB · Views: 6

Cyborg

Well Known and Active Forum User
VOC Member
Thanks again for posting that. It will (or should) help with my quest to jet the Comet’s flat slide. Unfortunately, I will also be tied up with other projects for a while.
1289F5EF-BCBD-4099-8D93-C3DD22E5E987.jpeg
 

Magnetoman

Well Known and Active Forum User
VOC Member
Again, Vincent content of this post is tangential. My goal was to see if a 1036 Concentric could be made to work well on a Gold Star without changing the spray tube or drilling the compensating air passage. That is, I wanted to determine if it could be made functional with only the simplest swapping of a few screw-in components in one of these 2-stroke bodies.

I made A/F measurements under two sets of conditions, "static" and "transient." I made the former by holding the throttle in a fixed position for ~5 sec. before changing the throttle to a different position. For less than ~1/3 throttle these were made on relatively level ground as well as when going uphill and downhill. Higher throttle settings only were made going uphill to keep the speed within reason given the conditions.

"Transient" measurements were made by snapping the throttle open. I use a 1.55 V battery in my throttle position sensor so the first graph shows a 15-sec. section of the data where I was in 2nd gear at ~1/3 throttle (the red curve), then snapped the throttle closed and back open, shifting successively into 3rd and 4th where I finally gave it full throttle. The magenta A/F curve shows brief lean pulses occurred roughly 1/2-sec. after each of these throttle movements, but none are excessive and I felt no hesitation from the engine. In fact, I couldn't be happier with how the engine felt throughout this entire run.

As can be seen from the first graph, even in regions where I kept the throttle constant the A/F curves fluctuate by ~+/-0.5 around a mean value. Added to this is the mean value at a given throttle setting at different times during the run (i.e. possibly under different conditions of, say, uphill one time and downhill another) could vary by nearly that much as well.

Data from two runs, with nominal '220' and '200' main jets, are plotted on the second graph. Although I watch the bar graph display of A/F ratio as I ride, it takes the better part of 2 hours after each run to find on the stored data the one or two-dozen time intervals worth examination, copy them to Photoshop, format, print, and then mark the AFRs and measure the corresponding relative throttle openings from the 0-1.55V signal from the throttle position sensor. I don't think I could have guessed what this curve would have looked like from just having observed the real-time bar graph display.

The first thing to note is changing from a '220' to a '200' should have leaned the previous ~12:3:1 at full throttle to 13.5:1. Instead, the mixture became richer by the equivalent of having changed a 220 to a 225 rather than a 200. I've mentioned several times over the years that I measured a number of '500' Amal main jets on my flow bench and found 25% of them to be as much as 3½ sizes too large or too small. The other 75% weren't perfect, having a spread of almost +/-1 jet size. Anyway, as these two jetting runs show, the '200' in it now is somewhat richer than the '220' that was in the first run.

I don't know if the '220' is smaller than its marked size, or the '200' is larger, but they're close enough that it makes sense to plot both sets of data on the same graph. Unfortunately, to measure the relative flow of the two jets would require me to dismantle the carburetor fixture currently on my flow bench, which I'm not inclined to do because I have more measurements I want to make. After plotting all the data I examined the outliers to see if I had made any mistakes, or if I had extracted any of them from graphs where I hadn't allowed the mixture to stabilize for ~4 sec. This eliminated three potential data points, but the three x's between 0.4 and 0.5 are real.

The scatter of the data points isn't due to experimental uncertainty in reading the values from the graphs, it is real. The same throttle setting under different conditions (e.g. accelerating on level road vs. climbing a hill) gives somewhat different AFR readings. Also, after the first run the AFR at idle was 12.2, but after the second it was 11.5 even though I hadn't touched the mixture screw in the meantime, and even though the idle "felt" the same.

For throttle settings between 0.2 and 0.5 the mixture is slightly leaner than the "optimum" 12-13 for maximum power. Raising the needle another notch isn't an option because that would make it much too rich in that region. However, there are two reasons I don't think this region is an issue. First, the 12-13 "rule" is just a guideline, and only time spent on a dyno would determine if more h.p. could be extracted with slightly different mixtures. Second, most time riding is spent cruising at mid-throttle so having a slightly leaner, but still rich, mixture will save a bit of fuel without costing any actual performance.

Trying to find reliable information on the "perfect" air/fuel ratio for gasoline, let alone E10 or E15, quickly takes one down a rabbit hole. If information is correct that 13.5:1 is the ideal ratio for maximum power then the present ~12:1 robs me of ~2% at full throttle. However, if 12.5:1 is the correct ratio, my current jetting essentially robs me of nothing other than the price of having to refuel 4% more frequently if I ride all the time at more than half throttle. On the other hand, if much of my time will be spent between 1/8 and 1/4 throttle the current jetting will provide better economy at the cost of only ~1% of the power in that range.

The performance with both the '220' and the '200' main jets felt perfect from idle to full throttle, as well as when snapping the throttle open from various initial positions. So, this work shows it is possible to configure a "2-stroke" Concentric to work well on a 4-stroke machine. Like, say, a Comet
AFR_8June19_01.jpg
AFR_9June19_04.jpg
...
 

Cyborg

Well Known and Active Forum User
VOC Member
Interesting.... the random relationship between the jet “size” and actual flow rate makes me wonder even more about the aftermarket jets that I purchased for the Mikuni. Have you ever had any Mikuni jets on your flow bench? Given the fact that there won’t be a flow bench in my future, I wonder if one could make a simple setup to check flow.

The .7 change in AFR at idle..... although not enough to feel any difference it still makes me wonder what caused it. Worn slide....global warming...as in all the good stuff vacated your gas tank in the Arizona heat?? It’s an Amal?
 

Magnetoman

Well Known and Active Forum User
VOC Member
Have you ever had any Mikuni jets on your flow bench? Given the fact that there won’t be a flow bench in my future, I wonder if one could make a simple setup to check flow.
No, I've never measured a Mikuni jet, and there's no simple way to check flow so you just have to cross your fingers and hope the jets you bought flow what is marked on them. I've thought of flow testing all of my Amal jets so I can sort them in to compartments by actual flow rather than by stamped number. But, that's unlikely ever to happen.

As a digression, a high-born Indian colleague (son in the line of the only Indian ever raised to hereditary British peerage) told me a story of his time spent receiving his Ph.D. from Cambridge. His possibly apocryphal (or possibly autobiographical) story was that a high-born Indian joined a research group at Cambridge and, in what is a pretty common practice, was given an initial menial job to begin familiarizing him with the lab. He was to sort loose resistors back into appropriate drawers, but when the professor returned later in the day he saw the resistors still strewn across the table. When the professor asked why he hadn't sorted them the student replied "the boy never showed up." I'm waiting for the boy to show up so I can supervise him doing the work to sort my jets.

The .7 change in AFR at idle..... makes me wonder what caused it.
I actually don't think that is significant. The carburetor has less than 500 total miles on it (so the Amal is less than 50% worn out...). The fact the mixture initially goes rich as soon as the throttle is cracked open means something like only a slight difference in tension on the throttle cable would account for the different idle readings.

However, what is significant is I installed a different main jet that should have been significantly leaner but instead was slightly richer. That's why it would be nice if the boy would show up to sort my jets so when I want a jet with greater/lesser flow I would know that the new jet actually flows more/less than the one it is to replace.
 

Magnetoman

Well Known and Active Forum User
VOC Member
I just finished modifying and jetting two Amal Concentrics for a Gold Star, requiring approximately 15 thirty-minute jetting runs totaling over 7 hours, covering ~200 miles, and resulting in ~160 pages of printouts of AFR data from my Innovate LM-1. At the risk of repeating a few things I've written before, some observations from this experience, in no particular order, are:

-- the bike "felt" like it ran fine for AFRs anywhere in the range ~10-15:1 so anyone who can correctly jet a carburetor by feel alone is a far better man than I.

-- watching the road while also mentally averaging the fluctuating AFR display and noting the throttle position would be hard enough, but to memorize just these two values for the at-least dozen throttle positions required to determine the full range of jetting is beyond my capabilities. Luckily, the Innovate LM-1 recorded the AFR, throttle position and rpm, freeing me to concentrate on potholes and possible sheriff's cars.

-- AFRs vary significantly depending on the conditions, e.g. accelerating up a hill, traveling on a level road, etc. This means that I actually needed more than a dozen readings to reliably determine the overall average shape of the AFR vs. throttle position curve for each configuration (hence the 160 pages of printouts).

-- because the AFR depends on conditions I will be adding an accelerometer to the "instrumentation package" in the future to make it easier to distinguish between AFR readings taken when accelerating, constant speed, or decelerating. In principle the rpm readings would be sufficient for this, but when analyzing 160 pages of data anything that makes the task easier is welcome. On this note, I can't imagine having done this without my Gerber Variable Scale and TI59 programmable calculator (vintage 1977) since I needed to normalize readings on the printouts from the throttle position sensor to the 1.56 V at full throttle.

-- it took a minimum of 15 sec. at full throttle while heading up a hill before the AFR stabilized at its final value. Attached is an example. I was going over 75 mph at the end of those runs which means during that time I covered ~1/4 mile. A shorter run reaching a lower speed would have given an erroneously high AFR reading.

-- at lower throttle settings the speed stabilized reasonably fast but I couldn't count on the AFR to stabilize until ~5 sec. had passed after I had reached that constant speed. Sometimes the AFR stabilized more quickly than this, but my default time at each throttle position was at least 5 sec.

-- a decade ago I measured a dozen new and used Amal jets of the same size on my flow bench and found 75% of them to be within +/-1 size of being correct. Unfortunately, the other 25% flowed too much or too little by as much as 3½ sizes. At one point when determining the jetting for the 1036 Concentrics I installed a jet marked two sizes leaner than the one that had been in it but the air/fuel gauge showed the nominally-smaller jet actually flowed a half-size richer. This was not the only time I ran into jet marking issues during this work. Keep this in mind if your jetting becomes too rich or too lean when changing to a jet whose marking makes it seem it "should" be correct.

-- wear or mis-manufacture of a needle jet by only 0.0005" is enough to completely wear out the jet (or make it too lean, if manufactured slightly too small). The needle jet in my Catalina Gold Star experienced that much wear during a 1200-mile ride so the needle jet definitely is not a set-and-forget component. As an aside, although the needle jet I used for my recent efforts is stamped 106' I previously had measured a large number of NOS jets using two different calibrated bore micrometers and found them to be 0.1065". This is how they were originally listed by Amal.
FullThrottle_240Main.jpg
 

Magnetoman

Well Known and Active Forum User
VOC Member
Cyborg said:
I'm feeling quite smug now that I've one upped a experimental physicist.
I would rather stab myself in the eye with a fork than attempt to analyze 44 minutes of data.
Tormented for over a month by knowing Cyborg had a newer model AFR sensor than me, I had no choice but to address that unacceptable situation. So, in the spirit of nothing exceeds like excess, I upgraded my perfectly fine AFR system to the next level. In addition to having the latest Bosch wide-band sensor, I'm now instrumented to log, for up to 566 hours (vs. 44 minutes) continuously if so inclined, the AFR, throttle position, rpm, acceleration, and exhaust gas temperature. I have one unused position on the data logger so could add road speed with another sensor, but that would be excessive. Or, would it? Hmm... Oh, and I have a separate, Canadian-made, aftermarket knock sensor controller I'll be testing to see if it provides useful information. It turns out the head-steady lug on a Gold Star couldn't be a more perfect location for mounting the sensor even had it been specifically designed for that purpose.

This weekend I installed the new "instrumentation package" on my Gold Star Catalina, whose Monobloc I already had jetted to be perfect using the old school method of "feel." However, in addition to testing the new instrumentation, this week should see the jetting get even perfecter.

Unfortunately, no instrumentation upgrade comes without collateral damage. Now I'll have to spend time learning how to export the 566-hours of data to Excel as well as write a program to plot AFR vs. throttle position rather than slowly extract this information manually. I already know from my previous measurements that the AFR at a given throttle position depends on whether the bike is accelerating under load, at a steady speed, or decelerating on a downhill stretch, so I'll want my Excel program to incorporate data from the accelerometer. A scatter plot of AFR vs. throttle position is exactly what is needed to understand the jetting over the full range of operation. See what you've forced me to do, Cyborg!

The weather looks like it will cooperate for a few jetting runs this week by dropping to the low 100s. Unless something gets in the way I should be able to make my first runs on Wednesday when the high will be only 101 oF.
 
Top