With 75,000 miles on my K75RT I am starting to see evidence of a rich fuel mixture in my plug readings and would like to see what can be done to clean things up a bit.  I know it's not oil fouling because she uses virtually no oil(1 qt/3000 miles).

Doing a valve adjust this week, then a good tuneup and off to find someone who will do a sniff on my exhaust.  I have a spare AFM and think it might be useful to do your mods to it.

I need to resume all I've done over the last week or so and it's only one device so I can only report one off results. I have not done a scientific study of exactly which element of the changes achieves what by isolating each variable with a full bank of tests.

Having said that....

1. Moved ceramic substrate by elongating fixation holes to push metal base closer to casing wall. ( plug side ).This gets a nice clear, wear-free path.
2. Reset copper wiper zero position. Initially set to corner of first green resistor rectangle, then fine tuned by road testing different settings.  Not sure how to quantify this other than road testing yet.
3. Using AFM bypass screw to set idle mixture as discussed above.
4. Some experimentation with ign. timing.

Output monitored by applying 6.0V between pins 5 and 8 and measuring pin 5 to 7. I ended up with about 0.5-1.0 mm of movement at the tip of the vane before the output changed. I don't think that is a figure to aim for, just a guide. I think performance under road test conditions has to be the definite criteria to tune for.

The wiper position does provide a means to fine tune the mixture across the range in a more subtle way than jumping teeth on the spring adjustment wheel. I suspect some tuning is in order after 20+ years of service. The main motivation of opening the unit was to get a clean track from a worn unit and recalibrate the wiper position afterwards.

I would recommend doing the torque test before moving anything so as to have a reference curve to refer to.That also provides a test of smoothness and linearity to quantify the condition of carbon track under working conditions before deciding whether to disturb it.

That's elevator version.

Thanks!  So you reduced the deadband a bit is that correct?  It seems like the deadband would cause a bit of a lean-out coming off of closed throttle. 

Being a lazy git, I am thinking of how I can streamline some of the torque testing.  I am thinking that it might be possible to designate a window where a specific torque delivers a specific output voltage.  Also thinking about how there is some way to use door position in the tuning, especially at the fully closed and beginning to open positions.

Need to think about putting together a test setup.

So you reduced the deadband a bit is that correct?

Hey you said you wanted it without commentary, now you want explanations.

You should find plenty of that in the flow meter thread. You could save some time by reading just the last page where I'd got most of my information and the methods sorted out. I'm more than willing to help but asking me start rewriting stuff because you don't want to read is going a bit far.

The last graph on that page shows the effect of different 'dead zones' compared to a non modified unit and stepping the preset on the spring tensioner. I think that is what you want to look at.

You could try measuring at the barn door as Charlie did ( post and his graph disappeared ). This will cover some questions if done carefully I went for measuring torque because it's a torque driven device, not displacement driven, and to start with I wanted to see how linear the output voltage was, or rather how close to logarithmic it was, just to gain understanding.

There would be some advantage to deciding from barn door testing whether it needed opening up. But only saving the question  of resealing it afterwards.

If you are just pushing on the end of the vane and measuring output voltage as a function of distance and not force then you are not testing possible mechanical problems. If there are tight spots in the mechanism you will miss them. Also I don't know whether this would be accurate enough to determine whether output was stable or not.

Charlie's graph had a flier. One that was plenty bad enough to cause noteworthy running problems if it was indeed due to bad contacts. It could just have been experimental error. That would need a few runs on a good unit to see how repeatable the results were. You need to ensure that you can get repeatable results first.

I'll give you some practical tips if you get that far.

Even if  you do my buckets and syringes method you will need to exercise lots of care to get stable graphs like I did. There is some experimental jitter in the graphs but not so much as to mask the results. Even quite small differences in dead zone produced clearly different curves.

I could improve the method accuracy but I think it worked well enough to achieve the required result.

Being a lazy git, I am thinking of how I can streamline some of the torque testing.  I am thinking that it might be possible to designate a window where a specific torque delivers a specific output voltage.

I think you'd need to have a bunch of measurements covering lift-off of the output. No single measurement would be accurate enough , you'd need at least half a dozen at and around lift-off to get a picture. If you want to take short-cuts you could probably drop the upper half of the range where they mostly seem to come together. 

re. what did I change ; my timing now back where it was before I started tuning but is not at the standard strobe mark. So although I did move it, doing so was not part of what got smother pick up.

That "standard strobe mark" is not an ignition timing mark but TDC. To set the ignition precicely you need to use a dial gauge to determine how far the piston has to travel to get to TDC.
The best way and it is very reliable to within 1° is to align these two marks.

Point-Seven-five wrote:Thanks!  So you reduced the deadband a bit is that correct?  It seems like the deadband would cause a bit of a lean-out coming off of closed throttle. 

Being a lazy git, I am thinking of how I can streamline some of the torque testing.  I am thinking that it might be possible to designate a window where a specific torque delivers a specific output voltage.  Also thinking about how there is some way to use door position in the tuning, especially at the fully closed and beginning to open positions.

Need to think about putting together a test setup.

If you read through what beamer has written on the flap opening resistance values, you can see some sort of logic to the spring pressure being lessened to deliver that sort of progression, the dead zone is like an area that allows the flap to start moving on an increasing throttle, so matching the dead zone to the spring set is going to help detail how the first couple of hundred to thousand revs are felt at the seat of the pants, and you'd have to tweak the Stop switch. Knowing how much dead zone you need and how much air is going to open the flap past the dead zone verses how many revs you got and how close to the stiocometric average you need to get off the line. That may be to 13:1 to 14:1, the old pumper carbs used to drop a wad of fuel in so the engine got down to 12.8:1 for a short period to give it that oomph, from stories I've read. Most people aren't dropping the clutch at over 6000rpm, its more like 1200 to 1500 and away, and in doing so get a true need for accurate fuel ratios to meet that load. There in the detail, is a way I think to do some detailed finessing. If you are so inclined to go to that enth degree to get something near perfect, than that area seems to be the place to play. Maybe a bit like VW with the American air tests, they set it up for the engine and only remapped a few to pass the testings, then back to proper fuel usage, we can do something like that.

Thanks Rick. What I meant is that I'm somewhat lower than that mark. I have not worked it out in degrees. I should have said timing mark, not strobe mark. I'll have look how much tomorrow.

... for the record, having checked: the plate is only slightly lower than the std position marked in the photo

If you look closely at the low end detail , the zero dead zone setting ( green crosses ) shows a smooth pickup. Finite dead zone has an angle when it breaks in.  Though the zero dead zone adjustment was too rich because it opened early and this was sub-optimal, I think the jerkiness in coming off closed throttle may be caused by too much dead zone.

The gentle rounded opening of the vane is happening off scale and so instead it comes in with a hit.I may try to get some more detail , like 1g steps around that zone.

It could also be due to slight resistance in the mechanical movement. Each horizontal interval on that graph is 2.6g at the lug. That would be about half that at the tip of the vane. If there is any resistance to movement it is going to be jerky coming off the stop.

There may be a case for a total strip and clean if the jerkiness can't be tuned out.

.75 wrote:Thanks!  So you reduced the deadband a bit is that correct?  It seems like the deadband would cause a bit of a lean-out coming off of closed throttle.  

There is now a full write up of the results of my investigation of AFM function and calibration in the how-to section.

I hope this will answer your questions without you having to plough through the old thread.

Thanks!  Nice work.  You have produced a lot of food for thought.