Sacred Screw Reset

[Kiwi_Roy]

Huzo asked me about 18 months back how one would go about resetting the sacred screw on a throttle body that had been messed with.
I explained to him that I am an Instrument Technician; measuring stuff is what we do for a living. Measuring gas flows is possible but it's tricky because you have to take temperature pressure and other factors into account, its always much easier to measure liquid flows, because for one thing you can see what's happening and you don't need an expensive flow meter you just need an accurate measure of volume or weight and a stopwatch for flow rate. I have been meaning to try this for a long time so today I put together a rig so that I could measure the flow through my Griso 1100 throttle bodies.

Then why do they use a flow bench in the factory I hear you ask? The factory has loads of cash and they don't want liquids leaking out all over the place.
What we all have is a source of water.
When an instrument designer has to specify a valve its standard to pick one by the quantity of water they will flow based on opening and pressure drop this is known as the Cv rating
Even valves that are designed to work with gasses are quantified by their Cv rating.
First of all I had to find an accurate method of measuring flow, I found a glass bottle with a fill line at 500 ML and checked its accuracy by weight, 1 Litre of water weighs 1 Kg, which is very convenient so I filled the container 8 times and verified the weight to be 4 Kg

I then proceeded to take a measurement of the Left Hand throttle body, lots of leaks and water later I found that the Left Hand throttle body passes half a litre of water in around 20 seconds (3 Litres/minute) sitting on the unmolested stop If I blip the throttle the valve may not come back to the exact setting but its always between 18.5 and 23 seconds to flow 500 ML
The next day I decided to give the throttle bodies a good scrub as I could feel a lip of grime right where the butterflies met the body, sure enough this lowered the time it took to flow 500 ml to more like 16 seconds

In theory someone else with a messed up throttle body off a 2 Valve Griso should be able to use a similar rig and re-set the sacred screw to give a similar result, I don't know if another 2 Valve engine like a Norge 1100 would use the same setting but I suspect it would. Obviously for other motors someone would have to front up with a pair of good unmolested bodies to measure.

I’m not sure about the sacred stop screw as I’m not about to touch mine but the bell crank screw which has a similar effect on the Right Hand throttle body changes in flow approximately 5% per flat (6 flats/rotation).
Seen here timing the bottle to fill to the 500 ML mark.

It's also possible to do a preliminary throttle balance by flowing through the Right Hand body and setting the Bell Crank until you get a similar flow to the Left Hand body, that's what I am doing in this picture. Actually since I had just done a balance a few days earlier no adjustment was required.

 

Conclusion: I think I can say with confidence that if I were to get another throttle body I could set it to better than 5%
Another person could reproduce my rig and get similar results, there is really only 1 critical dimension and even thats not so important.
 

[Huzo]

That’s really interesting mate...
I always mused about getting a known good TB and dumping a given volume of liquid and recording the readings over several runs and graphing the results to determine the accuracy.
Then I would think you could say that although you don’t know how much air will pass @ STP, you WILL know how much water will pass and you can then root around with the SS of the shagged set to replicate the result..
One could subsequently upset then reset the practice set.
You have departed significantly from my plan though in that, I proposed to set up a rig like yours and do what you did, but unlike me you actually got off your arse and did it...!
Touche’ Sir...!

[Kiwi_Roy]

Exactly, I probably have about 50 runs so far, 20 recorded.

I also took some measurements to see what effect turning the sacred screw would make, of course I didn't actually move it. I measured the thread pitch 0.8mm then I used my feeler gauges to offset the stop.
A 0.2 mm feeler, the equivalent of turning the screw in 1/4 turn lowered the fill time by over 1 second.

[Idontwantapickle]

Roy,

Nice work!
I have a question. In your experience would it be accurate, for the purposes of comparing two units, to measure pressure differential between the incoming and outgoing flow of air being drawn through the throttle body? As long as the conditions were similar and done back to back the absolute number wouldn't matter, would it?
Ok, that was two questions.
Thanks, Hunter

[Kiwi_Roy]

Roy,

Nice work!
I have a question. In your experience would it be accurate, for the purposes of comparing two units, to measure pressure differential between the incoming and outgoing flow of air being drawn through the throttle body? As long as the conditions were similar and done back to back the absolute number wouldn't matter, would it?
Ok, that was two questions.
Thanks, Hunter

Kind of like using an orifice plate with a U tube manometer to measure. I think you could make that work. how would you draw the air through perhaps a small computer fan or a vacuum cleaner.
the problem with orifice plates is you have to be so careful with the tapping points as its easy to create a pressure difference due to turbulence around an obstruction.
The reason I favour water is everyone has it available, its easy to measure and the conditions are easy to reproduce, the same differential pressure should give the same flow. Water too is fairly stable unlike air that changes in density with temperature.
At the point where the body is set up you can barely see daylight through the body.

[Idontwantapickle]

Kind of like using an orifice plate with a U tube manometer to measure. I think you could make that work. how would you draw the air through perhaps a small computer fan or a vacuum cleaner.
the problem with orifice plates is you have to be so careful with the tapping points as its easy to create a pressure difference due to turbulence around an obstruction.
The reason I favour water is everyone has it available, its easy to measure and the conditions are easy to reproduce, the same differential pressure should give the same flow. Water too is fairly stable unlike air that changes in density with temperature.

Roy,
I see your point about the water.
I was thinking about maybe a 2 foot straight tube roughly the size of the TB on either side and using a pressure differential gauge to compare the inlet and outlet with, as you mentioned, a fan or a vacuum to move the air.
I already have a differential gauge that I use to balance carbs etc. If a ever get my hands on an extra TB assy I think I'll try it.
You have inspired me!

Hunter

[lucky phil]

What exactly does the "sacred screw" control? the throttle blade stop? the relationship between the throttle body blades? is it on both TB's? does the Centiuro and V11 have them?

Ciao

[Kiwi_Roy]

What exactly does the "sacred screw" control? the throttle blade stop? the relationship between the throttle body blades? is it on both TB's? does the Centiuro and V11 have them?

Ciao

Phil, I think its only the CARC bikes that have the Sacred Screw, yes its the throttle stop on the Left Hand body, unlike the VII it must not be touched the right hand body is set by a bell crank to balance. The idle speed is set by a stepper motor valve that feeds air in downstream of the butterfly's. I think the practice has been to replace both bodies if they have been messed with.
Apparently the screw is set on a flow bench at the factory. Pete Roper gave us the goods on them a couple of years back.
If someone messes with the screw the stepper motor counters and you end up getting the mixture all messed up and get a burst from Pete that will make you shrivel in your boots.
Cheers
Roy

Roy,
I see your point about the water.
I was thinking about maybe a 2 foot straight tube roughly the size of the TB on either side and using a pressure differential gauge to compare the inlet and outlet with, as you mentioned, a fan or a vacuum to move the air.
I already have a differential gauge that I use to balance carbs etc. If a ever get my hands on an extra TB assy I think I'll try it.
You have inspired me!

Hunter

Hunter,
           If making a flowmeter was that simple every man and his dog would be making them, all you will measure is the differential pressure across a couple of points. Orifice plates are very tricky to set up.
If you could grab the gas that passes through the body and measure the volume you could make it work but how are you going to save it? How will you know that half of it hasn't leaked away?
The idea is to use a fluid that you can see, can save and can measure very easily.

Cheers
Roy

[SmithSwede]

Roy:  what a fascinating post.  Thanks for sharing. 

What is the purpose of the two clear hoses that come down from the top of the black tube?

[Kiwi_Roy]

Roy:  what a fascinating post.  Thanks for sharing. 

What is the purpose of the two clear hoses that come down from the top of the black tube?

The small one is the fill line, the larger one is the overflow at 500 mm, in operation the fill rate is set so it constantly overflows to provide a constant pressure across the butterfly opening.
Update:
I should say, I have no interest in trying to measure air flow, there are too many variables, I know how to do it with a large infusion of cash, the water rig was <$10 and time.

[lucky phil]

Phil, I think its only the CARC bikes that have the Sacred Screw, yes its the throttle stop on the Left Hand body, unlike the VII it must not be touched the right hand body is set by a bell crank to balance. The idle speed is set by a stepper motor valve that feeds air in downstream of the butterfly's. I think the practice has been to replace both bodies if they have been messed with.
Apparently the screw is set on a flow bench at the factory. Pete Roper gave us the goods on them a couple of years back.
If someone messes with the screw the stepper motor counters and you end up getting the mixture all messed up and get a burst from Pete that will make you shrivel in your boots.
Cheers
Roy

Ok thanks Roy. I'd heard of it but wasn't sure whet it was about.Glad the V11 doesn't have one or I'd need new TB's by now.

Ciao

[auldgit]

Phil, I think its only the CARC bikes that have the Sacred Screw, yes its the throttle stop on the Left Hand body, unlike the VII it must not be touched the right hand body is set by a bell crank to balance. The idle speed is set by a stepper motor valve that feeds air in downstream of the butterfly's. I think the practice has been to replace both bodies if they have been messed with.
Apparently the screw is set on a flow bench at the factory. Pete Roper gave us the goods on them a couple of years back.
If someone messes with the screw the stepper motor counters and you end up getting the mixture all messed up and get a burst from Pete that will make you shrivel in your boots.
Cheers
Roy

Hi, my '08 Nevada Classic has one too, identified in the workshop manual  Wondered why it was so-called and now I know! Cheers Pete

[tris]

That's neat Roy - BRAVO!

Its always struck me that at some point the plastic ends to the ball joints connecting the two TBs will wear out and need replacing.

But to replace them and still achieve the original "sacred setting" the length of the tie rod would need to be adjusted some how.

I suspect that this is the means to achieve that

[moto]

... Conclusion: I think I can say with confidence that if I were to get another throttle body I could set it to better than 5%
Another person could reproduce my rig and get similar results, there is really only 1 critical dimension and even thats not so important.

Roy,

Brilliant.

Moto

[EDITED to clarify my main point.]

[auzziguzzi]

Love your work, Roy.
An inspired concept.
John

Plastic 1 liter graduated cylinders cost only about US$12 on Amazon.

A good suggestion but it would more than double the cost of Roy's rig.

[SED]

Very cool Roy!

[Huzo]

I’m again moved to comment on the SS issue after perusing Roy’s excellent work..
I will do a diagram later but would you all like to consider this..
If you can run a flow test with a known good TB set against a suspect (or deliberately upset one) simultaneously from a common source, you don’t have to replicate your conditions for each run, because both TB’s sets, are experiencing identical feeds in real time..
Here’s a thought based on Roy’s model..
Grab a compressor and feed the open flow into a “Y” junction with one side feeding the good TB set and one feeding the bad ..Only hook up to the TB that has the sacred screw (blank the right hand one off from each pair.)
The compressor reservoir is performing as a plenum chamber, not to build pressure, there will be a gentle flow of air from the delivery hose, smoothed out by the reservoir.
The outlets from the plumbed TB’s, are fed into a dedicated manometer for each one.
The feed to the manometer needs a bleed hole.
The TB’s need to have their bleeds closed and the stepper circuit closed off.
Ideally if you had two unmolested sets, you could hook them both up and witness the equal manometer readings, since both circuits will be pressurised up to the point where the bleed hole in the supply pipe to the manometer, is releasing air at the same rate as it’s being supplied to the TB’s.
BUT BACK TO OUR TEST...
If the suspect TB is at a greater plate angle, it will flow more air from the Y junction and reflect a different pressure in it’s manometer.
Again then ideally, you could adjust the screw on the SUSPECT one to bring the manometers into unison.
You’ve to employ a bleed in the line between the TB’s and the manometer, or you’ll pressurise the whole lot and overflow the manometers.
This way I’m reasoning, means that you are converting flow rate into a visible pressure reading and because you are conducting a “side by side” test in real time, with the good and the suspect one, you obviate the need to graph or record rates.
The numerical value/s are irrelevant because you are matching the bad one with the known good one which already has the correct flow rate,
I’ll do a diagram if Roy is interested.

[Huzo]

Or...
Seems to me that supplying a known good TB and a rooted one simultaneously with airflow from a common source is not too difficult..
It’s when you get to the outlet and you try to compare the flow rate  at the outlet, that things get murky.
After I got up and took a piss, I thought maybe if my idea was altered such that you supplied the TB’s the same but  plumbed the outlet/s into a vertical transparent plastic column with a close fitting plastic ball and small drill holes up the side, so the ball would rise in the flow in both tubes, exposing more of the 1/16” holes on the way up, so equilibrium would be reached when the outflow from the tubes equaled the inflow and the ball would register a position.
Easy in initial tests to plumb both plastic tubes into a common flow source and adjust bleeds in supply line to calibrate the balls evenly...(you really never want to get caught out in polite company without your balls calibrated.. )
Then hook up to the control TB and the suspect one and bring the bad one into line with the good......
Dunno...

[Kiwi_Roy]

Sounds too complicated for me Huzo, go for it lol

Once you have the numbers from my water test you don't need to have 2 sets, all you have to do is reset your tampered sacred screw to reproduce my results.
All you need is a water column 500 mm tall and a measuring beaker of some sort.
A lab technician has offered to review the data and tell us how to achieve better results.
Actually Moto is a retired statistics professor, how lucky are we!.

[Huzo]

Sounds too complicated for me Huzo, go for it lol

I won’t have a bar of that Roy...
It’s just that you are up to your earholes in your project so no need to be distracted by my ramblings...
(PM read BTW...)
I was up at Pete’s the other day, I should have brought a known good set of TB’s and a rooted set home. Next time I’ll do it...

[Chuck in Indiana]

Uhh, Roy.. that gets TWO Attaboys..  

[Idontwantapickle]

Sounds too complicated for me Huzo, go for it lol

Once you have the numbers from my water test you don't need to have 2 sets, all you have to do is reset your tampered sacred screw to reproduce my results.
All you need is a water column 500 mm tall and a measuring beaker of some sort.
A lab technician has offered to review the data and tell us how to achieve better results.

Ah ha! Peer review! Then you can publish and rake in the dough! Once you get tenure we can't fire you, set for life!

[moto]

Roy's research here is potentially very valuable because it raises the possibility of saving $1,845.35, the current cost of a new throttle body assembly, when someone has messed with the sacred screw. For that to be plausible it is important to evaluate the quality of his results, which I decided to offer to do in the spirit of our forum.

I'm not a lab technician. Nonetheless Roy sent me his data and I worked it over and made this graph (click to enlarge):





There were 18 separate trials, or conditions, comprising 94 timings of how many seconds it took to fill a 500ml vessel. I converted his timings into flow rates in milliliters per second (ml/sec) with this formula:  ml_sec = 500/seconds. This makes the graph show larger values for larger throttle openings, etc. In working with the data I dropped some observations where Roy had reported a problem, and also found and corrected a typo by consulting him regarding the correct value.

These are the trials, noted on the graph, as he described them (edited somewhat):

1. Baseline before cleaning.
2. Blipped throttle plate.
3. Blipped throttle plate (again).
4. Second throttle body, for comparison.
5. Turned bell crank 1 flat (1/6 turn) clockwise
6. Turned bell crank 2 flats counter-clockwise
7. Turned bell crank 1 flat clockwise (to original position)
8. First run with new, square spout collector, after cleaning.
9. Original, round spout collector.
10. New baseline run, cleaned, with square spout collector
11. Blipped throttle plate.
12. Blipped throttle plate.
13. "No feeler gauge" condition (on sacred screw throttle body).
14. Added 0.15mm feeler gauge between tip of screw and stop.
15. Substituted 0.20mm feeler gauge.
16. Substituted 0.10mm feeler gauge (less open).
17. Substituted 0.05mm feeler gauge.
18. "Removed all feelers."

The graph is separated into three panels by the dashed lines: pre-cleaning; baseline runs after cleaning; and experiments with throttle openings. The computer program has drawn a single red line that connects the average flow rates for each of the 18 trials (the horizontal parts of the line). The main baseline run after cleaning the throttle body is trial 10, which resulted in a mean flow rate of 30.5 ml/sec over 19 observations.

Roy's measurements are excellent. The standard deviation of the flow rates during the 19 observations of trial 10 is only 0.189 ml/sec. If it is assumed that Roy can repeat this performance in the future, and if a normal model of his errors is used, the probability of him being within 0.37 ml/sec of the true value is about 95% on any single measurement. This would be a percentage error margin of +/- 0.37/30.5 = 0.012, or 1.2 percent, a good bit better than he suggested. (Assuming his measurements can be repeated independently of each other, his average measurement in, say, 10 repetitions would be even better, with a standard error of 0.06 and a percentage error margin of +/- 0.4 percent.

(The pooled standard deviation of all his observations around the separate means for the 18 different trials is 0.251 ml/sec, not much different from the trial 10 result.)

But the graph demonstrates Roy's measurement excellence without the formality of statistics. Comparing the left panel, pre-cleaning, with the center one clearly shows that Roy is able to detect the effect of cleaning with no ambiguity. Looking at the right panel, and reading the descriptions of the trials I listed above, shows that he is easily able to detect changes corresponding to a single flat of the adjuster on the bell crank, or a single rotation (one pitch) of the sacred screw (which he did not actually move). These are presumably the same magnitudes of changes that so-called mechanics introduce when messing with the screw. I would suppose that setting the sacred screw to achieve the appropriate flow using Roy's method would get it close enough to ideal that the stepper motor logic would not be upset. (The stepper motor's assumption about the throttle setting at rest is the cause of the problems with messed up screws, as I understand it.)

There are a couple of remaining worries. First, Roy really is an instrument calibration specialist who is able to work so carefully that statistics aren't necessary to evaluate his work. Others may need slightly better setups -- I already suggested using a graduated cylinder, in an earlier message in this thread that I deleted when it got no immediate response. Second, we are not quite sure that the 30.5 ml/sec passed by Roy's throttle body (when it is clean and when the screw is touching the stop) should serve as the standard. But probably it can. The value corresponds to filling a 500ml container in 16.4 seconds, which Roy measured to an accuracy (95% confidence interval) of +/- 0.2 seconds.

Roy's notes about his 18 trials should be read for a full appreciation of his work, but there is no point in me discussing them. He might want to, which I would find interesting.

I vote that Roy be tenured.

Moto

[Chuck in Indiana]

Also nice, Moto. Two Attaboys for you, too.   

[Huzo]

Is it fair to say, that because we are attempting to returned molested TB’s to standard, we need to establish the dump rate or flow rate of an unmolested TB ?
Then adjust the flow rate of the buggered one to the known flow rate of the standard...
Is that our aim...?

[Kiwi_Roy]

In a nutshell, thats it.

Start by giving them both a good scrub.
Establish the unmolested throttle bodies flow time (Time for a fixed volume to pass) if worn do several sets of flow readings, I have done this part for the Griso 1100
Substitute the molested throttle body and re-produce the same time per volume.
Note: Both Unmolested and Molested must have the Air Bleed fully closed, only the flow past the butterfly is to be read.

Although it would be best to measure both throttle bodies on the same rig provided the rig is similar it should be possible to reproduce the time to flow at a different location, it would be interesting to know what Moto thinks of the error possibilities. I chose 500mm WC as the pressure (500mm between throttle shaft and overflow)
I don't believe the cross sectional area of the water column has any effect but 1-1/2 ABS pipe is probably universal.
It would be interesting to know the air flow specs for the different models, I'm sure there is a way of relating one to another.

Looking at Moto's list 9 refers to this collector downstream of the butterfly the original round spout one using 1/2" tube was just a tiny bit restricted so I made another you can see the difference 8 - 9 -10 on the graph

8 & 10 use a square section tube with slightly larger cross sectional area.

[Huzo]

In a nutshell, thats it.

Start by giving them both a good scrub.
Establish the unmolested throttle bodies flow time (Time for a fixed volume to pass) if worn do several sets of flow readings, I have done this part for the Griso 1100
Substitute the molested throttle body and re-produce the same time per volume.
Note: Both Unmolested and Molested must have the Air Bleed fully closed, only the flow past the butterfly is to be read.

Although it would be best to measure both throttle bodies on the same rig provided the rig is similar it should be possible to reproduce the time to flow at a different location, it would be interesting to know what Moto thinks of the error possibilities. I chose 500mm WC as the pressure (500mm between throttle shaft and overflow)
I don't believe the cross sectional area of the water column has any effect but 1-1/2 ABS pipe is probably universal.
It would be interesting to know the air flow specs for the different models, I'm sure there is a way of relating one to another.

Empiricaaly, that really is it isn’t it Roy ?

[Kiwi_Roy]

Empiricaaly, that really is it isn’t it Roy ?

Sorry, your words are too big for my old brain to process lol

[Huzo]

Sorry, your words are too big for my old brain to process lol

Put in it’s most basic form, that really is it isn’t it Roy ?

[Gappy]

After reading all of that, I just got to say huh, sure glad my 1100 has carburetors.