Balance synching after loading Meinolf's bin

[rich888]

I've been looking around at TB balancing/synching.

There appears to be a few schools of thought about setting idle balance.

I was advised by Meinolf, after loading his bin file, that both air bleeds should be closed and that I should idle synch using both throttle stops.

This is obviously counter to most of the posts that say only use one throttle stop to set idle speed, or some that say to use the air bleed screws.

I am thinking that, after setting TPS, closing both air bleeds, I leave the throttle link bar disconnected and set the idle balance with just the throttle stops.

Then reconnect the throttle link bar making sure it's adjusted to keep the idle in balance.

Now set 3k rpm balance with the white knob.

 

I can't quite see how that won't change the idle balance?

Am I missing something?

 

 

[Meinolf]

Hi,

consider the following. The BIN contains two fuel (and ignition) maps, these are indexed by rpm and TPS and each rpm/TPS breakpoint contains fuel values. The main maps contain values for the left cylinder, the delta maps for the right cylinder contain the differences to the main maps. So a breakpoint in the main map might contain the value 100, the corresponding value in the delta map could be 10 (or -10). Which is added (or subtracted) from the main map value, so the value pair would be 100 and 110.

These fuel values would achieve a given Lambda/AFR at a given airmass going thru the engine at this specific breakpoint.

The airmass going thru the engine is determined by several engine specific characteristics and the opening of the throttle butterfly valve. The free area through which air can flow is described in a rather complicated cosine function, which also takes into account the diameter of the shaft and the thickness of the butterfly valve. The nature of sine/cosine functions is that they are not linear. Accordingly a very small rotation angle change from closed state would result in the same free area than a much larger angle change, say at halfway open.

The TPS breakpoints are staggered very closely at the beginning, the steps would (for example) be 2.0°, 2.6°, 3.2°,.... At higher settings the breakpoints much farther apart, say 45°, 55°, 67°, 81°. But, with well choosen breakpoints all steps would result in linear increase of airflow, or rather airmass, moving into the combustion chamber.

The essence is that at low TPS openings appearently minor changes have the same effect than much larger ones at bigger openings.

The BIN and the included maps are not very carefully designed in the first place. They are acceptable for, say, 95% of all engines, deliver poor results with 4% of the engines are are perfect for 1% of the engines.

Which, if one considers the priorities of a manufacturer, deviations from the blueprint due to mass production differences in all components, wear of said components and other changing parameters, such as gasoline mixture during the last 20 years, is not too bad.

So the question is how to accurately align TPS breakpoints. Opening angle is not practical to measure, so indirect airmass flow is chosen. Indirect because synchronization uses pressure as an analog to airmass, which together with the fuel generates a certain force with which the cylinder is moved downwards and results in a partial vaccum.

Springs are used to force the butterfly valves closed. The throttle stop screws act against the spring pressure and take up any loose play. Thus using the throttle stop screws for sync'ing gives a start non-variable position. My BINs contain fuel values in the main and delta maps tailored for a specific Lambda at any breakpoint. If the idle sync is not done as described the result will be that these fuel values at specific breakpoints no longer match. So, to use above figures, instead of having a value pair of 100 and 110 the value pair might be 100 and 120, because the next delta map breakpoint value is being used. The result would be different Lambda value for each cylinders. Which means the cylinders are working against each other.

Allowing a user to loading a different BIN with edited maps was not intended by Marelli. Instead CO trim and bypass screws were the intended methods to correct differences between the two cylinders. CO trim influences fuel delivery, bypass screws influence air delivery. The effect of both methods tapers of with increasing fuel values (CO trim) and increasing airflow (bypass screws).

So I've chosen to use fuel values which, with CO trim set to 0 and bypass screws closed, are on target. I could have chosen fuel values which provide the same Lambda with a CO trim of 100 and bypass screws 5 turns open, but why would I introduce 2 variables?

Cheers
Meinolf

[GuzziMoto]

I am not the expert Meinolf is. But my two cents is....

I am not a fan of the way you are supposed to balance the Guzzi motor using the air bleeds at idle and the rod that connects the throttle bodies at 3k rpm. I would prefer to do what Meinolf says, close the air bleeds and balance the throttle bodies by adjusting the rod between them. That is basically how every other motorcycle I have balanced the throttle bodies / carbs has done it. I also prefer to use one throttle stop to set the idle. I feel that is better as it eliminates a possibility of inducing variation in opening the throttles. If you use two separate stops and the linkage develops play in it you may find that it opens the throttle body the cable is attached to slightly before it opens the other throttle body. I think it is better to use the throttle stop on the side the cable is attached to.  The other side then should maintain the same relationship with respect to the first side regardless. If there is any play in the linkage it will always be held tight against the closed side of the play by the spring that is trying to close the throttle body.

If anything I say on this seems to contradict Meinolf, I would suggest you listen to Meinolf.

 

[rich888]

Hello again,

I understand what Meinolf is saying.

 

So to clarify, (and ignoring GuzziMoto as instructed )

You set idle synch with the connecting rod disconnected. Using the throttle stop screws ONLY.

You then adjust and connect the rod ensuring that both TB's are still on their stops.

Then set the 3k balance with the adjuster knob.

What I do not understand is if the two TB's are at rest on their stops, the rod is set to allow this and then rod length is adjusted to balance at 3k, one of the TB's must move off it's stop when returned to the idle position.

 

Am I being thick? I am I missing a neuance of the connecting rod?

[Meinolf]

28 minutes ago, rich888 said:

What I do not understand is if the two TB's are at rest on their stops, the rod is set to allow this and then rod length is adjusted to balance at 3k, one of the TB's must move off it's stop when returned to the idle position.

I am I missing a neuance of the connecting rod?

Hi,

your reasoning is quite correct. However, there's a lot of combined play in different bearings. And only with zero play would setting the connecting rod clash with the throttle stop screw setting. Taking a close look at the mechanics will make this more obvious.

Cheers
Meinolf

[Meinolf]

1 hour ago, GuzziMoto said:

a) ...do what Meinolf says, close the air bleeds and balance the throttle bodies by adjusting the rod between them

b) ....I also prefer to use one throttle stop to set the idle. I feel that is better as it eliminates a possibility of inducing variation in opening the throttles. If you use two separate stops and the linkage develops play in it you may find that it opens the throttle body the cable is attached to slightly before it opens the other throttle body.

Hi,

re a) My comments above refer to the idle sync. The running sync should be done as described in the service manual. This will take up the lash in the connections between the two throttle valves. Removing this lash has no impact on the breakpoint sync done with the idle sync procedure. As written in my previous post, this becomes much more obvious if the system is closely looked at and considering in which direction the lash works.

re b) If using one of the throttle stop screws will achieve both the synchronicity and the desired idle rpm, that's all right. I've found that getting both right typically requires both screws.

Cheers
Meinolf

[rich888]

3 minutes ago, Meinolf said:

Hi,

your reasoning is quite correct. However, there's a lot of combined play in different bearings. And only with zero play would setting the connecting rod clash with the throttle stop screw setting. Taking a close look at the mechanics will make this more obvious.

Cheers
Meinolf

OK Thanks Meinolf!

I thought I was going mad!

I will using both stop screws for idle and the white knob for 3k, which should tune out backlash/slack in the linkages...

 

One thing I did notice while setting TPS, the fully closed position with no throttle stop is quite variable.

For example, if you let the trottle close gently, the TPS may read 180mV. If you let the trottle snap shut, the TPS may read 150mV.

Trying to get consistant 157mV is impossible as the closed position is variable depending on how much force closes it.

I thought this may be due to dirt and soot inside the TB so I cleaned it. The bore and the edges of the butterfly. It is still showing some inconsistancy.

 

 

[Meinolf]

9 minutes ago, rich888 said:

For example, if you let the trottle close gently, the TPS may read 180mV. If you let the trottle snap shut, the TPS may read 150mV. Trying to get consistant 157mV is impossible as the closed position is variable depending on how much force closes it.

Hi,

the butterfly valves should be firmly seated against the throttle body. In fact the butterfly valve contact areas are slightly tapered (no sharp right angle edge) to ensure a consistent seating without hammering into the throttle body.

I've found that getting consistent (+/- 1 or 2mV) readings if using the same force to completely close the valves is standard. The variations you mention might well be due to worn out throttle shaft bearing(s). I've used standard bushings to replace the original ones if the play became to large.

Cheers
Meinolf

[rich888]

7 minutes ago, Meinolf said:

Hi,

the butterfly valves should be firmly seated against the throttle body. In fact the butterfly valve contact areas are slightly tapered (no sharp right angle edge) to ensure a consistent seating without hammering into the throttle body.

I've found that getting consistent (+/- 1 or 2mV) readings if using the same force to completely close the valves is standard. The variations you mention might well be due to worn out throttle shaft bearing(s). I've used standard bushings to replace the original ones if the play became to large.

Cheers
Meinolf

Yes could be that. I cannot feel any play, but the TPS is very sensitive. It picks up tiny movements.

The bike has 30,000 miles on the clock.

[Meinolf]

22 minutes ago, rich888 said:

... consistant 157mV...

Hi,

and just to close this off. The 157mV I recommend as TPS base setting stems from the TPS ADC look-up table in the BIN. Guzzi, or more likely Marelli, made a stupid mistake in this table because a wrong rounding was used.

The ADC in the 15M/RC is a 8bit device, translating the (analog) voltage coming from the TPS into 256 digital steps. 5V divided by 256 equals 0,01953V. The difference between 150mV (factory recommended value) and 157mV is quite small, in fact if falls under the graininess of the ADC function (look this up in the web if interested), but why use a wrong value if the correct one is known. Cheap blueprinting, as it is.

Cheers
Meinolf

[rich888]

2 minutes ago, Meinolf said:

Hi,

and just to close this off. The 157mV I recommend as TPS base setting stems from the TPS ADC look-up table in the BIN. Guzzi, or more likely Marelli, made a stupid mistake in this table because a wrong rounding was used.

The ADC in the 15M/RC is a 8bit device, translating the (analog) voltage coming from the TPS into 256 digital steps. 5V divided by 256 equals 0,01953V. The difference between 150mV (factory recommended value) and 157mV is quite small, in fact if falls under the graininess of the ADC function (look this up in the web if interested), but why use a wrong value if the correct one is known. Cheap blueprinting, as it is.

Cheers
Meinolf

Only 8 bit! (it's a 68HC11 isn't it?)

So 19mV steps with some non-linearity thrown in!

Does the throttle position reported on GuzziDiag relate directly to the raw TPS reading or is there some rounding?

[JBBenson]

1 hour ago, rich888 said:

One thing I did notice while setting TPS, the fully closed position with no throttle stop is quite variable.

For example, if you let the trottle close gently, the TPS may read 180mV. If you let the trottle snap shut, the TPS may read 150mV.

Trying to get consistant 157mV is impossible as the closed position is variable depending on how much force closes it.

I thought this may be due to dirt and soot inside the TB so I cleaned it. The bore and the edges of the butterfly. It is still showing some inconsistancy.

 

 

Clean the throttle bodies (especially upstream) thoroughly, as well as cleaning the butterfly pivots, inside and out.

I had the same problem you did, and it was a build up of gunk (probably from Ethanol in the fuel) that was not allowing the butterflies to completely close, or to close erratically.

[rich888]

11 minutes ago, JBBenson said:

Clean the throttle bodies (especially upstream) thoroughly, as well as cleaning the butterfly pivots, inside and out.

I had the same problem you did, and it was a build up of gunk (probably from Ethanol in the fuel) that was not allowing the butterflies to completely close, or to close erratically.

hello JB,

I did clean a lot of black sticky stuff off the throat taking extra care around the place where the butterfly seals. Awkward without taking the TB off, but possible with a mirror and Q-tips.

Initially the butterfly was sticky. You could feel it once disconnected.

As the valve is spring loaded, it can snap back, if you let it. More realistically, if you let it gently close, it does seat at more of less the same place.

Obviously this is a position it will never normally be in due to the stop screws, so the butterfly is seating onto somewhere 'new'.

 

[Meinolf]

1 hour ago, rich888 said:

Only 8 bit! (it's a 68HC11 isn't it?)

Does the throttle position reported on GuzziDiag relate directly to the raw TPS reading or is there some rounding?

Hi,

yes, the MCU is a 68HC11.

All ECUs (Marelli 15M/59M/5AM and Sagem 1000) whose code I've dissassembled use raw look-up tables for the digitalized input and a conversion look-up table.

The raw values are only used for out-of-bounds checks, most of the code uses the converted values.

Which makes sense, as most raw values are reversed - low raw values equivalent to high converted values - and using a look-up table saves some code and computing time to reverse the data. And look-up tables make it easier to adapt to new sensors, be it NTC, potentiometer or whatever.

Cheers
Meinolf

[Tomchri]

I have to read this even if the brain hurts,, another IPA. Happy to have controll over 40mm Dellortos,,,   and glad some body have control over the INTERWEB V11s. Ulversheim is not to far,,,

Cheers tom.