Governor modeling

[Nate S]

The governor is key to adjusting shift points at WOT but details of how it works are pretty poor from what I've seen. I thought I would try to understand it better so I disassembled a spare one, weighed the pieces, measured the spring and modeled the physics around how it functions. I then used that info (4700rpm unit) to predict what would need to be done for 5600 rpm and that matched a known 5600 rpm design pretty closely.

Below is a graph of gram force generated by the governor for 3 designs with known shift points. Note that changes other than the governor that change shift points will push these around. the governors will continue to be proportional to each other though. Pushing the 4700 rpm unit up 100rpm would also push the 5600 rpm unit up about the same amount.

The differences between the three are only in the mass of the inner and outer weights. The spring remains the same. Changing the spring can have odd results, changing only one weight can have odd results. If you aren't careful you'll move one shift point more than another. In fact without it modeled out it's almost guaranteed.

The lower part of the curve is a combination of inner weight, outer weight and spring weight. The upper part of the curve is primarily inner weight and less so, the spring rate. As you can see, changes to only the inner weight or spring will affect WOT shifts but have little effect on the part throttle shifts. Changes to the outer weight will affect part throttle shifts and won't do anything to WOT shifts.

I'm out of time right now. I'll add further detail in a follow up.

 

[WileERobby]

Oh.... from the title, I thought this had to do with New Jersey's Murphy and animal crackers.

 

[w.Hudson]

I am interested in this also. I would like to get a WOT shift points above 6000 rpm for my car

 

[Demonic]

Nice. Or you can just order a governor weight kit form A&A, among others.

http://www.aandatrans.com/Departments/Chrysler-727-Trans/Governor-Kits.aspx

 

[Wietse]

It would be nice to know what shift points to expect when changing the governor for say a 5800 governor.
Not just WOT but also part throttle and how the rear axle ratio would affect it.
You could tailor it to suit what you want, kind of, because it will be near impossible to map every scenario i guess.

Because, lets say you only swap the rear axle ratio from 3.23 to 4.10 i assume the governor shift point change to lower road speeds because of the higher centrifugal force caused by a lower ratio.
Who can say what happens with them when swapping axle ratio, the only value given are the WOT rpm shift points.

Interesting topic i am following!

 

[Nate S]

Nice. Or you can just order a governor weight kit form A&A, among others.

http://www.aandatrans.com/Departments/Chrysler-727-Trans/Governor-Kits.aspx

For sure, but that’s not who I am. Sorta like buying a built engine. Perfectly valid thing to do but it’s more fun to understand it.

 

[Nate S]

It would be nice to know what shift points to expect when changing the governor for say a 5800 governor.
Not just WOT but also part throttle and how the rear axle ratio would affect it.
You could tailor it to suit what you want, kind of, because it will be near impossible to map every scenario i guess.

Because, lets say you only swap the rear axle ratio from 3.23 to 4.10 i assume the governor shift point change to lower road speeds because of the higher centrifugal force caused by a lower ratio.
Who can say what happens with them when swapping axle ratio, the only value given are the WOT rpm shift points.

Interesting topic i am following!

I used road speed because it’s continuous. Tougher to make a chart with engine speed because it has 3 different sets of X axes. I’ll see if I can find a better way to display it. Of course the tranny doesn’t know what the rear axle ratio is, just shifts based on driveshaft speed. I happened to use 3.23 just because it’s common and what I’m working with.

The part throttle aspects are still dependent on the governor but very dependent on the throttle valve position and spring. It’s quite difficult to guess at those. WOT is super dependent on the governor and the throttle is bottomed out so many variables come out, making it more predictable.

Being that the lines on the graph are pretty proportional the whole way up, meaning the WOT shifts for the outer ones are 47:56 through most of the curve. The shifts would follow the same pattern. A part throttle 1-2 shift that was at 25 mph would move 56/47 up to 29.7. This assumes the curves keep the same shape. If you mess with the spring too much or the weights disproportionately it could move the part throttle shifts more or less than the WOT.

Another big variable is the carburetor and downshift linkage. Factory setups relied on a particular radius from the pivot to the throttle connection and therefore horizontal throw of the linkage pin (or ball). The common Holley upgrade and throttle bracket has a much larger radius, totally messing this up. Basically, if it’s set to bottom out at the end, the rest of the travel has the lever less compressed, making it short-shift. If you re-drill the Holley bracket so the pin is at the factory radius (usually 1.125”), the shifts move higher and the whole deal behaves better.

 

[Wietse]

Another big variable is the carburetor and downshift linkage. Factory setups relied on a particular radius from the pivot to the throttle connection and therefore horizontal throw of the linkage pin (or ball). The common Holley upgrade and throttle bracket has a much larger radius, totally messing this up. Basically, if it’s set to bottom out at the end, the rest of the travel has the lever less compressed, making it short-shift. If you re-drill the Holley bracket so the pin is at the factory radius (usually 1.125”), the shifts move higher and the whole deal behaves better.

That's a good explanation, actually i did not know this.

I figured the linkage mechanism might be setup up in a proportional way, where when increasing throttle the throttle pressure in the trans will increase proportionally.
This would make sense as it will react more "agressive" at bigger throttle openings to work against the governor pressure and assist with delaying the shift points to higher speeds.

And i think you can make it (more) proportional by looking at the upper bell crank linkage and modify the operation movement.
At that linkage you have horizontal travel and vertical travel, the vertical travel is what operates the trans linkage, the original position of the linkage "ball" would sit at the 2 o'clock position at rest with "as per book" adjustment using the positioning pin through the holes.
Playing with the position and linkage arm length would make it possible to change the throttle pressure build up to a linear or proportional increase.
You could use this to polish out any unwanted actions caused by the governor pressure.

 

[Nate S]

Some more..

A clarification of how it works. Line pressure in, 50psi idle, 100psi WOT goes past the right hand spool diameter to the middle of the spool. The area on the "governor pressure out" is greater than on the "in" side so it pushes the valve shut. As the output shaft spins, the weights pull away from the shaft, providing pressure to reopen the spool. As the force increases the system comes to a new equilibrium at a new pressure. The higher the force, the higher the pressure.

When it starts the weights are pulled up against the output shaft.

As it spins faster the force increases for a long time, eventually the outer weight pulls away from the shaft and the spring compresses beyond its initial preload. Soon thereafter the outer weight bottoms out on the snap ring. The full force of the outer weights contribution goes flat at that point.

The force of the inner weight continues to increase. It moves just a little bit though. It's the force that regulates the pressure not the physical motion per se. Eventually the inner weight bottoms out as well. The force provided by the whole system is enough to drive the governor pressure to overcome line pressure and springs and force the shifts.

 

[dvw]

This is good information. In my book I'd be only concerned with WOT. Now we can weigh the small governor weight. Then knew what percentage of the small weight needed to be removed per 100rpm increase. Someone could simply remove the small weight , weigh it , drill it and test.
Great explaination.
Doug

 

[Nate S]

Next post (probably Sunday) I’ll put in weights, springs etc.

 

[BSB67]

I have 5 or 6 inner weights all drilled to different depths/weights. When you Start trying to hit 6500 rpm, things get squirrelly. Ended up having to change the governor spring, and 1-2 shift valve spring too.

Because there are enough variables, if you think you've developed an accurate curve, it will only apply to your car, when everything else is held constant. Even temperature will change the shift point.

 

[Nate S]

I have 5 or 6 inner weights all drilled to different depths/weights. When you Start trying to hit 6500 rpm, things get squirrelly. Ended up having to change the governor spring, and 1-2 shift valve spring too.

Because there are enough variables, if you think you've developed an accurate curve, it will only apply to your car, when everything else is held constant. Even temperature will change the shift point.

It’ll apply to more than one car but I agree, the curves get steep out there in the very high rpm so little differences go quite far. At those speeds, by my spreadsheet you better be measuring in the .05 gram accuracy range and you better have calculated what you want. Guessing won’t work. The spring won’t do diddly. Only handles the transfer from one weight to another. If you get that wrong then the weights will be much more sensitive and probably never right.

It’ll apply to any car where you know your current shift points and know what governor you have.

 

[BSB67]

It’ll apply to more than one car but I agree, the curves get steep out there in the very high rpm so little differences go quite far. At those speeds, by my spreadsheet you better be measuring in the .05 gram accuracy range and you better have calculated what you want. Guessing won’t work. The spring won’t do diddly. Only handles the transfer from one weight to another. If you get that wrong then the weights will be much more sensitive and probably never right.

It’ll apply to any car where you know your current shift points and know what governor you have.

Sure, but you cannot tell what the shiftpoint will be by simply knowing the weight and spring. The spring does change WOT shift.

 

[Nate S]

Sure, but you cannot tell what the shiftpoint will be by simply knowing the weight and spring. The spring does change WOT shift.[I’m

Sure, but you cannot tell what the shiftpoint will be by simply knowing the weight and spring. The spring does change WOT shift.

The spring can change the WOT shift but you can’t tweak it with the spring very easily, in fact you’ll mess up other stuff, likely the other shift. You can tweak it with grams of weight added or removed. The physics here aren’t very complicated.

 

[BSB67]

Yes, the governor spring is a bigger, less predictable variable and when you change it, you're starting over in developing your predictive model. And yes, the physics is simple. But at higher WOT shift points you simply might not be able to get there weights alone. If you are starting with known results and only adjusting weights, its simple and predictable.

My point is that an inner weight in one car will not give the same rpm shift point in another car as line pressure, governor springs and shift springs are probably different. So the change is only predictable with a known starting point.

Good and fun post non the less.

 

[Nate S]

Yes, the governor spring is a bigger, less predictable variable and when you change it, you're starting over in developing your predictive model. And yes, the physics is simple. But at higher WOT shift points you simply might not be able to get there weights alone. If you are starting with known results and only adjusting weights, its simple and predictable.

My point is that an inner weight in one car will not give the same rpm shift point in another car as line pressure, governor springs and shift springs are probably different. So the change is only predictable with a known starting point.

Good and fun post non the less.

I’ll agree with that. You need to know what you’ve got. And if you aren’t starting with stock, it’ll be less predictable. The factory springs I’ve seen are very good quality, not just going to grab one that does the trick without lots of work. I’ll model some out at 6500 and see what’s needed. Question for you... do you have any detail on what you arrived at? I’d love to plug it in and see what the model says.

 

[BSB67]

Not here. And I don't think that I kept any of my math. I used known information from another trans and did the math to extrapolate the desired rpm, on the new trans and it was not even close. Its important to note that the factory used several different governor springs, and unless you know for certain the history and origin of a tranny, the spring is an unknown. The trans also started with a 600 rpm difference between the 1-2 shift and 2-3 shift.

 

[Nate S]

Further adventures...
I re-did the first chart to use RPM to make it more useful.

Then I took the model, using the 5600 RPM as a baseline and messed around with the springs.

I think it tells an interesting story. The spring obviously doesn't do anything to change down low. And lighter springs = shifts later. You'll note that the curves change though. With a single spring, changing the weights can move around the shift points in a fairly even fashion, where you started will be proportional to where you end. When you change the spring though the trajectory changes, one shift point will move farther than the other. If this is corrected for, further weight changes will have relatively linear behavior but additional spring changes will disproportionately move the two points. This bodes well for tweaking an existing governor but not so well for "one size fits all". That's why sometimes the A&A governors will throw the two out of whack and require a changed 1-2 or 2-3 shift spring.