Reading factory publications readily available on line would have answered a lot of the questions here but that appears to be too easy for some.
A lens actual focal length is rarely its marked nominal focal length. Due to production tolerances the actual focal length can vary 2 or 3 millimeters either side of the nominal value.
Graflex Corp. clearly states that each lens is optically measured then a production cam closest to it is measured focal length is chosen or a new cam made if one is not available. Some other camera manufactures call this the cookie cutter method. I do not know how close a lens has to be to the cam but probably less than a mm. The close focus distances will have the largest focus error on mismatched cam/lens focal length. There are 4 Pacemaker cams for marked 135mm lens, P31-132mm, P5-133.5mm, P6-135mm, P7-136.4mm.

The only adjustment that can be made to a Pacemaker Graphic Rangefinder is the infinity setting which on page 2 of the service instructions says a special cam .437 .0001 inch wide is required to set the rangefinder infinity. http://www.southbristolviews.com/pic...TRFService.pdf

Now, Mr. Precision, the cam follower arm will be a fixed distance from the cam base at any given focused distance regardless of cam marked focal length. If a factory matched lens/cam combination is focused at 15 feet and a totally different focal length matched set are installed on the same camera and focused at 15 feet the cam follower arm that rides on the slope of the cam and moves the movable mirror will be precisely the same height above the cam base. This will hold true for any focused distance.
A while back I had 3 Super Speed Graphics with factory matched 135mm Optar and cams. Two of the cams were the same P# and the third different.
There was no calibration information on the Graphic Rangefinder in a Super Graphic. 1 of the 3 bodies had a broken focus panel so I pulled the rangefinder from it. I used another body to establish the infinity point on all cams and measured it. This is the result: http://www.graflex.org/helpboard/viewtopic.php?t=4647
Next I established the cam height at the focused distances on the focus scale/flash calculator built into the camera and the relationship of cam movement from infinity to rail movement when the lens was focused. this is the result: http://www.graflex.org/helpboard/viewtopic.php?t=4653

Now with a Crown and a matched factory lens/cam one can measure the cam heights and find the bed to cam movement relationship and the procedure to make accurate cams from scratch for Pacemaker Graphics. A Crown is needed as the rangefinder is not functional when removed from the camera body as it is with the Super and trying to measure a cam follower arm height in a Speed would be a challenge. If you were to make all the measurements necessary and post the information you would make a lot of Pacemaker Graphic owners happy. Of course someone could send me a Crown with matched lens and cam and I will do all the work and post the information as I have in the past.

______

Ugh! Now you've got me jonesing for a working rangefinder for my Crown. The rangefinder cable broke soon after I got mine (top-mounted), and so I've been using it as a tripod-mounted view camera exclusively.
The Graphic Rangefinder uses a brass tube with two different length rods, one either side of the cam, 43 balls and 43 spacers, and another rod at the bed end. On the Pacemaker version the bed end has a plastic cap that should be a pressure fit. There are many reports on Graflex.org help board of the plastic cap coming off and the balls and spacers falling out. If the cap has come off and you have lost some or all of the balls and spacers then make new ones. I posted the measurements for the balls and spacers in a response on graflex.org help question at one time, I do not remember which one.

Question about a 2 x 3 Crown with side Kalart:
When the bed is dropped to the 2nd click to allow for the limited lens movements, is the optical range finder still accurate?
Neither the Kalart, Hugo Meyer, or Graphic Rangefinder are accurate with the bed dropped. Rangefinder operation is based on rail movement forward and dropping the bed pulls the rails forward more than they would be if the lens was focused with the bed at 90.