Packard Forum Index » V-8 (1955-56)

My 56 Patrician has been converted to use a floor shifter instead of the original push button shifting mechanism. The shifting motor, worm gear, actuator setup has been completely removed, and a cable has been installed, which has been routed through the floor under the front seat to a large and ugly B&M floor shifter.




I would like to remove the floor shifter, and get the pushbuttons working again. I haven't been able to find the missing original parts, so I have come up with a plan to use a microcontroller and linear actuator with potentiometer to get the buttons back in action. The pushbuttons are still on the steering column, and I have tested for continuity in the cables and determined that the buttons work. My plan is to use an optocoupler circuit to go down from 12v to 3.3v for the button inputs,, and I have a few relays set up so that I can turn on the reverse lights for backing up and enable engine start when in park. I plan on using a ESP32 microcontroller for the main logic. I haven't decided if I will install the linear actuator under the seat, since the cable has been pulled there, or if I will get a water proof one, that I would install directly on the shift lever of the transmission.

Here are my questions:

1) How long does it take to shift from P to R with the original setup?
2) How much force should be required to move the shift lever on the side of the transmission? Am I looking for a 30lb or 300lb linear actuator?

Thanks,
John

Attach file:

---

  1956-packard-patrician-55.jpeg (207.99 KB)

Talk to Mr PushButton

Hello John,

I actually had this idea in the past but didn’t pursue it since there would be 0 market demand. If you were going to do this you would want to use a servomotor not a linear actuator, as you can calibrate on position, not timing. These transmissions are sensitive to misalignment of the shifter arm. Side note, I don’t see the point of an optocoupler when a simple 2-resistor voltage divider for each button would suffice.

I really suggest to just buy a complete refurbished push button unit. I could supply most of it but Mr. Pushbutton would be a better source. Please send some photos under the hood showing what relays and linkages remain.

Another option is to get rid of the buttons and change to a column shift. I have a couple of spare columns.

P. S. The shifting force is basically a moderate thumb pressure. Less than checking belt tension.

Absolutely get in touch with Mr Pushbutton for his thoughts on the subject but if you want one of his rebuilt stock Packard units, unless things have changed in the last year or two I believe he may have an arrangement with one of the vendors to handle sales. The last time I spoke with someone buying a rebuilt unit they were also quite expensive -- and much more so without an old one to return.

For my two cents, I think the idea has merit as Packard actuators do have some issues that might be able to be eliminated with something plug in but a bit more robust. If you come up with something plug in that can use stock PBs, not only could it be used with a stock Ultramatic, possibly with a slight mechanical change and a bit of different programming it might also be adaptable to those who might want to or have already converted to the GM transmissions. There has even been some thoughts of using an 80s era Rolls Royce actuator for the GM conversions.

Regarding your question on strength: the original actuator was a bit weak in one area. It was fine for changing from one gear to another 99 percent of the time but if the car was parked on an incline or hard against a curb with tire pushback, with force of the tire or the weight of the car pushing against the parking pawl, the motor often did not have enough power to pull the pawl out of the parking gear. In those situations the car was locked in park until some means was found (often a tow truck) to remove the weight from the pawl.

On the speed, I doubt if anyone has measured it precisely but from this video posted by Mr Pushbutton it appears to take 2-3 seconds.

For whatever you come up with for a drive, since you are going to use a microcontroller setup to activate it, I think it would be good to go with some kind of direct optical encoder feedback off the actuator and have the microprocessor count pulses to determine exactly how many complete or partial revs of the motor it takes to determine where to stop. After spending years working on medical equipment that relied on precision potentiometers to determine positions of various items, I will disagree with Pgh Ultramatic on the point of trying to rely on a varying analog voltage to precisely control the stop points. Granted, it might be easier and maybe less expensive but even the most precision of pots do vary slightly in repeatability. A bit of corrosion on a connection, wiper, or in the suggested push button divider circuit can also affect a voltage enough that it could change stop points enough to matter. While the Ultramatic is not nearly as critical an application, the issue was noticeable enough in a clean environment so unless you get sealed components and take extra care with connectors I would wonder how the harsh conditions -- heat, vibration, dirt, grease, moisture, etc -- in a car would contend. The downside to using digital and microcontrollers is to remember that power in cars of that era was quite dirty. There was no consideration given to surges and spikes caused by relay or switch contacts operating and motor brush noise will be a serious issue. You will need to filter the heck out of things and even then, if a long run is involved, your 3.3 volts could be seriously affected.

Packard's fingers did a good job of finding the precise place to stop and I would try to duplicate that precision as much as possible. Optical sensors positioned much like the fingers might work but they would have to be mounted on a special new support which might be hard to place and also be weather resistant.

To go along with concerns about electrical position error, I would also suggest placing the actuator at the transmission to avoid introducing any mechanical error via slop from the cable and attachment points into the mix. Since the detents are part of the valve body, unlike the manual handle where you can feel if slop has developed or if you are fully in a position or not, if the motor stops where it thinks a gear is positioned but cable or mechanical slop adds an error and the motor has stopped with the valve between detents all kinds of problems will result.

No idea which type mechanisms you are considering but there are some rotary actuators that might also fit your requirements. Here is a actuator company with a good selection of types.

If you have not already looked it over, the original Autolite manual for the Packard system can be downloaded. Other than current draw, I don't remember if there are any other specs mentioned on the original motor parameters.

Quote:

Linear actuators are available with built-in position feedback.

Quote:

I didn't say that. I was ambiguous as to the type of servomotor, but a resolver feedback type would generally be best. You can digitize a resolver signal to feed digital inputs to whatever degree of accuracy, then program the controller to the correct shift point offsets as well as a global offset that is calibrated on installation.

You could simply have the resolver connect to the sector gear itself, though either the tuning would have to be sluggish, or the worm gear would need slight interference with the sector gear, to keep the motor from oscillating. A PID scheme, position to voltage, should be satisfactory.

You could be right and I'm sure there are different ways of controlling. Just depends on balance between cost and complexity and, of course, the Ultramatic can withstand a tiny bit of error so I am not saying absolute precision is needed. I have some bad tastes left by machines with analog only position sensing and I was going to detail the reasons. Decided no one here would really care and didn't want to hijack the thread so no detailed reply. Anything that is simple, plug in, and a bit more robust would probably be better than the status quo. It would be intersting to know what Bendtsen came up with since they advertised having a solution fof PB setups for those wanting GM trans.

Might be better to look at small starter motors.

And adapt one to the pb frame..

Those small starters should have more torque than the autolte..

Motor technology has improved significantly since the 50's with permanent magnet, reduction gearbox motors being common and cheap. They have the additional benefit of light weight and low inertia so I would expect that positioning could be precisely controlled.

True and a modern starter type motor might be made to fit but doubt you would want to try a real mini starter as I believe the case is grounded on those and that would require some mods. I believe the stock Packard motor is wound so it rotates a bit slower than a typical motor too so a regular starter type motor might be too fast. With only needing a reverse of polarity to swap directions, doubt the stock motor drive relay would work either so would probably need to rework that aspect.