Servos, including those used and 'costed down' in RC practice, may be attractive to consider as reliable alternatives for power switches, particularly those to be thrown in combination (e.g. to simulate interlocking or CTC). We have recently had very good discussion not only in 'best practices' for using servos physically to move and lock switch points but also in controlling them via a system like an Arduino or Raspberry Pi.
I'm starting this tbread at Mel Perry's suggestion as a place where all that discussion can find a proper home.
Thanks for the start Overmod! I’ll start with the drawing of my Tower Pro Servo SG90 mounted under a Atlas code 83 #6 Custom Line Turnout.Mechanically the servo works very good. The Evergreen ¼” square tubing worked great for strong but light weight mounting. I used .04” sheet styrene between the square tubing and the servo. The mounting leg on the servo needs to be removed. I used Amazing Goop to attach everything on this one because everything can be removed with a little elbow grease if needed.I drilled a #64 hole in the styrene sheet as the pivot point for the .03” music wire throw rod.My problem is the idle current to keep pressure on the point rails. The Tower Pro Servo motor heats up with a constant 50ma. When I remove the signal input to the servo it coasts back reducing the point rail pressure.I don’t have the needed “warm and fuzzy feeling” without pressure on the point rails. Mel My Model Railroad http://melvineperry.blogspot.com/ Bakersfield, California I'm beginning to realize that aging is not for wimps.
The main difference between Mel's mount and the type used by Tam Valley (and most of the other DIY and also 3D printed ones you can get on Shapeways) is that the others use a pivot point closer to the servo horn (the horn is the bit with the attachment holes for actuating cables, most servos come with 3 or 4 as part of the package, the shaft of the servo is splined with a threaded hole in the middle, you press fit a horn on the shaft, then run the supplied screw in. The various different patterns and sizes are mainly for various fitting with the servo's intended use in RC model planes).
This results in more movement on the turnout side, so when the servo reaches the setpoint in one direction or the other, with the points firmly against the stock rail, there is some bend in the actuating wire which can lkeep the points pressed tightly against the stock rails without the servo being forced back.
It appears frm what Mel reports that his design has little to no tension on the points if the servo actually comes to rest where it's not constantly fighting to move just a little bit more (stalled out) - which results in excess current draw and the servo getting hot - because they simply aren't designed to be used as stall motors.
I'll see if I can get another video of the one I have mounted directly under a turnout, suitable for digging a trench in a foam base under the turnout and dropping it all in from above - I would assume this is a vital criteria for Mel so he doesn;t have to ctry crawling around under his layout. I used the same thing under the layout with as much as 2 layers of 2" foam and had plenty of force to move Atlas turnouts and hold the points tightly closed, so at least while I still can, i will be installing them underneath, in the manner typical of other solutions like Tortoises.
Previously, I also used Tam Valley controllers. They work well, if you are not the DIY type. I am - so this time, I desgned my own. The lastest version has done away with the frog polarity relay - instead I will use the Tam Valley mounts with a microswitch actuated by the servo arm, which provides the same middle of thrown polarity swap that my relay controlled by the microcontroller did on my board. Leaving off the relay eliminates 2 relays, 2 transistors, 2 diodes, 2 resistors, and a pair of 3 position terminal blocks from my circuit. Parts cost alone makes up for the price of the specified microswitch for the Tam Valley mount, but it also allowed me to shrink the size of the PCB, and size is money. I have the first set of new boards, just have to build one up and try the new code which leaves out the relay control. But first I have to clean my bench off - there's currently a mini layout (3 pieces of flex and one turnout) plus a stack of Fast Tracks SweepSticks in two different radii and multiple fractions of circles, to use when I start laying track.
--Randy
Modeling the Reading Railroad in the 1950's
Visit my web site at www.readingeastpenn.com for construction updates, DCC Info, and more.
I’m thinking about using the Peco Spring. I modified several Atlas turnouts to operate like the Peco.It will take more power from the servo to overcome the spring but when the servo signal is removed the servo coasting or relaxing the pressure shouldn’t make any difference. The extra pressure will require more current to drive the spring points but it’s only momentary. The servo can draw well over 100ma momentarily without heating up, it’s the continuous drain that causes heating.My arthritis cooperating I’ll give that a try, can’t do much when everything hurts. I have several Pecos so I’ll try one of them before I modify an Atlas, much easier.I’m thinking it would be easier to modify the Arduino Sketch for two inputs than drop the signal to the servo.Still thinking, position indication will be lost if done in the Arduino with loss of power. Mel My Model Railroad http://melvineperry.blogspot.com/ Bakersfield, California I'm beginning to realize that aging is not for wimps.
This entire topic maybe a total waste. I had several new Tower Pro SG90 servos on my bench and decided to tinker around with them.Now I haven’t done anything with my turnout venture in months but today I found out that this batch of servos not only doesn’t get hot at 50ma but I put one in a vise so it couldn’t move and set the current at 70ma and no heat after 30 minutes.I’m not sure if I got a bad batch earlier or what. I checked the pressure needed to throw the points on a Peco SL-8352 and it measures 1.1 oz both directions to overcome the Peco spring. The new servos don’t get hot at 5 oz at 60ma and the fulcrum point on my mount will amplify the pressure on the points.I think my problem is solved. The point rails only need 2 oz for firm contact with the outside rails. I should be able to set the servos to 50ma and that should put 4 oz on the points, more than enough for the “warm and fuzzy” feeling on the Atlas turnouts.I’m going to continue making another mount and mount it and test it thoroughly before I call this project done and working. Rather made my day or for that matter my week. Mel My Model Railroad http://melvineperry.blogspot.com/ Bakersfield, California I'm beginning to realize that aging is not for wimps.
I store the position in the EEPROM so that when the system is powered down and then powered back up, it sets them back to the last known position. That includes setting the indicator LEDs. I can show you how I did that - I don't just write to one EEPROM location. There is a limit to the number of writes before the data cell gets unreliable. Since I write this every time a servo changes to a new endpoint, it wouldn't take all that long to use up the write cycles on a single memory location. So I have a routeine that loops through the entire memory range and gets the maximum utilization out of all the EEPROM - if you flipped the turnout 10 times a minute for 20 hours a week, it won't wear out the EEPROM for like 50 years - so I won't have to worry about it.
what's the drawback of operating the servo from one extreme to the other, 1ms - 2ms, using a short arm for a reasonable amount of travel?
greg - Philadelphia & Reading / Reading
GregEarly on when I started piddling around with the cheapo servos I found out that the repeatability of servo manufacturing wasn’t very good. I was surprised when I found out that only about one out of ten or so would do the advertised spec of 180°. A few at 170° but most under 160° so I keep the travel in or below the 120° range or less for compatibility. The 54° swing for my turnout mount works out the best keeping the movement using the longest throw arm that comes with the servo within 1¼” swing to clear the 1½” hole below the turnout. I also prefer the higher torque applied to the moving rails using the long arm on the servo with the long throw rod. That also keeps the linkage in the same range when I use the Du-Bro #167 bellcranks (¾”).Just a personal thing. When it isn't broke don't fix it. The mount works great as is.
Mel My Model Railroad http://melvineperry.blogspot.com/ Bakersfield, California I'm beginning to realize that aging is not for wimps.
tortoise machines have a fixed travel but also have an adjustable pivot point for controlling the amount of end travel.
would there be a need to for SW to store end points if the servo mechanism had an adjustable pivot?
looks like the forum will not display .png files
http://www.azatrax.com/image/tortoise-mounting.png
With a calculated pivot point it isn’t necessary to have it movable. When a Tortoise or like switch machine is used the mounting isn’t exact as the servo glued to the turnout, every servo switch machine mounting is the same.
The same Arduino sketch works on every serrvo because every servo mount is the same. Mel My Model Railroad http://melvineperry.blogspot.com/ Bakersfield, California I'm beginning to realize that aging is not for wimps.
gregclooks like the forum will not display .png files
you have to say "please", to the software.
Henry
COB Potomac & Northern
Shenandoah Valley
Here's a short video of the servos I use to control my Peco Code 83 turnouts. With the centering spring I only use momentary push buttons to activate the servo. Small control PC board in the servo is removed and I wire directly to the servo motor. Two sided tape secures it to the bottom of the layout. I found that 9V seems to be a good choice.
https://www.youtube.com/watch?v=Tyxa3nqnbhY
markie97https://www.youtube.com/watch?v=Tyxa3nqnbhY
+1