Subscriber & Member Login

Login, or register today to interact in our online community, comment on articles, receive our newsletter, manage your account online and more!

using servos as stall motors

1390 views
58 replies
1 rating 2 rating 3 rating 4 rating 5 rating
  • Member since
    July 2009
  • From: somerset, nj
  • 2,913 posts
using servos as stall motors
Posted by gregc on Sunday, February 23, 2020 8:14 AM

tortoise machines are easy to control since you simply apply voltage to move the motor until it stalls and don't have to worry about removing power.

has anyone tried the modifications in  Operate turnouts with servo motors  and Turnout Motors from Servos.

seems like a servo is a cheap source for a geared motor that can be controlled as simply as a Tortoise

greg - Philadelphia & Reading / Reading

  • Member since
    January 2009
  • 4,640 posts
Posted by RR_Mel on Sunday, February 23, 2020 11:24 AM

The Tower G90 servo draws high current in stall!!!!  A S90 servo in more than a slight pressure would be toast in a couple of minutes.  I’ve seen them draw close to 200ma.  50ma to 60ma moving the points then 20ma with a tiny bit of pressure, idle current is 14ma at 5 volts.
 
 
 
Mel
 
 
 
My Model Railroad   
 
Bakersfield, California
 
I'm beginning to realize that aging is not for wimps.
 
  • Member since
    July 2009
  • From: somerset, nj
  • 2,913 posts
Posted by gregc on Sunday, February 23, 2020 12:20 PM

the articles describe bypassing the servo electronics and driving the motor directly either with a small voltage (~0.5) or with higher voltage (9-12V) with 200-300 Ohm resistors which limit the stall current

greg - Philadelphia & Reading / Reading

  • Member since
    January 2009
  • 4,640 posts
Posted by RR_Mel on Sunday, February 23, 2020 12:54 PM

The only mod I’ve tried is changing one resistor to get them to go the advertised 180° swing and the 360° mod, making it a motor.  Working inside the servo isn’t easy for Shaky Mel.
 
That does sound interesting, when I think I can try that I’ll give it a go.  Won’t be today, I had a bad go with my diabetes this morning and I’m doing good to type this.
 
Mel
 
 
 
My Model Railroad   
 
Bakersfield, California
 
I'm beginning to realize that aging is not for wimps.
 
  • Member since
    February 2002
  • From: Reading, PA
  • 27,700 posts
Posted by rrinker on Sunday, February 23, 2020 1:02 PM

 I never took one apart to measure the resistance of the motor to see just what current it draws, but it's pretty high. That high current draw when one stalls because it can't move to the commanded position isn't the electronics. I've seen spikes of close to 1 amp on my power supply if I deliberately hold the servo. That too is the little SG90 size ones. A bigger one would only draw more current.

 If anything, using a servo as a stall motor would be like some of those Tortoise alternatives that DO safely work as stall motors, but instead of 15ma stalled, they draw 50ma or more. I guess if you are just controlling them with a toggle switch instead of electronics. Since it's so cheap to drive a servo the proper way using an Arduino, I don't really see a point in tearing them apart and then trying to use them in a manner they were never intended. It doesn't have to be anything fancy like my controller, Geoff Bunza has the circuit and code for a super simple one that just uses one pushbutton per servo and drives a bunch of servos.

                                   --Randy


Modeling the Reading Railroad in the 1950's

 

Visit my web site at www.readingeastpenn.com for construction updates, DCC Info, and more.

  • Member since
    July 2009
  • From: somerset, nj
  • 2,913 posts
Posted by gregc on Sunday, February 23, 2020 2:05 PM

the PCB is left in place to hold the motor in place, but is disconnected from the motor.

the motor is driven thru external resistors that limit the current to < 50ma.

 

rrinker
Since it's so cheap to drive a servo the proper way using an Arduino,

you don't need an arduino, just a toggle switch

greg - Philadelphia & Reading / Reading

  • Member since
    January 2009
  • 4,640 posts
Posted by RR_Mel on Sunday, February 23, 2020 3:11 PM

I would think 50ma is too high, the SG90s I have get pretty warm at 40ma when working as a stalled servo.  I try to keep the current under 25ma.  At 25ma it keeps a bit of pressure on the point rails but the servos buzz a bit.
 
The circuit draws 14ma so the motors on my servos are drawing about 11ma, no heat at 11ma.
 
Mel
 
 
 
My Model Railroad   
 
Bakersfield, California
 
I'm beginning to realize that aging is not for wimps.
 
  • Member since
    July 2009
  • From: somerset, nj
  • 2,913 posts
Posted by gregc on Sunday, February 23, 2020 3:34 PM

just pick resistor values that allow the motor to turn without damage

greg - Philadelphia & Reading / Reading

  • Member since
    February 2002
  • From: Reading, PA
  • 27,700 posts
Posted by rrinker on Sunday, February 23, 2020 3:42 PM

 What I was saying is, it's cheap enough to drive them the porper way anmd not stall them, is it worth it to use them in a mode they weren't intended for? Even if they are cheap, the effort involved in crawling under the layout to repalce them when they burn out seems to make it not worth the risk.

 Are those special unidirectional resistors that there are two in parallel? Big Smile

                                             --Randy

 


Modeling the Reading Railroad in the 1950's

 

Visit my web site at www.readingeastpenn.com for construction updates, DCC Info, and more.

  • Member since
    January 2009
  • 4,640 posts
Posted by RR_Mel on Sunday, February 23, 2020 4:42 PM

OK
 
I took a Tower SG90 apart and checked the motor resistance, 3.7Ω.
 
The motor turns pretty good at .5 volts, yes that’s ½ volt.  The arm travel is 180° at .5 volts @ 50ma no load, stall at .5 volts is 103ma.
 
I think I’ll stay with using an Arduino.
 
 
EDIT:
Travel time at .5 volts is 180° in 5 seconds.
 
EDIT:
It still works after I put it back together.
 
 
Mel
 
 
 
My Model Railroad   
 
Bakersfield, California
 
I'm beginning to realize that aging is not for wimps.
 
  • Member since
    July 2009
  • From: somerset, nj
  • 2,913 posts
Posted by gregc on Sunday, February 23, 2020 6:58 PM

rrinker
Are those special unidirectional resistors that there are two in parallel?

parrallel to split the wattage.   For the 9V and 80ma case, a 0.7W 100 Ohm resistor is required or 2 200 Ohm 1/3 W resistors

greg - Philadelphia & Reading / Reading

  • Member since
    January 2009
  • 4,640 posts
Posted by RR_Mel on Sunday, February 23, 2020 7:06 PM

A continuous 80ma will cook the motor in nothing flat, they get hot at 50ma.
 
 
 
Mel
 
 
 
My Model Railroad   
 
Bakersfield, California
 
I'm beginning to realize that aging is not for wimps.
 
  • Member since
    February 2002
  • From: Reading, PA
  • 27,700 posts
Posted by rrinker on Sunday, February 23, 2020 8:01 PM

More and more this just sounds like a bad idea. And I didn't even pull one apart and test it like Mel did. I have a good candidate though, one of the ones on my workbench doesn't seem to work any more - I thought my circuit was bad or my code, but I hooked up a different one and it works fine. 

 I know this was in MR a while back - I posted then that i didn't think it was a good idea.

                                 --Randy


Modeling the Reading Railroad in the 1950's

 

Visit my web site at www.readingeastpenn.com for construction updates, DCC Info, and more.

  • Member since
    January 2009
  • 4,640 posts
Posted by RR_Mel on Sunday, February 23, 2020 8:33 PM

In servo mode the SG90 can draw up to 80ma or more moving the points but with the points in position the current drops down into the safe range, in constant current mode it will draw the highest current the resistors allow.  Not good.  You would have to limit the current to a safe level of at least a max of 30ma or face motor burnout.  At 50 ma the motor gets hot enough to melt the plastic housing over time.
 
 
 
Mel
 
 
 
My Model Railroad   
 
Bakersfield, California
 
I'm beginning to realize that aging is not for wimps.
 
  • Member since
    July 2009
  • From: somerset, nj
  • 2,913 posts
Posted by gregc on Monday, February 24, 2020 6:57 AM

RR_Mel
e safe range, in constant current mode it will draw the highest current the resistors allow.  Not good.  You would have to limit the current to a safe level of at least a max of 30ma or face motor burnout.

a 170 Ohm (5V/0.03A) will limit the current to 30 ma.   Is there enough gearing to move points with that little current?    if not, maybe a bigger servo

greg - Philadelphia & Reading / Reading

  • Member since
    February 2002
  • From: Reading, PA
  • 27,700 posts
Posted by rrinker on Monday, February 24, 2020 7:35 AM

 It should be, unless you have particualrly stiff points (might not move a non-hinges #4 with code 100 rail, for example). I have one stuck to the bottom of a Peco turnout and while I haven't hooked my peak reading meter in line to see what peak current I hit, unless I try to hold it back, it throws the points, spring still in place, without exceeding much more than that, and that's my entire circuit with the ATmega328 and a relay, plus two LEDs on the control buttons. 

                                --Randy

 


Modeling the Reading Railroad in the 1950's

 

Visit my web site at www.readingeastpenn.com for construction updates, DCC Info, and more.

  • Member since
    January 2009
  • 4,640 posts
Posted by RR_Mel on Monday, February 24, 2020 8:38 AM

gregc

 

 
RR_Mel
e safe range, in constant current mode it will draw the highest current the resistors allow.  Not good.  You would have to limit the current to a safe level of at least a max of 30ma or face motor burnout.

 

a 170 Ohm (5V/0.03A) will limit the current to 30 ma.   Is there enough gearing to move points with that little current?    if not, maybe a bigger servo

 

Should be, when stalled the current rose considerably on my bench power supply set to .5 volts.  I didn’t check the amount of pressure on the moving arm at .5 volts.
 
 
 
Mel
 
 
 
My Model Railroad   
 
Bakersfield, California
 
I'm beginning to realize that aging is not for wimps.
 
  • Member since
    July 2009
  • From: somerset, nj
  • 2,913 posts
Posted by gregc on Monday, February 24, 2020 9:39 AM

max current is always with the motor stalled.   there's no BEMF so it's just the winding resistance that limits the current.

presumably the gearing amplifies the torque

greg - Philadelphia & Reading / Reading

  • Member since
    January 2009
  • 4,640 posts
Posted by RR_Mel on Monday, February 24, 2020 11:19 AM

The gear ratio appears to be about 12.8:1.
Four sets of 32Tx10T spur gears.
 
I’m a EEE not mechanical at all but I ran the gears through the formulas and came up with 80% efficiency with 4.6 gain.
 
You must figure in that it’s been 60 years since I received my degree so ±50% accuracy.
 
 
 
Mel
 
 
 
My Model Railroad   
 
Bakersfield, California
 
I'm beginning to realize that aging is not for wimps.
 
  • Member since
    July 2009
  • From: somerset, nj
  • 2,913 posts
Posted by gregc on Monday, February 24, 2020 2:03 PM

RR_Mel
The gear ratio appears to be about 12.8:1.

so when turning an arm < 1 revolution, the motor is turning < 13 revolutions

greg - Philadelphia & Reading / Reading

  • Member since
    January 2009
  • 4,640 posts
Posted by RR_Mel on Monday, February 24, 2020 2:57 PM

Doesn’t sound right!
 
There are four sets of gears each with 10T to 32T.  That is using magnifier glasses through the housing.  10T to 32T is 3.2:1.
 
I added the four ratios instead of multiplying so it’s 104:1
 
Remember I'm not mechanical.
 
EDIT:
It’s hard to tell without complete disassembly but there could be one more gear and that would up the ratio to 355:1.
 
EDIT:

TowerPro SG90 Servo

Specifications

Modulation:
Analog
Torque:
4.8V: 25.00 oz-in (1.80 kg-cm)
Speed:
4.8V: 0.12 sec/60°
Weight:
0.32 oz (9.0 g)
Dimensions:
Length:0.91 in (23.0 mm)
Width:0.48 in (12.2 mm)
Height:1.14 in (29.0 mm)
Motor Type:
3-pole
Gear Type:
Plastic
Rotation/Support:
Bushing
Rotational Range:
(add)
Pulse Cycle:
(add)
Pulse Width:
500-2400 µs
Connector Type:
JR
 
 
 
Mel
 
 
 
My Model Railroad   
 
Bakersfield, California
 
I'm beginning to realize that aging is not for wimps.
 
 
  • Member since
    February 2002
  • From: Reading, PA
  • 27,700 posts
Posted by rrinker on Monday, February 24, 2020 3:22 PM

 Is that 4 ets counting the output gear on the shaft and the ger ont he motor shaft? In an hour I'll be able to look at one myself and tell. 4 sounds right though, 3 sets of gears plus the motor gear and the final output gear. 

 Somewhere I learned how to calculate gear ratios - I think it was an MR article on remotoring/regraring, actually. In an issue that might have been before my time. Or possibly in the old Machinist's Handbook which I guess was my Dad's, I also have a Carpenter's Handbook that was my grandfather's. That one is all wrong now, actual vs nominal measurements for lumber have changed over the past 70 years or so.

 And if I want to learn to tie knots or sail - I have two US Navy Bluejacket's Manuals - one from 1943 and one from 1951. 

                             --Randy

 


Modeling the Reading Railroad in the 1950's

 

Visit my web site at www.readingeastpenn.com for construction updates, DCC Info, and more.

  • Member since
    January 2009
  • 4,640 posts
Posted by RR_Mel on Monday, February 24, 2020 3:30 PM

I can see 4 large (32T?) gears so small (10T?) on the motor and big on the output should be 104:1.
 
 
Mel
 
 
 
My Model Railroad   
 
Bakersfield, California
 
I'm beginning to realize that aging is not for wimps.
 
 
  • Member since
    January 2009
  • From: Maryland
  • 8,961 posts
Posted by ATLANTIC CENTRAL on Monday, February 24, 2020 4:13 PM

So what is the potential goal here? Save money over the Tortise?

Or save space?

Sheldon

    

  • Member since
    July 2009
  • From: somerset, nj
  • 2,913 posts
Posted by gregc on Monday, February 24, 2020 4:16 PM

RR_Mel
should be 104:1.

so ~50 rev in 0.3 sec (0.12 sec/60°) to turn the arm 180 deg.   150 rev/sec or 9,000 RPM seems reasonable

greg - Philadelphia & Reading / Reading

  • Member since
    January 2009
  • 4,640 posts
Posted by RR_Mel on Monday, February 24, 2020 4:30 PM

ATLANTIC CENTRAL

So what is the potential goal here? Save money over the Tortise?

Or save space?

Sheldon

 

I’m into the servos for the savings and ease of installation.  I’m not replacing all of my Atlas switch machines, only as they crap out.  My newest Atlas is almost 30 years old.  I went with a couple of Peco PL-10 and modified Atlas turnouts before I tried the servos.  The last two have been servos.  The servos take a smaller footprint or hole under the turnout than the PL-10 and don’t need the Peco spring mod.
 
At this point I have two PL-10s and two Tower Pro SG90 servos.  I like both the PL-10 and the servos better than the Atlas #65, pitiful when new at best.
 
So far the servo costs are About $10 for the Arduino controller and $1.50 each for servos.  I already had the 5 volt power supply that powers both the servos and the Arduino.
 
 
 
Mel
 
 
 
My Model Railroad   
 
Bakersfield, California
 
I'm beginning to realize that aging is not for wimps.
 
  • Member since
    January 2009
  • 4,640 posts
Posted by RR_Mel on Monday, February 24, 2020 4:31 PM

gregc

 

 
RR_Mel
should be 104:1.

 

so ~50 rev in 0.3 sec (0.12 sec/60°) to turn the arm 180 deg.   150 rev/sec or 9,000 RPM seems reasonable

 

Sound about right to me.  They really get after it.
 
 
 
Mel
 
 
 
My Model Railroad   
 
Bakersfield, California
 
I'm beginning to realize that aging is not for wimps.
 

 

  • Member since
    July 2009
  • From: somerset, nj
  • 2,913 posts
Posted by gregc on Tuesday, February 25, 2020 9:21 AM

ATLANTIC CENTRAL
So what is the potential goal here?

for me it's interesting.  i'd like to understand what makes the tortoise a stall motor.     and knowing multiple solutions to a problem can save time and avoid unnecessary hardship.

greg - Philadelphia & Reading / Reading

  • Member since
    February 2002
  • From: Reading, PA
  • 27,700 posts
Posted by rrinker on Tuesday, February 25, 2020 9:45 AM

 The Tortoise works like it does because the motor is wound such that it consumes 15ma at 12V with the rotor locked. Gonna make me remember all that DC motor theory now.... It is in itself a very low torque motor, so there are a nice pile of gears to serve as torque multipliers so it has enough power to move switch points.

 I used to have one that was me "experiment on" unit - doing such nasty things as getting a rhythm going and it will bounce off one stop and nearly coast to the other side, try to fore it too hard and the gears strip past one another, and I took it apart countless times. One thing about them is, they are pretty much indestructible. That poor abused one would still work fine when put back together - a little noisy, but ti would move back and forth just like it was supposed to.

 Pretty sure I even ran the motor with no gears installed, it's very quiet - allt he noise is the gears meshing. I never put a meter to it and checked the curent draw running free like that, or tested the winding resistance. Nor do  recall how many poles the motor had. 

 That low current means it's within the output level of many logic ICs, and other generally low power parts like comparators or op amps. And of course 15ma is generally well within the safety margin to run an LED. Those first Tortoise knockoffs drew a lot more current, meaning no simple series LEDs, and any drive circuit needed an extra stage to handle the current. SOme of those that are still around have newer models with much less current draw, but while better than their original model, is still 10-20ma more than a Tortoise.

 Servos are nice, when used as intended, because the power supply is independent of the control signal. The control signal draws nearly no current (hmm, have to try measuring that, my one meter has a low burden low ma range just for such things), so a driver can control a whole lot of servos without any special high current signal considerations.

 None of this really matters if you are simply using DPDT or SPDT toggles and no electronics, since even rather small toggle is usually rated for 1 amp or more. It all comes down to what the motor can stand without melting the windings.

                                       --Randy


Modeling the Reading Railroad in the 1950's

 

Visit my web site at www.readingeastpenn.com for construction updates, DCC Info, and more.

  • Member since
    January 2009
  • From: Maryland
  • 8,961 posts
Posted by ATLANTIC CENTRAL on Tuesday, February 25, 2020 10:29 AM

I programmed early PLC's to control assembly lines, pumping stations and other industrial applications in the 1980's. And I see the value in solid state logic for complex operations. We used PLC's to replace relay stacks that had dozens of inputs and relays in a single function.

But when the logic function is simple, a chain of only two, or three, or even four conditions, I don't really see the point. 

Multiple control locations require the same amount of input/output wiring no matter the logic method. 

I control complex interlockings from two different locations, with just a few $3 relays and LED lighted pushbuttons, using the same power supply for everything. Just seems so much simpler to me.

And I already have nearly 100 tortoise machines.....

I will still be curious what Greg felt the goal here was, saving money?

Sheldon

    

Subscriber & Member Login

Login, or register today to interact in our online community, comment on articles, receive our newsletter, manage your account online and more!

Search the Community

ADVERTISEMENT
ADVERTISEMENT
ADVERTISEMENT
Model Railroader Newsletter See all
Sign up for our FREE e-newsletter and get model railroad news in your inbox!