Better holding servo than Hitec 322?

lbenson

Senior Member
I'm trying an update of my Rivercam project ( http://www.picaxeforum.co.uk/showthread.php?13705-Rivercam-with-PICAXE-pan-tilt-control ) with a better camera and zoom. Unfortunately I'm finding with the heavier camera that there is jitter every few seconds when the tilt servo is outside of the range of about 145-155--I think because the weight of the camera. I've got it about as well balanced as I can front to back.

I'm using Hitec 322 servos. I tried a Hitec 422 but it was worse (not surprising--45.8 oz-in compared to 51 for the 322). Pan is stable.

I've done a lot of things to try to get rid of jitter from various sources, including serin commands, which I've replaced with hserin. I don't need very much tilt range--130 to 170 would probably be fine. I'm using sevropos after an initial servo command, and am moving one step at a time from one input point-of-focus to another. That goes smoothly.

Can anyone recommend a servo with better holding power? Or some other solution to my problem?

Edit: Running with 5 volts--would I do better with 6?
 
Last edited:

srnet

Senior Member
Any chance of a summary ?

A heavy camera is how heavy ?

Hitec 322 servos are not exactly heavy duty servos.
 

nickwoodrow

New Member
i'm looking at this for a similarish problem: http://www.ebay.co.uk/itm/121209419335?ssPageName=STRK:MEWAX:IT&_trksid=p3984.m1438.l2649 sail winch. I'm new to servos and am learning quickly bt I must admit I don't understand the "2 turns" bit! Surely it is either a "Normal servo" or a continuous one???? 6v is better i believe, isn't that the normal input voltage? this http://www.ebay.co.uk/itm/GS-5515MG-15kg-High-Torque-Throttle-Steering-RC-Servo-Metal-Gears-/330842758889?pt=UK_ToysGames_RadioControlled_JN&hash=item4d07c3dee9 runs between 4.8 - 7.2v the higher the volts the greater the torque! (I'll probably get slaughtered here later)! Just trying to help.
 

rossko57

Senior Member
I don't understand the "2 turns" bit! Surely it is either a "Normal servo" or a continuous one????
It's a winch. Winches pay out or pull in a line. The line can be all out, somewhere inbetween, or all in. It may take two or twenty turns of the winch drum between the extremes, depending on design. So no, its neither 'normal' nor 'continuous', its multiturn but within limits.
 

lbenson

Senior Member
Attached is a photo of my test rig with a Canon A590 IS with 4x optical zoom. It weighs 175 grams.

I intend to replace it with a Canon SX120 with 10x zoom--weight 245g.

Tomorrow I'll try to reduce the torque by reducing the height of the bracket which attaches the camera to the pan servo. I'll also add an arm to center the weight of the camera over the center of the pan servo wheel. And I'll see if I can get 6V DC somewhere.

Canon cameras can load an alternative firmware from SD--chdk ( http://chdk.setepontos.com ). Together with additional chdk-associated software, zooming, shooting, and downloading can be controlled externally--for instance using a small linux device like the Seagate Dockstar. I documented this here:

http://chdk.wikia.com/wiki/Ptpcam_in_headless_linux_Dockstar

and here:

http://chdk.wikia.com/wiki/Chdkptp_in_headless_linux_Dockstar_-_remote_control

Current pan and tilt code with hserin input. Input is "p" or "t" for pan or tilt, then "g###" for goto position (3 digits or 2 digits followed by non-digit).

Code:
' 08Srv_1D_New2 test servo
#picaxe 08m2

symbol bActiveServo = bit0 ' 0=pan,1=tilt

symbol panSpot = b13
symbol tiltSpot = b12
'symbol activeServo = b11
symbol activeSpot = b10
'symbol lastSpot = b9
'symbol otherServo = b8
'symbol otherSpot = b7
'symbol lastPanSpot = b6
'symbol lastTiltSpot = b5

'symbol b4=scratch
'symbol b3=scratch
'symbol b2=scratch

symbol cPanActive=0
symbol cTiltActive=1
symbol panServo = 2
symbol tiltServo = 4
symbol panMin = 75
symbol panMax = 225
symbol servoCenter = 150
symbol tiltMin = 75
symbol tiltMax = 225

pause 4000               ' wait 10 seconds
activeSpot = servoCenter 
panSpot = activeSpot       ' center
tiltSpot = activeSpot      ' center
'#rem
servo tiltServo,tiltSpot
pause 30	
bActiveServo=cPanActive
servo panServo,panSpot
pause 30
servopos tiltServo,tiltSpot
pause 30
servopos panServo,panSpot
pause 30
hsersetup B2400_4, %00001000 ' input true for M2 ("T"-bit2=0), disable hserout
pause 100
'endrem

main:
  do
'   serin 1,N2400,b1
   b1 = 0 ' invalid byte
   hserin b1
   if b1 = "p" or b1 = "t" or b1 = "g" then
'    sertxd(b1, " ")
    select b1
     case "p"
       bActiveServo=cPanActive
     case "t"
       bActiveServo=cTiltActive
     case "g"
'      serin 1,N2400,#activeSpot ' e.g., "pg75 "
'#rem
       b1 = 0 ' invalid byte
       do while b1=0 : hserin b1 : loop
       if b1 => "0" and b1 =< "9" then
         activeSpot = b1 - "0"
         b1 = 0 : do while b1=0 : hserin b1 : loop
         if b1 => "0" and b1 =< "9" then
           activeSpot = activeSpot * 10 + b1 - "0"
           b1 = 0 : do while b1=0 : hserin b1 : loop
           if b1 => "0" and b1 =< "9" then
             activeSpot = activeSpot * 10 + b1 - "0"
           endif
         endif
       endif
'#endrem
      if bActiveServo=cTiltActive then
        activeSpot = activeSpot min tiltMin
        activeSpot = activeSpot max tiltMax 
      else
        activeSpot = activeSpot max panMax 
        activeSpot = activeSpot min panMin 
      endif
      gosub servoOut
'      sertxd("+")
    endselect
   endif
  loop

servoOut:  ' smoothly move from lastSpot
' sertxd("spta ",#bActiveServo," ",#panSpot," ",#tiltSpot," ",#activeSpot,cr,lf)
  if bActiveServo = cPanActive then
    do while activeSpot <> panSpot
      if activeSpot < panSpot then
        dec panSpot
      else
        inc panSpot
      endif
'      gosub movePanServo
      gosub moveServo
    loop
  else
    do while activeSpot <> tiltSpot
      if activeSpot < tiltSpot then
        dec tiltSpot
      else
        inc tiltSpot
      endif
      gosub moveTiltServo
'      gosub moveServo
    loop
  endif
  return

movePanServo:
moveServo:
  servopos panServo,panSpot
  pause 30
  return

movetiltServo:
  servopos tiltServo,tiltSpot
  pause 30
  return
 

Attachments

Hemi345

Senior Member
I think I'd use some wider stock for the right angle piece that the tilt servo is mounted to. Then you can lay the tilt servo 'flat' to reduce the height of the channel piece (that the camera is coupled to the pan servo with) to just flat stock. That would get the center of gravity down quite a bit and allow you to put the pan servo centered under the lens rather than offset where the tripod mount is.
 

lbenson

Senior Member
I'm sorry, Hemi--I've been having trouble right along visualizing the permutations of this pan and tilt setup. Can you explain more fully?

Is the wider vertical right angle support piece head-on as in the photo or broadside? And the tilt servo goes how?

I had planned to put an arm on the tripod mount so that the lens (or center of gravity) of the camera would be directly over the center of the pan servo wheel.

Is the purpose of your rearrangement of the servos to get the wheel of the pan servo above the rest of the apparatus so that the raised bracket is not needed--just the centering arm? If so, could I not do that just by turning the tilt servo around 180 degrees, with the right-angle support as it is?

Thanks for the suggestions. By the way, I'm using your picaxe library for eagle--thanks for that too.
 

techElder

Well-known member
I don't think you have near enough hardware supporting that camera. Usually you would want the weight of the camera to be supported by bearings and just use the servos for motion.

The way you have it, I'm concerned that you are putting way to much "angular" force on the servo bearings. I doubt that they are designed to work with those forces.
 

G-Man

New Member
Use an offsetting weight. Place a similar weight below the camera tilt mechanism and that will mean that the mechanism is perpetually in balance. Now, when the camera is upright, it's "balanced" but the more you tilt it the more it'll be off balance. The same sort of affair as I'm describing is used on telescopes.
 

lbenson

Senior Member
@Texasclodhopper -- sounds reasonable. Can you point to examples?

@G-Man -- also reasonable, but I'm not sure I want to add more weight for the tilt servo to hold, even if it is better balanced.
 

erco

Senior Member
Don't add weight, redo your mechanism so the camera is hanging underneath everything, and try to keep the CG in line with your servo axes. And use ball-bearing, metal-gear quarter-scale servos, you are taxing the heck out of those poor little standard servos.
 

oracacle

Senior Member
yes i would either have the camera hanging below or, have it monted in a gimble like frame (something along the line of those gyroscope thing you see on tv that make people really very ill)

like this but with only 2 axis: https://www.youtube.com/watch?v=tjnYNpvZsQA

you can then have the servo alter it angle from the axis. also depending on how important it is you could reduce paralaxing (important for panoramic photography, maybe not here though)
 

lbenson

Senior Member
Erco--I tried first with the camera underneath, since that would work better for my ultimate mounting, but that was worse. But I didn't have the jitter bugs worked out, and no proper support or balancing. I'll try again taking into consideration the suggestions people have made.

Oracacle--whoa--makes me sick just to watch it.

I think I should make the pan servo connected to the support, and the tilt connected to the pan instead of the other way around as I have it now.
 

Hemi345

Senior Member
I'm sorry, Hemi--I've been having trouble right along visualizing the permutations of this pan and tilt setup. Can you explain more fully?

Is the wider vertical right angle support piece head-on as in the photo or broadside? And the tilt servo goes how?

I had planned to put an arm on the tripod mount so that the lens (or center of gravity) of the camera would be directly over the center of the pan servo wheel.

Is the purpose of your rearrangement of the servos to get the wheel of the pan servo above the rest of the apparatus so that the raised bracket is not needed--just the centering arm? If so, could I not do that just by turning the tilt servo around 180 degrees, with the right-angle support as it is?

Thanks for the suggestions. By the way, I'm using your picaxe library for eagle--thanks for that too.
Yeah, on 2nd glance, it looks like you could flip the tilt servo upside down and get the camera placement closer to the axis of the tilt servo's shaft. Then you wouldn't need the rectangle piece between the camera and the pan servo. Just flat piece of aluminum fastening to the camera's tripod mount and then to the servo horn on the pan servo.


Code:
   --------
   |   O   |
   --------
       v--^
  [T]>[P]
 

lbenson

Senior Member
Thanks to all, especially those who focused my attention on the need for having the axles of the servos in line with each other and with the center of the lens. Here, attached, is my second try, supported from the top with the pan servo on top and the tilt servo beneath.

This is much better, with no floppiness within the tilt range that I am looking at. There is still some jitter that seems anomalous to me--every few seconds at tilt of 150 but not 152 or 148. At 150 everything is about as close to vertical as I can get it.
 

Attachments

G-Man

New Member
@Texasclodhopper -- sounds reasonable. Can you point to examples?

@G-Man -- also reasonable, but I'm not sure I want to add more weight for the tilt servo to hold, even if it is better balanced.
You wouldn't necessarily be "adding more weight", you'd be offsetting the weight of the camera. It will relieve the servo of the stress of fighting against gravity. that is how motorized telescopes get away with such small motors moving such large amounts of weight:

 

oracacle

Senior Member
they also spend alot of time and money on R&D (notice the offset verticle pivot in your example) to get the counter balance distance and size correct, and unless you get it correct you will cause more issues than you solve - if the shaft is too long with much wait you will have the smae problem currently being experienced to short with too little wait you will just exaserbate the current issue.

however if you can post some of the maths involves in calculating the correct size weight and length of shaft in relation to the pivot point point i am sure we can design something that would be acceptable.
 

AlbertZ

Senior Member
I'm trying an update of my Rivercam project ( http://www.picaxeforum.co.uk/showthread.php?13705-Rivercam-with-PICAXE-pan-tilt-control ) with a better camera and zoom. Unfortunately I'm finding with the heavier camera that there is jitter every few seconds when the tilt servo is outside of the range of about 145-155--I think because the weight of the camera. I've got it about as well balanced as I can front to back.

I'm using Hitec 322 servos. I tried a Hitec 422 but it was worse (not surprising--45.8 oz-in compared to 51 for the 322). Pan is stable.

I've done a lot of things to try to get rid of jitter from various sources, including serin commands, which I've replaced with hserin. I don't need very much tilt range--130 to 170 would probably be fine. I'm using sevropos after an initial servo command, and am moving one step at a time from one input point-of-focus to another. That goes smoothly.

Can anyone recommend a servo with better holding power? Or some other solution to my problem?

Edit: Running with 5 volts--would I do better with 6?
Use a Hitec 645MG. 107 oz-in of torque @ 4.6 volts. These are used for steering servos on 1/10 scale RC cars. The 322 & 422 are way too puny.
 

G-Man

New Member
they also spend alot of time and money on R&D (notice the offset verticle pivot in your example) to get the counter balance distance and size correct, and unless you get it correct you will cause more issues than you solve - if the shaft is too long with much wait you will have the smae problem currently being experienced to short with too little wait you will just exaserbate the current issue.

however if you can post some of the maths involves in calculating the correct size weight and length of shaft in relation to the pivot point point i am sure we can design something that would be acceptable.
The counterweight is adjustable up and down the shaft to achieve balance by means of a simple thumb screw. And if by "notice the offset verticle pivot in your example" you mean the angle of the pivot, that has nothing to do with the rotation and everything to do with aligning the telescope to the axis of rotation of the planet so that stars can be more easily tracked.
 

oracacle

Senior Member
and were does ont get these adjustable weight for for the camera with adjustable thumb screw and shaft to fit it upon.

msot of us dont have the work shop or materials to make such items, so it case of you add the weight, you get it wrong and its worng and causes issues.

also the ofset pivot will affect the counter balance, it mean that less will be needed, it also one of the only way that can achieve the the counter balnce in aligment with the scope.

the idea is mostly impracitacle due to the difficulties of correct implimentation. things like the centre of gravity for the camera would have to be found, which could also mean not rotation around the optical centre of the lense. dont forget this is camera, it a different animal to scope. heck with a telephoto lense on my camera the centre of gravity is way infront of the optical centre, the camera mass is offset to the right due the placement of the battery . by using counter weights, you wold have to use several to correctly counter the offsentre nature of a cameras mass (i would need a weight to the left and either below with variation for diffent lense, and settings (inclusive of zoom and focus, which both change the optical centre), or to the rear with similar adjustment)- a scopes mass tend to be along the centre and fairly evenly distrubuted along it entire length making it easy to counter balance.

counter balancing is a fine art and unless you can accuratly account for every variable you will run into trouble.
 

techElder

Well-known member
Hey, guys. Its a tiny camera in an upper window on a ledge that is 6 inches wide. Get a grip ...
 

oracacle

Senior Member
yeh do have a grip, part of the reason i was trying to explain why counter balance would be far too much even if you were to put the engineering chalanged of getting it right to one side
 

rossko57

Senior Member
Well to be fair, rigging weights is less of an engineering challenge than the pan/tilt mechanism. For test purposes looking to see if the jitter is weight/creep induced, it could be lolly sticks and blobs of putty.
 

G-Man

New Member
Well to be fair, rigging weights is less of an engineering challenge than the pan/tilt mechanism. For test purposes looking to see if the jitter is weight/creep induced, it could be lolly sticks and blobs of putty.
Or duct tape and some washers or coins...

Fine, man don't take the advice. That's up to you. I'm telling you it would work. If you could fab a pan/tilt mount adding some weight to the bottom of it HAS TO BE a piece of cake in comparison. It also doesn't have to be that fine tuned. ANY amount of offsetting weight (less than or equal to the camera's weight) will relieve the servo from needing to hold some of that weight with braking force... What is happening is the PWM pulse comes into the servo and sets it to the desired position. Then the pulse drops and the servo slips some because it is no longer under power. Then the next pulse comes in a resets it back to the desired position. This is what is causing the oscillation in the mount. But like I said, it's your build and I'm just giving you advice which you can take or leave as wish.
 

flyingnunrt

Senior Member
Have you thought about ganging up a couple of servos.
Quite a common thing to do with large RC models.
But you have to pay attention to getting the geometry exact or they can fight against each other a bit.
ganged servos.jpg
 

lbenson

Senior Member
Thanks for the suggestions. With the servo axes aligned with each other and the middle of the lens, and the camera mounted beneath the servos as shown in post 17, I am now sufficiently satisfied with the pan and tilt motion to move on to the rest of the project.

I am looking at a full range of pan values, 75-225, and a range for tilt of about 130-150. If it turns out on testing that there are specific settings at which jitter appears, I can avoid them. If the problem appears larger, I'll try some of the counter-weighting suggestions, or bigger/better servos.

As mentioned, these are not really big cameras--Canon A590 IS @ 175 grams or Canon SX120 @ 245g, though I may sometime want to try my Canon SX40HS at 595g--that may take more muscle and better balancing.

Integration with the zoom and remote shooting with chdk is my next challenge.
 
Top