Society of Robots - Robot Forum

Mechanics and Construction => Mechanics and Construction => Topic started by: Bart_at_Haunted_High on November 11, 2011, 07:34:11 PM

Title: 360 Servo Hack and the POT function
Post by: Bart_at_Haunted_High on November 11, 2011, 07:34:11 PM
I have a question specific to the function of the built in POT in the servos, and the 360 deg hack where the pots are removed and jumpered with resisters.


If I wanted to limit the rotation to (4) 360 deg turns, could I use an external drive gear with the correct ratio , driving the POT with long leads from its original internal location, to limit the turns on the servo itself?

What Im trying to accomplish is high speed linear motion using a 360 deg hacked servo, driven with conventional PWS servo control.  Will this do what I'm hoping it will?


Title: Re: 360 Servo Hack and the POT function
Post by: Soeren on November 11, 2011, 08:48:23 PM
Hi,

What Im trying to accomplish is high speed linear motion using a 360 deg hacked servo, driven with conventional PWS servo control.  Will this do what I'm hoping it will?
Depends on how you define "high speed", but I wouldn't use that description in relation to servos, as they're typically around 60RPM give or take. With "PWS", I assume you mean PDM a.k.a. a standard hobby servo signal?

Why not stick to your first thread and give us the numbers we need to help you?
Title: Re: 360 Servo Hack and the POT function
Post by: Bart_at_Haunted_High on November 11, 2011, 08:58:55 PM
Thats just it..

I dont have any numbers.
Except that the linear length I need is about the full length of a guitar neck. so for arguement lets say 18 inches.

As for speed, how qyickly does one change frets from one to another when they're playing power chords in a rock song.?

A slide down the neck from top to bottom fret should take 2 second. ( 1 second would be ideal )
Thats seems fast, and most ( say 98% of the time ) fret movement is only a few inches every 1/2 second to every 1 second.
Title: Re: 360 Servo Hack and the POT function
Post by: Bart_at_Haunted_High on November 12, 2011, 09:32:56 AM
Yes, PDM.... I'm still learning all the jargin.  Thank's for being kind to the mistakes of a newbie ;-)
Hi,

What Im trying to accomplish is high speed linear motion using a 360 deg hacked servo, driven with conventional PWS servo control.  Will this do what I'm hoping it will?
Depends on how you define "high speed", but I wouldn't use that description in relation to servos, as they're typically around 60RPM give or take. With "PWS", I assume you mean PDM a.k.a. a standard hobby servo signal?

Why not stick to your first thread and give us the numbers we need to help you?

Title: Re: 360 Servo Hack and the POT function
Post by: Soeren on November 13, 2011, 07:44:59 AM
Hi,

Thats just it..

I dont have any numbers.
Except that the linear length I need is about the full length of a guitar neck. so for arguement lets say 18 inches.

As for speed, how qyickly does one change frets from one to another when they're playing power chords in a rock song.?

A slide down the neck from top to bottom fret should take 2 second. ( 1 second would be ideal )
Thats seems fast, and most ( say 98% of the time ) fret movement is only a few inches every 1/2 second to every 1 second.
Seems like you could come up with some number after all  ;)

Then there's the rest of the info that I asked you about and since you seem unwilling to follow up on that post, I have repeated it here...
Quote
We need some info/metrics on guitar as well as hand to help.
Like...
Is it a real guitar that should remain functional?
If not, what is it made from and what are the dimensions of the fret board?
The hand is made of what?
And weighs (approximately) how much?
Of what material?
Doesn't even take numbers to answer those, but they sure would help toward a solution.
And please remember, it's really not us who need this info, it's you, as you need to post the info to get the help you want!

The solution depends a lot on whether you can modify this guitar (neck) or not and on the weight of the hand/arm that needs to be moved.

If it's to look as "real life" as possible, remember that the hand movement time is not the time in between each note played, but rather swift in the up/down movement.
The time to move over the entire fretboard shouldn't be radically different from going eg. 4 frets - I didn't get the impression of the rock band playing slow mourning music?
Title: Re: 360 Servo Hack and the POT function
Post by: Bart_at_Haunted_High on November 13, 2011, 09:14:50 AM
We need some info/metrics on guitar as well as hand to help.
Like...
Is it a real guitar that should remain functional?
No, Not at all.. I'll probably manufacturer the body and neck around what ever infrastructure that is created for the movement, and not the other way around.
If not, what is it made from and what are the dimensions of the fret board?That is yet to be determined
The hand is made of what
The hand and arm are going to be pre existing parts from a light weight blown foam PVC Skeleton, which will have loosely fitted hinge points at shoulders, elbows and wrists for very free movement.
And weighs (approximately) how much? Oh Gosh, I don't have that info yet, but it can be had in a week or two when I get access to those parts. Less that 2 pounds?    I don't know.
Of what material?
 a light weight blown foam PVC  

The solution depends a lot on whether you can modify this guitar (neck) or not and on the weight of the hand/arm that needs to be moved.

If it's to look as "real life" as possible, remember that the hand movement time is not the time in between each note played, but rather swift in the up/down movement. YES
The time to move over the entire fretboard shouldn't be radically different from going eg. 4 frets - I didn't get the impression of the rock band playing slow mourning music?  Correct in your thinking
Title: Re: 360 Servo Hack and the POT function
Post by: Soeren on November 13, 2011, 10:39:16 AM
Hi,

No, Not at all.. I'll probably manufacturer the body and neck around what ever infrastructure that is created for the movement, and not the other way around.
Now we're cooking ;D

The DIY guitar approach means you can make (route) or build a channel on the back side of the neck to get the mechanism out of (frontal+/-) sight.

If the motor is placed in the body of the guitar and a pulley screwed into the head stock, a wire can be used to transfer the rotational movement into a linear one very cheaply.
To make "stops", either the motor can be pulse on for short periods (experiment when the setup is in place), or you could use a "flag" mounted to the wire and passing slotted optocouplers, either one for each fret, or a single and several flags. The same goes for end-stops.

I'd guess a suitable speed (total distance) should be max. 0.2s to 0.3s.
2.5m/s equates to 0.183s for the 18", so for a motor pulley with a diameter of 1", it means 31.3 revolutions at slightly over 10kRPM - oops, clearly not practical, so a larger pulley is needed and perhaps less speed demand.
There's a limit to how large a pulley diameter is possible, but a 4" diameter is still hide-able and would reduce the need to ~2500RPM and a 6" pulley would reduce it to ~1700RPM.
However, the larger the diameter, the more torque is needed.
While I don't know your skeleton arm, I don't think the needed torque to move it back and forth will be much above 1..2oz, but that you'll need to verify yourself (fishing weight scales or similar).
Assuming a 4" pulley have a radius of 2", so if the required force to move the hand is 2oz, the motor would need a torque of greater than 2*2=4 ozf-in - Double that would be prudent, so 8 ozf-in would be the goal (with a 4" pulley and a speed of 2500 RPM loaded).

If you'll settle for less speed, you will need less torque of course.

The pulley system needs to be firm and well made. The pulley on the head might need a ball bearing, as it needs to be smaller and will thus rotate that much faster.


If it's hard to follow what I described, I cn make you a drawing.
Title: Re: 360 Servo Hack and the POT function
Post by: Bart_at_Haunted_High on November 13, 2011, 03:50:41 PM
I'm not so foolish to think I'm following everything, and not foolish enough to refuse a helping hand. heck yes, I would greately appreciate a drawing.

You totally lost me on the flagging , or stopping at each fret.
Title: Re: 360 Servo Hack and the POT function
Post by: Bart_at_Haunted_High on November 13, 2011, 03:58:55 PM
I just got my hands on an older Japaneese Stratocaster knockoff guitar that I thought would make a good model for this project. and after taking a quick measurement, Total Linear Motion up-and-down the neck is more like 11 to 12 inches. That is a good number to work with for this project. I'm confortable with that.
Title: Re: 360 Servo Hack and the POT function
Post by: Bart_at_Haunted_High on November 13, 2011, 04:06:06 PM
wait, so the average unhacked servo turns a bit over 180 degrees.
If I want 11 inches of travel, I would need a drive pully mounted on a heavy duty servo of about 22 inches in Circumference? or about 7" Diameter?  Would that work?

( My head hurts )
Title: Re: 360 Servo Hack and the POT function
Post by: Soeren on November 13, 2011, 06:38:57 PM
Hi,

I'm not so foolish to think I'm following everything, and not foolish enough to refuse a helping hand. heck yes, I would greately appreciate a drawing.

You totally lost me on the flagging , or stopping at each fret.
Flag, see attached sketch.
Frets, the "bands" down a guitar neck.

Thee sketch is very loose and just goes to show how I'd see things go together. The optocouplers would mount in the neck of course.

Title: Re: 360 Servo Hack and the POT function
Post by: Soeren on November 13, 2011, 06:47:08 PM
Hi,

wait, so the average unhacked servo turns a bit over 180 degrees.
If I want 11 inches of travel, I would need a drive pully mounted on a heavy duty servo of about 22 inches in Circumference? or about 7" Diameter?  Would that work?
I'm repeating myself here... You cannot use a servo for this.

Take a look at the sketch and let's do the numbers for the 11..12" tomorrow, it's nearing 2 a.m. here and I have an appointment on the hospital for the next two days (just blood samples tomorrow though, but a fair bit of travel time), so I have to pull some Z's soon.


( My head hurts )
A bottle of Scotch will sort that out in 30 minutes (one way or the other)  ;D
Title: Re: 360 Servo Hack and the POT function
Post by: Bart_at_Haunted_High on November 17, 2011, 06:51:38 PM
The Scotch worked too well, what few braincells I had working on this problem.... I destroyed.

Now I have to start all over.
Title: Re: 360 Servo Hack and the POT function
Post by: Soeren on November 20, 2011, 11:06:22 AM
Hi,

Now I have to start all over.
Just say when... (you're rebooted) :)
Title: Re: 360 Servo Hack and the POT function
Post by: Bart_at_Haunted_High on November 24, 2011, 07:18:37 AM
Musical choices are all up for grabs now, but I like some of the rockabilly/psycobilly offerings out there with a halloween theme to them. It also has that raw "garage band" sound to it.
Easily identifies the number of instruments without a lot of synthasis or overdubbing, at most a little heavy on the reverb and echo at times, but a real retro feel. 1950s -  1960s.

Yes, the timeline may seem long but we will eat up every minute from now till September 2012 getting this ready.

I use a software called  Brookshire VSA http://www.brookshiresoftware.com/vsa_overview.htm (http://www.brookshiresoftware.com/vsa_overview.htm) for syncing the audio to the servo movement.  and the servo driver board of choice is a Lynxmotion SSC-32http://www.lynxmotion.com/p-395-ssc-32-servo-controller.aspx (http://www.lynxmotion.com/p-395-ssc-32-servo-controller.aspx)

Along with the guitars I will also need to add some automation to a standard 5 piece drum set. I just picked it up yesterday ( even has a cowbell .... snnicker )
Title: Re: 360 Servo Hack and the POT function
Post by: Bart_at_Haunted_High on November 24, 2011, 07:29:09 AM
Soeren,

You had also asked about hand position on the neck, and how the hand not only slides up-and-down the neck, but how the grip changes on the neck depending on the chord being played.... at this point I'm willing to surrender some of that to keep it simpler.  If the hand has one grip but can slide up and down the neck I'll be happy.

I think I'd put more emphasis on the other "hand - arm" combo, so I can have a look for picking ( wrist movement ) and Strumming ( elbow movement ) .

Also the guitar itself would be mounted axially thru the "Guitar Body" to the Skelleton waist or hip area so there is some action to the guitar body itself. ( slightly rocking up and down )

Ideally I'd also like to add some side-to-side twist at the torso of the skelleton or similar motion, as it can only add to the "show". Otherwise the figures may still look too "static".

First step though, has to be a workable sollution for the guitars.  I'm toying with switching out the bass guitar for a stand up bass, but the solution for the neck of the instrument would still be the same.
Title: Re: 360 Servo Hack and the POT function
Post by: Soeren on November 28, 2011, 02:40:18 PM
Hi,

Nice software (if you have it), but with a little more work and a 5.1 (or7.1) sound setup and running only stereo, it' possible to use a track for timing and perhaps another for triggering the events.

Anyway, how are your options for making the mechanical work on the neck?
If you'd rather use a timing belt and the associated pulleys, the gearing will just have to be adapted (no problem).

Just discovered that I got the earlier calculations 10 times too large - that's what happens when you get too lazy to grab a calculator and have the boob tube running in the background - oh well, lesson learned ;D

Did you get around to find the force needed to move the hand?
Title: Re: 360 Servo Hack and the POT function
Post by: Bart_at_Haunted_High on November 29, 2011, 05:30:53 PM
I have Brookshire VSA 4.0, and will probably upgrade to the latest version buy programing time for this project.

The force must be ounces, but untill I build up a mock up and check the resistance, I won't know a true number. So if your original calculations were off by a factor of 10... I'm back to asking if an RC servo will do the trick. ( Don't throw anything at me, just asking )
LOL

Bart
Title: Re: 360 Servo Hack and the POT function
Post by: Soeren on November 29, 2011, 06:18:40 PM
Hi,

I have Brookshire VSA 4.0, and will probably upgrade to the latest version buy programing time for this project.
Do you really need that?
Or to elaborate... Is the new version that much better that it's worth throwing all that money after it, rather than at some hardware for the project?


So if your original calculations were off by a factor of 10... I'm back to asking if an RC servo will do the trick. ( Don't throw anything at me, just asking )
LOL
Not throwing a thing (not even a fit), but I am typing the following really slow  ;)
I arrived at 10,000 RPM where it should have been 1,000RPM.
Fast servos are around 150ms for 60° and this equates to less than 67RPM (which I mentioned already). If you want to spend major cash, you might get up to around 100RPM - Perhaps a picture paints a 1000 words...
(http://That.Homepage.dk/Img/ServosAreWeaklings.jpg)
The tiny motor on the left/front is a servo motor (from a broken servo). The other motor is a 9V motor, surplus originally intended for cordless drills and sans gearbox. It's something around that size you should aim for.