Teensy and design mods

[Boguinhos]

Hi Collin!
Nice work you have been developing on top of the already great job isaac879 did.

I noticed that you raised up the tilt axis and read on YT that you were focusing on using the teensy instead of the arduino.
How did you managed with it? is it worth it? I kinda lost track on knowing where you at in this moment.
I really like this design but i was hoping to see if i had any chance on having it controlled by VISCA IP.
But i haven't found any kind of "visca to ptz" library, only to make ptz rs232 controllers. And "reverse engineering" it would be too much for what i can chew. But i'm guessing that the arduino won't even have the amount of hardware resources to plug in a w5500 or similar. ( i'm not needing the slider part).
Oh, and , by the way, also read that you ditched the hall sensors. did you regret it?

Thanks.
João

[cbradburne]

Hi João,

Sorry I've only just spotted your comment. I mustn't be receiving notifications. Anyway, better late than never!!

Yes, I raised the tilt axis to accommodate a longer camera. Video cameras tend to be much longer than DLSRs. Keeping the tilt axis at the original height would've meant restricting the tilt angle as the camera would've hit the mount if tilted too far forwards or backwards.

For my own mount I am using Teensy 3.2 boards as the Nano has very little program storage and is slow, but mostly there is a library built specifically for the Teensy called "TeensyStep" https://github.com/luni64/TeensyStep in which you can control multiple stepper motors with acceleration. Isaac's code tries to do acceleration but it never worked correctly for me. For me, when running with Isaac's code, the motors would start the move in the wrong direction, then slow down and move the right way, and then stop abruptly with no deceleration. Also the extra speed of the Teensy means there is plenty of overhead and greater stepping speed.

Worth noting though that I've also focused on the video side and removed the "photo" side. So I've made it so I can set 6 different positions and you can move to any of those 6 positions directly, with acceleration. Rather than having a list of positions and only being able to move between them in order, like the original design.

I haven't looked at VISCA IP. To be fair, I'm fairly new to the PTZ game. Do you already have some VISCA gear?

The hall sensors can be added quite easily so there's no chance of regret :) Like I've said previously, you only need hall sensors to home the mount, and if you can't permanently store the camera positions between power cycles, then there's no point ever homing.

Did you build one of these mounts?

I'm hopeful Isaac will return to it when he has the time. I've just been trying to help people out with it when I can as I know I had a lot of questions before and while I was building mine.

Thanks for getting in touch,
Colin

[Boguinhos]

Hi Collin,
Yeah i wrote this before i've started to comment on your youtube videos, so actually we have "spoken" before.
About teensy:
I've used some variety of arduinos and esp before, never any teensy and i really didn't have the need for it so i am not aware of the full potential that it has. I agree on the acceleration part as i've been checking out the code and trying to understand it and i've noticed that acceleration is only available when pan or tilt move is executed individually , even with the multistepper lib, and that when you recall positions it has some simulation of acceleration that does not really feel like having much effect. For my needs it's ok as i don't really want it to move a lot, just for correcting shots but sure it would be great that i could smooth the stored movements when recalled. Your "shot recall" of 6 positions approach is also what i would be looking for. Is your code for teensy posted here? i think i didn't find it. would like to try it!

I have telemetrics system at my work which i don't think that works on VISCA protocol but the remote could control VISCA devices. i would find it to be interesting if i could somehow make this and the remote work together. Not that i would use it in production but it could be a simple and cheap enough solution for some cameras we have on ceiling that are static but could use that little remote adjustment. But that would mean to port this out to some controller with ethernet support and enough spare pins for the 2 TMC drivers.
Now i understand why the hall sensors are not of much help. These are digital sensors that as soon as they detect some magnet they switch state, meaning that it wont be precise enough for this job.

Yes i did build these mounts and i've been changing and testing design. I've doubled the pan motor base width and reinforced the zone where the tilt structs sit, added more lenght on the camera plate so i could put wider cameras, made a much sturdier camera plate, and last mod was to make the tilt sprocket that is attached to the camera plate twice the size so i could have more tilt torque. All of this cause the original design could handle a FS5 MK2 with a 35 samyang lens with metabones weight but a) the tilt would skip steps if speed was too high because of the lack of acceleration and b) the whole structure wobbled in movement also partially due to the lack of deceleration on the movement.

Oh, and also made a one part inner bearing print and a design that had much less tolerance on the bearing balls so that they couldn't allow vertical movement on the ptz baseplate.

Currently i'm able to use a lighter G series 105 mm full zoomed and do some movements without having much noticeable vibration.
I think my tilt gear is somehow crooked from the printing cause if i move fast they don't feel aligned at some rotation angles but i think that it was my fault that it might have somehow warped when ripping them off from the printer bed.

Thank you so much for your help and share!