Jump to content


Photo

crystal motor for a K3


  • Please log in to reply
13 replies to this topic

#1 Michael Collier

Michael Collier
  • Basic Members
  • PipPipPipPip
  • 1262 posts
  • Gaffer
  • Los Angeles, CA

Posted 26 March 2007 - 04:30 PM

I just bought my K3 from ebay and while I am waiting to get it, I wanted to get a jump on modifying it to run sync for sound. My plan is to take a torque motor, attach a hall sensor on the shaft and drive the camera. The motor will be conected to a custom microcontroller (xtal controlled of course) with digital input of running speed. First is this possible? I want to fully remove the spring mechanism and replace the drive with an electronic system. Programing the chip is easy enough. I have a 40 Mhz chip and programer on its way, so actually I am thinking of other features to add to it (maybe jam/generate timecode for slate purposes? Intervolometer? maybe mo-cap, mocon ability? I am not sure.) The question for you is how could a motor mate with the drive components within the camera, and how much torque aprox. should I be looking for in a motor to drive this thing? Also I assume there are mechanical speed regulators inside that can be removed for a crystal controlled motor? The microproccessor can control speed, so direct drive would be preferable. Not concerned with a 1:1 drive, the program can compensate.

good idea? fool hardy? I am not spending a lot on it, so no need bash the sense in doing this, I just think it might be a fun project.

Discription is good, pics are better. I won't have my K3 for a week or so, but I wanted to get started with this.

Next project: Dual HD-SDI flash storage system to replace those damn awful decks (who has the time to lug those 80K monstrosities around? maybe a 601 input for uncompressed analog video too. who knows. Built in video playback too! look for that for around 15-25K for a 1hr/720 Gb digi-mag version (1080p uncompressed up to 60fps, plus HS version planned). Might just start a rent-only company like Panavision.)
  • 0

#2 Nick Mulder

Nick Mulder
  • Basic Members
  • PipPipPipPip
  • 1023 posts
  • Other
  • Auckland, New Zealand

Posted 26 March 2007 - 05:26 PM

I just bought my K3 from ebay and while I am waiting to get it, I wanted to get a jump on modifying it to run sync for sound. My plan is to take a torque motor, attach a hall sensor on the shaft and drive the camera. The motor will be conected to a custom microcontroller (xtal controlled of course) with digital input of running speed. First is this possible? I want to fully remove the spring mechanism and replace the drive with an electronic system. Programing the chip is easy enough. I have a 40 Mhz chip and programer on its way, so actually I am thinking of other features to add to it (maybe jam/generate timecode for slate purposes? Intervolometer? maybe mo-cap, mocon ability? I am not sure.) The question for you is how could a motor mate with the drive components within the camera, and how much torque aprox. should I be looking for in a motor to drive this thing?


When you say hall effect sensor are you just hacking your own encoder ? If so I recommend getting/making a proper optical encoder that uses a periodically (like a thousand times a rev) interrupted IR LED read by a phototransistor on the other side - much more resolution than just a hall effect reading a magnet changing polarity once a rev (or however many times you can get magnet to do that yerself)...

anyways, yes its totally possible - but you've got a bit of coding to get around even just for P 'proportional' response - let alone a full PID implementation - If you've got a chip and the know-how then go for gold - there is some info here pertaining to the OOPIC if that is is your chip of choice (its kinda like a souped up Stamp) - anyways, its especially nice that if you use a chip it will be available for other related applications ... flash sync, motion control-esque applications where the camera rate is the 'sync' for model moves etc...

If not tho I'd look up Phase Locked Loops - Its how many xtal cct's were made in the days, and still are ...

and yep - youd want to remove the mechanical governor - It was surprising how small a motor you need once its gone, just the friction in the system left which usually soaked up all the potential advantage of the momentum a governor generated ... If you are doing quick start/stops and ramping then you'll want to over rate the motor so it has enough torque left over to stop itself...

self-learned/faulty/mistaken/contradictory ramble ends here :rolleyes:
  • 0

#3 Michael Collier

Michael Collier
  • Basic Members
  • PipPipPipPip
  • 1262 posts
  • Gaffer
  • Los Angeles, CA

Posted 26 March 2007 - 06:13 PM

Well sensor design hasn't been finalized. Easily done is a rotary encoder (esp since it would be traveling in a known direction, simplifying the readings.) I don't think I need thousand times resolution, but at least 36 pulses per rev would be nice. Then I can compare that to the clock counting how long it took to turn 10 degrees and slow or speed the motor accordingly.

Chip of choice is the PIC from microchip (similar to the OOPIC, but without all the extra stuff on the board) I programed my first with a circuit I found in popular electronics (no-parts programer) I bought one that can accomodate the chips up to 40 pins, so I should be set (plus I can get free samples of these chips) Not sure which chip in particular I'll use, probably the smallest one I have coming that will do the job, since this won't be a comutational intensive app.

I thought about PLL but I have the chips for free, I have an LCD readout and there is somethign cool about entering in the exact speed down to +/- .001 accuracy. Plus the unit should be able to calculate iris change should I ever want to impliment speed ramps, etc.

Now, proportional I understand, what is 'Full PID' method? Is this anything like proportional but with error correction? Such as after a frame is exposed, analize if the frame phase is still locked on (IE opening shutter just as the 1/24 interval passes each time)

I assume that removing the gov and other un-needed parts might reduce the units noise? would it ever get quiet enough to shoot blimped for sync sound?

also while researching the flash-uncompressed HD deck, I found out they are going to make a 3gb SDI link. That means in the future uncompressed 1080p 10bit over one coax cable! interesting stuff.
  • 0

#4 Nick Mulder

Nick Mulder
  • Basic Members
  • PipPipPipPip
  • 1023 posts
  • Other
  • Auckland, New Zealand

Posted 26 March 2007 - 06:48 PM

Well - it sounds like we are in same neighbourhood with regards to knowledge (swings and roundabouts I'm sure)  -  but to answer the original query - yeh man go for it !  

I think a lack of answers and motivational words from others is an indication that you (and myself) are starting to chip away at a grossly overpriced industry that shrouds itself in arcane techno-babble or more often the 'wall of silence' so as to discourage people from getting a cut of the limelight after having developed their own 'super-duper filmy thingy-bob'...

I'm constantly amazed how easy it is to make yourself a version of something that would otherwise cost at least 10 times as much on the market ...  I'd like to think that I'd keep all my developments 'open-source' - thing is I'm still yet to come up with anything new (still saving a poop load of $$$ and learning in the meantime) but maybe when I stumble across my own aforementioned SDTTB (super-duper filmy thingy-bob) then I may just develop a shroud of my own ...

We'll see  :ph34r:

But yeh - I say dont worry about what others say, if you have an idea go for it - even if you end up with a redundant time waster, you'll probably have learned to spot when things are heading that way - A valuable resource, not learned by someone telling you not to do it at cinematography.com because they are jealous of your initiative...

Still though, I'm not referring to all and sundry here - its a great resource with many a unique character with their associated off-the-radar skillsets who are more than happy to spill their IP guts -  just be careful you dont get indoctrinated into the mostly semi-pro with an eye on pro methodology that many seem to 'strictly' adhere to ...  'nah, man dont do that, its not in the manual' etc...

Noise, governor - yep imagine so, they spin pretty fast

LCD - cool cool! just bought a 20x4 one myself for my Bolex EL and H-16 timelapse thingy-bob (see the one with heaps less functionality on eBay for $1000)

PID - Proportional Integral Derivative (excuse the wikipedia link addiction I have)

Proportional >> where am I minus where should I be = how much I need to slow down or speed up to make where am I = where should I be = 0 (error)

Integral >> looking at what has happened in the past (I understand this conceptually or at least how to implement this one the least)

Derivative >> looking at the difference in actual and desired velocity as another error calculation (again, less understanding of this)

but its all doable - just need to get down to reading up implementing calculus in the discrete world of code as opposed to the continual reality of the world around us (which is discrete anyway, but not so that we have the ability for our senses to notice)

wowsa, now we're talking particle physics - time to stop rant

HD - whats that ?? B)

Edited by Nick Mulder, 26 March 2007 - 06:50 PM.

  • 0

#5 Michael Collier

Michael Collier
  • Basic Members
  • PipPipPipPip
  • 1262 posts
  • Gaffer
  • Los Angeles, CA

Posted 26 March 2007 - 10:36 PM

Hmmm calculus eh? That won't work, these chips aren't capable of floating point (and even division is tough with these) Maybe a simple method but torque probably won't be a simple constant. I assume as the motion in the camera changes, so too does the force required to move. Perhaps I can store a count value for each 10 degree motion seperatley, and then set the controler to both accomodate a global error correction (over the history of the entire take) Then store a local error correction number (36 in total) to tell the motor how to adjust for expected resistence ahead. Then what ever over shoot or undershoot would feed back into the global error number which would speed or slow the mechanism to compensate for error.
  • 0

#6 Nick Mulder

Nick Mulder
  • Basic Members
  • PipPipPipPip
  • 1023 posts
  • Other
  • Auckland, New Zealand

Posted 26 March 2007 - 10:57 PM

I think you've already got the gist of PID already B) at least either the I or the D - heh

yeh - floating point, thats what I was hooting on about the discrete vs. continuous jazzforms - I get around it by scaling integer Word values - so for instance 65535 = 655.35 - yeh, you loose a few dingdongs but you gain in flipdipses
  • 0

#7 Antonio Bunt

Antonio Bunt
  • Basic Members
  • PipPipPip
  • 134 posts
  • Cinematographer

Posted 26 March 2007 - 11:13 PM

I like to build things for my cameras, I am waiting for my K3 too but I want to build a motor. Can you share like diagrams or so? I am also worried since the K3 purrs like an overgrown cat. It has been discussed, but how can one cope with the noise?
PS: You just lost me when you started talking about HD! ;)
  • 0

#8 Nick Mulder

Nick Mulder
  • Basic Members
  • PipPipPipPip
  • 1023 posts
  • Other
  • Auckland, New Zealand

Posted 26 March 2007 - 11:39 PM

I like to build things for my cameras, I am waiting for my K3 too but I want to build a motor. Can you share like diagrams or so? I am also worried since the K3 purrs like an overgrown cat. It has been discussed, but how can one cope with the noise?
PS: You just lost me when you started talking about HD! ;)

Yeh man - HD - some sort of fancy pants anamorphic stereo 70mm ??

hmmm, cct diagrams - well, thing is we are talking about microcontrollers which in effect are pretty simple beasts in terms of wiring up (power, gnd - go!) all the conceptual stuff is where the magic (or blue smoke) happens - so you'll get a bunch of modernist architecture-esque logic diagrams and wotnot you'll need to turn into code... Do you know C ? do you know elementary calculus and methods of how to effectively 'discretize' it so that the little chips can be succesfully Hogwarted ?

As for noise - I find it totally goes away when you immerse it in water - so what I often do, is buy a Bolex on eBay and fill up my underwater housing with water and then install camera inside funky focusing and damn heavy but workable for a load or two, smart cookies will even develop the film in the said housing - either that or tape socks to it

(have a search here a the forums for 'blimp' 'barney' or 'noise' you'll get a bunch of prescripted discussions we roll-played for such occasions)

Edited by Nick Mulder, 26 March 2007 - 11:42 PM.

  • 0

#9 Dennis Kisilyov

Dennis Kisilyov
  • Basic Members
  • PipPipPip
  • 152 posts
  • Other

Posted 27 March 2007 - 11:20 AM

I suggest you ask Clive Tobin for the info, I don't know how inclined he would be to release the specs for his K3 sync motor, but as far as I understand a considerable amount of machine-ing goes into the install.

Which would mean that most of the clockwork that is K3 would be kept, the shutter, the big gear. So installing such a motor would only do sync, it would not reduce the noise level any, as the spring motor itself does not make nosie, the gears do :-).

I'm certain his motors shipped with some type of nstallation instructions.
  • 0

#10 Olex Kalynychenko

Olex Kalynychenko
  • Basic Members
  • PipPipPipPip
  • 868 posts
  • Cinematographer
  • Dnepropetrovsk, Ukraine

Posted 27 March 2007 - 02:15 PM

I just bought my K3 from ebay and while I am waiting to get it, I wanted to get a jump on modifying it to run sync for sound. My plan is to take a torque motor, attach a hall sensor on the shaft and drive the camera. The motor will be conected to a custom microcontroller (xtal controlled of course) with digital input of running speed. First is this possible? I want to fully remove the spring mechanism and replace the drive with an electronic system. Programing the chip is easy enough. I


Good idea.
I hope, you will have good result.
But, after a many years my experience on repair of different original systems of speed control ( analog, pulses XTAL, pulses control with calculation, PID control ) and after a long time of create of crystal sync speed control, i can tell you, this is not easy problem.
And the problem not on programming, the problem on create of mathematical model of PID control with real mechanical system.

After you to be tired with playing, and , if you not lost of idea of film making, i will recommend you to buy Kinor-16 SX-2M camera with electrical motor and to order of modification of motor on crydstal syn speed control.

Kransogrosk-3 born with mechanical spring motor, and 30 m of film rolls.
The elephant not be a flying elephant, if you add of wings,
The elephant born to be of elephant, not to be of bird.

Yes, I know about modified Krtanosgorks-3 with Clive tobin crystal sync speed electrical motor.
If you wish, you can build similar camera.
  • 0

#11 Clive Tobin

Clive Tobin
  • Basic Members
  • PipPipPipPip
  • 402 posts
  • Industry Rep
  • Spokane Valley, WA, USA

Posted 27 March 2007 - 09:12 PM

[quote name='Dennis Kisilyov' date='Mar 27 2007, 08:20 AM' post='163286']
I suggest you ask Clive Tobin ...I'm certain his motors shipped with some type of nstallation instructions. /quote]

Nobody asked me exactly but I still have them posted on my website at http://www.tobincine...com/page50.html .

Most people did not actually remove the spring, and just disabled the governor, owing to the difficulty and genuine risk of personal injury of going further.

p.s.- A microprocessor based system without a dedicated crystal circuit will never reach the standards of accuracy required for double system sound. Our motors ran within 3 PPM (0.0003 %) of correct speed.

One young punk hotshot was selling purported crystal-less crystal control boards on Ebay, maybe last year, claiming you could crystal control any camera for peanuts and not enrich us evil manufacturers. Besides the instructions being totally inaccurate, nonspecific to any camera model, and incomplete, and with no sensors furnished to sample the camera speed and phase, it turned out his wonderful PIC based circuit was only accurate to 1% if you were lucky, and worse with any temperature change. I very much doubt that any purchaser was able to actually get something working.

In case nobody has noticed, selling crystal motor equipment lately is NOT the way to get rich. This is why we are now selling telecine equipment instead.
  • 0

#12 Michael Collier

Michael Collier
  • Basic Members
  • PipPipPipPip
  • 1262 posts
  • Gaffer
  • Los Angeles, CA

Posted 27 March 2007 - 09:25 PM

Thanks for the info clive! I have no intentions of building this to put true sync motor makers out of business, this is more for fun and to get back into electronics, since I can see having working knowledge being somewhat benefitial on set, esp. in my early career.

I plan to clock the PIC with a crystal, not an RC circuit as that kid probably did. I can get crystals down to +/- 1 or 2 PPM, and since the PIC is a 1 clock cycle per instruction unit, its easy to track the time (also I can route the clock cycle to a programable pre-scaler and set the limit to trip an interupt to keep proper time) should this become too difficult, I can build a seperate crystal oscilator. I have yet to find info as to weather the temp can affect a crystal. I don't think it does, but I need to know in case I need to compensate for that.

If I disable the spring (wasn't planning to remove it anyways, nor my garage door spring) would it be easy to re-engadge it? Also a big question I have is how to mate the motor to the drive. Is there a standard spline shaft I can use? or maybe a gear exposed that can be mated too? I assume I have to open mine up and see. Are there any resources that can help me take apart the K3 and get it back together again? I want to investigate first hand how the motor/drivetrain connect.
  • 0

#13 Will Montgomery

Will Montgomery
  • Sustaining Members
  • 2030 posts
  • Producer
  • Dallas, TX

Posted 28 March 2007 - 11:29 AM

Kransogrosk-3 born with mechanical spring motor, and 30 m of film rolls.
The elephant not be a flying elephant, if you add of wings,
The elephant born to be of elephant, not to be of bird.


I laughed for about 5 minutes when I read this. :D

I do agree though. As an intellectual challenge, do it by all means, but there are other cameras out there what will acomplish what you want for the same amount of money once you consider all the time you'd put into it.
  • 0

#14 Clive Tobin

Clive Tobin
  • Basic Members
  • PipPipPipPip
  • 402 posts
  • Industry Rep
  • Spokane Valley, WA, USA

Posted 28 March 2007 - 07:44 PM

... I have yet to find info as to weather the temp can affect a crystal. ...
If I disable the spring (wasn't planning to remove it anyways, nor my garage door spring) would it be easy to re-engadge it? Also a big question I have is how to mate the motor to the drive. Is there a standard spline shaft I can use? or maybe a gear exposed that can be mated too? I assume I have to open mine up and see. Are there any resources that can help me take apart the K3 and get it back together again? I want to investigate first hand how the motor/drivetrain connect.


Some of this is addressed in the link I already gave.

Of course a crystal is affected by temperature. You select the other circuit components to minimize the temperature coefficient at the expected operating temperature range. With the usual AT-cut crystal you just need some N or negative tempco of the capacitors. Even if you don't do it the speed should be close enough for less than full-spool length shots.

You can't really disengage the spring motor, but if you remove the winding handle it will just passively idle or coast when driven separately. The spring housing has a gear ring outside it to drive the takeup spool so you can't just remove it. Removing the spring from the housing is very likely to get out of control and to lop off any protruding body parts and save you the expense of a nose bob, sex change or vasectomy later. See the warnings in the link.
  • 0


Rig Wheels Passport

Gamma Ray Digital Inc

Aerial Filmworks

Media Blackout - Custom Cables and AKS

Technodolly

Opal

Glidecam

Visual Products

rebotnix Technologies

Metropolis Post

Ritter Battery

FJS International, LLC

The Slider

Willys Widgets

Abel Cine

Wooden Camera

CineLab

Tai Audio

CineTape

Paralinx LLC

Broadcast Solutions Inc

Media Blackout - Custom Cables and AKS

Aerial Filmworks

Visual Products

Glidecam

Paralinx LLC

Technodolly

Willys Widgets

The Slider

rebotnix Technologies

Abel Cine

Broadcast Solutions Inc

CineLab

FJS International, LLC

CineTape

Ritter Battery

Wooden Camera

Opal

Metropolis Post

Gamma Ray Digital Inc

Tai Audio

Rig Wheels Passport