Jump to content


Photo

Is there alternative to Crystal Sync?


  • Please log in to reply
43 replies to this topic

#1 Matthew W. Phillips

Matthew W. Phillips
  • Sustaining Members
  • 1792 posts
  • Other

Posted 04 October 2009 - 03:10 PM

I have been getting heavily into Electronic theory lately and have been talking with some fellows who are quite knowledgable of circuits and whatnot. Anyhow, I have been trying to work on a circuit that puts out a 24Hz pulse for the purposes of driving a step motor for 24fps. However, the other day I had a talk with a fellow at an Electronics store who said that crystal is becoming largely outdated in most other sectors of timing other than film and are instead using a 555 timer that is phase locked. He said it's actually more accurate and also cheaper.

Crystal is expensive and then, because you have to divide the signal (crystals only come in MHz so you must digitally divide the resonance), it becomes a bit pricey. What do you guys think?
  • 0

#2 Matthew W. Phillips

Matthew W. Phillips
  • Sustaining Members
  • 1792 posts
  • Other

Posted 04 October 2009 - 08:50 PM

Okay, I managed to get a working circuit going for a 24Hz tone on a 555 timer. Right now, it just outputs to a green LED that flickers at 24 pulses per second. The output part could just as easily power a stepper motor, providing the proper angles were calcuated (or gears used) to get the right rotation for each frame of film. As far as I'm concerned, the hardest part is over.

Anyone have any ideas on whether they agree or disagree that the 555 timer with phase lock is better (or at least as good) as a crystal oscillator?

Here is a picture of my 24Hz timer prototype.

Posted Image
  • 0

#3 Paul Bruening

Paul Bruening

    (deceased)

  • Sustaining Members
  • 2858 posts
  • Producer
  • Oxford, Mississippi

Posted 04 October 2009 - 10:13 PM

Hal and John seem to be the most knowing of these things though I remember a few helpful folks here and there when I was blundering through this issue.

Steppers are driven through a controller then a driver. Some steppers have built-in drivers which makes it easier. You can get ball park units and use a computer to adjust them or you can engineer a specific controller. When someone like Bruce G'day McNaughton does it he already knows the engineering and makes-up a smaller board that matches all the parts in the process and it plain works.

I went toward the PWM direction which uses a variable square wave to control speed. I haven't proven if it works, yet. I couldn't round up an Ebay-cheap way to monitor precise speed like 23.976. Frankly, I just wasn't smart enough or trained enough to tackle it.

I can say that about the time you think you've solved it, something like pulse feedback goofs up all your plans. All the same, I wish you all the luck in solving your challenge. If you can make it work I'll be first in line asking you how you did it.
  • 0

#4 Keith Walters

Keith Walters
  • Sustaining Members
  • 2219 posts
  • Other
  • Sydney Australia

Posted 04 October 2009 - 10:58 PM

I have been getting heavily into Electronic theory lately and have been talking with some fellows who are quite knowledgable of circuits and whatnot. Anyhow, I have been trying to work on a circuit that puts out a 24Hz pulse for the purposes of driving a step motor for 24fps. However, the other day I had a talk with a fellow at an Electronics store who said that crystal is becoming largely outdated in most other sectors of timing other than film and are instead using a 555 timer that is phase locked. He said it's actually more accurate and also cheaper.

Crystal is expensive and then, because you have to divide the signal (crystals only come in MHz so you must digitally divide the resonance), it becomes a bit pricey. What do you guys think?

I think the fellow in the electronics store has little or no idea what he is talking about. The 555 timer was introduced by Signetics in 1972, things have moved on from there somewhat...

I also think you will need to put in quite a bit more study on electronics theory before you start trying to design phase-locked-loop circuits, particularly where motors are concerned.

To address some of your questions/statements

1. Stepper motors are rated in degrees per step eg "1.8 degrees/step", which means the motor makes 1/200th of a rotation each time you pulse it. That means you would need to drive it with a frequency of 200 x 24 = 5,000Hz to make it run at 24 fps, not 24Hz.

2. Virtually all older cameras (that I presume you want to work with) use some form of belt drive between the motor and the mechanism, so the motor is not going to be actually running at 24 revolutions per second anyway.

3. Most cameras use either ordinary brushed DC motors, or more recently, AC synchronous motors with a variable AC frequency drive. The reason for this is simply that while stepper motors have a lot of advantages, they also tend to be extremely noisy in operaton. They come into their own when you want a motor that can run extremely slowly, and also "retrace its steps" in reverse. The Lynx motors retrofitted to the Fries Mitchell 35III are an excellent example of this. The 1930s ultra-precise film transport combined with a computer-controlled stepper motor are a powerful combination and are still widely used today for stop-motion work. However, because of the high torque and way they are coupled to the camera, the motors scream at 24fps and above.

4. The thing about the crystals is somewhat mangled. It is indeed possible to use just one crystal to generate a huge range of frequencies, all with the accuracy and stability of the original crystal. But you still need at least one crystal and this requires some complicated digital circuitry.
I think what you friend in the electronics store is alluding to is that, in the past, when digital electronics was not so advanced, it was usually cheaper to just get individual crystals made for the different frequencies, which also avoided having a lot of old-fashioned power-hungry TTL computer chips in your device.
Nowadays, digital circuitry is much more affordable and hardly uses any power, so it is far more cost-effective to use just the one crystal, with a digital frequency synthesizer. This can be an off-the-shelf unit, or more likely these days, a custom-programmed microcontroller. You can still get custom crystals made, and they aren't all that expensive, but a digital synthesizer can often use a common dirt-cheap crystal frequency that is already available off the shelf.

If you want to experiment with a stepper motor, you might be able to salvage one from a junked laser printer. The laser beam is scanned across the photodensitive drum by an octagonal mirror mounted on a stepper motor. In most cases this motor is driven by a specialized driver chip, which with any luck will be mounted on its own separate circuit board. All you have to do to make it run is apply a frequency to the input pin of the chip. If you can identify the chip's type number, you will possibly be able to download a datasheet from the manufacturer's website.
  • 0

#5 Matthew W. Phillips

Matthew W. Phillips
  • Sustaining Members
  • 1792 posts
  • Other

Posted 05 October 2009 - 01:09 AM

I think the fellow in the electronics store has little or no idea what he is talking about. The 555 timer was introduced by Signetics in 1972, things have moved on from there somewhat...


Okay, you show your ignorance here, Keith. He actually has an engineering degree and knows quite what he is talking about. Why does it matter when they were introduced? I never said the use of them was new...I said that people are using them more than they used to. Also, the 555 timer hasn't been stagnate over the years, it has evolved. They aren't all the same despite their name. They all have similar functions but some are more reliable than others.

I also think you will need to put in quite a bit more study on electronics theory before you start trying to design phase-locked-loop circuits, particularly where motors are concerned.


I am studying as I go. What's wrong with that? I'm not building the next RED camera, I'm just hobbying around. Why do you always have a snobbish tone?

To address some of your questions/statements

1. Stepper motors are rated in degrees per step eg "1.8 degrees/step", which means the motor makes 1/200th of a rotation each time you pulse it. That means you would need to drive it with a frequency of 200 x 24 = 5,000Hz to make it run at 24 fps, not 24Hz.


This is ridiculous. First off, why would you times 200 x 24? 24 is the desired amount of frames you wish to move per second. 200 is the amount of rotations a 1.8 degree step motor would make for a 360 degree rotation. See how you are mistaken? Try wiring an LED to a 24Hz pulse. The light blinks 24 times a second. This denotes that 24 times per second the circuit is outputing current to the LED. A step motor cannot move less than it's step. Therefore, even if you had a 1.8 degree step motor, and wished to do one full rotation per second, you'd be at 200 Hz, not 5khZ...I don't know how you got that figure. Unless you are assuming that I'm trying to make a frame move by a full step motor rotation. Actually, I want to move a frame with a step, not a full rotation, otherwise I wouldnt use a step motor.

Secondly, step motors are made with varying degrees of rotation. The old surplus store near my house has a ton of DC step motors, the most interesting of which is 7.5 degrees. Do the math...360/7.5 = 48...48 is 24 x 2 which makes this a very attractive idea for toying with since the math is easy. I even found one supplier online that offers a 15 degree step motor which would result in a 24 Hz frequency (asuming I could get one step to equal one frame movement.)

2. Virtually all older cameras (that I presume you want to work with) use some form of belt drive between the motor and the mechanism, so the motor is not going to be actually running at 24 revolutions per second anyway.


I don't recall even mentioning what camera I was working with or that I was even working with a camera. I could be building a projector or God knows what. I'm researching the issue of syncronization at the moment.

3. Most cameras use either ordinary brushed DC motors, or more recently, AC synchronous motors with a variable AC frequency drive. The reason for this is simply that while stepper motors have a lot of advantages, they also tend to be extremely noisy in operaton. They come into their own when you want a motor that can run extremely slowly, and also "retrace its steps" in reverse. The Lynx motors retrofitted to the Fries Mitchell 35III are an excellent example of this. The 1930s ultra-precise film transport combined with a computer-controlled stepper motor are a powerful combination and are still widely used today for stop-motion work. However, because of the high torque and way they are coupled to the camera, the motors scream at 24fps and above.


AC synchronous motors are a terrible idea for a camera despite the fact that they are 60 Hz (supposedly.) And if you know as much as you act like you do, Keith, then you know why. The power grid (at least here in the States) mandates that all sync must be accomplished by the end of the day, but often times there is a spike in the frequency near midnight to meet that deadline, thus proving that the rest of the day operated at varying frequencies that lagged behind. I know people who have seen this in action watching the monitors.

Step motors are not necessarily loud at higher frequencies. I talked with the owner at the store near my house and he showed me quite a bit of them in action at varying frequencies and some where quite impressive in how quiet they were. Put them in a decent casing and they may be capable of quiet operation.

I'm not discounting a DC brushed motor...I'm just saying that step motors are useful in that they would work better for film shutters (at least in theory) because of their physical stoppage.

4. The thing about the crystals is somewhat mangled. It is indeed possible to use just one crystal to generate a huge range of frequencies, all with the accuracy and stability of the original crystal. But you still need at least one crystal and this requires some complicated digital circuitry.
I think what you friend in the electronics store is alluding to is that, in the past, when digital electronics was not so advanced, it was usually cheaper to just get individual crystals made for the different frequencies, which also avoided having a lot of old-fashioned power-hungry TTL computer chips in your device.
Nowadays, digital circuitry is much more affordable and hardly uses any power, so it is far more cost-effective to use just the one crystal, with a digital frequency synthesizer. This can be an off-the-shelf unit, or more likely these days, a custom-programmed microcontroller. You can still get custom crystals made, and they aren't all that expensive, but a digital synthesizer can often use a common dirt-cheap crystal frequency that is already available off the shelf.


I got a 555 timer for $1.69 at Radio Shack. It's sortof hard to be cheaper than that. Crystal oscillators (with everything needed to utilize them properly) are not as cheap.

If you want to experiment with a stepper motor, you might be able to salvage one from a junked laser printer. The laser beam is scanned across the photodensitive drum by an octagonal mirror mounted on a stepper motor. In most cases this motor is driven by a specialized driver chip, which with any luck will be mounted on its own separate circuit board. All you have to do to make it run is apply a frequency to the input pin of the chip. If you can identify the chip's type number, you will possibly be able to download a datasheet from the manufacturer's website.


Thank you for offering something useful to me, Keith.
  • 0

#6 Ian Cooper

Ian Cooper
  • Basic Members
  • PipPipPipPip
  • 469 posts
  • Other
  • England

Posted 05 October 2009 - 02:11 AM

Without wishing to wade into heated discussion, but from your photo it seems as if your current 555 timer circuit is relying on an R-C network to set its frequency. ...If this is the case then just be careful as the actual frequency will drift quite a bit (although perhaps not enough to be too noticable on a 'scope,) and the stability will also be affected by temperature. Even modern 555 timers themselves aren't exactly well known for their stability or accuracy, and capacitor tolerance and temperature stability varies from poor through to dreadful dependant upon type.

Personally (speaking as a R&D electronics engineer, not as someone in the film industry) if you want to easily come up with a circuit having the accuracy of a crystal oscillator, then you're better off starting your design with a crystal oscillator of some type. An inverter gate, two resistors, two caps and a crystal may work out a few pence/cents more expensive than a 555, but a digital divider circuit is a lot easier to get working reliably than trying to get a phase locked loop designed to keep a 555 running any where near as accuratly as a crystal will from the start.

Of course I do realise part of the fun of experimenting on 'hobby' projects is going about things the harder way just to prove it is possible, so best of luck which ever path you choose to follow! :)

...and if you aren't using an RC network on your 555, or have otherwise considered the drift issues then please accept my apologies in advance. B)
  • 0

#7 Olex Kalynychenko

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

Posted 05 October 2009 - 02:31 AM

However, the other day I had a talk with a fellow at an Electronics store who said that crystal is becoming largely outdated in most other sectors of timing other than film and are instead using a 555 timer that is phase locked. He said it's actually more accurate and also cheaper.

Crystal is expensive and then, because you have to divide the signal (crystals only come in MHz so you must digitally divide the resonance), it becomes a bit pricey. What do you guys think?

The guy, who answer you have superficial knowledge of automatic control system.
The automatic control system do not pulses or value of voltage only - this is complex mathematical modelling of process.
Afer you will have mathematical model, you can building of system of automatic control on any components. This can be analog computational algorithm ( operational amplifier) or digital automaton or microporcessor.

The speed control - this is complex problem and this is problem can be describe like PID control.

If you study of initial systems of speed conrtol, yes, this systems had block of frequency comparison
and analog conversion of control of DC motor.

That's why, the 555 timer with low precision of parameters of output signal, pase locked ship - technology of last century.
Yes, this will work, but with low value of parameters of speed control.


About stepper motor and motion control.
This is verygood idea, but, you must draw attention on magnetic characteristics of stepper motor,
the time of dead time, inertia and other.

If you wish build of cheap system, 555 chip will good choose, but, do not wait of high technical parameters of speed control of this system.

All modern compoters included of generator of main frequency and this is crystal sync generator ( or
built-in generator of processor ).
Any case, this generator can run on quartz resonator, or on RC components.
  • 0

#8 Ian Cooper

Ian Cooper
  • Basic Members
  • PipPipPipPip
  • 469 posts
  • Other
  • England

Posted 05 October 2009 - 02:36 AM

...or more recently, AC synchronous motors with a variable AC frequency drive.


AC synchronous motors are a terrible idea for a camera despite the fact that they are 60 Hz (supposedly.) And if you know as much as you act like you do, Keith, then you know why. The power grid (at least here in the States) mandates that all sync must be accomplished by the end of the day, but often times there is a spike in the frequency near midnight to meet that deadline, thus proving that the rest of the day operated at varying frequencies that lagged behind. I know people who have seen this in action watching the monitors....



Variable speed drives are totally independant of the incoming power supply and its frequency. The first thing they do is convert the AC to DC irrespective of its frequency. The DC is then converted back to AC at any frequency so desired with 'crystal' accuracy. These days this is generally achieved through the use microprocessors. If critical motor speed is important then an encoder feedback device will be used to form a closed-loop system.
  • 0

#9 Matthew W. Phillips

Matthew W. Phillips
  • Sustaining Members
  • 1792 posts
  • Other

Posted 05 October 2009 - 03:37 AM

Without wishing to wade into heated discussion, but from your photo it seems as if your current 555 timer circuit is relying on an R-C network to set its frequency. ...If this is the case then just be careful as the actual frequency will drift quite a bit (although perhaps not enough to be too noticable on a 'scope,) and the stability will also be affected by temperature. Even modern 555 timers themselves aren't exactly well known for their stability or accuracy, and capacitor tolerance and temperature stability varies from poor through to dreadful dependant upon type.

Personally (speaking as a R&D electronics engineer, not as someone in the film industry)


Thank you for your input and help. I was originally looking for input regarding the stability of 555 timers vs. Crystal oscillation. I wasn't looking for Keith to insult me or imply I am not capable of eventually creating a phase locked loop. With enough practice and study, I think people can do pretty much anything they want. I just feel like Keith often spends more time telling others what they can't do instead of doing something himself. It's easy to be an armchair naysayer but harder to contribute and do something to show that you yourself are worth your salt.

Anyhow, yes, the current circuit is a basic R-C network. I was using this to start from to get a gist for the 555 timer. Whether or not it's as stable as crystal might have been the wrong question for me to ask. The actually key is...will the variation actually cause a drift that would be perceiveable when considering dialog syncing with film? I have researched 555 timers quite a bit lately and I have no reason to believe that they are going to drift to the level where you will notice a dialog drift. Even a single frame slip here and there will not result in a perceiveable off-sync. I have found this to be true even with old Super 8 cameras that used a motor system, in many cases, that was inferior to even the 555 timer. Some don't even have a timer at all but just motor and gear systems that approximate film speed by the average RPMs. Even with those, syncing is very possible and almost always works for the duration of a Super 8 cartridge.

I have read Ian, and you can correct me if I'm wrong or misread, that adding capacitors around the +VCC and the Ground can help keep the frequency stable on account of absorbing power during spikes and releasing it during dips. Therefore, my question is...do you think 555 timer would really have that great of a variance to affect syncing ability?
  • 0

#10 John Sprung

John Sprung
  • Sustaining Members
  • 4635 posts
  • Other

Posted 05 October 2009 - 03:53 AM

I have been getting heavily into Electronic theory lately and have been talking with some fellows who are quite knowledgable of circuits and whatnot. Anyhow, I have been trying to work on a circuit that puts out a 24Hz pulse for the purposes of driving a step motor for 24fps. However, the other day I had a talk with a fellow at an Electronics store who said that crystal is becoming largely outdated in most other sectors of timing other than film and are instead using a 555 timer that is phase locked. He said it's actually more accurate and also cheaper.

Crystal is expensive and then, because you have to divide the signal (crystals only come in MHz so you must digitally divide the resonance), it becomes a bit pricey. What do you guys think?


The 555 timer -- at least as I remember it from 30 years ago or so -- is just a way of getting a square wave from the time constant of a resistor and capacitor. It doesn't have any way of phase locking. For that, you need a PLL -- Phase Locked Loop -- chip. The CD4046 was the PLL back in those days. You can certainly use 555's as clock sources, provided that you don't need a really precise clock. It would work fine for a standalone system, if you don't need real time for anything. No way would it be adequate for sync sound.

What a phase locked loop does is it controls a variable oscillator to sync exactly with a reference signal. So, you 'd still need a source for that reference frequency, which is usually a crystal oscillator and frequency divider. The guts of an old time crystal camera motor are basically this:

You have a crystal oscillator and frequency divider from which you get reference frequencies. The PLL controls the speed of the DC motor by varying the duty cycle of the power it sees. This is done by amplifying a sawtooth wave so that it clips, and just moving the offset up or down. It gets its motor speed reference from an LED and phototransistor that produce a square wave as a disc with holes in it attached to the motor shaft passes between them -- a light chopper.

Back when I built a crystal motor for grins, you could get all the chips and the crystal for under $20, so say under $200 in today's money. I don't know chip prices now, but I doubt that they've risen higher than inflation.




-- J.S.
  • 0

#11 Matthew W. Phillips

Matthew W. Phillips
  • Sustaining Members
  • 1792 posts
  • Other

Posted 05 October 2009 - 04:11 AM

No way would it be adequate for sync sound.
Back when I built a crystal motor for grins, you could get all the chips and the crystal for under $20, so say under $200 in today's money.


$200 is very pricey for such a thing, as far as I'm concerned.

I found this page from the University of Delaware that did admit that Crystal is superior to 555, however, is says that 555 is fine for applications under 1% accuracy. For anything under 0.1% accuracy, however, they recommended Crystal or digital implementation.

If this is true (and I believe it is since it's being taught at Udel), then I would imagine under 1% accuracy would be sufficient for syncing. 1% would be one frame off every 100 which would work out to 36 drifted frames on a 3600 frame Super 8 camera, assuming you had a 2.5 minute take @ 24fps. This would be a 1.5 second drift which would be noticeable, but it would be far superior to the systems on Super 8 cameras. This is disregarding the fact that the frequency will most likely both spike and dip here and there balancing out the overall result. I still don;t believe you would have a perceiveable drift. I guess no way to know for sure without testing the theory.

Here is the link to the article...
http://www.physics.u.....timer lab.pdf
  • 0

#12 Keith Walters

Keith Walters
  • Sustaining Members
  • 2219 posts
  • Other
  • Sydney Australia

Posted 05 October 2009 - 06:07 AM

AC synchronous motors are a terrible idea for a camera despite the fact that they are 60 Hz (supposedly.) And if you know as much as you act like you do, Keith, then you know why. The power grid (at least here in the States) mandates that all sync must be accomplished by the end of the day, but often times there is a spike in the frequency near midnight to meet that deadline, thus proving that the rest of the day operated at varying frequencies that lagged behind. I know people who have seen this in action watching the monitors.


For your information, that is exactly how all the 24V Arri cameras (535, 435, Arricam etc) work. (I'm not too sure about the later Panavisions, but I think they use something similar).
The drive circuit takes in 24V DC and converts it into 3-Phase AC which then drives the synchronous motor. The AC frequency is produced by a precision crystal controlled oscillator/

I got a 555 timer for $1.69 at Radio Shack. It's sortof hard to be cheaper than that. Crystal oscillators (with everything needed to utilize them properly) are not as cheap.


So, considering the 555 chip came out in 1972, it's sort of odd that so many cameras were made after that date, that used crystal control.
In reality, what they more or less have is the equivalent of your 555 oscillator, with an extremely comples crystal oscillator system driving it.

Anyway, have fun.
  • 0

#13 Gary Lemson

Gary Lemson
  • Basic Members
  • PipPipPip
  • 155 posts
  • Other
  • Santa Cruz Mountains

Posted 05 October 2009 - 07:53 AM

...Crystal is expensive and then, because you have to divide the signal ...


Actually, crystals are quite inexpensive these days. Some can be had for under a $1.00. Check out Mouser Electronics , Newark, or Digikey.

This problem can be easily solved with a micro-controller (and crystal). There are some really inexpensive devices from Microchip Corp (and others) available. It would help to have some programming background to use, however, they also have inexpensive training kits and tutorial material available, and there is loads of beginner info online.
  • 0

#14 Paul Bruening

Paul Bruening

    (deceased)

  • Sustaining Members
  • 2858 posts
  • Producer
  • Oxford, Mississippi

Posted 05 October 2009 - 10:20 AM

Since so many people are in the mood to tackle this kind of topic: How do I read rotary resolutions of 1/1000 of 1 fps? What type of motor control style would be obliged to that kind of resolution and still be variably controlled? At least, help me get into the right ball park.

As always, I am grateful for any and all help in advance.

Edit: I should have said 1/1000 of each rotation of any fps.
  • 0

#15 John Sprung

John Sprung
  • Sustaining Members
  • 4635 posts
  • Other

Posted 05 October 2009 - 01:09 PM

I found this page from the University of Delaware that did admit that Crystal is superior to 555, however, is says that 555 is fine for applications under 1% accuracy. For anything under 0.1% accuracy, however, they recommended Crystal or digital implementation.

If this is true (and I believe it is since it's being taught at Udel), then I would imagine under 1% accuracy would be sufficient for syncing. 1% would be one frame off every 100 which would work out to 36 drifted frames on a 3600 frame Super 8 camera, assuming you had a 2.5 minute take @ 24fps. This would be a 1.5 second drift which would be noticeable, ....


Actually, many people can see lip sync being out by one frame. (I'm not one of them, everything looks out of sync to me when I try to pay attention to it..... ) So, to be good enough for film use, you'd want to be within half a frame even if you shot a whole 2000 ft mag of three perf. Ballpark, say one frame in an hour. If you want to use your crystal to clock time code generators and run time of day, make it more like one frame per day. You normally jam your code generators together first thing in the morning, and again coming back from lunch.

So, arithmetic:

(24 frames/sec)x(60 sec/min)x(60 min/hr) = 86,400 frames/hr. 1/86,400 = 0.00115% = 11.5 parts per million.

(86,400 frames/hr)x(24 hr/day) = 2,073,600 frames/day. 1/2,073,600 = 0.00005% = 0.5 parts per million.

A dozen parts per million is quite doable, half a PPM takes extra care. That's why the typical digital camera's built in TC isn't a good as you get from the Lockit type boxes that we use.

On a DIY crystal oscillator, it should be no problem at all to get 50 PPM, which would be fine for takes of a couple minutes or less.




-- J.S.
  • 0

#16 Paul Bruening

Paul Bruening

    (deceased)

  • Sustaining Members
  • 2858 posts
  • Producer
  • Oxford, Mississippi

Posted 05 October 2009 - 01:12 PM

...or you can bloop the head and tail and fudge it in post. Bruce put a head and tail bloop in Frankenmitchel so I could put sound on anything.
  • 0

#17 Matthew W. Phillips

Matthew W. Phillips
  • Sustaining Members
  • 1792 posts
  • Other

Posted 05 October 2009 - 01:29 PM

...or you can bloop the head and tail and fudge it in post. Bruce put a head and tail bloop in Frankenmitchel so I could put sound on anything.


I do head and tail with my current camera.

And John, I'm not trying to argue but I find it hard to believe that anyone could detect one frame off on a 2000' mag. Like you said, when people try hard to pay attention, anything looks off sync. But to notice it during playback of one frame off...that is hogwash, I'm sorry.
  • 0

#18 John Sprung

John Sprung
  • Sustaining Members
  • 4635 posts
  • Other

Posted 05 October 2009 - 01:36 PM

How do I read rotary resolutions of 1/1000 of 1 fps? What type of motor control style would be obliged to that kind of resolution and still be variably controlled?


What you want to do is put a light chopper on the fastest shaft in the system, which is usually the motor. Putting it directly on the motor also steers you clear of any gear lash issues. Suppose your motor runs two revolutions per frame. That works out to 2880 RPM. Give it a light chopper with one hole per degree, and you have 17,200 Hz as your control frequency.

The PLL outputs an analog control voltage. If you want manual variable speed, the easy thing to do is merely switch over from the PLL output to a potentiometer voltage divider, and dial in whatever voltage/speed you want.



-- J.S.
  • 0

#19 John Sprung

John Sprung
  • Sustaining Members
  • 4635 posts
  • Other

Posted 05 October 2009 - 01:52 PM

... I find it hard to believe that anyone could detect one frame off on a 2000' mag.


Human perception of sync is instantaneous. A frame or two out looks the same to us whether it's a two second cut or at the end of two hours. It's only for our efforts to maintain sync that the duration matters.

If you have access to some kind of computer editing system, you can do a test. Make several copies of a take in a row. Put the first one in sync, then pull up a frame of track and let the second have sound a frame early, then two frames early, and so forth. Do another run with sound going an additional frame late each time. Run it at normal speed, and you'll get an idea how well you see sync. BTW, most people are more sensitive to sound early than sound late. That's because the speed of sound is much slower than light, so depending on the distance, we normally see things with a slight sound delay. Resolution is also important. We've had things that looked OK on the Avid turn out to be out of sync on the big screen.




-- J.S.
  • 0

#20 Paul Bruening

Paul Bruening

    (deceased)

  • Sustaining Members
  • 2858 posts
  • Producer
  • Oxford, Mississippi

Posted 05 October 2009 - 01:59 PM

What you want to do is put a light chopper on the fastest shaft in the system, which is usually the motor. Putting it directly on the motor also steers you clear of any gear lash issues. Suppose your motor runs two revolutions per frame. That works out to 2880 RPM. Give it a light chopper with one hole per degree, and you have 17,200 Hz as your control frequency.

The PLL outputs an analog control voltage. If you want manual variable speed, the easy thing to do is merely switch over from the PLL output to a potentiometer voltage divider, and dial in whatever voltage/speed you want.



-- J.S.


Thanks John,

That's generally where I was heading. PWM with 20 turn pot and read off the motor shaft. I can't remember what the math turned out to but I could get to exactly 23.976 off of an 1800 RPM rated motor. I just don't know enough about it beyond dumb guessing and luck.

Matthew,

I know blooping is an unsophisticated way to do it. But, it's damn handy when you never know what your sound gear might end up being. Plus, everything in motor and voltage variations on both picture and sound can go wonky and still line up. I went with the auto blooping so I wouldn't waste all that footage on hand slates. That's something that matters when you're stuck with short ends.
  • 0


Abel Cine

Paralinx LLC

Tai Audio

Metropolis Post

Rig Wheels Passport

Technodolly

Aerial Filmworks

Gamma Ray Digital Inc

Visual Products

CineTape

Opal

FJS International, LLC

Glidecam

Willys Widgets

Ritter Battery

Wooden Camera

rebotnix Technologies

Broadcast Solutions Inc

CineLab

The Slider

Media Blackout - Custom Cables and AKS

Wooden Camera

Tai Audio

Glidecam

Abel Cine

Broadcast Solutions Inc

CineTape

Media Blackout - Custom Cables and AKS

Gamma Ray Digital Inc

Rig Wheels Passport

The Slider

CineLab

rebotnix Technologies

Aerial Filmworks

Ritter Battery

FJS International, LLC

Paralinx LLC

Willys Widgets

Visual Products

Opal

Metropolis Post

Technodolly