AVS Forum banner

1 - 20 of 48 Posts

·
Registered
Joined
·
887 Posts
Discussion Starter #1 (Edited)
Hi!
The screen is made from rolled square steel tubes, about 1x1". Curve radius is about 11m, and the width of the screen about 2,8m. The fabric itself is not DIY, and is a fine mesh woven material.

The masking system has two tubular motors, which are controlled with relays from one controller each. The received IR is common between the controllers, but nothing else. On each motor there is a quadrature encoder that gives feedback for the position of the mask. The controllers are based on PICAXE 18M2 micro-controllers. These handles IR-decoding, reading/decoding encoder feedback, storing and recalling presets and jog. 10 presets can be stored, and the masks can be jogged simultaneously or individually. I can post images and details later, here is a short video that shows how it operates:

 

·
Registered
Joined
·
887 Posts
Discussion Starter #3 (Edited)
Welding up the frame about two years ago.



Attaching thin wooden strips at the back of the frame to have something to staple the screen material to.



Screen material stapled to the back of the frame. Material has much give, and if it is tensioned "sideways" it will bow out, since the frame is curved.



False wall with flush mounted speakers. Wall is covered with DMD sound transparent fabric. Wall has about 12cm thickness of absorbent material.



Screen mounted to sides of the false wall. I used 4 angle brackets with elongated holes, to be able to adjust the screen slightly in/out.



Motors and aluminum tubes. Cut a notch in the tubes to lock the motor. Motors are very plain with mechanical endstops. They are 220V, 30RPM, 3Nm, about 100W. The brand is TESC. End of line items at $12 each on eBay:D Got 4 since I planned to build a flat screen with a 4-way masking system.



Close up of attachment of the encoder. It is mounted on a piece of aluminum strip. The small spring keep the (LEGO...)gears engaged without play.



Motors are mounted on two pieces of wood that is screwed to the wall with two angle brackets each. On each side of the screen I have attached 20mm plastic tubes to make a smooth edge for the masking material to slide on. Recently I replaced these tubes with longer ones, since I needed to have the masking material higher up on the motor tubes.

 

Attachments

·
Registered
Joined
·
4,939 Posts
Nice!! Very impressive. Lego gears and all.
 

·
Registered
Joined
·
887 Posts
Discussion Starter #5 (Edited)
Nice!! Very impressive. Lego gears and all.
Thanks! The Lego gears are quite durable, and the encoder rotates very easily, so I don`t expect any trouble from this part of the system:D

So far the PAD running iRule seems one of the weak parts. It slows down after a little use, and gets less and less responsive. The IR-codes that I use for the jog-buttons are "repeat on hold", which doesn`t work well in iRule so far, at least in my setup. I get the "spinning wheel" after a while in the iRule app when I press one of these buttons. From a regular remote it works every time. I can reprogram my hardware to not use repeat on hold if nothing else work out.

This is a screenshot that show the iRule-page for the mask and lenslift control. The upper row of buttons are the presets, and the lower part are the jog buttons. On the left are common buttons for most of the iRule pages. The AR-buttons are macros that control the masks, lens lift for A-lens, PJ lens memory, and anamorphic stretch. The background image is made by M Borner, and copied from one of the iRule threads.

 

Attachments

·
Registered
Joined
·
530 Posts
Looks to work wonderfully. I wish I had your skills to make such a system. I currently don't have masking so it would make my system much better.


I have a DIY flat to curved 157/200 inch diagonal screen with a Panamorph UH480. I used one track linear actuator on each end to curve the screen and one to move the "sled" that has the UH480 mounted on it. Hooked up a wireless DC control with remote so all I have to do is push the remote button to curve the screen and move the lens into position and then press the stretch button on my PJ remote and I am ready to watch a scope movie in less than 30 seconds.
The AT screen is 1x4 wood framed and secured in the middle top with a French cleat and middle bottom with a wooden dowel attached to my baffle wall. It curves very evenly and nicely. The linear actuators on the ends keep the screen level horizontally.
Total cost for the system was about $350.


Moto
 

·
Registered
Joined
·
887 Posts
Discussion Starter #7
Hi! Sounds very interesting, I have never seen a screen like that. Also seems that the automation works well. I started to use iRule, because I moved everything out of the HT-room, and I needed to remote control some devices with rs232, some with ir, and some with ip. I even use a remote relay to turn on a HTPC. rs232, ir and relay is done via a Global Cache gateway. Except the small problems with the PAD, iRule works well. With the macro buttons I go from 16:9 to 2,35:1 pushing only the 2,35 button on the left of the shown iRule screen. I made a short video that show what happens:

 

·
Registered
Joined
·
887 Posts
Discussion Starter #8 (Edited)
For some reason the attachment disappeared, so I uploaded it again..

The electronics is simple, thanks to the microcontroller. Can be built on a veroboard or similar. The bottom circuit is the controller schematics for one motor. In addition there are three other circuits that are shared for all (currently two) controllers:

The upper circuit is a 5V powersupply/ trigger turn-on. The powersupply is a simple 7805 3-terminal regulator. The components inside the blue dashed line is the trigger circuit.

The next circuit is the IR-receiver. It has a LED020 receiver/demodulator, that outputs the "baseband" IR-signal to the microcontroller.

The next circuit is a retriggeable "one-shot". It will output a high signal to the microcontroller as long as there is IR-activity. With this signal I can jog as long as I press a jog-button, and stop jogging when the key is released. Without this this would not be possible, as the microcontroller can not wait for IR, and count encoder pulses at the same time. The signal also works as an emergency stop between presets. The mask would stop if an arbitrary button is pressed.

It would still work without this signal, and a small change in S/W, but then a press on a jog-button will start jogging, and a second press would stop jogging.

 

Attachments

·
Registered
Joined
·
887 Posts
Discussion Starter #12
That is extremely simple! Thanks for the info. I may have to put it into use one day.
Thanks! I`ll make the S/W available, if anyone wants to try this out. The editor that is needed to program the microcontrollers is free, all you need is the editor, a serial cable, and a Windows PC with a serial port, or a usb->serial converter.

The editor can be downloaded from here. There is also a shop, where you can get the microcontrollers, and some other components that I used. I don`t think you can get the microcontrollers elsewere.
 

·
Registered
Joined
·
1,719 Posts

·
Registered
Joined
·
887 Posts
Discussion Starter #15
Looks good, I like seeing do-it yourself controller. Actually I am trying to make my own and having issues with it, you see pretty good with electronics, any chance you could take a look at my thread? I am thinking I might have to start over, maybe something like you are using.


http://www.avsforum.com/forum/117-2...731818-masking-control-help.html#post28445778
Hi! I have read your building thread, and I remember being impressed, especially by the masking mechanics. Very nice HT-room also! I didn`t read your masking help thread before, but I just had a look at it now. I`m no electronics expert, but I learned a few things the hard way myself.. I noticed that my motors generate nasty voltage spikes when the mechanical end switch is reached. This is a bad thing when using a microcontroller, as it easily hangs or restarts, if the supply voltage drops below a certain level. A more careful layout of the board could probably have taken care of the problem. The layout I`m using now was built for a four-way system, with 4 microcontrollers, and a separate board with all the relays on it, se here, post #44 . I will probably make a smaller layout with everything on one board for my current two motor system.

I modified the relay board with RC-snubbers from the relay output and to ground. This took care of most of the spikes. On the schematics I posted in this thread the RC-snubbers consist of C4 + R6, and C5+R7. I made my own first, but later I bought some integrated ones, which was more effective, see here.

Another issue that has room for improvement in my design is proper grounding. If I make a new layout I would at least keep the ground for the relay drivers (Q2, Q3) separate from the ground for the micro controllers. Both would then connect to a common ground point in the box.

You could also try to reduce the value of your pull down resistors, if the latches can handle a little more current. Maybe some ferrites on the motor power leads, or a AC-filter would help.

Else, a nice thing with the encoder is that I can set my presets freely, and change them from the remote control very easily.
 

·
Registered
Joined
·
887 Posts
Discussion Starter #16 (Edited)
Gunnar ;
awesome job all around, thx for posting.
Q for you. When you run the masks in 16:9 do you also apply some EQ to the Mains R/L for the small loss of hi freq thru the black out material? Or not really noticeable so no need?
Thanks!
No, so far I have tried to finish the room, I did some image calibration, but have not started with the audio part yet. You can just see the speakers behind the screen on the below picture (masking not yet installed). The L/R speakers will, as you say, be behind the masks in 16:9 format. So far I think it sounds rather good, but it can probably use some eq. The mask material is audio transparent. It is called DMD, and I got it from here



Rear view:

 

Attachments

·
Registered
Joined
·
1,719 Posts
Hi! I have read your building thread, and I remember being impressed, especially by the masking mechanics. Very nice HT-room also! I didn`t read your masking help thread before, but I just had a look at it now. I`m no electronics expert, but I learned a few things the hard way myself.. I noticed that my motors generate nasty voltage spikes when the mechanical end switch is reached. This is a bad thing when using a microcontroller, as it easily hangs or restarts, if the supply voltage drops below a certain level. A more careful layout of the board could probably have taken care of the problem. The layout I`m using now was built for a four-way system, with 4 microcontrollers, and a separate board with all the relays on it, se here, post #44 . I will probably make a smaller layout with everything on one board for my current two motor system.

I modified the relay board with RC-snubbers from the relay output and to ground. This took care of most of the spikes. On the schematics I posted in this thread the RC-snubbers consist of C4 + R6, and C5+R7. I made my own first, but later I bought some integrated ones, which was more effective, see here.

Another issue that has room for improvement in my design is proper grounding. If I make a new layout I would at least keep the ground for the relay drivers (Q2, Q3) separate from the ground for the micro controllers. Both would then connect to a common ground point in the box.

You could also try to reduce the value of your pull down resistors, if the latches can handle a little more current. Maybe some ferrites on the motor power leads, or a AC-filter would help.

Else, a nice thing with the encoder is that I can set my presets freely, and change them from the remote control very easily.

Thanks for the ideas I will try them out and see what happens. Originally I had a position tracker on the motor but it would not make up for over run of the motor. Basically it would just shut the motor down at the correct time but the motor would coast an extra 1/8" or so each time, well after 8 times it was off 1". Does what your are using track the live position of the motor, if say someone pushed the masking by hand and opened it further does the controller see that?
 

·
Registered
Joined
·
887 Posts
Discussion Starter #18 (Edited)
Thanks for the ideas I will try them out and see what happens. Originally I had a position tracker on the motor but it would not make up for over run of the motor. Basically it would just shut the motor down at the correct time but the motor would coast an extra 1/8". or so each time, well after 8 times it was off 1". Does what your are using track the live position of the motor, if say someone pushed the masking by hand and opened it further does the controller see that?
No, the controller would not see that. But the motor can`t be turned by hand. Not sure what type of gear is used. I shut off power to the encoders when the masks are standby. This is done with the Q1 transistor on the schematic. I do this because the light source in the encoder is said to weaken when left on all the time.

In the software I have a predefined offset that will shut off the motor a curtain number of pulses before reaching the target count (the selected preset). I continue counting encoder pulses a while after powering off the motor to make sure no pulses are lost. The motor stops very quickly, so the error is small, and the error does not accumulate, since the encoder sits on the driven axis. The encoder is 128CPR, and with quadrature decoding of the two pulse trains I get 512 changes of state pr motor revolution. I think this was the encoder I used.

This translates to about 3.2 pulses pr mm of mask travel. Without losing any pulses the error would be the same every time, and can be tuned with the offset. So far I have not noticed any change in the preset positions. If necessary I could make a calibration routine. Eg I could use 2,35:1 or 1,33:1 as the master position, and calculate the other presets as offsets. In case of error I would have to set only the master position again.

I don`t want to use the mechanical endstop switches in the motors for anything but safety stops, outside the preset range, in case something goes wrong..

If your masks can be moved by hand this would probably not work well in your system.
 

·
Registered
Joined
·
1,719 Posts
No, the controller would not see that. But the motor can`t be turned by hand. Not sure what type of gear is used. I shut off power to the encoders when the masks are standby. This is done with the Q1 transistor on the schematic. I do this because the light source in the encoder is said to weaken when left on all the time.

In the software I have a predefined offset that will shut off the motor a curtain number of pulses before reaching the target count (the selected preset). I continue counting encoder pulses a while after powering off the motor to make sure no pulses are lost. The motor stops very quickly, so the error is small, and the error does not accumulate, since the encoder sits on the driven axis. The encoder is 128CPR, and with quadrature decoding of the two pulse trains I get 512 changes of state pr motor revolution. I think this was the encoder I used.

This translates to about 3.2 pulses pr mm of mask travel. Without losing any pulses the error would be the same every time, and can be tuned with the offset. So far I have not noticed any change in the preset positions. If necessary I could make a calibration routine. Eg I could use 2,35:1 or 1,33:1 as the master position, and calculate the other presets as offsets. In case of error I would have to set only the master position again.

I don`t want to use the mechanical endstop switches in the motors for anything but safety stops, outside the preset range, in case something goes wrong..

If your masks can be moved by hand this would probably not work well in your system.

It doesn't easily move by hand but it could be if you pushed hard enough. I also have manual switches on the screen wall that I can hit to open or close quickly with out grabbing remotes and powering anything up. The other tricky part of my system is that my masking closes completely to cover the entire screen when not being used, this means that when it closes the masking panels run into each other. So everything has to be real précises because if it is off and the panels hit with out shutting the motor off, well it will smoke the motor I am guessing.


My ideal system is one that monitors the position of the masking at all times so it knows where it is. When shutting the masking it would shut unit a mechanical switch is tripped, as the panel hit each other. Upon hit that switch it would also recalibrate to make up for any discrepancies. So every time you shut down the theater they close and recalibrate automatically. A magnetic encoder would not have the problem with being on all the time I am guessing. Biggest problem is I don't have enough time to mess around with it.
 

·
Registered
Joined
·
887 Posts
Discussion Starter #20
I think the center switch, calibration and the wall switches could be implemented in a system like mine. There are several unused inputs/outputs on the microcontrollers that can read the status of the center switch, use as wall switches etc.. If you wanted to monitor the position all the time, I think you need an additional decoder, and when the controller want to move the mask it would get the current position from the separate decoder. Both the separate decoder and the controller could share the same encoder.

BUT as you say, time is a limiting factor. I started experimenting with microcontrollers for this purpose several years ago.

What material did you use for your masks? I can`t find pictures in you building thread that show how the material is rollled up, and how the mask leading edge is guided? This part could need some improvement in my system. I use a cheap curtain track with small nylon slides that slides in the track. It works ok, but I think it would be better to have "riders" with wheels in the track, to decrease friction. The DMD-material has a lot of give, and the top would sag, if I try to close the masks completely. It doesn`t help to increase the tension in the bungee/wire, it still sags. It still sags a bit, when the mask move to 1.33:1, but the sag is behind the upper trim, so it doesn`t show. Annoying anyway..
 
1 - 20 of 48 Posts
Top