3D Modeling; The Future

[inobu]

The research cabin looks great! Is this the same typ e that was used on the super rail cargo?

 

About the OLO: No need to sync constantly, just count the commands from the end position like most home made cnc-s. Then you only need step up and down and two end position microswitches. (this was used in the c64 floppy too, 1st seek was always full length regardless of position until the mech clicked at the end, then it started counting)

This is where the shown configuration fails. 

 

Your example is based on an internal system. The CPU initiates the command and the instructions are executed based on the clock pulse of the processor within its own system. In olo's case the phone has its processor and the olo driver has its. The processors of each unit operates based on its own clock speed (1 giga Hz or what ever) this determines when and how fast the instruction will be carried out. Without synchronization one "script" will execute faster then the other after a few seconds you will have a clump of resin.    

 

Look at your calculations 0.035mm at the highest resolution is fast. The commands has to be packetized 

and read which takes time also. Because the communication path is wifi the overhead adds another issue.

This type of communication is inherently flawed as packets are commonly lost which brings into play error detection and error protection (Frame Check Sequence and Cyclic Redundancy Checks). This is where the overhead and payload and synchronization comes into play. These are just a few factors which have me wondering. 

 

Every other manufacture has to spool their data to their machines but these guys can use a wifi connection and iphone to process a 3D file, drive a SoC ARM processor and stepper motor on a wireless battery system for a few hours.  ok if you say so.

 

Inobu  

[kvp]

I would like to mention that you seem to overengineer the whole thing. Stepper motors have a known step time and distance. This means if you send a step command every 20-30 seconds (the time for one layer to solidify), then it's much slower than the reaction time of the motor. Also the communication channel seems to be serial data over analog audio with FSK signalling. (according to the documents online) This means the keying speed of an average morse operator. In a modern environment, your TV's IR remote control is sending data faster.

 

Printing procedure: turn screen off, send step signals to seek to the bottom of the pool, turn screen on, light layer, turn screen off, send step signal, wait for step signal to complete (worst case timing), repeat from turn screen on until all layers are done. The first layer will be much thicker thanks to the mechanical inaccuracies of the box and the pattern on the puller plate, but from then, it's a very simple and slow process.

 

The whole thing is a slow process and 1 or 2 seconds extra wait on a near minute long lighting session doesn't really matter. The key is that stepper motors always step the same distance and if you always wait for the worst case step time, you only have to count the steps. Most hobby CNC systems work this way and most errors come from driving the motors too fast, which could lead to step loss. In this case the step time is several seconds per single step, while a hobby CNC has several hundred steps every seconds.

 

ps: Hardware wise the main step control processor can be an 8 pin PIC chip, used for led throwies and keychain remote controls. The biggest power usage would come from the stepper motor, which has to make several hundred steps a day. A hobby CNC does the same in a second, so power requirements for the two should be around the same. (one whole day of printing could need the same power a CNC uses every second) The phone is just playing back a low framerate video that was downloaded from the server. It's one long photo, then darkness with a step command sound, then one long static photo again, then one command sound again. Even the cheapest media players could do it, but the phones are required for their high resolution screens. This puts all logic and work on the cloud servers and can be done before the actual printing begins.

 

ps2: Manual operation: if you are very patient, then using pre printed (or hand drawn) projector foil, mirror routed sunlight and hand stepping would be possible with a home made contraption: insert foil onto tray, lock it in place, uncover sunlight hole, wait, cover sunlight hole, unlock and remove foil, turn a single full rotation on the puller plate drive crank, repeat for all foils

[inobu]

Where does it say that they are using FSK? I gotta read this

 

 

For those reading along. FSK is Frequency Shift Keying.... If you recall modems that sound that you hear when the modem connected is FSK.

 

 

Inobu

[kvp]

Two pages back there was a link to the project q&a and it's there. It was even explained here. This is kind of a low tech project for a high tech resin.

[dabsan]

the details are sweet! handpainted or airbrushed?

Hand painted, very carefully :)

[dabsan]

Nice work Debsan! it came out very nice!

 

jeff

Thank you Jeff, I think I will put it up for sale on shapeways.

[inobu]

Hand painted, very carefully :)

That's good for hand painted. You get a Harder and Steenbeck infinity air brush and you will do wonders.

 

Inobu

[dabsan]

That's one of the best feelings! To see your own designed, printed and finished model. I hope it ignited the spark to create more!

Thanks Toni, yes I will be making something else soon.

[dabsan]

That looks nice. The 3D printing is nothing. It is the creation of the file that garners the accolades.

 

Inobu

Many thanks Inobu :)

[inobu]

How does OLO's app work?

 

Once you have selected a model that you would like to print from within the app it will be sent as an STL file to our cloud servers. There, it will be processed into the individual layers required to create the model. It is then converted into a .olo file, a vector layer file format with a synchronized audio component. It is a small file is only about 1/6 the size of a normal STL file. This file is then downloaded from the servers and saved locally on the phone. The phone must have notifications turned off, disconnected from wifi, and be placed in Airplane Mode so as not to cause interruptions to the print job. Then the printing begins!

 

Operative word. synchronized audio component    

 

See it requires synchronization but they have no means for audio communication between the two components. The assumption was made that it was wifi but I guess that's not the case. Only two analog inputs, mic and headset and their operation video shows neither being connected. 

 

lol.....

 

If you want to believe.....believe.

 

Inobu  

[inobu]

Many thanks Inobu :)

There is a company in France you might want to check into. Last year there was a guy I ran into that was prototyping shells and his was close to a mold finish. I cannot remember the company's name. I'll have to ask around to get in touch with the guy.

 

Inobu

[kvp]

Synchronised means it's a black and white video file with the audio track containing the stepper commands. On the printer side, you'll need an active headset mic, a pullup resistor and a tone detector circuit, like a pic mcu with an adc (the 12f510 is 77 cents at digikey) and a piezo buzzer for the reverse direction. Very 70ies level hobby electronics level tech. I would use a data rate of 1 to 4 bits per sec to be very reliable. (see r2d2 droid talk for an example)

[inobu]

You got to be kidding. lol ...............black and white video file?  that is called facsimile or Fax as what we called it. It just gets better and better.
 
 
So you are saying this system uses an active headset mic and piezo buzzer as a means to transmit data back an forth.
 
So this is what you are implying is going to happen? without the dialing of course

 

 
lol
 
Inobu Edited April 14, 2016 by inobu

[HantuBlauLOL]

b/w video file is for the resin mask, while the audio is for controlling the stepper, as simple as that. cmon..

[inobu]

Someone is going to OLO all the way to the bank.

 

Inobu

[cteno4]

That's the whole point of making it as simple as possible using the phone as much as possible for any processing. Preprocessing on their cloud system keeps all that more complex coding very centralized rather than doing all that on a bunch of platforms. Phone software just has to play the slide show and beep after each layer. the audio means no physical connection needed between the phone and printer. Only need a very simple processor to step the print plate up a notch when it gets each beep when the layer is printed!

 

This all is why I see this as more a novelty printer for larger and less high resolution things rather than a hardcore hobbyist printer.

 

But with prices dropping so fast on high res screens more will come all built as one. While Linux computer is $20 now to do what the phone is doing!

 

Cheers

 

Jeff

[dabsan]

Hi, here are some photographs of the painted ZX45A-1 Research Cabin, I am very pleased with the look of it.

 

 

P1000329.jpg

 

P1000328.jpg

 

P1000330.jpg

 

P1000331.jpg

Shapeways link :-

https://www.shapeways.com/product/723T4K364/japan-rail-freight-research-cabin

[kvp]

And i've found a picture of it on the M250 super rail cargo:

http://www.isok.jp/rail/car/fc/pict4672.jpg

Does the cabin fit on the model?

Edited April 15, 2016 by kvp

[dabsan]

Hi, I don't have a M250 super rail cargo train to try it on but it should fit ok. Would be great to see it on someones model.

I am planning on running it more like this :-

 

 

Might even include a figure on the back :)

[inobu]

The tabs don't line up. This is a Kato Koki 19D (top) and Kato M250 flat car (bottom). 

 

 

 

container size is outer dim

Inobu

Edited April 15, 2016 by inobu

[dabsan]

The tabs don't line up. This is a Kato Koki 19D (top) and Kato M250 flat car (bottom). 

 

 

 

container size is outer dim

Inobu

Hi Inobu, thanks for your photograph, good to see the comparison between the wagons.

I took a few snaps of the measurements, the inside of the container is pretty tightly fitting and I am sure it would grip on to that M250 wagon base even though the tabs are suituated in different places from the koki, I originally designed it for fitting a koki. Wish I had a M250 wagon to try it on.

 

 

Edited April 15, 2016 by dabsan

[inobu]

I think it would because the tabs are inside of the koki tabs.

 

Inobu

[cteno4]

ive been able to clip the tomytec containers onto the m250. the one clip is a bit tight but it works w/o seeming to bend the clip or container side much. little hit with the dremel can make a slot easily if needed!

 

i guess the little indent for the tab could just run down the length of the container to make it so it could get snapped with any tab location.

 

ill be ordering one, just looking if there are any other bits i want on shapeways! kind of have most of all of tonis bits now. thanks for doing this debsan!

 

jeff

Edited April 15, 2016 by cteno4

[dabsan]

ive been able to clip the tomytec containers onto the m250. the one clip is a bit tight but it works w/o seeming to bend the clip or container side much. little hit with the dremel can make a slot easily if needed!

 

i guess the little indent for the tab could just run down the length of the container to make it so it could get snapped with any tab location.

 

ill be ordering one, just looking if there are any other bits i want on shapeways! kind of have most of all of tonis bits now. thanks for doing this debsan!

 

jeff

Thanks Jeff, there are some great items on Shapeways and it's a real pleasure when finding something unique and special on there. I very much appreciate your interest in my model. I have started another model, this time something in Z Gauge, will tell more later :)

[dabsan]

I think it would because the tabs are inside of the koki tabs.

 

Inobu

Thanks Inobu, yes I was thinking the same thing, appreciate your help.