Talk:Generation 6 Electronics

From RepRapWiki
Jump to: navigation, search

A couple of things here. The TI part aside Allegro is still much better. My background is an electronics designer. First thing of note is capacitor decoupling. This is important and most designers leave them out because they are the most expensive parts. If you want these parts to work they must be decoupled especially at the motor bus side at each chip. The inductance of the long leads can build up harmonic noise that creates rough travel. The other issue is heat. These components have Mosfets built into them. At 1 or 2 amps they dump a lot of heat. This needs properly laid out boards with ample ground plane to dissipate that heat. 3 amps takes a rather large heat sink, 2 amps a smaller heat sink, 1 amp the large areas of copper on a 2 oz board.

That 2 oz requirement. This is motor technology and we are starting to drive these at 2 amps anything over an amp needs 2 oz boards. Lastly Allegro has it all worked out with 16 step technology the motors run cooler because of a rough sine wave. 32 steps is over kill.

So going back to Allegro, the biggest deals of space on the board are ample copper , decoupling at the IC with 100uF for the low frequency bypass (that squeal you hear is basically low frequency to the computer world), and no one board with everything on it. Proper motor board layout is decoupling of the motor outputs to the control IC. May not be a big deal to everyone, but I want long life out of these motors and drive circuitry. I agree the motherboard needs a refit. I like the idea of ready to plug and play and the USB port integrated. Maybe a daughter card arrangement that way you smoke the output card you keep the expensive part, the motherboard.

With all of that I have designed up a single motor controller board. I am toying with putting on the extruder circuits here but want to keep the costs low and non standard boards cost more. Right now it is all three X,Y,Z outputs with opto interfaces. The connectors are standard Molex no miniature connectors where the crimping tool cost $500. The signal connector is 12 pin. I have added in a reset that resets all motors back to zero at start up. This is done on-board with a simple timer. No big deal simply wait 10 seconds for the power supplies to settle, with the pin low the motors go to their home positions, the timer after 10 seconds takes the pin high and the step and direction functions are as they exist right now.

I will sell these for what they cost in parts. I am not interested in making a profit only that the electronics gets simpler and is cost effective. My goal is under $90 and a goal to include the extruder circuit on this board too with heater bed capacity (all power electronics on one board). The stepper motor chip is the Allego 4983 and will already be soldered to the board as this component is tough to solder but it has the ball technology under its belly that makes for the proper heat transfer to take place (this is very important). Whatever I can get the components at, can get the boards made up to my specs at, a minute of my time to wave solder the chips in place, and any taxes or fees are needed by third parties and viola a circuit that should work for many moons to come. Needs 12 volts at 10 amps max capacity to run. If the heater is included that will be a different voltage to keep the currents down. Have to think about that one but probably leave it up to you all to provide the power. 24 volts minimum, and can certainly switch AC, but now we are talking UL and CSA certification. 24 volts DC should be okay.

BTW not to be a spoiler but everyone should know that heating ABS too high a temp out gases cyanide gases. PVC out gases hydrochloric acid, which will etch any metal you have hanging around and needless to say not good to breath; it will etch you too. If you do use those materials that smell acrid (I know you do), make sure not around anyone and not inside. Take a laser machining class and you will find out the only things really okay to heat up plastic wise are acrylics, polyethylene, and Delrin. So just be mindful, you only have one pair of lungs.

I will have the board ready in about two weeks for myself and I will send two more out for free for peer review. Let me know if you would like a free power board as I have described I will have two available you will have to agree to evaluate it. Please do copy it, only know this I will always be selling these at cost. I hope this inspires others here to do the same. We want Rep Rap everywhere well then lets do that and stop making profits off such a tender and fragile new technology.

I will have the schematics and the board info up in a while. I only use Express Board. It is free and very easy to use for layout. I have a special method I use to make Gerber plots and have tehse made eslwhere.

Whew that was a lot! --Chris Parkinson 00:51, 27 October 2010 (UTC)

Update, Having worked hard to get this into a small board for the first promo price I packed it well with about 100 components. What we now have is a single power board 3.8" X 2.5" 4 layer board with the advantage now of a daughter board that has the brains in it. The daughter board will be simple and nothing more than an existing copy thus far done of the motherboard technology. I will add in buffering for added protection to the processor. This also aids in driving the power mosfets. I now have also created the driver for the heater bed with the extra thermistor. I speced this out at 48 VDC at 2.5 amps, which will get you to the 100 watts you need.

There is a power board with all the power electronics and a interconnection to a daughter board with the processor and other signal based items ie: opto sensors, USB to RS232, etc... With proper buffering this arrangement will help save the computational integrity of the system if a motor shorts out and takes the power board with it.

The board will need a heat sink as the universality of the drivers allows up to 3 amps to be driven that means you can run a wide range of stepper motors with this driver and the fact that it is 16 steps will run smoother with less jaggies in your final parts. That being said a lot of heat will be dumped and time for a heat sink and maybe a fan. I will look into this further when I build the proto up. I have ordered most of the parts and I can get almost everything else locally cheap (I live in Silicon Valley, California), with the exception of the driver chips. Allegra got a little in front of themselves with not enough inventory and am backordered until 3rd week of November. The only way to get the small size and performance here is by this chip alone. Any other will make the board grow in size and the price goes up.

So here is my dilemma. I was hoping to have three boards up and running by the second week in November. My issue is I am also in Law School and have finals in December. I will have no time until the first of the year to work on this with this Allegra backorder (So goes electronics BTW). Anyone willing to help? I will have the boards ordered by the end of this week, and can send anyone interested schematics and PC board layouts. My point is to drive the entire electronics to under $100. I think in large quantities I can get the whole electronic circuit for under $60 in kit form. And again no profit to be made nothing but a great circuits that truly are plug and play with that LED Blink happening and the system linking every time to the PC ready to run the motors and build parts.

So anyone who wants to help with this please post here otherwise I will have to wait unit the new year as my academic priority comes first. --Chris Parkinson 05:57, 28 October 2010 (UTC)

Chris, sounds like a fun candidate for the Generation_8_Electronics to me. Tell, the super-secret developer mailing list, since you're a dev doing dev, and also tell (IRC) channel #RepRap If you are unfamiliar with IRC you can use your web browser. Just click to join the channel.

Glad to have you on board! --Sebastien Bailard 07:25, 28 October 2010 (UTC)

Is there a possibility to manage a heated bed with these electronics? So is it possible to connect a second thermistor and is there a second output to connect a relay or something similar? I see some unused connections next to the usb-connector. I hope those can be used for this.

Tman 11:26, 2 September 2010 (UTC)

Yes this is possible, we have connectors for later addons. So you would need new firmware and an extra small PCB that is connected via the extra RS485 port.

We will develop a heated bed addon PCB in the near future also. So customers that buy our electronics have the ability to upgrade in the future. Just buy installing new firmware and buying/plugging in the extra PCB. I hope that answers your question.

User:Camiel 14:51, 2 September 2010 (UTC)

  • Perhaps we could connect an extruder controller V2.2 via the RS485 and with a UTP cable. The extruder board can be used to control a heated bed, other extruders and a DC motor for something else? Do you think this would require a lot of change to the firmware, seeing that the basis for the extruder controller to act like a slave is already in the firmware? --Matthew 09:05, 6 October 2010 (UTC)

I really would like to see the details, so I am waiting for the detailed schematics of this board.

I too would like to see the details (schematics, parts lists, etc). Is this open source work or will it be licensed to your company? --Matthew 06:50, 15 September 2010 (UTC)

Matthew, everything in RepRap is open source to the extent of everyone's ability. Camiel will be putting the files up once he starts selling. Both because he's a nice guy, and because we're RepRap and he's a RepRap developer--Sebastien Bailard 16:29, 15 September 2010 (UTC).

  • Excellent, thanks Sebastien. I'm looking forward to see how it's all implemented and which MOSFETS, drivers and MCU's are used.--Matthew 06:39, 22 September 2010 (UTC)

Matthew, everyone, you should hang out in, the super-secret developer mailing list.--Sebastien Bailard 09:17, 22 September 2010 (UTC)

The posted specs say Atmega644. Is this correct, or is it the 664p like the Gen. 3 (Makerbot) electronics. The two chips seem to be virtually identical, except that the 664 does not have the second serial port used for by the RepRap firmware for RS485. Camiel mentioned that the RS485 port is still present, so it should be the 664p, right?

Also, this appears to lack the sd card slot found on Gen. 3. I know the 5D code still does not have support for it, but a guy can hope. Is this correct, or does it perhaps have a microSD slot on the back like the Techzone version? --Whosawhatsis 18:50, 19 September 2010 (UTC)

I consider it as inappropriate to nominate a man, who is solely working for the purpose of his own business, as "lead developer". There are dozens of people working equally or more for a better RepRap, no need to alienate these. Traumflug 20:08, 4 October 2010 (UTC)

I would say he's the lead developer of this design, since it appears he's invested his time and money into the development. --Buback 21:32, 4 October 2010 (UTC)

Traumflug, Camiel is an entrepreneur, but he's also one of the many entrepreneurs who really wants to make RepRap better. It is unfair to characterize his actions as only working for the purpose of his business. He wants reprap to achieve its myriad public and covert goals and wants to make a living at it. He's in it for love of art, and also love of making house payments.

I personally regard him as one of our lead developers, since he's actually uploading files and distributing atoms, but I'm not sure if the subject is worth arguing about. He is currently fuzzy as to what files are required, but I am loudly educating him on the subject: ("All the eagle files, Camiel, all the eagle files".) --Sebastien Bailard 22:09, 4 October 2010 (UTC)

  • Sebastien, Unknown: as I fit into all these qualifications as well, please call me a lead developer as well. Just like rocketscientist, nophead, VDX, peer, wakeer, Stoffel15, the MakerBot guys, the TechZone guys, ... and many others. Traumflug 09:08, 5 October 2010 (UTC)

Is there a chance that a 3d model might be available? STL perhaps? Would be great for throwing into a assembly to see how things fit. If not its easy enough to make one from the supplied info.--Isaac Olson 18:43, 25 October 2010 (UTC)

Regarding optical endstop sensors: the Vishay TCST2300 may be a better choice than the TCST2103, especially for the Z axis sensor, because it has tighter resolution, specd at 0.2 mm vs. 0.6 mm. Downsides of this alternative: the pullup resisters R1, R7 and R15 may have to be increased to about 15K due to lower current transfer ratio. The larger pullups would also make the sensor inputs more susceptible to noise, so twisting the wires might be necessary. Has anyone researched Z endstop accuracy? --Jan Kok 07:25, 26 November 2011 (UTC)