For those folks who have been following this thread, the previous post pertaining to the final draft of the layout has been updated to include an additional image and other information.
Simply for the sake of discussion and referring to the layout of Post #121....
What can be done to make this layout more versatile? There is a wide variety of stepper drivers in the marketplace and through some more design and circuitry, I am almost certain that this controller could be improved to be compatible with a larger portion of them. In order to make it more compatible with a larger portion of stepper drivers, this is what I propose:
The board should include a single two pin jumper setting and the necessary circuitry to allow the Propeller to send 5VDC signals to the STEP, the DIRECTION, and the DISABLE input pins of the stepper driver. In other words, through a single two pin jumper setting, the Propeller could either output 3.3VDC or 5VDC signals. Or perhaps these should be individual settings, with four two pin headers, since a person or machine may have several different types of stepper drivers.
Each of the DISABLE output circuits of the layout (IO pins 0, 21, 24, and 25) should have the necessary circuitry to either provide a ground, a 3.3VDC signal, or a 5VDC signal, with the distinction of a 3.3VDC or a 5VDC signal, being determined by the jumper setting(s) mentioned in item 1.
I have been reading this thread, and others, relating to the use of the Propeller for 3D Printing electronics. I too am interested in doing this although my primary difficulty at the moment is whether I can justify using the Propeller based on the following Pros and Cons:
Pros:
The mutiprocessing paradigm would benefit 3D printer control for things like synchronized moves and extrusion.
The ability to program in a choice of languages e.g. C, Spin and PASM offers flexibility (although I am not convinced of the need to use C++).
Being able to multi-thread within a cog could also be useful for orthogonal processing e.g. monitoring similar inputs.
The performance to cost ratio of the Propeller is a big incentive.
Cons:
Basic I/O is limited so the use of e.g. SPI peripherals would be mandated.
I envisage that on-chip code space would be insufficient requiring the ability to place code in external memory and execute it efficiently.
The lack of on-chip peripherals makes alternative devices look more appealing. The Propeller must be chosen on the basis of additional benefit after including the cost of required external hardware.
Use of the Propeller requires a different programming mindset which has to be considered for software development and ongoing maintenance.
I am approaching the implementation of 3D Printer electronics differently to established practice. My paradigm:
1) The controller only processes primitives, it has no understanding of GCode. A separate PC Host application can import GCode but, long-term, it is envisaged that a custom slicer will be used to generate primitives.
2) The controller is responsible for maintaining safety of the 3D Printer. Methods will be provided to monitor filament presence, stepper motor fault conditions, movement and temperature safety limits.
3) Printing over USB and from SD card will be supported although use of a FAT filesystem on the SD card is not envisaged (for reduced overhead and licensing issue avoidance).
4) In a departure from established practice, I aim for printer independence in the primitive data transfer implementation i.e. a set of build commands should be independent of material in use and absolute speed specification (relative speed specification is permitted).
5) Operational profiles should be retained in the 3D Printer electronics. The PC Host application may optimize printing based on knowledge of a connected printer's configuration but this should not be mandatory.
6) Sufficient printer hardware options should be supported e.g. 5 extruders (for separate single or 4-colour printing or support with up to 4-colour printing). PID temperature control of extruder(s) and heated build platform.
7) A single-board design without the need for jumpers or current setting potentiometers. The intention here is for an install-and-forget solution with operational configuration and tuning provided by firmware.
I have investigated implementing this using a twin Microchip PIC solution. This relies significantly on the integrated peripherals of the two PICs (one handles USB/SD card and temperature control - the other movement and user interface). Implementing a similar scheme using the Propeller could reduce the number of microcontrollers to one but requires the addition of I/O expanders, ADC and DAC converters (SPI for adequate performance).
My biggest concern in adopting the Propeller is that of programming for a large-ish application. The twin-PIC solution would afford me 128kB of distributed code space and I assume that a Propeller implementation would need a similar amount of code space. I understand that this can be achieved in Propeller-based designs but I am not sure how expanded code space translates into usability on the cogs. Am I right in thinking that code has to be paged into the cog under the control of some driver/kernel and this has to be done in less-than-cog-ram-size chunks? If so, I could envisage some overhead issues will need to be contended with.
So, I believe 3D Printer electronics can be implemented with a Propeller (although I would not attempt an all-in-one solution like Marlin or similar firmware). In fact, I think the distributed paradigm of firmware and smarter PC Host application would fit a Propeller-based design better.
I hate to say it, but if you are thinking of a four extruders, you are going to need a lot of IO pins just for the stepper drivers alone. The cost of the extra peripherals are almost negligible, just breaking the peripheral components down to the chips and passive components, and in my design, these peripherals only go to items that do not require a large amount of speed, so it is okay with me. However, I would not even think about a four extruder design, without adding another Propeller chip, which at that time, you may as well eliminate the expansion chip. Looking back now, if I had gone with a dual Propeller design, and especially if I wanted more than one extruder, I would have had one Propeller chip just to run the stepper motors with the IO pins, cogs, and stepper driver software, but at that point, there is the requirement of serial communication between the two chips.
That being said, I have written SPIN drivers for my steppers in the past, and the speed is definitely there if you use counters effectively. SPIN is quite easy to learn and very effective for programming the Propeller. I do not believe there is a need for C, however in the long run, I may be required to use PASM, but I do not think so. Time will tell.
The question is whether or not the software will fit.
However, I would not even think about a four extruder design, without adding another Propeller chip, which at that time, you may as well eliminate the expansion chip.
I am not so sure about that. The Propellor could be programmed to accommodate multiple extruders over one cog. You could compute the requirement for steps and step the motors in parallel as needed. More motors does not necessarily require more cogs.
Looking back now, if I had gone with a dual Propeller design, and especially if I wanted more than one extruder, I would have had one Propeller chip just to run the stepper motors with the IO pins, cogs, and stepper driver software, but at that point, there is the requirement of serial communication between the two chips.
You could implement SPI communications between Propellers but I am not sure that two would be required. My current printer uses an ATmega1284p clocked at 16MHz giving, I believe, 16MIPS performance. A propeller clocked at 5MHz can achieve 160MIPS across all cogs.
That being said, I have written SPIN drivers for my steppers in the past, and the speed is definitely there if you use counters effectively. SPIN is quite easy to learn and very effective for programming the Propeller. I do not believe there is a need for C, however in the long run, I may be required to use PASM, but I do not think so. Time will tell.
I would use PropGCC to implement the bulk of the firmware. Computing movement paths is intensive and I doubt that Spin could provide the required performance. I aim to use TI DRV8825 stepper drivers which are capable of 32x microstepping. I believe that the Propeller is one of the few low cost processors (other than an ARM device) that could achieve a 32x step rate.
The question is whether or not the software will fit.
In 32kB I am sure it would not. External code expansion would certainly be required and I have yet to determine how, and how efficiently, this is supported in the available languages.
Have you looked at the Propeller Project Board? I find I can lay out circuits in a cleaner fashion than on the Proto Board. And I also like the SMT areas for those type of chips, pads on the back for an addition regulator, and pads for a uSD connector on the back also.
I looked at it earlier when considering the options and opted for the proto board, because I am very familiar with the proto board and I thought it had more prototyping space. However, looking back now, that might have been a better choice because of the 16VDC VIN, but I already have the proto boards, so I am kind of locked in.
I am glad you found this interesting enough to take a serious look at it
I would be interested in seeing an illustration or a pic, if you have one handy.
I would be interested in seeing an illustration or a pic, if you have one handy.
Bruce
Of prior installations? The company that I was subbing for monitors this forum, and they will sue me if I publish any of my designs, which they now say they own,(not that they are earth shattering).
Pretty much 80% came out of the OBEX.
I haven't' done a personal project on the Project Board yet, but Ken just sent me on to put on the Prototype QuadRover that I got last year.
I just ordered the remaining missing pieces to get this board underway. I don't know how soon I will get all the parts, but I should soon have a prototype built and ready for various tests.
It appears that I will have to make my own daughter board, instead of using perf board, because according to my calculations, the hot end heater and build platform heater are both going to require quite a bit of current
To be perfectly honest, I am really looking forward to getting this board built and experimenting with it, because I have been wanting a similar board for a very long time.
Instead of waiting on the remaining parts, I decided to start building the board last night. At this point, I have completed most of the difficult soldering. Let me state just for the record, that it was much easier drawing this thing up, as compared to actually making all the necessary connections. In the illustration, it appears as though there is plenty of room to work, but in reality, the workspace is quite cramped. However I must say that it is actually quite nice to have the illustration as a guide. The actual board looks very similar to the illustration, with a few minor changes. Although the illustration does not show it, there are a lot of solder bridges on the bottom of the board, and as mentioned, space is cramped, so if you try to construct this board, be careful and watch out for shorts, testing and verifying everything.
Additionally, if you intend to attempt this project on a Propeller Proto Board, you should have pretty decent soldering skills, a good soldering iron, an eye for detail, a solder sucker, a multi-meter, an aligator clamp, and plenty of light.
The top of the board looks very nice, but the bottom is quite a different story. When I get finished with it, I will post a photo. Actually as the 3D printer project goes forth, there may be several photos throughout various threads.
My remaining parts should be here tomorrow afternoon and I should finish it tomorrow evening sometime, with the exception of the ADC channels and daughter board.
It does not appear that I will finish the board this evening, because I had to take a break for a while, and I will probably get back to it, in about an hour or so. There is however a major setback. When ordering my parts, I forgot to order the 10 pin ninety degree header for plugging in the Adafruit ADC breakout board in a vertical position. What a bummer. Oh well, I will do what I can.
With the exception of the circuitry for the two ADC channels required for the hot end and build platform thermistors, the Proto Board is all soldered up and ready for some serious testing. Regarding my previous post, after checking around for a ninety degree header, I opened an old parts pouch for an OEM BS2 and to my amzement, there was a ninety degree header So the Adafruit breakout board is also soldered up and installed.
Over the next several days, I will be creating the following temporary testing cables to allow attachment of all the peripherals:
In the provided illustration, you will find that two revisions have been made to the Proto Board layout, which should be apparent from viewing the illustration. However by referring to one of the revisions, the three pin headers going to the ADC channels allow me to finish the Proto Board with out additional circuitry and since I will be having remote boards at the heated build platform and the extruder, this should work out well.
Out of the four channels provide by the Adafruit ADC breakout board, I will only be using two of them, one for the hot end thermistor and the other for the heated build platform thermistor. The three pin headers going from the ADC will provide VDD and VCC going to remote circuit board locations, and the remaining pin will bring back the reading to the ADC. In the illustration below, you will also notice a small schematic. This schematic represents the circuitry that I will have on the remote circuit boards for reading the thermistors. R1's value will be dependant upon the thermistor utilized, however, although not shown in the schematic, R1 will be attached to the remote circuit board with a pressure terminal, to allow for quick changes of R1. In addition to allowing quick changes for experimentation purposes, the pressure terminals will allow for quick changes of R1, when changing filament types from PLA to ABS, because different readings are required.
As mentioned earlier in this thread, when utilized as a controller for a 3D printer, a daughter board can be attached to the SVGA port area of the Proto Board. When I drew up the previous illustrations, I did not take all things into consideration, therefore the depth of the daughter board must be increased to allow for placement of all the necessary components. And it is noteworthy that at this point I am still making assumptions, so please be cautious.
Before I begin this discussion, let me first set down some facts:
The illustrations are only meant to be a pretty good guide in wiring this board up. As mentioned earlier, I had to make a few alterations, to get things to fit properly, especially since my board required an additional (4) transistors and (4) resistors for the DISABLE circuitry required by my stepper drives.
The extruder fan, the hot end heater, and the build platform heater, will all be powered by 12VDC, each being controlled by a Propeller IO pin via IRF3708 MOSFET. Although currently being discussed, it is my intention to make a seperate board for these items and attach it to the SVGA port area of the Proto Board. A four wire cable will go from the Proto Board to this seperate board, supplying three signals and a ground.
In previous illustrations, the daughter board seemed to be of sufficient size, but in fact, it was only a mere 7/8" X 1-5/16", which is not very big at all, especially when (3) TO-220AB packages (IRF3708) must be installed.
Another consideration is that the hot end heater and the build platform heater are high amperage items. It is my intention to allow 2A for the hot end heater and 20A for the build platform heater, with the additional extruder fan only requiring 0.25A. The 20A for the build platform heater is just a guess at this point and not actually a fact. However, the parts shown on previous illustrations of the daughter board are believed to be insufficient due to the high amperage demands.
Additionally, it is worth noting that during the construction of my controller board, it became necessary to move the female header for the 10 pin ADC breakout board 0.20" inches in the direction of the SVGA port area.
To sum it up and as mentioned, the previous suggested size of the daughter board was insufficient, so I am now suggesting a new size, based upon the available room after Proto Board construction. Additionally, different types of terminals will now be utilized to accommodate the higher amperages required, although I may still be dreaming. At this point nothing is carved into stone, except for the room available to accommodate the daughter board, without crowding the Proto Board, which will now be 1-3/16" X 1-5/16". Please refer to new illustration below for a size perspective.
As seen in the EAGLE board layout below, I can definitely get all the components to fit, but getting an adequate trace width for the heated build platform, will be quite a different story. I may need to reconsider my strategy and maybe have two stacked daughter boards, one for the heated build platform and the other for the extruder fan and hot end heater.
EDIT: I suppose I could try a board fill and see what it looks like.
EDIT: I think I may work with the layout a little more before abandoning the single daughter board concept.
EDIT: It is starting to look like it might just work out.
In the attachment below, you will find the Eagle files for the daughter board. I would imagine these files would work just fine for smaller heated build platforms, but I do not believe it will work for mine, especillay for printing with ABS filament, and perhaps not even PLA, because my platform will be 12 X 12. Either way, I did the best I could with the space available. I will give it a chance and see if I can lift the copper off the board
I might also try elongating the board to go over the programming port of the Proto Board if this does not work.
Anyhow, here are the files. You will have to use the rats nest tool to get the copper pours.
EDIT:
I might also try elongating the board to go over the programming port of the Proto Board if this does not work.
LOL... I can't do that, because of the header and future cable for the user interface.... DUH
EDIT: Instead of downloading the Eagle files below, please refer to Post #141 for the latest Eagle daughter board files.
Instead of guessing why not very simply look up the correct trace width for the current requirement. For higher currents it is best to place large pads, do not make the connections in eagle, but rather draw a bunch of rectangles and lines on the board view to define the trace. If the rectangles cover part of the pads, that is all that is needed to make a copper connection. Place the mosfets on the board very close to the screw terminals for the shortest trace length. Route signals normally in eagle that are control signals like gate, gate pulldowns, gate input resistors. The photo you have isn't going to work for any real load. If needed you can manually route high current traces on top and bottom to get more copper. Other options include not using a trace, but rather use some large hole pads and jumper heavy gauge wire around the board as in 16g, 18g etc.
The picture in the previous to last post does not include copper pours, however the Eagle files in the last are basically all copper pours. If I had to guess from my research, I would say that board in the Eagle files would handle 5-7A, if not a little more.
Edit: 5-7A on one mosfet
2A on another
and about .5A on another
Other options include not using a trace, but rather use some large hole pads and jumper heavy gauge wire around the board as in 16g, 18g etc.
Yea, I really didn't like that option, but it was better than any of my other options. With a little rearrangement of the components and one jumper wire, I should now have somewhere in the vicinity of 15 amps available for the build platform heater. In addition to gaining amperage, I also get to keep the daughter board concept, which should make for a nice, tidy, and compact Proto Board, that will be ready for some serious work.
For those that may be interested, I have attached new Eagle files for the daughter board. There are three copper pours, two on the bottom layer and one on the top layer. All the copper pours must be utilized, otherwise you will have to widen several traces to gain amperage.
EDIT: Please note that the attached files are for 2 oz. copper boards.
EDIT: I should have done this before uploading the file and I apologize. I just performed a DRC check, which showed 8 errors. To resolve all the errors, set the grid at 0.0125" and lower JP1 and JP2 one grid space (0.0125"). Adjust wires as necessary, no major headache, and then run DRC and it should report 0 errors.
Okay folks, not attempting to beat this subject into the ground, but here is the final draft of the daughter board, unless someone reports an error. Key points and changes are as follows:
Eagle 6.6.0 DRC reports 0 errors.
Board dimensions: 1.2875" Wide X 1.2" Deep.
0.9875" Distance between the centers of the mounting holes.
0.625" Distance from front of board to the center of the mounting holes.
The outer diameter of the mounting pads has been increased to 0.2"
All traces have been widened to a minimum of 0.016".
Isolation has been widened to 0.016".
The two extra pads for the jumper wire have been removed, because the jumper wire is not necessary.
SUMMARY
This daughter board is intended to mount to the SVGA port mounting holes of the Propeller Proto Board, with the main terminal edge of the daughter board being flush with the edge of the Proto Board. It is designed to contain three MOSFETs(IRF3708), with each MOSFET controlling a different amount of current, and with each MOSFET being controlled by a Propeller IO pin. In addition to the three Propeller IO pins, a ground must also be supplied to the daughter board, and a four pin header has been added to the daughter board to allow connection between the daughter board and the Proto Board. Based upon trace widths and copper pours, this board should be able to supply the following currents, according to the component name and based upon the copper thickness of the board utilized:
1 Ounce Copper Clad PCB
Q1 - Maximum current unknown, guessing 12.6A (Based upon copper pour)
Q2 - 2.45A Maximum current (Based upon trace width)
Q3 - 1.5A Maximum current (Based upon trace width)
2 Ounce Copper Clad PCB
Q1 - Maximum current unknown, guessing 20A (Based upon copper pour)
Q2 - 3.9A Maximum current (Based upon trace width)
Q3 - 3A Maximum current (Based upon trace width)
EDIT: IMPORTANT NOTE - Considering that the positive supply voltage is wired in series from Q1 to Q2 to Q3, maximum current levels should be avoided, if all of them will ever be powered simultaneously.
*** IMPORTANT
Use only non-conductive stand-offs for mounting this daughter board to the Proto Board.
These files were generated with EAGLE 6.6.0 software.
I must admit that I thought it would be impossible to create a single sided daughter board without any jumpers, but I did it.
I decided to try and design a single sided board, because I only needed (1) small board and for the options I wanted for a two sided board, it was going to cost me $60+, plus I would have to wait. With a single sided design, I can at least try to make my own at a fraction of the cost, plus I have all the materials and chemicals.
PLEASE NOTE THAT THE SCHEMATIC HAS CHANGED FROM THE TWO SIDED BOARD IN THE PREVIOUS POST!!!!
Eagle 6.6.0 DRC reports 27 errors (These errors are due to traces being to close to the board edge and various isolation errors).
Board dimensions: 1.2875" Wide X 1.2" Deep.
0.9875" Distance between the centers of the mounting holes.
0.625" Distance from front of board to the center of the mounting holes.
The outer diameter of the mounting pads has been increased to 0.2"
All traces have been widened to a minimum of 0.016".
Isolation has been widened to 0.016".
The two extra pads for the jumper wire have been removed, because the jumper wire is not necessary.
SUMMARY
This daughter board is intended to mount to the SVGA port mounting holes of the Propeller Proto Board, with the main terminal edge of the daughter board being flush with the edge of the Proto Board. It is designed to contain three MOSFETs(IRF3708), with each MOSFET controlling a different amount of current, and with each MOSFET being controlled by a Propeller IO pin. In addition to the three Propeller IO pins, a ground must also be supplied to the daughter board, and a four pin header has been added to the daughter board to allow connection between the daughter board and the Proto Board. Based upon trace widths and copper pours, this board should be able to supply the following currents, according to the component name and based upon the copper thickness of the board utilized:
1 Ounce Copper Clad PCB
Q1 - Maximum current unknown, guessing 12.6A (Based upon copper pour)
Q2 - 2.45A Maximum current (Based upon trace width)
Q3 - 1.5A Maximum current (Based upon trace width)
2 Ounce Copper Clad PCB
Q1 - Maximum current unknown, guessing 20A (Based upon copper pour)
Q2 - 3.9A Maximum current (Based upon trace width)
Q3 - 3A Maximum current (Based upon trace width)
EDIT: IMPORTANT NOTE - Considering that the positive supply voltage is wired in series from Q1 to Q2 to Q3, maximum current levels should be avoided, if all of them will ever be powered simultaneously.
*** IMPORTANT
Use only non-conductive stand-offs for mounting this daughter board to the Proto Board.
These files were generated with EAGLE 6.6.0 software.
It took me the whole day, but I finally got around to making my film positive for the single sided daughter board. I suppose I should mention that since the artwork is crowding the board edge, I made the positive 1/8" larger on all sides, and I will do the same for the board prior to exposure. After etching, I will trim the board down to the intended size.
As mentioned, I will be making this board according to previous experiments and using my unique exposure cylinder, but just so you have a general idea, here is the overall scheme:
PCB Material: Philmore DATAK Premier Series (1/16" 2oz. double sided FR-4 board with positive acting photo-resist film)
Exposure Size: 1.54" X 1.45"
Final Board Size: 1.29" X 1.20"
Positive Medium: Premium Matte Laser Film (0.004") Top Coated With Acrylic Krylon Crystal Clear Satin (#51313)
Exposure Lamps: (1) New 60W/120V Sylvania A15 Ceiling Fan Light Bulb (60A15/CL/2PK/BL)
Exposure Time: 1 minute and 30 seconds
Developer Mixture: 4 grams of Rooto Crystals of Household 100% Lye Drain Opener to 1 pint of tap water
(available at ace hardware)(caustic soda/sodium hydroxide)
Developer Temperature: 106 degrees fahrenheit
This scheme should provide a fairly nice board, however it is the manual drilling that worries me, but on the bright side, it is only 35 holes
The next step is to accurately cut my PCB blank. I wonder how long that will take me Needless to say that I am not very enthused, but that will all change if this board turns out nice, then I will be all smiles
Well that process took a lot longer than expected. It has been such a very long time since I messed with my PCB equipment, and the last time I used everything, I did not take the time to clean everything up properly. That will teach me to clean up after myself
Anyhow, everything is now ready to go:
A positive film has been made
All the equipment has been properly cleaned
A new batch of developer has been mixed
A new batch of etchant has been mixed (First time I am trying this stuff)
Three PCB blanks have been cut to the exposure size (Hopefully I will get it right the first time)
Okay, I guess that is about it. I should be ready to go.
Let me state for the record that I am trying two new things this time around:
I am trying a new etchant that I have never used before.
In the past, I tried to get the board into the etchant as quickly as possible after developing, while the photo-resist was still warm and pliable. This time I will let the photo-resist cool a few minutes and harden, before putting it into the etchant.
Will it be shame or glory? I will let you know in about 20 minutes
The board has been etched and it turned out very nice, except that I accidentally scratched the resist, of one of the signal traces, with my plastic tweezers, going from Pin #2 on the four pin header to R3. I believe there is still enough copper there to do the job, so I am going forward with this board, instead of shooting for perfection, and doing it over. Besides that, the copper looks beautiful.
As a side note, the shelf life of my Tinnit electroless plating solution has expired, and it is now garbage, so I guess bare copper will have to do.
Let me state for the record that I am trying two new things this time around:
I am trying a new etchant that I have never used before.
In the past, I tried to get the board into the etchant as quickly as possible after developing, while the photo-resist was still warm and pliable. This time I will let the photo-resist cool a few minutes and harden, before putting it into the etchant.
With the exception of the scratch mentioned above, I am very happy with the board, however as far as the two new things I tried, I did not particularily like the etchant, because it scares the bejeezers out me, and as far as letting the board cool, I will not do that again, because it took forever to develop. And what I said above was not actually what I meant. I meant to say that immediately after exposure, I like to start developing, while the board is warm.
So now I am off to hand drill my holes and cut the board down to it's final size, and providing that turns out alright, I will solder it up.
IdBruce,
Did you use the presensitized board "Philmore DATAK Premier Series (1/16" 2oz. double sided FR-4 board with positive acting photo-resist film)
that is over 2 years old?
I use iron on transfer method on some of my simple circuits, but always wanted to try photosensitive board material.
Shelf Life of 6 months would not be very cost effective for my small minor projects.
Yes, I used the old board material and it came out beautiful, except for the scratch I mentioned. But just so you know, I took good care of those boards, by keeping them in a cool and secure location, and loosely packed. You certainly don't want to be putting scratches or heavy weight on the boards, or store them in a high heat area.
Some guys get pretty decent results with iron on transfer, but I never had very good results with it. Now that I have my system figured out, the positive acting photo-resist film works well for me. You know you have it just right when you pull it out of the developer and all you see is highly defined traces, pads, and holes.
I like the Philmore-DATAK boards, because it doesn't require UV lighting and the board is exposed in a minute and half. Everyone has their preferences though and some folks swear by the UV lighting. As for me, I built a UV exposure box years ago, but now I have a much more simple and inexpensive setup, and I wouldn't even think about firing up the old one, even though it is pretty nice also. However with the old one, I could only do one side at a time. With the new setup, I can do double sided exposures, no sweat, as long as I keep everything clean
If you want to try positive acting photo-resist film, I would highly suggest Philmore-DATAK boards, and follow the clues I left above.
The daughter board has been soldered up, attached to the Proto Board, and wired to the Proto Board. Additionally, I put heatsinks on both Q1 and Q2. I doubt that Q2 needed a heatsink, but I had a small clip-on for a TO-220 package, so I slipped it on anyhow. I put a much larger heatsink on Q1, however, I am not sure whether it will disapate enough heat, because at this point, I am really unsure of the amperage demand for the heated build platform and I am having a little trouble deciphering datasheet mumbo jumbo. With the heatsinks added, possible continuity between the tabs, and I am assuming also the drain of the MOSFETs, are dangerously close to one another. I will have to monitor that situation closely, just in case the tab is connected to the drain, because the datasheet is not very clear on the TO-220 package.
I am now done with the Proto Board and daughter board, so I can concentrate on other issues. When I get a photo, I will post it here, so you can see what the end result looks like. One thing is certain, it no longer looks like the image below.
Comments
What can be done to make this layout more versatile? There is a wide variety of stepper drivers in the marketplace and through some more design and circuitry, I am almost certain that this controller could be improved to be compatible with a larger portion of them. In order to make it more compatible with a larger portion of stepper drivers, this is what I propose:
I have been reading this thread, and others, relating to the use of the Propeller for 3D Printing electronics. I too am interested in doing this although my primary difficulty at the moment is whether I can justify using the Propeller based on the following Pros and Cons:
Pros:
The mutiprocessing paradigm would benefit 3D printer control for things like synchronized moves and extrusion.
The ability to program in a choice of languages e.g. C, Spin and PASM offers flexibility (although I am not convinced of the need to use C++).
Being able to multi-thread within a cog could also be useful for orthogonal processing e.g. monitoring similar inputs.
The performance to cost ratio of the Propeller is a big incentive.
Cons:
Basic I/O is limited so the use of e.g. SPI peripherals would be mandated.
I envisage that on-chip code space would be insufficient requiring the ability to place code in external memory and execute it efficiently.
The lack of on-chip peripherals makes alternative devices look more appealing. The Propeller must be chosen on the basis of additional benefit after including the cost of required external hardware.
Use of the Propeller requires a different programming mindset which has to be considered for software development and ongoing maintenance.
I am approaching the implementation of 3D Printer electronics differently to established practice. My paradigm:
1) The controller only processes primitives, it has no understanding of GCode. A separate PC Host application can import GCode but, long-term, it is envisaged that a custom slicer will be used to generate primitives.
2) The controller is responsible for maintaining safety of the 3D Printer. Methods will be provided to monitor filament presence, stepper motor fault conditions, movement and temperature safety limits.
3) Printing over USB and from SD card will be supported although use of a FAT filesystem on the SD card is not envisaged (for reduced overhead and licensing issue avoidance).
4) In a departure from established practice, I aim for printer independence in the primitive data transfer implementation i.e. a set of build commands should be independent of material in use and absolute speed specification (relative speed specification is permitted).
5) Operational profiles should be retained in the 3D Printer electronics. The PC Host application may optimize printing based on knowledge of a connected printer's configuration but this should not be mandatory.
6) Sufficient printer hardware options should be supported e.g. 5 extruders (for separate single or 4-colour printing or support with up to 4-colour printing). PID temperature control of extruder(s) and heated build platform.
7) A single-board design without the need for jumpers or current setting potentiometers. The intention here is for an install-and-forget solution with operational configuration and tuning provided by firmware.
I have investigated implementing this using a twin Microchip PIC solution. This relies significantly on the integrated peripherals of the two PICs (one handles USB/SD card and temperature control - the other movement and user interface). Implementing a similar scheme using the Propeller could reduce the number of microcontrollers to one but requires the addition of I/O expanders, ADC and DAC converters (SPI for adequate performance).
My biggest concern in adopting the Propeller is that of programming for a large-ish application. The twin-PIC solution would afford me 128kB of distributed code space and I assume that a Propeller implementation would need a similar amount of code space. I understand that this can be achieved in Propeller-based designs but I am not sure how expanded code space translates into usability on the cogs. Am I right in thinking that code has to be paged into the cog under the control of some driver/kernel and this has to be done in less-than-cog-ram-size chunks? If so, I could envisage some overhead issues will need to be contended with.
So, I believe 3D Printer electronics can be implemented with a Propeller (although I would not attempt an all-in-one solution like Marlin or similar firmware). In fact, I think the distributed paradigm of firmware and smarter PC Host application would fit a Propeller-based design better.
Regards,
Neil Darlow
I hate to say it, but if you are thinking of a four extruders, you are going to need a lot of IO pins just for the stepper drivers alone. The cost of the extra peripherals are almost negligible, just breaking the peripheral components down to the chips and passive components, and in my design, these peripherals only go to items that do not require a large amount of speed, so it is okay with me. However, I would not even think about a four extruder design, without adding another Propeller chip, which at that time, you may as well eliminate the expansion chip. Looking back now, if I had gone with a dual Propeller design, and especially if I wanted more than one extruder, I would have had one Propeller chip just to run the stepper motors with the IO pins, cogs, and stepper driver software, but at that point, there is the requirement of serial communication between the two chips.
That being said, I have written SPIN drivers for my steppers in the past, and the speed is definitely there if you use counters effectively. SPIN is quite easy to learn and very effective for programming the Propeller. I do not believe there is a need for C, however in the long run, I may be required to use PASM, but I do not think so. Time will tell.
The question is whether or not the software will fit.
Actually, it would permit 5 extruders and, yes, it is a lot of I/O. Multiple Microchip MCP23S17s would each give 16 I/O per chip over a SPI interface.
I am not so sure about that. The Propellor could be programmed to accommodate multiple extruders over one cog. You could compute the requirement for steps and step the motors in parallel as needed. More motors does not necessarily require more cogs.
You could implement SPI communications between Propellers but I am not sure that two would be required. My current printer uses an ATmega1284p clocked at 16MHz giving, I believe, 16MIPS performance. A propeller clocked at 5MHz can achieve 160MIPS across all cogs.
I would use PropGCC to implement the bulk of the firmware. Computing movement paths is intensive and I doubt that Spin could provide the required performance. I aim to use TI DRV8825 stepper drivers which are capable of 32x microstepping. I believe that the Propeller is one of the few low cost processors (other than an ARM device) that could achieve a 32x step rate.
In 32kB I am sure it would not. External code expansion would certainly be required and I have yet to determine how, and how efficiently, this is supported in the available languages.
Regards,
Neil Darlow
Have you looked at the Propeller Project Board? I find I can lay out circuits in a cleaner fashion than on the Proto Board. And I also like the SMT areas for those type of chips, pads on the back for an addition regulator, and pads for a uSD connector on the back also.
http://www.parallax.com/product/32810
Jim
I looked at it earlier when considering the options and opted for the proto board, because I am very familiar with the proto board and I thought it had more prototyping space. However, looking back now, that might have been a better choice because of the 16VDC VIN, but I already have the proto boards, so I am kind of locked in.
I am glad you found this interesting enough to take a serious look at it
I would be interested in seeing an illustration or a pic, if you have one handy.
Bruce
Of prior installations? The company that I was subbing for monitors this forum, and they will sue me if I publish any of my designs, which they now say they own,(not that they are earth shattering).
Pretty much 80% came out of the OBEX.
I haven't' done a personal project on the Project Board yet, but Ken just sent me on to put on the Prototype QuadRover that I got last year.
Jim
I just ordered the remaining missing pieces to get this board underway. I don't know how soon I will get all the parts, but I should soon have a prototype built and ready for various tests.
It appears that I will have to make my own daughter board, instead of using perf board, because according to my calculations, the hot end heater and build platform heater are both going to require quite a bit of current
To be perfectly honest, I am really looking forward to getting this board built and experimenting with it, because I have been wanting a similar board for a very long time.
Additionally, if you intend to attempt this project on a Propeller Proto Board, you should have pretty decent soldering skills, a good soldering iron, an eye for detail, a solder sucker, a multi-meter, an aligator clamp, and plenty of light.
The top of the board looks very nice, but the bottom is quite a different story. When I get finished with it, I will post a photo. Actually as the 3D printer project goes forth, there may be several photos throughout various threads.
My remaining parts should be here tomorrow afternoon
It does not appear that I will finish the board this evening, because I had to take a break for a while, and I will probably get back to it, in about an hour or so. There is however a major setback. When ordering my parts, I forgot to order the 10 pin ninety degree header for plugging in the Adafruit ADC breakout board in a vertical position. What a bummer. Oh well, I will do what I can.
Over the next several days, I will be creating the following temporary testing cables to allow attachment of all the peripherals:
In the provided illustration, you will find that two revisions have been made to the Proto Board layout, which should be apparent from viewing the illustration. However by referring to one of the revisions, the three pin headers going to the ADC channels allow me to finish the Proto Board with out additional circuitry and since I will be having remote boards at the heated build platform and the extruder, this should work out well.
Out of the four channels provide by the Adafruit ADC breakout board, I will only be using two of them, one for the hot end thermistor and the other for the heated build platform thermistor. The three pin headers going from the ADC will provide VDD and VCC going to remote circuit board locations, and the remaining pin will bring back the reading to the ADC. In the illustration below, you will also notice a small schematic. This schematic represents the circuitry that I will have on the remote circuit boards for reading the thermistors. R1's value will be dependant upon the thermistor utilized, however, although not shown in the schematic, R1 will be attached to the remote circuit board with a pressure terminal, to allow for quick changes of R1. In addition to allowing quick changes for experimentation purposes, the pressure terminals will allow for quick changes of R1, when changing filament types from PLA to ABS, because different readings are required.
As mentioned earlier in this thread, when utilized as a controller for a 3D printer, a daughter board can be attached to the SVGA port area of the Proto Board. When I drew up the previous illustrations, I did not take all things into consideration, therefore the depth of the daughter board must be increased to allow for placement of all the necessary components. And it is noteworthy that at this point I am still making assumptions, so please be cautious.
Before I begin this discussion, let me first set down some facts:
- The illustrations are only meant to be a pretty good guide in wiring this board up. As mentioned earlier, I had to make a few alterations, to get things to fit properly, especially since my board required an additional (4) transistors and (4) resistors for the DISABLE circuitry required by my stepper drives.
- The extruder fan, the hot end heater, and the build platform heater, will all be powered by 12VDC, each being controlled by a Propeller IO pin via IRF3708 MOSFET. Although currently being discussed, it is my intention to make a seperate board for these items and attach it to the SVGA port area of the Proto Board. A four wire cable will go from the Proto Board to this seperate board, supplying three signals and a ground.
- In previous illustrations, the daughter board seemed to be of sufficient size, but in fact, it was only a mere 7/8" X 1-5/16", which is not very big at all, especially when (3) TO-220AB packages (IRF3708) must be installed.
- Another consideration is that the hot end heater and the build platform heater are high amperage items. It is my intention to allow 2A for the hot end heater and 20A for the build platform heater, with the additional extruder fan only requiring 0.25A. The 20A for the build platform heater is just a guess at this point and not actually a fact. However, the parts shown on previous illustrations of the daughter board are believed to be insufficient due to the high amperage demands.
- Additionally, it is worth noting that during the construction of my controller board, it became necessary to move the female header for the 10 pin ADC breakout board 0.20" inches in the direction of the SVGA port area.
To sum it up and as mentioned, the previous suggested size of the daughter board was insufficient, so I am now suggesting a new size, based upon the available room after Proto Board construction. Additionally, different types of terminals will now be utilized to accommodate the higher amperages required, although I may still be dreaming. At this point nothing is carved into stone, except for the room available to accommodate the daughter board, without crowding the Proto Board, which will now be 1-3/16" X 1-5/16". Please refer to new illustration below for a size perspective.EDIT: I suppose I could try a board fill and see what it looks like.
EDIT: I think I may work with the layout a little more before abandoning the single daughter board concept.
EDIT: It is starting to look like it might just work out.
I might also try elongating the board to go over the programming port of the Proto Board if this does not work.
Anyhow, here are the files. You will have to use the rats nest tool to get the copper pours.
EDIT:
LOL... I can't do that, because of the header and future cable for the user interface.... DUH
EDIT: Instead of downloading the Eagle files below, please refer to Post #141 for the latest Eagle daughter board files.
The picture in the previous to last post does not include copper pours, however the Eagle files in the last are basically all copper pours. If I had to guess from my research, I would say that board in the Eagle files would handle 5-7A, if not a little more.
Edit: 5-7A on one mosfet
2A on another
and about .5A on another
Edit: On 2 Oz. copper boards that is...
Yea, I really didn't like that option, but it was better than any of my other options. With a little rearrangement of the components and one jumper wire, I should now have somewhere in the vicinity of 15 amps available for the build platform heater. In addition to gaining amperage, I also get to keep the daughter board concept, which should make for a nice, tidy, and compact Proto Board, that will be ready for some serious work.
For those that may be interested, I have attached new Eagle files for the daughter board. There are three copper pours, two on the bottom layer and one on the top layer. All the copper pours must be utilized, otherwise you will have to widen several traces to gain amperage.
EDIT: Please note that the attached files are for 2 oz. copper boards.
EDIT: I should have done this before uploading the file and I apologize. I just performed a DRC check, which showed 8 errors. To resolve all the errors, set the grid at 0.0125" and lower JP1 and JP2 one grid space (0.0125"). Adjust wires as necessary, no major headache, and then run DRC and it should report 0 errors.
PDF files have been uploaded to my previous post. Some additional notes have been added to that post as well.
1 Ounce Copper Clad PCB
2 Ounce Copper Clad PCB
EDIT: IMPORTANT NOTE - Considering that the positive supply voltage is wired in series from Q1 to Q2 to Q3, maximum current levels should be avoided, if all of them will ever be powered simultaneously.
*** IMPORTANT
These files were generated with EAGLE 6.6.0 software.
I decided to try and design a single sided board, because I only needed (1) small board and for the options I wanted for a two sided board, it was going to cost me $60+, plus I would have to wait. With a single sided design, I can at least try to make my own at a fraction of the cost, plus I have all the materials and chemicals.
I will attempt to make this single sided board, according to the experiments of this thread: http://forums.parallax.com/showthread.php/135950-ATTN-PCBers-SUBJECT-New-website-and-exposure-box-design-for-your-creative-needs
PLEASE NOTE THAT THE SCHEMATIC HAS CHANGED FROM THE TWO SIDED BOARD IN THE PREVIOUS POST!!!!
1 Ounce Copper Clad PCB
2 Ounce Copper Clad PCB
EDIT: IMPORTANT NOTE - Considering that the positive supply voltage is wired in series from Q1 to Q2 to Q3, maximum current levels should be avoided, if all of them will ever be powered simultaneously.
*** IMPORTANT
These files were generated with EAGLE 6.6.0 software.
As mentioned, I will be making this board according to previous experiments and using my unique exposure cylinder, but just so you have a general idea, here is the overall scheme:
Exposure Size: 1.54" X 1.45"
Final Board Size: 1.29" X 1.20"
Positive Medium: Premium Matte Laser Film (0.004") Top Coated With Acrylic Krylon Crystal Clear Satin (#51313)
Exposure Lamps: (1) New 60W/120V Sylvania A15 Ceiling Fan Light Bulb (60A15/CL/2PK/BL)
Exposure Time: 1 minute and 30 seconds
Developer Mixture: 4 grams of Rooto Crystals of Household 100% Lye Drain Opener to 1 pint of tap water
This scheme should provide a fairly nice board, however it is the manual drilling that worries me, but on the bright side, it is only 35 holes
I have attached several photos of my prototyping table, which is where I will expose, develop, and etch the boards. It is actually a pretty sweet setup and works quite well for small projects such as this one. To learn more about this table, visit this thread: http://forums.parallax.com/showthread.php/139308-Construction-Of-A-PCB-Prototyping-Table
The next step is to accurately cut my PCB blank. I wonder how long that will take me
Anyhow, everything is now ready to go:
- A positive film has been made
- All the equipment has been properly cleaned
- A new batch of developer has been mixed
- A new batch of etchant has been mixed (First time I am trying this stuff)
- Three PCB blanks have been cut to the exposure size (Hopefully I will get it right the first time)
Okay, I guess that is about it. I should be ready to go.Let me state for the record that I am trying two new things this time around:
- I am trying a new etchant that I have never used before.
- In the past, I tried to get the board into the etchant as quickly as possible after developing, while the photo-resist was still warm and pliable. This time I will let the photo-resist cool a few minutes and harden, before putting it into the etchant.
Will it be shame or glory?The board has been etched and it turned out very nice, except that I accidentally scratched the resist, of one of the signal traces, with my plastic tweezers, going from Pin #2 on the four pin header to R3. I believe there is still enough copper there to do the job, so I am going forward with this board, instead of shooting for perfection, and doing it over. Besides that, the copper looks beautiful.
As a side note, the shelf life of my Tinnit electroless plating solution has expired, and it is now garbage, so I guess bare copper will have to do.
With the exception of the scratch mentioned above, I am very happy with the board, however as far as the two new things I tried, I did not particularily like the etchant, because it scares the bejeezers out me, and as far as letting the board cool, I will not do that again, because it took forever to develop. And what I said above was not actually what I meant. I meant to say that immediately after exposure, I like to start developing, while the board is warm.
So now I am off to hand drill my holes and cut the board down to it's final size, and providing that turns out alright, I will solder it up.
Did you use the presensitized board "Philmore DATAK Premier Series (1/16" 2oz. double sided FR-4 board with positive acting photo-resist film)
that is over 2 years old?
I use iron on transfer method on some of my simple circuits, but always wanted to try photosensitive board material.
Shelf Life of 6 months would not be very cost effective for my small minor projects.
Thanks
Garyg
Yes, I used the old board material and it came out beautiful, except for the scratch I mentioned. But just so you know, I took good care of those boards, by keeping them in a cool and secure location, and loosely packed. You certainly don't want to be putting scratches or heavy weight on the boards, or store them in a high heat area.
Some guys get pretty decent results with iron on transfer, but I never had very good results with it. Now that I have my system figured out, the positive acting photo-resist film works well for me. You know you have it just right when you pull it out of the developer and all you see is highly defined traces, pads, and holes.
I like the Philmore-DATAK boards, because it doesn't require UV lighting and the board is exposed in a minute and half. Everyone has their preferences though and some folks swear by the UV lighting. As for me, I built a UV exposure box years ago, but now I have a much more simple and inexpensive setup, and I wouldn't even think about firing up the old one, even though it is pretty nice also. However with the old one, I could only do one side at a time. With the new setup, I can do double sided exposures, no sweat, as long as I keep everything clean
If you want to try positive acting photo-resist film, I would highly suggest Philmore-DATAK boards, and follow the clues I left above.
The daughter board has been soldered up, attached to the Proto Board, and wired to the Proto Board. Additionally, I put heatsinks on both Q1 and Q2. I doubt that Q2 needed a heatsink, but I had a small clip-on for a TO-220 package, so I slipped it on anyhow. I put a much larger heatsink on Q1, however, I am not sure whether it will disapate enough heat, because at this point, I am really unsure of the amperage demand for the heated build platform and I am having a little trouble deciphering datasheet mumbo jumbo. With the heatsinks added, possible continuity between the tabs, and I am assuming also the drain of the MOSFETs, are dangerously close to one another. I will have to monitor that situation closely, just in case the tab is connected to the drain, because the datasheet is not very clear on the TO-220 package.
I am now done with the Proto Board and daughter board, so I can concentrate on other issues. When I get a photo, I will post it here, so you can see what the end result looks like. One thing is certain, it no longer looks like the image below.