Thoughts on a multiple Stamp project
Bill Chennault
Posts: 1,198
All--
If a person REALLY wanted to save real estate space and needed multiple·BS2s (pick a number, say 25), each performing a·simple* function, couldn't you use surface mount components and eliminate the DB9, the transistors, and the "gang" resistor (what do you call that big long one?), and MAYBE use a regulator appropriately sized to handle all the components involved (except any "external" ones)? Plus, one might somehow eliminate the 20 pin connectors. (But, I don't have any ideas in that regard.)
Each SMD EEPROM would need to be programmed via an EEPROM programmer prior to SMD soldering, I suppose.
*The "simple" function may be to read two position sensors and control·two·tiny motors. Not really a single function, but not time consuming or complex in any event. Of course, such a function would not require 16 GPIO, either. So maybe the connectors could be reduced in size quite a bit.
EDIT: Ok. I see the SMD BS2 is discontinued. So, my real estate savings is rapidly disappearing!·And, looking at Parallax's 4k EEPROM, it appears to be a DIP device with no mention of an SMD device. So, maybe this whole idea is basically not worth the cost of tooling up to implement it.
--Bill
ps An alternative that WOULD work would be to simply buy 25 OEM BS2s and some (what? 24 pin SIP sockets?). Such an arrangement would not be HUGE, it would just be a LOT bigger.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
You are what you write.
Post Edited (Bill Chennault) : 3/12/2007 7:39:14 PM GMT
If a person REALLY wanted to save real estate space and needed multiple·BS2s (pick a number, say 25), each performing a·simple* function, couldn't you use surface mount components and eliminate the DB9, the transistors, and the "gang" resistor (what do you call that big long one?), and MAYBE use a regulator appropriately sized to handle all the components involved (except any "external" ones)? Plus, one might somehow eliminate the 20 pin connectors. (But, I don't have any ideas in that regard.)
Each SMD EEPROM would need to be programmed via an EEPROM programmer prior to SMD soldering, I suppose.
*The "simple" function may be to read two position sensors and control·two·tiny motors. Not really a single function, but not time consuming or complex in any event. Of course, such a function would not require 16 GPIO, either. So maybe the connectors could be reduced in size quite a bit.
EDIT: Ok. I see the SMD BS2 is discontinued. So, my real estate savings is rapidly disappearing!·And, looking at Parallax's 4k EEPROM, it appears to be a DIP device with no mention of an SMD device. So, maybe this whole idea is basically not worth the cost of tooling up to implement it.
--Bill
ps An alternative that WOULD work would be to simply buy 25 OEM BS2s and some (what? 24 pin SIP sockets?). Such an arrangement would not be HUGE, it would just be a LOT bigger.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
You are what you write.
Post Edited (Bill Chennault) : 3/12/2007 7:39:14 PM GMT
Comments
If it were real estate space primarily driving your choices, how about an 8-pin PIC or AVR processor running off their internal clock? You've got 6 I/O pins to work with and there wouldn't be any external components needed except for the regulator and capacitors for the power supply.
After all, why would anyone choose something as complicated and slow as a Stamp when something simple and fast and cheap would do?
The answer lies in looking at your application and choosing the appropriate tool for the job. Parallax has done a nice job in producing a processor with a lot of built-in functionality for complex control purposes (I2C,SPI,1-Wire support, direct LCD and serial support). The language and programming process is simple and straightforward with easy tools to assist with debugging. The basic arithmetic is 16-bit rather than 8-bit and there are arithmetic operators needed for common control functions.
Development costs are high with one-of-a-kind devices and often anything you can do to make that process simpler and easier will pay back in the end. On the other hand, amortizing the development cost of a fairly simple application across even only 25 identical devices may favor something cheap and tiny like a PIC or AVR or TI's new micro-wonder. Maybe you do still want a Stamp involved, but as a controller/coordinator for the array of special purpose controllers.
The SMD BS2 Interpreter is still available…Please see the following link. The longer resistor package is a SIP Resistor Pack. On the OEM they are individual rather than bussed which have a common line. The OEM Modules would require 20-pin SIP Sockets…In you first paragraph you mentioned no DB9 and the support parts so how would you program each individual unit for upgrades to the firmware?
http://www.parallax.com/detail.asp?product_id=PBASIC2CI/SS
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Chris Savage
Parallax Tech Support
·
Mike, as always, thank you for the gentle instruction. I am learning a lot about microcontrollers, but have almost an infinity to go! I have worked my way through the most-excellent WAM and am ALMOST finished with the equally good StampWorks. The problem—at this time--is that I do not know enough yet about microcontrollers to generalize what I do know to other devices.
·
However, I am an experienced code-cutter and that part of the microcontroller experience is beginning to come together for me.
·
I have looked at other microcontrollers and am quite impressed with their abilities, especially the DiosPro. It is quite a machine. (I have two of those neat AX-12 gearmotors? servos? awaiting something fast enough to talk to their firmware. Alas, that something will not be a Stamp. (However, a Propeller would doubtless do the job at 1Mb/s, which is the INITIAL default baud rate! Sounds dumb to me, but there it is.)
·
The reason my thoughts are Stamp-centric is really very simple. (Attention, Chris!) The unparalleled support from Parallax is vital to the layman. I think it must also be important to the pro, as I see a lot of high-level dialogue flowing back and forth between forum users and Parallax employees.
·
My initial questions involving the reduction of real estate space involved a more immediate goal of mine on the way to development of an eight-legged walker ala Edgar Rice Burroughs <g> . I need to develop an articulated spine. My theory is that nature does it best and why either ignore nature or try to best it? I had thought of using a Stamp to control EACH vertebra of the articulated spine. (A cat has 26 to 27 vertebra.) Using spherical bushings, one can emulate an articulated spinal column well if you limit their overall travel.
·
I may begin thinking about a flexible shaft running through the entire length of connected spherical bushings using some kind of offset cam arrangement to position one relative to the next one. (However, although this might work, there would have to be a clutch arrangement to allow selective vertebra to move independently of others. Oh, well. Of course, a “clutch” could be VERY broadly defined! [noparse][[/noparse]Even as I write, this sounds overly complicated.])
·
While we are talking about ME, I used to be rather heavily involved in genetic programming, albeit at a fairly amateurish level. However, genetic programming—programs that learn by themselves—are real and can actually produce real results. For example, I have written programs that produce code that will derive the quadratic equation from SCRATCH, in an amazingly short time. (A few hundred thousand programs, or so (actually less), is all it takes to close in on an acceptable result.)
·
Now, what I would like to do is have a five degree of freedom desktop arm—highly overbuilt—that would allow me to, say, pick up a steel bb and place it in a container. I should not say “me”; I should say a “genetically” derived program that is initially tested for suitability on a PC—a perfectly good platform for this work—with the few hundred that passed individually downloaded to a Stamp for actual operation of the arm. The ones that evaluated better than the cutoff criteria would be passed back to the PC, mutated such that their numbers were increased a thousand-fold and the process would begin over. Since actual physical processes would take place, I imagine it would take at least a few days to “evolve” a program that will pick up a bb and drop it in a container with consistency.
·
A Stamp would be perfect for such a simple arm.
·
Just some thoughts as I fish for ideas on the net.
·
--Bill
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
You are what you write.
I agree the BS2 and Parallax support make a wonderful team. However, it's not the be-all and end-all of microprocessor products. It makes a GREAT onesy-twosy prototype. It makes a very flexible and capable processor for a robot.
However, when you start talking about 25 processor applications, then going with the 'native' PIC has huge cost savings associated with it. Plus, somebody doing a 25 processor solution had better have enough experience with small processors that they can 'go off' on their own with less support than Parallax offers.
www.microchip.com has 'native' PIC resources, as well as links to other PIC compilers.
I agree, completely. That is why I am here, thinking, discussing, and planning instead of in my shop turning and burning. Oh! I almost forgot! And stealing ideas from you guys!!! [noparse]:)[/noparse]
--Bill
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
You are what you write.
If you can isolate an individual node for example, if I lift my hand off the table my conscious master controller says lift your hand. I really don't know how many muscles are involved in that movement but I imagine the individual nodes you talk about would have that sensor input and be able to stabilize its movement from a single command from the master controller. It might have it's own multi servo array it controls rather than having to send all the signals from the master controller to every node.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Think Inside the box first and if that doesn't work..
Re-arrange what's inside the box then...
Think outside the BOX!