PDA

View Full Version : Propeller Chips - Put on Breadboard, PCB or What?



Humanoido
12-09-2009, 09:45 PM
I'm looking for alternate ways to put together large numbers of Propeller chips without using solderless breadboards or etched printed circuit boards. The method should not involve soldering directly to the chip pins.

humanoido

Graham Stabler
12-09-2009, 10:04 PM
Veroboard and IC sockets for the DIP version of the Prop.

StefanL38
12-09-2009, 10:14 PM
schmartboards ?

PropModule 1x1 (http://gadgetgangster.com/find-a-project/56?projectnum=161) ?

You said that you want that EVERY prop can contact to every other prop.
Therefore you need a bus. now it depends on what you want to do.

just do internal calculations ? the bus could be 16bits or 24 bits wide.

Have a lot of IO-pins. 8bit bus.

But anyway what do you want to do ?
Is this a project like somebody wants to build a motorcycle around a V12-engine "BIG BLOCK" because he has fun with building it ?

or do you have a certain and useful target ?

If you just want to have raw calculation-power some kind of a high-end server-PC does this off the shelf
or you use a linux-PC-cluster. (Means buy a lot of old and used PCs that can run a smaller linux-system on them

best regards

Stefan

mctrivia
12-09-2009, 11:12 PM
Never thought about it but 1x1 would work well. You could use long pins or stripped wire as bus. New modules will not have headers so you can easily stack like that.

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
24 bit LCD Breakout Board now in. $21.99 has backlight driver and touch sensitive decoder. (http://forums.parallax.com/showthread.php?p=848975)

kwinn
12-09-2009, 11:49 PM
Use hot melt glue to stack the chips and wire wrap wire soldered to machined pins to connect everything up.

Duffer
12-10-2009, 01:54 AM
It sounds like you're talking about "freeforming" a number of Props together. Check out the following link that describes freeforming to create BEAM robot projects. There are some very artistic freeform circuits. With some care, I guess you could piggy-back several DIP Props in this freeform style. Connect "common legs" for power, ground, I2C, etc. "Splay" the I/O pin legs a bit and solder on female headers. It sure would make an interesting looking "multi-bug" contraption.

http://www.solarbotics.net/library/techniques/freeform.html

Also see: http://www.beam-wiki.org/wiki/Freeforming_BEAM_circuits·There are some good tutorials in the Resources section at the bottom of the above page.

Duffer


Post Edited (Duffer) : 12/9/2009 7:07:36 PM GMT

mikediv
12-10-2009, 03:02 AM
Duffer that's wild when you think about it ,, it would not be that hard to piggy back a few 40 pin props and just bend out the pins you need to separate, I bread board or wire wrap but I could whip together 4 40 pin prop chips and then put the whole stack into a 40 pin socket and have a 4 prop chip?? pretty cool now if I could just learn to program so I could actually use what I build lol

kwinn
12-10-2009, 03:28 AM
I was not aware the technique had a name, but "freeforming" is an apt description. I used it to build a simple data collection/telemetry circuit using cmos logic many years ago. It had to go in a small cylindrical area of a model rocket.

VIRAND
12-10-2009, 05:38 AM
Someone suggested and I agree with using either the serial pins or another pair, either with a two wire bus TX and RX,
or a ring, where the data goes from TX to RX of the next Propeller. The other method may be faster but has the problem
of two Propellers talking at the same time, which is not a problem with the Ring.

Serial controller networks can be implemented so that lets say the first byte or two of a "packet" tells who should listen
(and optionally who talked). 0 or FF could mean "all propellers listen". I think the internet used to be like that, with less than
255 IP addresses.
It is hard to imagine using up more Propellers than that by yourself, but if need be, using more net addresses is not hard to do.

When full speed is necessary, so is a Parallel Bus. Any claim that serial can possibly be faster (e.g. SATA) is unbelievable
without well specified testable evidence such as a very simply explained lossless compression code used with QAM.
Speed is as cheap as pins, and extra slow pins are cheap using shift registers and other I/O expansion techniques.

As for physically connecting them, I'm not going to recommend making a sandwich between two perfboards with
chip sockets on them, and Propeller pins sticking out horizontally in the middle, since I did something like that once
and it is almost impossible to get so many pins in sockets that way.

mctrivia
12-10-2009, 06:38 AM
Serial can be faster. But only when you are pushing the limits.

SATA works by sending bits in each direction 1 at a time. because it uses both the signal and inverse of the signal and is well sheilded the cable is not easily effected by stray emi. Also because there is only 2 wires carying the signal they can easily be precisely trimed to have the exact same propogation delay. This allows extremely high(up to 6GBs) clockrates.

Paralel on the other hand uses only 1 wire per signal and is not as well sheilded so is effected more by emi. also with 80 wires(not all signal but I don't remember exact number) it is much harder to keep propagation delays the same. if you were to send the data at the same rate as sata the bits may not arrive in the correct time slot.

also SATA is faster because it was better designed.
*Only 1 device per cable
*Lower voltage so faster cheaper parts can be used


Yes if you could build a parallel cable as well as the SATA cable then it would be faster but doing so would be difficult and expensive.

SATA is faster because the serial data is sent at a much higher clock rate. if you were to try and send data paralel that fast you would have problems with propogation delay on the wires.



The prop on the other hand is no where near the speed needed to make serial better then parallel.

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
24 bit LCD Breakout Board now in. $21.99 has backlight driver and touch sensitive decoder. (http://forums.parallax.com/showthread.php?p=848975)

Humanoido
12-11-2009, 10:01 AM
What if the chips were not mounted linear but in a three dimensional arrangement that would utilize the inside of a multi dimensional hyper sphere? Or for illustrative purposes, a single sphere. So there could be some creative arrangements here. While the chip could be affixed, it could also be suspended, but I am not sure how to do the suspending.

kwinn
12-11-2009, 10:11 AM
One of the reasons for SATA is that as the speeds get higher it is much more difficult to keep all the bits in a parallel interface aligned time wise. Small differences in transistor characteristics and on chip path lengths cause the bits to arrive at the output pins at slightly different times. This timing jitter puts an upper limit on the speed of a parallel bus.