A
Abstract Dissonance
- Jan 1, 1970
- 0
I went ahead and came up with a solution to my problem of trying to switch
components in all possible series and parallel configurations... a graphic
of the basic idea is at
http://www.geocities.com/abstract_dissonance/circuits2.jpg
Basicaly I would have a matrix of switches.. each switch though contains 7
possible ways to connect the 4 wires at a junction(excluding rotational
symmetry). In the graphic you see the 7 paths and several sample circuits
that can be generated by using certain paths. I believe I can generate all
series/parallel combinations with atmost (2n)^2 switches(but may be much
less. For 2 switches it requires at most 8 junctions to get all 12
series/parallel combinations.
My idea was simple to create a switch with all these paths in it and I could
electronically select the path by some means to to configure the array to
get any combination I desired.
The 8th circle is all the "paths" stacked together but each one rotated
slightly and there would be a non-conductive material between each
"path"(the grey stuff) which, ofcourse, would be conducting. I figured that
I could make rotary switch out of this and then rotate the switch with some
small motor that is controlled ultimately by a MCU so I can select the
different combinations.
Another method for the switch would be to stack them all directly on
top(without the rotation) and be able to select the paths by pushing in and
out of the shaft... it would require all the "rotational" symmetry to be put
back in though... 15 or so paths instead of 7.
I initally was going to do this with relays but it would end up, from what I
have figured out, to be quite expensive and very bulky. Another thing I
looked into was a crosspoint switch but I have not been able to determine if
they can route to all the possible wire combinations and they tend to be
quite expensive too.
My requirements is mainly that I can switch from one combination to another
in < 1 sec at most but I'd be happy with about 0.5s. The switching voltages
and currents are quite low too but I don't see this as much of a problem
unless I could miniaturize the switches(which won't happen for a long time
anyways).
My main questions are: Is there another method to do this? I mainly need
relay like action(not solid state stuff but that it could be an option if it
is cheap).
The second problem I'm having is trying to actually rotate the switch to
select the appropriate path. I want this things to be cheap and I'm not
sure how I can switch the paths easily with a very simple mechanism. I
guess I'll have to dive into how stepper motors work and see if I can find a
way to do it effectively and cheaply but any ideas or comments would be
appreciated on this.
I have several ideas about how to actually create the "wafers"(the circular
disks that each contain a path and will be stacked like plates to form the
switch) but maybe someone can offer an efficient method. My inital idea was
simply to cut out the paths by hand with some thin copper and place them on
some cardboard(probably not) or plastic "punches" about 1in in diameter for
and put a plastic shaft down the center for rotation... this would atleast
give me a prototype to work with. Another method I thought would be to just
get some conductive paint and use that to paint on the paths. I could make
templates for each path and probably even get them to be pretty small(maybe
even a 1/4in in diameter which would be good enough for me).
For my project it would require about 144 of these switches and so if I used
some transistors or relays then the cost would be several hundred
dollars(from my initial look into it). It would be nice if there was a much
better way to do this but I can't seem to find anything that is appropriate
and this is the best thing that I can think of. (My main drive is that I
cannot alter the original signal to any appreciable degree(which means that
solid state stuff is most likely going to be out of the question... not
because it might only affect the single by 1 part in 1M but because of other
reasons).
Anyways, Any comments would be appreciated.
Thanks,
AD
components in all possible series and parallel configurations... a graphic
of the basic idea is at
http://www.geocities.com/abstract_dissonance/circuits2.jpg
Basicaly I would have a matrix of switches.. each switch though contains 7
possible ways to connect the 4 wires at a junction(excluding rotational
symmetry). In the graphic you see the 7 paths and several sample circuits
that can be generated by using certain paths. I believe I can generate all
series/parallel combinations with atmost (2n)^2 switches(but may be much
less. For 2 switches it requires at most 8 junctions to get all 12
series/parallel combinations.
My idea was simple to create a switch with all these paths in it and I could
electronically select the path by some means to to configure the array to
get any combination I desired.
The 8th circle is all the "paths" stacked together but each one rotated
slightly and there would be a non-conductive material between each
"path"(the grey stuff) which, ofcourse, would be conducting. I figured that
I could make rotary switch out of this and then rotate the switch with some
small motor that is controlled ultimately by a MCU so I can select the
different combinations.
Another method for the switch would be to stack them all directly on
top(without the rotation) and be able to select the paths by pushing in and
out of the shaft... it would require all the "rotational" symmetry to be put
back in though... 15 or so paths instead of 7.
I initally was going to do this with relays but it would end up, from what I
have figured out, to be quite expensive and very bulky. Another thing I
looked into was a crosspoint switch but I have not been able to determine if
they can route to all the possible wire combinations and they tend to be
quite expensive too.
My requirements is mainly that I can switch from one combination to another
in < 1 sec at most but I'd be happy with about 0.5s. The switching voltages
and currents are quite low too but I don't see this as much of a problem
unless I could miniaturize the switches(which won't happen for a long time
anyways).
My main questions are: Is there another method to do this? I mainly need
relay like action(not solid state stuff but that it could be an option if it
is cheap).
The second problem I'm having is trying to actually rotate the switch to
select the appropriate path. I want this things to be cheap and I'm not
sure how I can switch the paths easily with a very simple mechanism. I
guess I'll have to dive into how stepper motors work and see if I can find a
way to do it effectively and cheaply but any ideas or comments would be
appreciated on this.
I have several ideas about how to actually create the "wafers"(the circular
disks that each contain a path and will be stacked like plates to form the
switch) but maybe someone can offer an efficient method. My inital idea was
simply to cut out the paths by hand with some thin copper and place them on
some cardboard(probably not) or plastic "punches" about 1in in diameter for
and put a plastic shaft down the center for rotation... this would atleast
give me a prototype to work with. Another method I thought would be to just
get some conductive paint and use that to paint on the paths. I could make
templates for each path and probably even get them to be pretty small(maybe
even a 1/4in in diameter which would be good enough for me).
For my project it would require about 144 of these switches and so if I used
some transistors or relays then the cost would be several hundred
dollars(from my initial look into it). It would be nice if there was a much
better way to do this but I can't seem to find anything that is appropriate
and this is the best thing that I can think of. (My main drive is that I
cannot alter the original signal to any appreciable degree(which means that
solid state stuff is most likely going to be out of the question... not
because it might only affect the single by 1 part in 1M but because of other
reasons).
Anyways, Any comments would be appreciated.
Thanks,
AD