Efficient Switch Synchronization for Remote Switches

[fredreload]

So I got a switch here that I want to synchronize it to another switch some distance away, when I flip one of the switch on, I want the other switch to turn on as well as fast as possible with as little delay as possible. One thing I can do is to have a laser matching these two switches so when one turns on the other one also turns on, but then I have a lot of these switches so I would have multiple lasers for each pair of switch. Is there another way to do this to synchronize the switches from a distance away? Even if I connect each pair of them with a wire I would need multiple wires as well.

 

[Buzz Bloom]

Hi fred:

I offer the following as a suggestions. Each switch has a clock. The problem then reduces to synchronizing the clocks. I will discuss that after dealing with getting the switches to all turn on together.

From your description I assume that the collection of switches are stationary with respect to each other, so there are no simultaneity issues. I also assume there is only one master switch, and when this is activated, the process for having all switches turn on together starts. (If there are multiple master switches, then a much more complicated process is needed.) I also assume the time required to send a signal between the master switch and the other switches are significantly different.

Let M be the master switch. Let T be the maximum of the time requites for a signal to travel from M to each of the other switches. When the master switch is activated, the switch's internal logic calculates the sum of the current time and T (and perhaps also a small safety increment), and transmits this time value to all the switches. Then all the switches turn on when each of their respective internal clocks agree with the transmitted time.

The easiest way to synchronize all the identical clocks is when they area are all close together. They can each be connected with a wire, all of idenetical lengths, to a common signal source. All the closkes are set to zero, and when they receive a signal via the wire all the clocks start to run. The the individual swithces and clocks can be moved to their respoective destinations, but at a speed suficiently slow that relatavistic effects that effect their "ticking" speeds are negligible.

If I have misunderstood the constrains you intended for this problem, I apologize.

Regards,
Buzz

 

[fredreload]

Hi,
Thanks for your reply. It is a good idea to time each switch with a clock. I do not want the switches to be turned on together at the same time, they each will be switched on and off at a different speed. I think I will change and explain the experiment in a different way. Let's say I have multiple wires each carries a signal. The signal running in the wires all have different speed, the first wire maybe 130m/s, second wire 150m/s and so on. When the signal reaches the end of the wire, I want to carry that signal on to the second wire. My idea is that I need to have a laser at the end of the first wire where the signal travels to, then as soon as the laser detects the signal it would create a signal where it is "linked to" to carry on that signal on the other wire like below. I would need to do this for a great number like for probably 100 billion (First Wires) and 100 billion (Second Wires), so I will probably need 100 billion lasers to connect them all. But let's say I can take a scan of all 100 billion wires at the same time to keep track of all the signals, but then I will still have to do a search to find the wire it is paired to (probably with a map) and create a signal at that wire using laser again. So either way it seems I need 100 billion lasers that would help me automatically create a signal at the other wire, or do a scan on all the wires and manually create a signal on every pairing wire when the signal reaches the end of the wire. But like you mentioned if I can get the timing on all the signals, along with their speed, I can create a signal on the second wire when I need to and synchronize it, but that still mean I need 100 billion clocks all connecting to a master clock?
First Wire Laser Second Wire
-*-------- ------>---- ----------
---------- ------>---- ------*---
---*------ ------>---- ----------
---------- ------>---- *---------

 

[Buzz Bloom]

Hi Fred:

I apologize that I am not getting a picture in my head from your description of the setup.

1. I don't understand the relationship between lasers and wires. Do you mean that you are using optical cables rather than electrical wires?
2. I don't understand what you mean by "as soon as the laser detects the signal." Do you have in mind some kind of laser that performs as a detector? If so can you explain this in more detail? If not, please explain what this means.
3. Do you mean that the connections between the switches are limited to a single path starting from a master switch to a 2nd switch, then a path connecting the 2nd to a 3rd switch, etc., with all the switches connected serially in this manner?
4. The following two quotes confuse me as to what you mean by "synchronize".

So I got a switch here that I want to synchronize it to another switch some distance away

I do not want the switches to be turned on together at the same time, they each will be switched on and off at a different speed.

Regards,
Buzz

 

[fredreload]

Hi,
1 and 2 together, yes I am expecting this laser to be able to detect such a signal in the first wire, sort of like a sensor detecting laser bouncing back, and then create another signal at the second wire again using a laser(using laser to create a signal).
3. It would probably be a parallel connection, but I don't think a master switch can get the timing of something running at 150m/s and 130m/s.
4. Part1. Make that switch a wire, so a signal jumps from the first wire to the second wire when it travels to the end.
Part2. The signal or (current) in the wire will be running in different speed for say 150m/s or 130m/s.

 

[Buzz Bloom]

Hi Fred:

I am still confused. You seems to be asked for advice concerning the design of a system of signalling and switching elements. I do not get a clear picture of the functional specifications for this system. I also don't understand the relevance of lasers, and why you want to use lasers rather than electrical signals over wires.

This is what I suggest you do.
1. Describe a configuration of switches and wires. Simply your description of the problem you are try to solve to a few, say 5 switches. Pick one way to hook together these switches. Also discuss a few alternative ways that would also be OK just to convey the concept of the range of alternatives. If there is no restrictions on how the switches are wired to each other, say that.
2. Describe what behavior you want from the system. This involves (a) defining all possible input patterns to the master switch, and (b) the resulting desired behavior of all the other switches. Do not bother to describe any behavior that is incidental to the desired behavior.
3. Describe any additional constraints. The speed of signals not being the same on different wires is an example.

I hope this is helpful.

Regards,
Buzz

 

[berkeman]

The signal running in the wires all have different speed, the first wire maybe 130m/s, second wire 150m/s and so on.

How do you accomplish that? And how can the signals be traveling so slowly?

When I first started reading your thread, I thought of simple 3-way light switches...

http://www.askmediy.com/wp-content/uploads/3-way.jpg