HI, I'm new to Arduino, trying to build a project and control 1000 independent relay.
I could't find any solution so far, I'd appreciate someone to share an experience for this goal
thanks!
You will most likely run into problems with power consumption with this many relays. And you may have trouble simply with the physical construction of the hardware (number of wires, weight, bulk).
Douglas_Jr:
HI, I'm new to Arduino, trying to build a project and control 1000 independent relay.
I could't find any solution so far, I'd appreciate someone to share an experience for this goal
thanks!
Well that is no different an application then any project needing lots and lots of external output pins. So strings of external shift registers whos outputs drive transistors which drive relays. It's not so much complex as it is just shear compounded repeating of components.
You might want to step back and explain to us what the relays will be controlling, what the final load voltage (ac or dc) and current draw is. When dealing with 1,000 duplicated output circuits, anything that saves a component or two per circuit pays hugh dividends. Power requirements alone when you want to control 1,000 anything can be very expensive and bulky.
Lefty
KE7GKP:
There are many ways of doing that. If you care to reveal the details of your project we may be able to offer some suggestions that would suit your project.
Ok 
while my project may seem a bit funny for someone, I can confirm that I'm pretty serious and I'm going to create a custom size vending machine with hundreds of cells. The idea is to open a small barrier door in front of the cell so product may fall down to conveyor.
I'm ready with software coding as well as any mechanical construction works, this doesn't worry me.
So I think i may need a really small electric lock or strike that'll just release the barrier for a second. I'm not sure about which component to choose for electric lock too honestly.
So you don't want a relay but a solenoid. You only need to activate one at a time so that cuts down on your power requirements. But even with this you will need to be a good electronics engineer to get that many shift registers or multiplexers working. For a start look up what decoupling is all about.
You could use a LED driver like the TLC5925 or STP16C596 which will sink up to 100ma at 10s of volts. These act like shift registers but are constant current; that would at least eliminate sticking resistors on each line. There are numerous manufacturers of chips like these which will have the same pinouts. A 16 channel chip is around $1.50.
Modern Device has a kit that implements these to provide 64 outputs (and the ability to string on extra boards) that you can use as an example. At $20 for 64 outputs that's only 31 cents an output.
Thanks guys
I'm still in process of researching
see what i got: http://mondomatrix.com/info/?page_id=325
http://www.mondomatrix.com/images/power_matrix_network_big.jpg
seems to be easy like i want, controller is on backorder though.
any ideas on similar solution maybe?
Before you start looking at controllers the first thing you need to do is to find out what current each solenoid will take. Only then can you make a sensible decision about controllers.
Well, i'll have one solenoid activated at a time.
If you need to drive <=1024 outputs but only one at a time then the best setup is a matrix like arrangement of 2*32 outputs. There is no need for stringing shift registers to 1024 outputs. Especially this reduces the number of required driver transistors to 64 which is far better than 1000 circuits. It also reduces overall circuit size and wiring a lot. Another solution would be to have individually adressable driver circuits per solenoid. This would keep the parts count up but reduce wiring efforts even further. It is a good idea to work out the number beforehand and see if wiring or circuit costs will dominate. And don't forget the maintenance cost. My gut feeling says that stringing shift registers is the most expensive approach.
Thanks!
based on the post replies and difficulties, I'm starting to think about some track actuators that'll push the product from shelf instead of having solenoid opening the barrier It's a lot of money though but If I create an XY track actuator with push to Z Axle, that should do the job i bet.
looking at Linear Track Actuators engineered by Firgelli Automations now
any ideas are highly appreciated!
Well, i'll have one solenoid activated at a time.
Yes you did sat that before, I was asking because you need to know where to put the drive capacity.
Anyway a mechanism sounds so much better. It will take a bit longer to actuate. For ideas look at the reprap project, it is designed for 3D printers but the mechanism is just what you need.