Home made alarm system regulations

[aptsys]

What Sam wants to do isn't particularly complicated as far as things go, it should be quite possible to make something similae to the ADE biscuit system with a very basic 8-pin microcontroller and a couple of external components. If he has access to the right equipment then he could make the whole thing surface mount and condense it down to no bigger than a one pence piece.

Furthermore, we're assuming the system will have the necessary electronics housed within each sensor. It doesn't have to be done that way. For example, each room could have a concealed panel housing the control and ID electronics, powering all the sensors in that room and interfacing back to the main computer/panel via RS485. The possibilities really are limitless on a bespoke system.

I agree, the possibilities of making an alarm system that's totally customisable make the prospect a lot more entertaining - the only possible difficulty is making keypads that don't look like a maplin project box with some buttons on it. I guess buying a keypad and replacing the PCB with your own is an option but then the cost starts going up.

A small 6 pin SOT-23 PIC10F could be programmed to operate as an addressable node making it easy to fit into existing detectors in a similar way to iD biscuits

 

[bernardgreen]

For low speed control or monitor of remote locations I often use DTMF signalling. Cheap but tidy looking keypads then come ready made as telephones.

 

[onion_soup]

If I DO manage one, anyone instrested in the designs?

Sam.

How long will it take for you to design/make ?

Will you provide technical help/support?

 

[bernardgreen]

A small 6 pin SOT-23 PIC10F could be programmed to operate as an addressable node making it easy to fit into existing detectors in a similar way to iD biscuits

There are times when small is not beautiful. There are advantages in using "old and large" packages for components where reliability of operation is essential.

 

[plugwash]

I tend to think it would work better to build concentrators that to try and mess arround with tiny biscuits in each sensor. In each concentrator I would then place a reasonablly large pic (I would probablly go for the 18f2520, the 18F series are the easiest pics to work with IMO and 10 analog inputs isn't bad for a 28 pin part) a voltage regulator for power and a RS-485 line driver chip

This would then connect to the sensors with EOL wiring and the central controls over RS485 (i'd probablly use cat5 for the RS485 links with one pair carrying data and the remaining 3 carrying 12V power for the concentrators and active sensors).

the hardware for the concentrators should be pretty trivial (I could probablly knock up a PCB layout in a day), the tricky bit will be the software but if you stick to a speak only when spoken to protocol it shouldn't be too hard.

 

[bernardgreen]

I tend to think it would work better to build concentrators that to try and mess arround with tiny biscuits in each sensor.

I totally agree. That approach removes almost all restrictions on the type of sensors to be used.

 

[breezer]

good luck, but i can see this is not looking like it will be cheap

 

[i_am_fubar]

Hey, sorry I've been away (generic DIY work). But yea, I understand it won't be easy of cheap... but its sometihng to make to teach myself about multi-node networks.

Shall look into some designs, thanks for the advice.

Sam.