Email: Password: Remember Me | Create Account (Free)

Back to Subject List

Old thread has been locked -- no new posts accepted in this thread
Joseph Hebert
12/16/11 06:36
Read: 362 times
Okemah, Ok
United States


 
Msg Score: +1
 +1 Good Answer/Helpful
#185117 - cost and reliability
Responding to: Andy Neil's previous message
Wow! I can't recall once seeing language become such an impedance as in this thread. Of course, that may be just because I wasn't really looking.

Andy, when Jeckson clarified "low cost" by saying "low budget," he wasn't answering you. He was answering Per.

As noted, cost is relative here.

- Cost as in imported special components.
- Cost as in time to implement.
- Cost as in consuming extra processor resources to monitor the sensor.
- Cost as in special cabling needed.
- Cost as in more board space for sensor normalization.


Jeckson, I believe that when Richard or Andy ask for a number, they're asking about cost per unit, not projected sales volume. How much (or how little) do you want to spend on each alarm sensor?

That said, I have more that I want to share with you. You mentioned developing your own shock sensor (magnet on a spring or something). I can not urge you strongly enough to choose another approach. That approach has two fatal flaws, the first of which is cost. It always costs more to reinvent the wheel, (which is to say, to develope something from scratch that is already refined and available for purchase). Moreover, simply in cost of components it will cost more than a simple reed and magnet switch.

But the more serious cost of this approach will be unreliability. Inevitably you will end up with a sensor that is either too sensitive (triggering an alarm condition if someone just bumps the rack), or so insensitive that it will be defeatable simply by moving the box slowly and carefully enough. You are going to find that there is no "Goldilocks zone" where the sensitivity is "just right." That is not to say the "zone" will be very small and hard to find. The "zone" will simply not exist. You will find that once the sensor is damped enough to carefully remove the box without tripping the alarm, you will still be able to trip it by bumping the rack.

It occurs to me that you might believe you want the sensor to be hypersensitive. You might think it's a good thing for the circuit to detect even bumping the rack. The opposite is true. If you put in anything as sensitive as that, it will only be a matter of time before your circuit becomes a nuisance to be routinely ignored until it's convenient to go and confirm that it was false alarming again.

The simplest solution would simply be to mount a micro-switch on the rack that is contacted by the case. If the case is removed (even a fraction of an inch) the contacts trigger an alarm, but as long as the case is securely mounted to the rack the circuit will be completely insensitive to shock induced false alarms.

Now if it is also true that your device is networked you can have an even more secure circumstance, an even higher reliability, redundancy. In other words, you can use the physical switch contacts to trigger an alarm, and then confirm the condition by loss of network data.

One more thing. Keeping in mind what Per said about letting your circuit current get too low, do a bit more research on "alarm circuits," this time looking specifically for "end of line (EOL) supervision," just in case you're also concerned with tampering, or someone trying to circumvent the alarm contacts.

Good luck,

Joe

List of 28 messages in thread
TopicAuthorDate
Freeze..thief protection      JecksonS Ben      12/10/11 07:27      
   Alarm Sensors      Joseph Hebert      12/10/11 11:06      
      Earth magnetic field sensor      Bert Van Den Berg      12/19/11 10:36      
         inexpensive solid state magnetic field      JecksonS Ben      12/20/11 01:36      
         It means..      JecksonS Ben      12/28/11 04:27      
            ???      Per Westermark      12/28/11 06:11      
               same      JecksonS Ben      12/31/11 04:12      
   Too vague!      Andy Neil      12/10/11 12:04      
      Cost - and what type of protection/detection      Per Westermark      12/10/11 13:21      
      How low is "low" cost?      Andy Neil      12/11/11 07:33      
         OK..great..maybe Links      JecksonS Ben      12/11/11 23:41      
            Still no idea what you actually want      Andy Neil      12/12/11 14:12      
      What exactly?      JecksonS Ben      12/12/11 23:53      
         Bounce        Per Westermark      12/13/11 00:59      
            OK      JecksonS Ben      12/14/11 00:32      
         If it's networked      Andy Neil      12/14/11 13:43      
            hard to handle      JecksonS Ben      12/15/11 01:19      
               Come on ... it requires a number        Richard Erlacher      12/15/11 08:42      
                  numbers      JecksonS Ben      12/16/11 01:02      
            cost and reliability        Joseph Hebert      12/16/11 06:36      
               cost      JecksonS Ben      12/17/11 03:44      
                  Does that mean you don't know what your price target it?      Richard Erlacher      12/17/11 08:57      
                  soft base      Joseph Hebert      12/17/11 10:11      
                     manufacture products      JecksonS Ben      12/18/11 03:07      
                        shock sensors        Joseph Hebert      12/18/11 16:17      
                           advice      JecksonS Ben      12/19/11 00:49      
                              Capacitive solutions can be projective      Per Westermark      12/19/11 01:00      
                                 metal..      JecksonS Ben      12/19/11 01:09      

Back to Subject List