hooking up my SECU16 to my Ocelot
- Thread starter electron
- Start date
BraveSirRobbin
Moderator
Yes, good point, just use a "fail safe" timer and open it a few seconds after the longest travel time the garage door has (in either closed or opened travel).
You can also trigger this off of your monitoring contacts (you are probably monitoring the open and closed position with contact closures as inputs to the SECU16). Before doing this IF statement I would probably also use a two second timer so you don't get into a race condition (immediately triggers because the contacts haven't opened yet) or a conditional if (if open monitor is "closed" then trigger on the next closed monitor).
You can also trigger this off of your monitoring contacts (you are probably monitoring the open and closed position with contact closures as inputs to the SECU16). Before doing this IF statement I would probably also use a two second timer so you don't get into a race condition (immediately triggers because the contacts haven't opened yet) or a conditional if (if open monitor is "closed" then trigger on the next closed monitor).
jrfuda
Active Member
Actually, it's just the change from open to closed that triggers the door to toggle its current position, so the contact/relay only has to be closed for less than a second (If the door is closed, closing the circuit opens it and vice versa). So timing is not neccessary.Think about it, when you send a command from the wireless opener in your car, you're only sending it for a fraction of a second. So my "failsafe" opens the relay whenever it sees it as closed.
When I set it up, I assumed there wasn't enough current going through it to be concerned with, so I didn't test it. The limit of my testing was using a piece of wire to simulate the closing of the circuit by the SECU16 relay, to ensure I was using the correct pait of contacts/screws on the garage door controller. It can't hurt to use one of BSR's relay, just to be safe, though. (I've done the same for my Ocelot/SECU16 watchdog for my server, the SECU16's relays are wired directly to the PC's power and reset buttons without realys... I'mnot experieinceing any problems there either)
Prior to setting-up the garage door, my wife and I used to leave the garage door open once in a while when we shouldn't (at night or when we're not home). Now, HS alerts us when the door is open for >5 minutes after sunset or >10 minutes when the house is vacant for >10 minutes. The first is by voice command, the second is by email to my work and personal addresses + my cell. I can then, when away from home, close the door via the web interface, RDP, or email. I even have an email/text message "template" saved in my cell phone that I can send to HS to have it close the door.
When I set it up, I assumed there wasn't enough current going through it to be concerned with, so I didn't test it. The limit of my testing was using a piece of wire to simulate the closing of the circuit by the SECU16 relay, to ensure I was using the correct pait of contacts/screws on the garage door controller. It can't hurt to use one of BSR's relay, just to be safe, though. (I've done the same for my Ocelot/SECU16 watchdog for my server, the SECU16's relays are wired directly to the PC's power and reset buttons without realys... I'mnot experieinceing any problems there either)
Prior to setting-up the garage door, my wife and I used to leave the garage door open once in a while when we shouldn't (at night or when we're not home). Now, HS alerts us when the door is open for >5 minutes after sunset or >10 minutes when the house is vacant for >10 minutes. The first is by voice command, the second is by email to my work and personal addresses + my cell. I can then, when away from home, close the door via the web interface, RDP, or email. I even have an email/text message "template" saved in my cell phone that I can send to HS to have it close the door.
ok, I ran more alarm wiring to control the garage door and everything is working without any relays. The button has about 6VDC when not pushed, when I push the button to open the door, it drops to 0VDC. I have to turn the relay ON for at least 1000ms or it won't work (I guess the tiny circuit board in the switch converts the momentary 'click' to the 1 second click or something), but otherwise it works great. Will the SECU16 unit be ok like this, or do I have to get a relay? Will it send 24VDC from the output relays to the button? Thanks!
Guy Lavoie
Active Member
If its working ok as-is, then there should be no worry.
Guy Lavoie
Active Member
What exactly do you mean by "how far the garage door is open"? Normally garage door openers only open it fully, to one position. To detect one of several positions, you would need a way to convert its position to some kind of analog value. Not obvious to do with such a long throw device like a garage door. One way would be with a multi turn potentiometer that could somehow be geared to the chain or motor driving the door. As an example, Phigets has a 10 turn sensor that could probably be rigged up. The hard part is matching the number of turns on the sensor to the actual travel.
sometimes my garage door doesn't open all the way (not sure why), so I want to know if the door is ajar, fully open or closed. I am checking the appdig forums out now, and found references to a project John Warner did, but I can't find it, the board is so slow, and the search engine won't let me search all forums at the same time. His project sounds like what I am looking for.
As far as tracking the garage door position goes...
There are many different ways to do this, all requiring different types of hardware and interface.
The ones I like best probably wouldn't use the Ocelot or SECU16 - at least not directly. The idea of watching the travel of the door useing a potentiometer like Guy suggested is probably one of the simplest to implement. You will need to work out the mechanics of rotating the pot as the door travels. You will need to be able to calibrate it - and also probably need to reset the open and closed values stored every time your open/close switches are triggered (to account for slip, etc.). Using mechanical sensors like this often leads to "zeroing" errors otherwise - always a problem that the robotics guys have when they navigate by dead reckoning.
You can also substitute an optical encoder for the potientiometer. You probably won't be able to use the Ocelot for this without extra hardware between them.
This idea got me thinking about some other ways to do this. What about putting optical coding on the track the door runs in? This would be a stripe along the track with one or more substripes of altering black and white regions. It would even be possible to develop a code such that you always new the absolute position of the door based on the current black/white pattern. If you are willing to live with a resolution of a few inches, you wouldn't need that many sensors/stripes. This assumes that you want to use encoding that specifies the specific location.
If you are willing to have the optics only give you direction (or a local position which repeats several times along the track), you can get away with fewer strips. Since the door moves pretty slowly, you can keep track of position based on counting changes up or down.
You can do this with only one stripe of alternating black and white regions if the regions are different lengths (progressively longer, for example). But if you do this, you will also need to take into account variations in door speed.
None of this would be particularly difficult to implement. The hardest part is probably making sure the sensors always track correctly over the patterned stripe.
This got me thinking about what I think is the coolest (and maybe silliest) solution I've come up with so far.
Mount UPC barcodes from your favorite products (cereal, soup, etc.) along the track. A different product should be used for each position. The number of different labels/products required is defined by the desired resolution of the system.
Then, mount a CueCat on the garage door. As the door opens and closes, the CueCat will scan these barcodes and they will be entered into the computer as keystrokes. All you need is a program that knows the order of the products and can match the codes to that order. You will know if the door is opening or closing by the order the codes are received and you will know the closest position (within your resolution) because you know the last code passed and the direction the door is traveling. The position will be between that last code and the next one.
I would recommend using the PS/2 version of the CueCat for this because you can probably run the cable a longer distance the the USB version.
There are many different ways to do this, all requiring different types of hardware and interface.
The ones I like best probably wouldn't use the Ocelot or SECU16 - at least not directly. The idea of watching the travel of the door useing a potentiometer like Guy suggested is probably one of the simplest to implement. You will need to work out the mechanics of rotating the pot as the door travels. You will need to be able to calibrate it - and also probably need to reset the open and closed values stored every time your open/close switches are triggered (to account for slip, etc.). Using mechanical sensors like this often leads to "zeroing" errors otherwise - always a problem that the robotics guys have when they navigate by dead reckoning.
You can also substitute an optical encoder for the potientiometer. You probably won't be able to use the Ocelot for this without extra hardware between them.
This idea got me thinking about some other ways to do this. What about putting optical coding on the track the door runs in? This would be a stripe along the track with one or more substripes of altering black and white regions. It would even be possible to develop a code such that you always new the absolute position of the door based on the current black/white pattern. If you are willing to live with a resolution of a few inches, you wouldn't need that many sensors/stripes. This assumes that you want to use encoding that specifies the specific location.
If you are willing to have the optics only give you direction (or a local position which repeats several times along the track), you can get away with fewer strips. Since the door moves pretty slowly, you can keep track of position based on counting changes up or down.
You can do this with only one stripe of alternating black and white regions if the regions are different lengths (progressively longer, for example). But if you do this, you will also need to take into account variations in door speed.
None of this would be particularly difficult to implement. The hardest part is probably making sure the sensors always track correctly over the patterned stripe.
This got me thinking about what I think is the coolest (and maybe silliest) solution I've come up with so far.
Mount UPC barcodes from your favorite products (cereal, soup, etc.) along the track. A different product should be used for each position. The number of different labels/products required is defined by the desired resolution of the system.
Then, mount a CueCat on the garage door. As the door opens and closes, the CueCat will scan these barcodes and they will be entered into the computer as keystrokes. All you need is a program that knows the order of the products and can match the codes to that order. You will know if the door is opening or closing by the order the codes are received and you will know the closest position (within your resolution) because you know the last code passed and the direction the door is traveling. The position will be between that last code and the next one.
I would recommend using the PS/2 version of the CueCat for this because you can probably run the cable a longer distance the the USB version.
BraveSirRobbin
Moderator
Guy:
Can't you string a few series resistors together as a voltage divider, then have multiple magnetic contacts "short" or "close" particular resistors, thus giving you an overall voltage change based on the number of resistors shorted or opened?
I know you will need to use several contact closures, but this method would let you only use one SECU16 input.
Something like the schematic below (warning, I did this before my morning coffee so its probably not correct)! :lol:
Can't you string a few series resistors together as a voltage divider, then have multiple magnetic contacts "short" or "close" particular resistors, thus giving you an overall voltage change based on the number of resistors shorted or opened?
I know you will need to use several contact closures, but this method would let you only use one SECU16 input.
Something like the schematic below (warning, I did this before my morning coffee so its probably not correct)! :lol:
