Accenta 'Comms' Help Needed

[Gabs]

ANY HELP IS GREATLY APPRECIATED 

 

Would anyone be able to help me in finding out how a keypad talks to the main panel, I have an oscilloscope asd I have connected it to the 'Comms' wire, when the Accenta keypad is connected to the panel. This wire is at a constant 5 volts, with dips in this voltage when a button is pressed, or the panel needs to send the keypad information. 

 

 

This first image shows the data line between the panel and keypad when 1 has been pressed, these dips in the 5 volt signal were created by the keypad, for the panel to recognise. 

 

 

The above image shows what happens when the 'PROG' key is pressed. The highlighted in red waveform is the keypad sending the Prog signal to the panel, the one in green is the panel responding to the keypad telling it to light up all LEDs.

 

What I do not understand is how the keypad can create a dip in the 5v line if the panel is constantly giving out 5v, same with the panel, how can it then create a drop in the voltage in the data line, if the keypad is constantly giving out 5v. The only way I thought it could be done was the panel shorts out the keypads data line, and the same with the keypad, it shorts out the 5v of the keypad to create 0v in the 2 way data line. Just wondering if this is correct? If anyone knows. I have tried the sound line, however it doesn't carry any data, just 'beeps' and alarm tones.

[sixwheeledbeast]

Having never attacked an Accenta with a scope before I am just theorising, however...

 

On a Accenta you have one signal line for data transmission therefore the transmission is serial.

As you say there are no separate Clock and Data lines to avoid data collision but what's to say a single non-zero return (NRZ) signal cannot be two way?

You could still transfer signals asymmetrically, maybe down to encoding for example, you could have a stop (handover) bit. The last bit on the scope images seems similar so this maybe the handshake bit for the reply or even the actual handover of the supply.

Both will need the ability to drive 5V signals to each other, although I would imagine the keypad will supply it most of the time. That way if the voltage is missing (hence using NRZ) for a period of time the keypad could be declined offline and tampered.

 

If you think about it these non-LCD basic alarm keypads are fairly dumb, they sit and wait for input so in most cases the messages to be sent towards the panel are call and response (the response changing/updating the display with data from panel).

The only time the panel would need to call the keypad into action (send data the keypads way) would be in an alarm situation, hmm I wonder what that sound wire could also be used for...

[james.wilson]

you can put an lcd keypad on them bit I've never put a scope on one

[Gabs]

1 hour ago, sixwheeledbeast said:

Having never attacked an Accenta with a scope before I am just theorising, however...

 

On a Accenta you have one signal line for data transmission therefore the transmission is serial.

As you say there are no separate Clock and Data lines to avoid data collision but what's to say a single non-zero return (NRZ) signal cannot be two way?

You could still transfer signals asymmetrically, maybe down to encoding for example, you could have a stop (handover) bit. The last bit on the scope images seems similar so this maybe the handshake bit for the reply or even the actual handover of the supply.

Both will need the ability to drive 5V signals to each other, although I would imagine the keypad will supply it most of the time. That way if the voltage is missing (hence using NRZ) for a period of time the keypad could be declined offline and tampered.

 

If you think about it these non-LCD basic alarm keypads are fairly dumb, they sit and wait for input so in most cases the messages to be sent towards the panel are call and response (the response changing/updating the display with data from panel).

The only time the panel would need to call the keypad into action (send data the keypads way) would be in an alarm situation, hmm I wonder what that sound wire could also be used for...

 

Wow, thanks for the reply, think I understand it  I understand that, but still am unsure how the keypad can drop the 5v that the panel is sending it, when a button is pushed. I would have thought that the Comms was always at 0v, and would only become active with 5v when a button is pushed, the Accenta has a different Tamper loop, with the comms disconnected, but everything else left in place, none of the keypad buttons work, as it is only this line the signal is send down. I have an early Accenta LCD keypad, and it works with his panel, using exactly the same drops in voltage, so it can recognise them and then displays vague messages on the screen, not as good as the G4 one. If I was designing an alarm, I'd have made it like this, If one or 2 were pressed, then the button number would correspond with the number of 5V Square waves are sent. Can you think of any reason, the ADE didn't do it like this? Or have any alarm manufacturers done it like this? 

Thanks, 

Gabs

 

[sixwheeledbeast]

Ah yes you are correct about the tamper loop, although the theory is still correct.

Have a look into NRZ serial transmission for other benefits of having the line high all the time.

Also think about what could happen if there was an intermittent short on the multi cored cable for example.

 

Transmitting decimal numbers via a type of pulsing binary is very inefficient. Your too young to remember how long it took to dial 999 on a pulse phone...

You can send a hex value in a nibble, for example this would cover all 15 keys no problem, but imagine an easier ways using rows and columns which keypads generally are.

As an example, divide your keypad into col (1-3) and row (1-5). that's 3x5 each axis (We will start counting at 1 here not 0). Send binary X, binary Y and a check and/or stop bit, all can be done with binary in a bytes worth of data.

Key 0       001 001 1 0 
Key 1       001 010 1 0
...
key 7       010 011 0 1
...
Key Prog    011 100 0 1
Key Set     011 101 1 0

All the keys would be sent in the same amount of time as sending an 8 via pulse.

Using BCD you could likely make that smaller again, the lesser amount of bits needed to transmit to get the message over the quicker the communication.

 

[Gabs]

1 hour ago, sixwheeledbeast said:

Ah yes you are correct about the tamper loop, although the theory is still correct.

Have a look into NRZ serial transmission for other benefits of having the line high all the time.

Also think about what could happen if there was an intermittent short on the multi cored cable for example.

 

Transmitting decimal numbers via a type of pulsing binary is very inefficient. Your too young to remember how long it took to dial 999 on a pulse phone...

You can send a hex value in a nibble, for example this would cover all 15 keys no problem, but imagine an easier ways using rows and columns which keypads generally are.

As an example, divide your keypad into col (1-3) and row (1-5). that's 3x5 each axis (We will start counting at 1 here not 0). Send binary X, binary Y and a check and/or stop bit, all can be done with binary in a bytes worth of data.

Key 0       001 001 1 0 
Key 1       001 010 1 0
...
key 7       010 011 0 1
...
Key Prog    011 100 0 1
Key Set     011 101 1 0

All the keys would be sent in the same amount of time as sending an 8 via pulse.

Using BCD you could likely make that smaller again, the lesser amount of bits needed to transmit to get the message over the quicker the communication.

 

 

Thanks for the reply, reason for me asking all this is that I am going to have a go at building my own panel, baed around an Arduino board: https://en.wikipedia.org/wiki/Arduino 

 

I will have a look at that type of serial transmission, hopefully I can have a go at programming something to recognise it. I can see a problem with the speed and efficiency, according to my oscilloscope, when pressed, the keypad sends the data out in a burst that lasts only 24 ms in total, so very fast. Do you know if all alarm panels use this method? Or is it just ADE panels? WHere have you gotten your knowledge on this from, did you design alarm panels at one time?

 

Thanks,

Gabs 

[sixwheeledbeast]

All systems are different, manufacturers will have there own protocols to suit there kit.

Galaxy is a standard 485 bus for example, some maybe clock and data or Wiegand derived like the majority of access control, there are many more ways it's all down to the manufacturer.

 

I suppose I have a fairly broad knowledge of a few things computing and electronics being one, some from school and college but a lot self taught. They have always my thing.

Back in the days of C&G 1851, data transmission methods where part of the security system exam, not sure if this is still the case.

I certainly have never done anything near a professional layout or designed panel but built a good few different circuits on stripboard in my time.

 

What you can learn now with the relative cost of SBC's like Arduino and RPi is a lot more is based on GPIO and code, over the logic with discretes and IC's that I started on.

 

I can only suggest you start basic adding functionality as you go on, if that be in logic gates or whatever code you choose.

Make a sensor trigger a sounder. Then make able to be armed and disarmed. Then eventually work up to security of the transmission and tamper protection.

Here is a link you may find useful about the way a keypad would work from my post above for example http://www.circuitbasics.com/how-to-set-up-a-keypad-on-an-arduino/

As you can see the computing power of an SBC is many times greater that of the Accenta you have on the bench so it's quite different practically.

[Rulland]

3 hours ago, sixwheeledbeast said:

 

Make a sensor trigger a sounder. Then make able to be armed and disarmed. Then eventually work up to security of the transmission and tamper protection.

 

Or just buy a system that has been fully tested by the manufacturers, and does the job, just saying!.

Unless it's Texecom, then the engineers are the guinea pigs, lol 

#Adrian#John

Edited December 27, 2018 by Rulland
Thpo

[sixwheeledbeast]

It's not about buying it tho, it's about what you can learn from making it. Many of the questions Gabs has asked are about reverse engineering and trying to understand what and why, I think that's a good thing.

Practical engineering skills seem to be at a low across all fields, I'd imagine many of the employers here will be thinking Gabs could make a better engineer than some they have had on the books for years.

 

[Gabs]

3 hours ago, sixwheeledbeast said:

All systems are different, manufacturers will have there own protocols to suit there kit.

Galaxy is a standard 485 bus for example, some maybe clock and data or Wiegand derived like the majority of access control, there are many more ways it's all down to the manufacturer.

 

I suppose I have a fairly broad knowledge of a few things computing and electronics being one, some from school and college but a lot self taught. They have always my thing.

Back in the days of C&G 1851, data transmission methods where part of the security system exam, not sure if this is still the case.

I certainly have never done anything near a professional layout or designed panel but built a good few different circuits on stripboard in my time.

 

What you can learn now with the relative cost of SBC's like Arduino and RPi is a lot more is based on GPIO and code, over the logic with discretes and IC's that I started on.

 

I can only suggest you start basic adding functionality as you go on, if that be in logic gates or whatever code you choose.

Make a sensor trigger a sounder. Then make able to be armed and disarmed. Then eventually work up to security of the transmission and tamper protection.

Here is a link you may find useful about the way a keypad would work from my post above for example http://www.circuitbasics.com/how-to-set-up-a-keypad-on-an-arduino/

As you can see the computing power of an SBC is many times greater that of the Accenta you have on the bench so it's quite different practically.

 

Thanks, I was going to start by making a key-switch panel, based around a microprocessor. I have made one around transistors before:

 

 

This was on a hand etched board, that I made. I used a paint marker to draw on the tracks, for all of the components. I will have a look at the site though. I have some CMOS chips, I may use to make an alarm, to panel. I made some bell box comfort LEDs swell, this way. I like the rewarding feeling of building circuits. The circuit pictured above is. A very basic transistor latch, it would need a relay to be able to switch any decent siren on, however its correct draw is practically zero. I was thinking of trying to mimmic the Accenta's keypunches of 4 and 9 (SOS) with an Arduino, this would be a good way of testing if I can use my own. Do you know what type or brand of microprocessor is used in most alarm panels, PIC or ATMEL. I think the Accenta has a Pic Microcontrollers but not sure, the NVM chip can't be found on the G3, so it may be held in the microprocessor.