Weird intermittent failure of DSC zone expanders

[TinkerBear]

I'm trying to install a DSC PC1864 panel with 32 zones in my house as part of a renovation.  I'm putting sensors on doors + windows + motion, which is probably overkill, but I'm retired so it's partly about keeping myself occupied.  As the entire DSC Power Series is discontinued and out-of-stock, I got one used off eBay with 4 PC5108 zone expander boards (40 zones).  (Both alarm panel and RFK5500 keypads have been reset to factory defaults, several times.)

 

The install is progressing (slowly) along with the renovation - I've got 6 zones connected right now, the wire/sensors going in as the contractor tears apart the walls for window replacements and insulation.  Just for testing, I've connected all 4 zone expanders and wired each zone with a single 5K6 resistor, and enabled as zone type 03.  I can test each zone individually by shorting with tweezers or disconnecting the resistor, and the panel responds.  (I've now got the door chime enabled for all the zones with dummy resistors, just to make testing easier.)

 

Here is where things start to get weird... Zones 9-21 will (at some point after a day or two) spontaneously report that they are open.  All of them at once, when they have resistors solidly connected to them.  That's all of zone expander 1 and part of zone expander 2.  The zones remain open until the panel is power-cycled, or the zone is physically shorted or opened.

 

My first theory was that some of my zone expanders were broken or marginal, so I swapped their addresses around to verify.  The problem recurred... still on zones 9-21, even with those zones on different zone expander boards.  This suggests that the zone expander boards aren't at fault.

 

My second theory was that the alarm's main board has some kind of memory fault, and the zone failure was happening within the main board, not coming from the zone expanders.  I installed a KeyBus protocol dumper ( https://github.com/taligentx/dscKeybusInterface ) on a microcontroller and logged the protocol exchange.  The bogus zone openings are definitely coming from the zone expander boards, and not simply being imagined by the alarm panel.

 

My third try... was to throw money at it.  I ordered and installed an Eyezon UNO5108 zone expander (a new "compatible" zone expander) and installed it as zone expander 1.  It... still fails.  It reports zones 9-16 as open at the same time as zone expander 2 (still a DSC PC5108) reports zones 17-21 as open.  The only slight difference is that only zones 12 and 16 remain latched "open" after on the UNO5108.  (But on either the UNO5108 or PC5108, each zone can still be cleared by shorting or disconnecting it briefly.)

 

I am currently at a loss to explain this.  I'm hoping someone with more experience can shed some light?

 

I guess my next steps are to eye everything else in the system with suspicion.  Like the RFK550 keypads with wireless receiver... could random wireless data be causing this?

Edited September 28 by TinkerBear

[al-yeti]

55 minutes ago, TinkerBear said:

I'm trying to install a DSC PC1864 panel with 32 zones in my house as part of a renovation.  I'm putting sensors on doors + windows + motion, which is probably overkill, but I'm retired so it's partly about keeping myself occupied.  As the entire DSC Power Series is discontinued and out-of-stock, I got one used off eBay with 4 PC5108 zone expander boards (40 zones).  (Both alarm panel and RFK5500 keypads have been reset to factory defaults, several times.)

 

The install is progressing (slowly) along with the renovation - I've got 6 zones connected right now, the wire/sensors going in as the contractor tears apart the walls for window replacements and insulation.  Just for testing, I've connected all 4 zone expanders and wired each zone with a single 5K6 resistor, and enabled as zone type 03.  I can test each zone individually by shorting with tweezers or disconnecting the resistor, and the panel responds.  (I've now got the door chime enabled for all the zones with dummy resistors, just to make testing easier.)

 

Here is where things start to get weird... Zones 9-21 will (at some point after a day or two) spontaneously report that they are open.  All of them at once, when they have resistors solidly connected to them.  That's all of zone expander 1 and part of zone expander 2.  The zones remain open until the panel is power-cycled, or the zone is physically shorted or opened.

 

My first theory was that some of my zone expanders were broken or marginal, so I swapped their addresses around to verify.  The problem recurred... still on zones 9-21, even with those zones on different zone expander boards.  This suggests that the zone expander boards aren't at fault.

 

My second theory was that the alarm's main board has some kind of memory fault, and the zone failure was happening within the main board, not coming from the zone expanders.  I installed a KeyBus protocol dumper ( https://github.com/taligentx/dscKeybusInterface ) on a microcontroller and logged the protocol exchange.  The bogus zone openings are definitely coming from the zone expander boards, and not simply being imagined by the alarm panel.

 

My third try... was to throw money at it.  I ordered and installed an Eyezon UNO5108 zone expander (a new "compatible" zone expander) and installed it as zone expander 1.  It... still fails.  It reports zones 9-16 as open at the same time as zone expander 2 (still a DSC PC5108) reports zones 17-21 as open.  The only slight difference is that only zones 12 and 16 remain latched "open" after on the UNO5108.  (But on either the UNO5108 or PC5108, each zone can still be cleared by shorting or disconnecting it briefly.)

 

I am currently at a loss to explain this.  I'm hoping someone with more experience can shed some light?

 

I guess my next steps are to eye everything else in the system with suspicion.  Like the RFK550 keypads with wireless receiver... could random wireless data be causing this?

What cable you using, how far apart is everything? 

[MrHappy]

DSC control weren't very poplar in the UK... The git hub thing looks interesting.

 

Personally I'd put every part of the alarm system controls onto to a bench, all circuit linked with resistors & see if it plays up ?

 

As you doing something daft like thinking the alarm control panel is meant to power every single item of the system ?

 

 

[james.wilson]

What's the power like? 

Good battery. Check the 12v is behaving as it should and not too high or low when this occurs

[TinkerBear]

On 29/09/2024 at 13:01, al-yeti said:

What cable you using, how far apart is everything? 

 

Using four-conductor tinned alarm wire, no shielding, two short runs 5m each.  I've looked at the signals with an oscilloscope, and it looks pretty clean to me.

[TinkerBear]

11 hours ago, james.wilson said:

What's the power like? 

Good battery. Check the 12v is behaving as it should and not too high or low when this occurs

 

Battery is a little old - 2019 - but the voltages look fine.  I don't currently have a way to track the voltage for the multiple days it takes for a failure to happen.  I guess I could bodge together something with an Arduino... but that's probably a last resort?

 

As far as power to the house - we do get frequent power failures here, so we have solar plus battery backup.  There is a small gap if power fails, but I get multiple notifications if that happens, and it hasn't happened when the alarm freaks out.

[TinkerBear]

14 hours ago, MrHappy said:

DSC control weren't very poplar in the UK... The git hub thing looks interesting.

 

Personally I'd put every part of the alarm system controls onto to a bench, all circuit linked with resistors & see if it plays up ?

 

As you doing something daft like thinking the alarm control panel is meant to power every single item of the system ?

 

Well, the alarm panel is meant to power every other item in the system.  I've done the power budget, and I'm nowhere near maximum.

 

The panel is in the same room as my workbench, so I've got a keypad and some of the zone expanders on my bench.  The zones that fail are unfortunately in the box on the wall.  I could change the addressing to put those zones on the ones on my bench... but I'm not sure what to look for.  I get a protocol capture wherever the bits are - and those zones are reporting open.

[TinkerBear]

I'm starting to suspect the RFK5500 keypad.  The "RFK" version is a combined PK5500 keypad plus a PC5132 wireless receiver on the same PCB, but I believe they still operate independently.  I'd reset the alarm panel and the keypad, but I hadn't reset the wireless receiver.  I'm not sure how the wireless functions work -- if the wireless zones are reported to the panel the same as zone expander events are -- but if they were, it would explain what I'm seeing.

 

I've now done a [996] wireless module factory default, and... now the "PC5132" has disappeared from the module inventory [903].  I guess it needs to be set up again to appear?

 

Now it'll just be a couple days to a week to see if the problem recurs.

[MrHappy]

57 minutes ago, TinkerBear said:

 

Well, the alarm panel is meant to power every other item in the system.  I've done the power budget, and I'm nowhere near maximum.

 

Rated at 1.7ah 

 

-110mA for the main pcb

-140mA  4 PC5108

 

1450mA left for the radio expander(s) detection & sounders.... ect ?

 

what voltage do you have at the expender which plays up

 

have you swapped the expander with one which does not play up & addressed it & fitted in the place of the one which does play up ? 

 

 

 

 

[al-yeti]

On 29/09/2024 at 00:05, TinkerBear said:

I'm trying to install a DSC PC1864 panel with 32 zones in my house as part of a renovation.  I'm putting sensors on doors + windows + motion, which is probably overkill, but I'm retired so it's partly about keeping myself occupied.  As the entire DSC Power Series is discontinued and out-of-stock, I got one used off eBay with 4 PC5108 zone expander boards (40 zones).  (Both alarm panel and RFK5500 keypads have been reset to factory defaults, several times.)

 

The install is progressing (slowly) along with the renovation - I've got 6 zones connected right now, the wire/sensors going in as the contractor tears apart the walls for window replacements and insulation.  Just for testing, I've connected all 4 zone expanders and wired each zone with a single 5K6 resistor, and enabled as zone type 03.  I can test each zone individually by shorting with tweezers or disconnecting the resistor, and the panel responds.  (I've now got the door chime enabled for all the zones with dummy resistors, just to make testing easier.)

 

Here is where things start to get weird... Zones 9-21 will (at some point after a day or two) spontaneously report that they are open.  All of them at once, when they have resistors solidly connected to them.  That's all of zone expander 1 and part of zone expander 2.  The zones remain open until the panel is power-cycled, or the zone is physically shorted or opened.

 

My first theory was that some of my zone expanders were broken or marginal, so I swapped their addresses around to verify.  The problem recurred... still on zones 9-21, even with those zones on different zone expander boards.  This suggests that the zone expander boards aren't at fault.

 

My second theory was that the alarm's main board has some kind of memory fault, and the zone failure was happening within the main board, not coming from the zone expanders.  I installed a KeyBus protocol dumper ( https://github.com/taligentx/dscKeybusInterface ) on a microcontroller and logged the protocol exchange.  The bogus zone openings are definitely coming from the zone expander boards, and not simply being imagined by the alarm panel.

 

My third try... was to throw money at it.  I ordered and installed an Eyezon UNO5108 zone expander (a new "compatible" zone expander) and installed it as zone expander 1.  It... still fails.  It reports zones 9-16 as open at the same time as zone expander 2 (still a DSC PC5108) reports zones 17-21 as open.  The only slight difference is that only zones 12 and 16 remain latched "open" after on the UNO5108.  (But on either the UNO5108 or PC5108, each zone can still be cleared by shorting or disconnecting it briefly.)

 

I am currently at a loss to explain this.  I'm hoping someone with more experience can shed some light?

 

I guess my next steps are to eye everything else in the system with suspicion.  Like the RFK550 keypads with wireless receiver... could random wireless data be causing this?

So as mrH have physically swapped board locations rather than just addressing