From nick at rcpt.to Sun Feb 3 00:01:21 2008 From: nick at rcpt.to (Nick Bannon) Date: Sun, 3 Feb 2008 00:01:21 +0900 Subject: [tech] Coke machine lights, empty detection, fans Message-ID: <20080202150007.GA15018@ucc.gu.uwa.edu.au> There've been a few Coke machine niggles: * the blinkenleds weren't visible - fixed * slot 0 and 5 neons were burnt out - fixed * slot 0 and 6 empty detection is broken - partly diagnosed [TRS] and I put the blinkenleds back into place in the coin slot, 'till the next time someone pushes a coin through anyway. A couple of wires had come off and were resoldered. Slot 0 and 5 had dead 90V neons - now all the empty indicator lights work! The neons aren't directly used for software empty detection anymore, but it's nice to see them, especially when dispensing through the snack machine panel. There's 6 more spare neons inside the pocket inside the machine. Dunno what their lifespan is meant to be, it might be normal or they might get a little overdriven? Slot 0 has burnt out before, slot 5 may well have been an original globe. The neons all have a common rail; and they'll be turned on by 100VAC (through current limiting resistors). The common side is about 60V away from chassis ground and the switched side (when it's on) is about 40V away from chassis ground. That 100VAC signal is cut down to size by a fat capacitor/diode/resistor to drive an optocoupler, on the isolation boards. The resistors get _hot_. The 4N38A opto on slot 0 is burnt out and gets warm. Slot 6 empty detection sometimes works, possibly a dry joint. The new empty detection system is sound in principle, but either it's not perfectly in spec and eventually things can burn out, or we're just unlucky. On the other side of the optocoupler are the coke brain input lines - slot 6 through to 0, 5V and GND. (grounded to the chassis of the whole machine) If an input line is at 1.6V or more, or if it breaks or is allowed to float (at about 1.9V), it's detected as not-empty. If it's pulled low, 1.5V or less, it's empty (thanks [SGC] for the tests!). The slot 0 and 6 input lines work fine, the slot 0 and 6 neons light up fine, so it's definitely the isolation boards that are the problem. I guess we should resolder the slot 6 joints and replace the slot 0 opto. If it burns out again, we can try something different. ::-) [MTL] and I inspected the fans recently - there were originally two mains ones; and one broke. The 12V replacement works fine, so we should probably go 12V-only and clean up some unnecessary mains wire. Nick. -- Nick Bannon | "I made this letter longer than usual because nick-sig at rcpt.to | I lack the time to make it shorter." - Pascal From nick at ucc.gu.uwa.edu.au Mon Feb 4 21:05:53 2008 From: nick at ucc.gu.uwa.edu.au (Nick Bannon) Date: Mon, 4 Feb 2008 21:05:53 +0900 Subject: [tech] Coke machine lights, empty detection, fans In-Reply-To: <20080202150007.GA15018@ucc.gu.uwa.edu.au> References: <20080202150007.GA15018@ucc.gu.uwa.edu.au> Message-ID: <20080204120553.GA32676@ucc.gu.uwa.edu.au> On Sun, Feb 03, 2008 at 12:01:21AM +0900, Nick Bannon wrote: [...] > The resistors get _hot_. The 4N38A opto on slot 0 is burnt out and gets > warm. Slot 6 empty detection sometimes works, possibly a dry joint. TODO: The slot 0 optocoupler is now in a socket - if someone has a 6 pin optocoupler to hand *TURN THE MACHINE OFF* then replace the opto and see if it fixes slot 0! I resoldered the capacitor on slot 6, looks like it's working again. TODO: > [MTL] and I inspected the fans recently - there were originally two > mains ones; and one broke. The 12V replacement works fine, so we should > probably go 12V-only and clean up some unnecessary mains wire. Nick. -- Nick Bannon | "I made this letter longer than usual because nick-sig at rcpt.to | I lack the time to make it shorter." - Pascal From nick at ucc.gu.uwa.edu.au Mon Feb 4 21:28:06 2008 From: nick at ucc.gu.uwa.edu.au (Nick Bannon) Date: Mon, 4 Feb 2008 21:28:06 +0900 Subject: [tech] Coke machine lights, empty detection, fans In-Reply-To: <20080204120553.GA32676@ucc.gu.uwa.edu.au> References: <20080202150007.GA15018@ucc.gu.uwa.edu.au> <20080204120553.GA32676@ucc.gu.uwa.edu.au> Message-ID: <20080204122806.GC32676@ucc.gu.uwa.edu.au> On Mon, Feb 04, 2008 at 09:05:53PM +0900, Nick Bannon wrote: [...] > On Sun, Feb 03, 2008 at 12:01:21AM +0900, Nick Bannon wrote: > > The resistors get _hot_. The 4N38A opto on slot 0 is burnt out and gets > > warm. Slot 6 empty detection sometimes works, possibly a dry joint. > > TODO: The slot 0 optocoupler is now in a socket - if someone has a 6 > pin optocoupler to hand *TURN THE MACHINE OFF* then replace the opto > and see if it fixes slot 0! Nah, it won't last long. There's something around 5V AC going into that opto input. Nick. -- Nick Bannon | "I made this letter longer than usual because nick-sig at rcpt.to | I lack the time to make it shorter." - Pascal From jacques at chester.id.au Tue Feb 5 02:38:51 2008 From: jacques at chester.id.au (Jacques Chester) Date: Tue, 05 Feb 2008 02:38:51 +0900 Subject: [tech] Favour Message-ID: <47A74DAB.4030003@chester.id.au> If anyone is in the clubroom tomorrow / today, could they reboot troppo? She seems to have carked it again. From zanchey at ucc.gu.uwa.edu.au Tue Feb 5 11:53:09 2008 From: zanchey at ucc.gu.uwa.edu.au (David Adam) Date: Tue, 5 Feb 2008 11:53:09 +0900 (WST) Subject: [tech] Favour In-Reply-To: <47A74DAB.4030003@chester.id.au> References: <47A74DAB.4030003@chester.id.au> Message-ID: On Tue, 5 Feb 2008, Jacques Chester wrote: > If anyone is in the clubroom tomorrow / today, could they reboot troppo? She > seems to have carked it again. Done. Xul was up and responding to keyboard input but not the network, so I did a clean restart. [DAA] From zanchey at ucc.gu.uwa.edu.au Sun Feb 10 00:16:51 2008 From: zanchey at ucc.gu.uwa.edu.au (David Adam) Date: Sun, 10 Feb 2008 00:16:51 +0900 (WST) Subject: [tech] /home enlarged, Mussel/Mermaid network changes Message-ID: Nick B, Patrick C and I made a couple of changes this evening: * enlarged /home on Martello to 100GB * moved Mermaid and thus Mussel to the gigabit switch, giving them higher speed network links. As a result, printing is currently unlikely to work. No progress has been made on the Manbo-Musundo migration (Triple M?). David Adam UCC Wheel Member zanchey@ From nick at ucc.gu.uwa.edu.au Sun Feb 24 18:32:57 2008 From: nick at ucc.gu.uwa.edu.au (Nick Bannon) Date: Sun, 24 Feb 2008 18:32:57 +0900 Subject: [tech] Coke machine lights, empty detection, fans In-Reply-To: <20080202150007.GA15018@ucc.gu.uwa.edu.au> References: <20080202150007.GA15018@ucc.gu.uwa.edu.au> Message-ID: <20080224093257.GA25045@ucc.gu.uwa.edu.au> On Sun, Feb 03, 2008 at 12:01:21AM +0900, Nick Bannon wrote: [...] > The resistors get _hot_. The 4N38A opto on slot 0 is burnt out and gets > warm. Slot 6 empty detection sometimes works, possibly a dry joint. I've replaced the opto. A possible cause of slot 0 burning out could be that the voltage dropping capacitor on that line has failed short circuit... I don't have a replacement to hand, but a temporary (?) fix is that I've added a ~27K resistor to the opto LED input. The opto LEDs only need 2-4mA to light up, so dropping the 100V input voltage with resistors is not tooooo wasteful. We'll see if it lasts. Slot 6's still seems good after its resoldering, so all the empty detection lines are currently working! TODO: > [MTL] and I inspected the fans recently - there were originally two > mains ones; and one broke. The 12V replacement works fine, so we should > probably go 12V-only and clean up some unnecessary mains wire. Nick. -- Nick Bannon | "I made this letter longer than usual because nick-sig at rcpt.to | I lack the time to make it shorter." - Pascal