From jurg.hengstler at gmx.net Wed Nov 1 07:31:09 2006 From: jurg.hengstler at gmx.net (=?iso-8859-1?Q?J=FCrg_Hengstler?=) Date: Wed, 1 Nov 2006 00:31:09 +0100 Subject: [tech] Sony monitor Message-ID: hello adrian there is a post about a sony monitor switch. I have this sony tv for 20 years and now it into a problem with just that sdg5p switch. Did you ever solve your problem? Jurg, Stockholm Sweden From harrymc at ucc.gu.uwa.edu.au Sun Nov 12 21:09:49 2006 From: harrymc at ucc.gu.uwa.edu.au (Harry) Date: Sun, 12 Nov 2006 21:09:49 +0800 Subject: [tech] Status of snack machine Message-ID: <45571D1D.3080909@ucc.gu.uwa.edu.au> So Nick, Cameron, and Harry braved the snack machine internals on Friday evening and found: 1. Checked PSU through up to motor driver ICs. PSU all looks ok. 2. Measured 24v high-side motor driver (UCN5890A) and found three stuck at 24V; pins 9, 10, and 12(?). No stuck pins of low-side driver (UCN5841A). 3. Disconnected three stuck pins for P8 (connector to motor row/column matrix). Tried to dispense from motor not connected to "stuck" circuits but no response. This could be just the high-side motor driver (UCN5890A) being dead and the low side driver is actually ok. 4. Connected a motor directly to 24V and it operated ok. TODO: A. Check motor matrix with meter to confirm all diodes in matrix are ok. B. Observe serial clocking and data signals on both drivers are bring driven ok by MCU C. Check if low-side driver outputs are driving low so driver is ok. D. Find high-side (and possibly low-side) drivers and replace. Spares of these would be handy because they look hard to find. E. Clean goop from melted redskins from rack that may be stalling motors F. See if the bernard code needs any more stall amd overcurrent sensing to disable motors if overloaded and report faults. msh and zarquin searched the net for the parts and the UCN5890A is discontinued and looks hard to find :-( Need to check if Mike Langan at Worldwide has some squirrelled away somewhere. All for now Harry From harrymc at ucc.gu.uwa.edu.au Mon Nov 13 20:50:14 2006 From: harrymc at ucc.gu.uwa.edu.au (Harry) Date: Mon, 13 Nov 2006 20:50:14 +0800 Subject: [tech] Octal high side driver from STMicro Message-ID: <45586A06.6080302@ucc.gu.uwa.edu.au> http://www.stmicro.fr/stonline/products/literature/ds/1416.pdf Also SGS-Thomson parts listed on ChipDocs http://www.chipdocs.com/datasheets/datasheet-pdf/SGSThomson-Microelectronics/L98.html From harrymc at ucc.gu.uwa.edu.au Tue Nov 14 14:54:00 2006 From: harrymc at ucc.gu.uwa.edu.au (Harry) Date: Tue, 14 Nov 2006 14:54:00 +0800 Subject: [tech] Octal high side driver from STMicro In-Reply-To: <45586A06.6080302@ucc.gu.uwa.edu.au> References: <45586A06.6080302@ucc.gu.uwa.edu.au> Message-ID: <45596808.80602@ucc.gu.uwa.edu.au> Harry wrote: *gah* the earlier ones were low side :-/ The right one: http://www.st.com/stonline/products/literature/ds/11456/vn808cm-e.htm# Datasheet: http://www.st.com/stonline/products/literature/ds/11456.pdf And DigiKey have them: http://www.digikey.com/scripts/dksearch/dksus.dll?KeywordSearch&Site=US&Vendor=497&MPart=VN808CM-E The 74HC595 is _almost_ an appropriate shift register to partner with the driver to replicate the failed part. The issue is that the gate signal tristates the outputs (so they could be setup to turn off by adding eight pull-down resistors) _but_ the data can only propogate from the shift register to the output latches on a rising RCLK (see data): http://www.fairchildsemi.com/ds/MM/MM74HC595.pdf It may be easier to use a 3.3volt regulator and a small Xilinx CPLD like: http://www.xilinx.com/bvdocs/publications/ds091.pdf This is overkill but the CPLDs don't get much smaller these days :-/ If I get time, I'll draw a circuit for the solution. Harry From harrymc at ucc.gu.uwa.edu.au Tue Nov 14 15:04:40 2006 From: harrymc at ucc.gu.uwa.edu.au (Harry) Date: Tue, 14 Nov 2006 15:04:40 +0800 Subject: [tech] Octal high side driver from STMicro In-Reply-To: <45596808.80602@ucc.gu.uwa.edu.au> References: <45586A06.6080302@ucc.gu.uwa.edu.au> <45596808.80602@ucc.gu.uwa.edu.au> Message-ID: <45596A88.2010207@ucc.gu.uwa.edu.au> Harry wrote: > It may be easier to use a 3.3volt regulator and a small Xilinx CPLD like: > http://www.xilinx.com/bvdocs/publications/ds091.pdf > This is overkill but the CPLDs don't get much smaller these days :-/ Err. Apologies for the one-sided conversation here. You could step back to the past and use a good old 22V10 (SO24 package). I didn't realise that they were still in use. They are just __very__ much faster these days :) http://www.latticesemi.com/dynamic/view_document.cfm?document_id=120 Harry From nick at ucc.gu.uwa.edu.au Tue Nov 14 16:47:33 2006 From: nick at ucc.gu.uwa.edu.au (Nick Bannon) Date: Tue, 14 Nov 2006 16:47:33 +0800 Subject: [tech] Octal high side driver from STMicro In-Reply-To: <45596A88.2010207@ucc.gu.uwa.edu.au> <45596808.80602@ucc.gu.uwa.edu.au> References: <45586A06.6080302@ucc.gu.uwa.edu.au> <45596808.80602@ucc.gu.uwa.edu.au> <45596A88.2010207@ucc.gu.uwa.edu.au> <45586A06.6080302@ucc.gu.uwa.edu.au> <45596808.80602@ucc.gu.uwa.edu.au> Message-ID: <20061114084733.GA26872@ucc.gu.uwa.edu.au> On Tue, Nov 14, 2006 at 02:54:00PM +0800, Harry McNally wrote: > The right one: > http://www.st.com/stonline/products/literature/ds/11456/vn808cm-e.htm# Great! > Datasheet: > http://www.st.com/stonline/products/literature/ds/11456.pdf > > And DigiKey have them: > http://www.digikey.com/scripts/dksearch/dksus.dll?KeywordSearch&Site=US&Vendor=497&MPart=VN808CM-E > > The 74HC595 is _almost_ an appropriate shift register to partner with > the driver to replicate the failed part. The issue is that the gate > signal tristates the outputs (so they could be setup to turn off by > adding eight pull-down resistors) _but_ the data can only propogate from > the shift register to the output latches on a rising RCLK (see data): > http://www.fairchildsemi.com/ds/MM/MM74HC595.pdf > > It may be easier to use a 3.3volt regulator and a small Xilinx CPLD like: > http://www.xilinx.com/bvdocs/publications/ds091.pdf > This is overkill but the CPLDs don't get much smaller these days :-/ [...] On Tue, Nov 14, 2006 at 03:04:40PM +0800, Harry wrote: > Err. Apologies for the one-sided conversation here. You could step back > to the past and use a good old 22V10 (SO24 package). I didn't realise > that they were still in use. They are just __very__ much faster these > days :) > > http://www.latticesemi.com/dynamic/view_document.cfm?document_id=120 'Course, we'd still need to learn how, and buy/make/borrow a programmer... http://members.surfeu.de/matthias.prinke/electronics/galprog.en.htm http://www.s-line.de/homepages/bosch/indexen.html Maybe it's time we bought a GALEP-4 (Linux version still on it's initial "alpha" version from August 21, 2003) or some USB equivalent: http://www.smartcom.co.uk/menus/main.asp?QTopMenu=Programmers 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 harrymc at ucc.gu.uwa.edu.au Tue Nov 14 17:32:15 2006 From: harrymc at ucc.gu.uwa.edu.au (Harry) Date: Tue, 14 Nov 2006 17:32:15 +0800 Subject: [tech] Octal high side driver from STMicro In-Reply-To: <20061114084733.GA26872@ucc.gu.uwa.edu.au> References: <45586A06.6080302@ucc.gu.uwa.edu.au> <45596808.80602@ucc.gu.uwa.edu.au> <45596A88.2010207@ucc.gu.uwa.edu.au> <45586A06.6080302@ucc.gu.uwa.edu.au> <45596808.80602@ucc.gu.uwa.edu.au> <20061114084733.GA26872@ucc.gu.uwa.edu.au> Message-ID: <45598D1F.30704@ucc.gu.uwa.edu.au> Nick Bannon wrote: > On Tue, Nov 14, 2006 at 03:04:40PM +0800, Harry wrote: > >>Err. Apologies for the one-sided conversation here. You could step back >>to the past and use a good old 22V10 (SO24 package). I didn't realise >>that they were still in use. They are just __very__ much faster these >>days :) >> >>http://www.latticesemi.com/dynamic/view_document.cfm?document_id=120 > > 'Course, we'd still need to learn how, and buy/make/borrow a programmer... > > http://members.surfeu.de/matthias.prinke/electronics/galprog.en.htm > http://www.s-line.de/homepages/bosch/indexen.html > > Maybe it's time we bought a GALEP-4 (Linux version still on it's initial > "alpha" version from August 21, 2003) or some USB equivalent: > http://www.smartcom.co.uk/menus/main.asp?QTopMenu=Programmers I have a Sunshine (well clone) that will do 22V10s given access to an MSDOS box with an ISA slot. But I realised later that a 24 pin DIP version of the 22V10 would be easier so the part could be removed from a socket on the daughter board for programming; or even a ZIF socket like the EPROM. For ease of programming's sake, it may be better to use the Xilinx part with JTAG port and put a JTAG interface on the daughter board so it can be reprogrammed in-situ. And I have a JTAG programmer that is almost finished (as a prototype anyway) for programming .. Harry From nick at ucc.gu.uwa.edu.au Fri Nov 17 13:23:56 2006 From: nick at ucc.gu.uwa.edu.au (Nick Bannon) Date: Fri, 17 Nov 2006 13:23:56 +0800 Subject: [tech] Status of snack machine In-Reply-To: <45571D1D.3080909@ucc.gu.uwa.edu.au> References: <45571D1D.3080909@ucc.gu.uwa.edu.au> Message-ID: <20061117052356.GH26872@ucc.gu.uwa.edu.au> On Sun, Nov 12, 2006 at 09:09:49PM +0800, Harry wrote: [...] > 2. Measured 24v high-side motor driver (UCN5890A) and found > three stuck at 24V; pins 9, 10, and 12(?). Terry Banks at VendPro ( http://www.vendpro.com.au/ ) put me onto Vendors Exchange - and they have the part, and other assorted spares, for our Rowe 5900! FWIW, Rowe International Inc. has no Australian agent. They sold their vending business to Automatic Products International Ltd.'s affiliated company Gross-Given Manufacturing Co. http://www.veii.com/pages/zoomInParts.asp?backpage=search.asp&partID=51273 PartNumber : VE8304 Price : $30.00 (each) Description : DRIVER CHIP FOR AP COF/SNAK UCN5890A In Stock : YES > No stuck pins of low-side driver (UCN5841A). (They don't seem to have that part listed, perhaps it fries less often! ) 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 harrymc at ucc.gu.uwa.edu.au Fri Nov 17 14:35:43 2006 From: harrymc at ucc.gu.uwa.edu.au (Harry) Date: Fri, 17 Nov 2006 14:35:43 +0800 Subject: [tech] Status of snack machine In-Reply-To: <20061117052356.GH26872@ucc.gu.uwa.edu.au> References: <45571D1D.3080909@ucc.gu.uwa.edu.au> <20061117052356.GH26872@ucc.gu.uwa.edu.au> Message-ID: <455D583F.4070703@ucc.gu.uwa.edu.au> Nick Bannon wrote: > http://www.veii.com/pages/zoomInParts.asp?backpage=search.asp&partID=51273 > > PartNumber : VE8304 > Price : $30.00 (each) > Description : DRIVER CHIP FOR AP COF/SNAK UCN5890A > In Stock : YES Good hunting Nick. Finding a number of parts for spares would have been desirable but it is cheaper that $100US but both costs are insignificant compared to diagnosis and resoldering effort _or_ time and prototyping costs of a re-engineered solution. Does it warrant buying two ($60) so we have a spare in the bottom of the machine in case spares have vanished next time around ? >>No stuck pins of low-side driver (UCN5841A). > > (They don't seem to have that part listed, perhaps it fries less often! ) Yes. We haven't tested the part other than it was not in a "shorted outputs" state so it may well be ok. Harry From grahame at angrygoats.net Fri Nov 17 15:17:52 2006 From: grahame at angrygoats.net (Grahame Bowland) Date: Fri, 17 Nov 2006 15:17:52 +0800 Subject: [tech] Status of snack machine In-Reply-To: <455D583F.4070703@ucc.gu.uwa.edu.au> References: <45571D1D.3080909@ucc.gu.uwa.edu.au> <20061117052356.GH26872@ucc.gu.uwa.edu.au> <455D583F.4070703@ucc.gu.uwa.edu.au> Message-ID: On 17/11/06, Harry wrote: > Nick Bannon wrote: > > > http://www.veii.com/pages/zoomInParts.asp?backpage=search.asp&partID=51273 > > > > PartNumber : VE8304 > > Price : $30.00 (each) > > Description : DRIVER CHIP FOR AP COF/SNAK UCN5890A > > In Stock : YES > > Good hunting Nick. > > Finding a number of parts for spares would have been desirable but it is > cheaper that $100US but both costs are insignificant compared to > diagnosis and resoldering effort _or_ time and prototyping costs of a > re-engineered solution. One question that occurs to me; how do we know that when we put the chip in it won't be immediately cooked again because the fault is actually somewhere else? Do we have enough information to verify that the chip is the only problem - it seems kind of unlikely it'd spontaenously fail without receiving some sort of nasty jolt. From harrymc at ucc.gu.uwa.edu.au Fri Nov 17 15:47:59 2006 From: harrymc at ucc.gu.uwa.edu.au (Harry) Date: Fri, 17 Nov 2006 15:47:59 +0800 Subject: [tech] Status of snack machine In-Reply-To: References: <45571D1D.3080909@ucc.gu.uwa.edu.au> <20061117052356.GH26872@ucc.gu.uwa.edu.au> <455D583F.4070703@ucc.gu.uwa.edu.au> Message-ID: <455D692F.9020900@ucc.gu.uwa.edu.au> Grahame Bowland wrote: > One question that occurs to me; how do we know that when we put the > chip in it won't be immediately cooked again because the fault is > actually somewhere else? Do we have enough information to verify that > the chip is the only problem - it seems kind of unlikely it'd > spontaenously fail without receiving some sort of nasty jolt. It's a valid concern and we had a list of tests in a previous email. On reflection the tests can be improved upon .. First step is to fit the chip and test without the motor matrix cables plugged in. The chip only drives out on those cables so I don't expect failure when the board is powered up this way. Next, use a CRO to test that the clocking and data signals are flowing through the high side (failed) part and low side part (they are daisychained so serial data has to flow through the pair correctly). Next is to measure the cables going to the motor matrix to confirm there are no shorted diodes in the matrix or in the cam switch matrix too. We started that with my TX3 meter using the diode tester setting but we didn't confirm that Nick's meter will do that as well; it has a diode test position but it needs checking. Otherwise finding a meter that tests the diodes for correct forward voltage (around 0.5volts we measured on some) rather than a short or open circuit. Another task is to inspect the receptacles that each of the trays plug into to confirm that neither the tray plug or socket on the back panel that receives them is damaged and may be shorting. Along the way, a task is to clean all of the melted Redskin gunk in the trays and visually inspect all of the trays for any obstruction that will stall the screws. Nick did a short test of operating one of the motors (using clippy leads) directly from the current limited side of the 24volt motor power supply. That test could be conducted on all motors to confirm there are no frozen tray screws. Only after all of those tests should the motor matrix be plugged in and the driver chip tested under load. That is the way of diagnosis and repair. You don't just change the failed "safety shutdown computer power supply line fuse" at the nuclear power plant and throw the big 2GW circuit breaker to ONLINE ;-) Snack machine requires similar care. Harry From trs80 at ucc.gu.uwa.edu.au Mon Nov 20 21:37:22 2006 From: trs80 at ucc.gu.uwa.edu.au (James Andrewartha) Date: Mon, 20 Nov 2006 21:37:22 +0800 (WST) Subject: [tech] netbooting machines Message-ID: Maroon wasn't booting after the cleanup, investigation revealed the network card's BIOS wasn't running. Putting it in a different slot helped, although by that point I'd written an etherboot floppy as well. It then booted, but couldn't find the nfs server to mount its root filesystem. Googling suggested removing IPAPPEND 1 from the pxelinux.cfg, which worked. I've also updated the kernels on the blackboxes to 2.6.18, as hal refused to work with 2.6.12 (hence the need to modprobe usbmouse on boot). This went ok, however the script that bind mounted /mnt/{ip} didn't work because ifconfig needs /proc mounted, so I added a manual mount. For sheer pendantry I also installed a patched version of mount that allows the use of a symlinked /etc/mtab using the patch from http://gentoo-wiki.com/index.php?title=HOWTO_Read-only_root_filesystem#mtab Oh, and they have a crontab of @reboot /bin/mount -a because nfs filesystems are mounted when a network card in /etc/network/interfaces is brought up, which doesn't happen because the kernel does it. -- # TRS-80 trs80(a)ucc.gu.uwa.edu.au #/ "Otherwise Bub here will do \ # UCC Wheel Member http://trs80.ucc.asn.au/ #| what squirrels do best | [ "There's nobody getting rich writing ]| -- Collect and hide your | [ software that I know of" -- Bill Gates, 1980 ]\ nuts." -- Acid Reflux #231 / From trs80 at ucc.gu.uwa.edu.au Mon Nov 20 23:15:26 2006 From: trs80 at ucc.gu.uwa.edu.au (James Andrewartha) Date: Mon, 20 Nov 2006 23:15:26 +0800 (WST) Subject: [tech] netbooting machines In-Reply-To: References: Message-ID: On Mon, 20 Nov 2006, James Andrewartha wrote: [blackboxes] They also don't shut down cleanly, no idea what's going on there. I've compiled up the spca5xx driver for the logitech webcam (which is plugged into pitch) but it looks a bit arse and the control program segfaults. -- # TRS-80 trs80(a)ucc.gu.uwa.edu.au #/ "Otherwise Bub here will do \ # UCC Wheel Member http://trs80.ucc.asn.au/ #| what squirrels do best | [ "There's nobody getting rich writing ]| -- Collect and hide your | [ software that I know of" -- Bill Gates, 1980 ]\ nuts." -- Acid Reflux #231 /