**** BEGIN LOGGING AT Tue Mar 18 02:59:57 2008 Mar 18 03:22:35 Changing them is pretty easy Mar 18 03:24:49 Let me know if the cap mods help. Last time I had the Toppy hooked up (about 5 months ago), it started having intermittent hangs, which I suspected to be PSU related. Mar 18 04:37:14 peteru_: I'm taking the multimeter home too, and will make some measurements. Mar 18 07:34:54 voltage went from 14.3 to 12.4, 37.6 to 30.6, 21 to 17.1, 9.15 to 7.73, 4.5 to 5.1, 3.3 to 3.3, 19.7 to 15.3 Mar 18 07:35:07 and now the 500GB disk works :-) Mar 18 07:35:21 peteru: there's the diffs Mar 18 07:35:28 tonymy01: thx for the tip Mar 18 07:36:07 Wow! Those PSU are dangerous. Mar 18 07:39:42 I replaced all six caps Mar 18 07:40:05 Also, having the disk as master didn't work - I had to have it as cable select. Mar 18 07:40:26 (that concerned me for a while, having done all the soldering and the disk still didn't work :-( Mar 18 07:43:57 the only supply that is regulated is the 5V and the 3.3V. When the pulse mode modulation can't cope with the load on the failing 5V caps any more, it increases the PWM to try to maintain 5V, which raises everything else that isn't regulated. Mar 18 07:44:00 so I'm copying data from the old 80GB tempy toppy disk to the new 500GB permanent topfield disk Mar 18 07:44:20 the big killer is the 14+ volts on the HDD 12V supply Mar 18 07:44:33 Hi guys Mar 18 07:44:57 and also the 30V supply feeds the tuners, so a voltage too high on that for too long fries either the tuner, or a transistor feeding it Mar 18 07:45:08 So, that was like $4 in components at retail prices. Mar 18 07:45:10 gday eric Mar 18 07:45:59 Which means that Topfield could have used decent quality caps at an additional cost of $0.05 per unit. Mar 18 07:46:15 peteru: yeah, I bought two sets, one 16V and one 25V, but the 25V ones were too big, so my total outlay was $10 :-) Mar 18 07:46:19 but i think they were pretty good Mar 18 07:46:47 just i guess the excessive heat and reasonable load was enough to deteriorate them over time Mar 18 07:47:14 my toppy is very hot 24x7x365 Mar 18 07:47:40 it's a strategy of mine - treat gadgets hashly, then you're forced to upgrade to new cool stuff. Mar 18 07:48:10 As of about 2 years ago, better PSUs tend to use the solid state kind of caps. New mobos also tend to have them for power conversion. Mar 18 07:58:58 Are they 10mm diameter? Mar 18 07:59:12 less than that Mar 18 07:59:18 the 25V ones probably are Mar 18 07:59:25 that Rod got I mean Mar 18 08:00:38 Hmmm, I was just looking up a few alternatives in the Farnell cattledogue and most of the better quality high-temperature caps are either 10 or 12.5mm diameter. Mar 18 08:00:58 bbl Mar 18 08:01:22 maybe they are, dunno Mar 18 08:01:28 never measured them Mar 18 08:02:35 http://www.jaycar.com.au/productView.asp?ID=RE6304&CATID=&keywords=1000uF&SPECIAL=&form=KEYWORD&ProdCodeOnly=&Keyword1=&Keyword2=&pageNumber=&priceMin=&priceMax=&SUBCATID= Mar 18 08:02:51 10x16 from jaycar, so they must be 10mm Mar 18 08:03:06 oops, that is 10v Mar 18 08:03:09 The Vishays are rated for 8000 hours at 125C, but are 12.5mm diameter Mar 18 08:03:42 No point spending $5 per capacitor if they are not going to fit :-) Mar 18 08:03:54 10x20 for the 16V ones Mar 18 08:04:44 yeah, the 25V ones are 12.5x20, and they apparently don't fit Mar 18 08:04:45 Mind you, the Australian stock for the Vishay caps is 5, UK stock is healthier at 1348 Mar 18 08:07:40 Rubycon caps are only 4000 hours at 105C, 10mm for about $2 each. Mar 18 08:07:59 At those prices, one might as well go with the Jaycar caps. Mar 18 08:08:47 my PSU voltages were still within about 5%, and my Toppie came from the 1st October 2003 batch Mar 18 08:09:07 but I replaced the 5V and 12V ones as precaution, since I had it in bits anyway Mar 18 09:14:42 the 25V ones were about half as wide again Mar 18 09:14:49 they didn't fit Mar 18 11:09:47 hmm the names and the svc data isn't in order in a svc.dat file Mar 18 11:10:14 and I've two entries that share the same name data Mar 18 11:11:00 +found Mar 18 11:11:08 7 Guide Mar 18 11:11:44 one of the records has a funny value in the first field of the svc data though Mar 18 11:11:49 0x7 Mar 18 11:12:14 I only usually see 1 for radio stations of 0 for the rest Mar 18 11:16:33 73,1798,1,177500,4115,1282,1315,7,"Seven Network","7 Guide" Mar 18 11:16:36 77,1798,1,177500,4115,1282,1318,0,"Seven Network","7 Guide" Mar 18 11:16:50 lcn,unk1,unk2,mhz,onId,tsId,svcId,radio,rf_name,name Mar 18 11:19:34 rf_name? I'd say it may be either a length byte or some flag field. Mar 18 11:20:02 7 == strlen("7 Guide") Mar 18 11:23:14 the others are all 0 though Mar 18 11:23:21 The 1798 may be a PID for PAT Mar 18 11:23:23 except the radio channels that are 1 Mar 18 11:23:42 yeah, in that case flags are more likely. Mar 18 11:24:36 200,1810,1,226500,4112,545,550,1,"ABC","ABC DiG Radio" Mar 18 11:24:36 201,1810,1,226500,4112,545,551,1,"ABC","ABC DiG Jazz" Mar 18 11:24:54 70,1798,1,177500,4115,1282,1316,0,"Seven Network","7 HD Digital" Mar 18 11:25:24 what do PID and PAT stand for? :) Mar 18 11:28:40 program association table? Mar 18 11:29:13 PID = packet id Mar 18 11:29:26 PAT - you got it Mar 18 11:35:40 Cool! STMicro have announced a 3.4Gbit HDMI active equaliser IC. That means that it should be possible to use really long HDMI1.3 cables, up to 20m and still do 1080p in 16-bit deep colour. Mar 18 11:36:37 It may even be possible to construct active cables, for runs longer than 20m Mar 18 11:36:47 cool Mar 18 11:38:31 It's also a 3-to-1 HDMI switch. Mar 18 11:38:39 is 1080p 444 Mar 18 11:40:05 I have a blu-ray movie which seems to be encoded at about 10Mib the actors faces smear in all the dark scenes Mar 18 11:40:14 The deep colour in HDMI 1.3 can be YUVxy or RGB. Obviously RGB is sampled uniformly and I think YUVxy is as well. Mar 18 11:40:26 ah k Mar 18 11:42:12 I think I'll need to do a bit more research into the svc.dat before I try one of my modified files on the wiz Mar 18 11:42:28 Traditionally there were many issues with colour subcarrier filtering, but also with relative positioning of the chroma and luma samples. It's important to know whether the subsampled colour is interstitial or not, otherwise it is not possible to rescale the image and preserve proper colour. Mar 18 11:43:18 interstitial? Mar 18 11:43:35 is the the location of the colour to the luma? in a block of 4 pixels? Mar 18 11:43:45 as you scale, the chroma and luma end up moving relative to each other and you end up with the colour siting next to the black and white picture, instead of being overlayed. Mar 18 11:43:59 right Mar 18 11:44:50 well, you can have co-sited colour sampling (you take the chroma sample at the same place as the luma samples) or interstitial (you sample chroma somewhere in between the luma samples) Mar 18 11:46:01 the idea behind interstitial sampling is that the colour comes from a location that is geometrically at the centre of the luma samples and thus the most likely "correct" colour for that area. Mar 18 11:47:37 and interlace makes this even more interesting, because you need to know the temporal location of your samples as well. Mar 18 11:48:03 the interlace order? Mar 18 11:51:15 some codecs will encode odd and even fields separately, whereas other codecs combine them to form a frame. it's possible that the various adjacent chroma and luma samples came from different points in time. Mar 18 11:54:53 I'm off now Mar 18 11:55:35 * Yuv422 hopes he won't be dreaming about interlaced video Mar 18 11:55:40 cya **** ENDING LOGGING AT Tue Mar 18 21:11:42 2008 **** BEGIN LOGGING AT Wed Mar 19 01:22:24 2008 **** ENDING LOGGING AT Wed Mar 19 02:59:57 2008