With the recent rearrangement of computer roles in my place, I’ve been left without a dedicated studio workstation. I made an attempt at running my home photo gallery on my old home office P2, but that was a failure. The computer was just too slow to handle Apache, MySQL, PHP and Gallery2 with a reasonably snappy response. So I repurposed the P2 machine to just running HomeSeer for my X10 home-automation.
WARNING: This blog contains writings known to the state of California to cause cancer and birth defects or other reproductive harm.
Archive for the 'Tech' Category
Page 4 of 7
I got a 60GB Video iPod from my girlfriend for Christmas in 2005. I’ve been having a great time with it.
I bought a pair of DLO HomeDocks for it, one to use at home and one at the office. They’re a pretty good value for the money, compared to Apple’s equivalent offering (which requires assembling a whole bunch of separately purchased pieces to get the same functionaliy that the DLO HomeDock has out of the box.)
It’s a pretty good deal… but not good enough.
The sound quality from these things is absolutely appalling. We’re talking class-action lawsuit kind of appalling. The sound is so bad, it should never have made it past the first usability testing, let alone been unleashed upon unsuspecting customers. And they have the nerve to say on the box “Play your iPod’s music through your home stereo and hear how good your iPod really sounds.” Well, if a constant barrage of digital bleeps, buzzes, clicks and whirrs underlying every song sounds good to you, maybe you’ll agree with that statement. I would say I got a couple defective units, but apparently they’re all like this.
Not one to be defeated so easily, I thought I would take one apart to see what makes it tick. I have a feeling it may be possible to clean up the sound with some careful mods.
I get this feeling, because in my experience, this kind of digital noise leaks into audio in only a few ways:
All of these problems, once correctly identified, can be rectified. Power supplies can often be cleaned up with additional filtering at the supply, and decoupling capacitors at each load. Sometimes, bus-topology supply or ground wiring can allow a noisy digital load with high switching currents to affect the supply and ground levels for all components downstream of it. This can be fixed by running separate supply leads back to a relatively noise-free central distribution point, making a star topology.
As a last resort, it is probably possible to just bypass the electronics in the HomeDock that are creating these problems.
So far, I haven’t been able to determine a whole lot in my examination of the innards of the HomeDock, but I have determined a few things:
So, that’s about all I’ve been able to glean so far. If I have any success in cleaning up the sound of this abominable piece of crap, I’ll post it here.
My front porch lights stopped working some time ago. Today I was able to narrow down the fault to the Leviton HCS10 X-10-controlled wall-switch. The relay in the switch was clicking, but no power was coming out on the blue lead.
Figuring I had nothing to lose, I decided to attempt to repair the module. They’re bloody expensive. Even if the repair failed, I’d still get to look around inside.
It should go without saying, but in this litigious world, it does not: doing this may burn down your house. Or it might electrocute you. It will definitely void your warranty. Do not do this if you are someone who doesn’t know a diode from an op-amp. You have been warned.
It was not easy getting the thing apart. Popping of the switch plate is easy, it’s just four plastic tabs. The rest of the job was very difficult, however. It appeared to be held together by eight screws, four on the front and four on the back. But turning the screws had no effect. I was able to turn them, but nothing came apart. In the end, out of sheer frustration, I just put a soldering iron onto the screw head, and melted it apart. BUT DO NOT DO THIS! There is a better way, which I discovered too late.
The trick, which would have saved me a lot of time had I realized it earlier, is that the screws are all reverse-threaded for some reason. Too keep people out, maybe? But then why use a standard Phillips head? A tamper-resistant or one-way screw might have worked better.
Anyway, inside the module there are two PC boards, mounted in a plastic frame, and connected together by a SIP header. I used my multimeter to probe around. I found that the blue controlled lead did have a good connection with the relay. But no connection from the black “hot” lead to the relay. I couldn’t see the PCB tracks that should have been carrying the power, they were apparently under the relay. I had to remove the relay to trace the connection back from there.
To remove the relay, it was necessary to separate the two PCBs. This required desoldering the SIP header that connects them. It was not easy, even with the help of a vacuum solder-sucker. They never get out all the solder. I had to keep pressure separating the boards (bending them, even), and then go back and forth with the soldering iron on the header pins. On each pass across the header, a little bit more pin comes out. With much patience, the boards come apart eventually.
Probably easier, if you have a spare SIP header of sufficient length, to just cut the old one with wire cutters and replace it later.
With the two PCBs apart, it is then possible to desolder the relay, which is much easier to desolder than the SIP header.
When I removed the relay, what did I see under it? A broken trace!
That’s the track that carries power from the hot lead to the relay. No wonder no power was coming out.
I fixed the broken track by soldering a bridge of wire across the gap. Replaced the relay, reassembled the module PCBs in their frame, and tested again on my bench. Working!
Putting the entire module back together wasn’t entirely easy, because I had melted it apart. But I was able to melt it back together again. Would have been better, if I had known the screws were reverse threaded, but there it is. It seems reasonably solid, anyway.
Replaced the module in the wall, and now the porch lights work again.
Certain aspects of the Harmony 628 (and, I suspect, other Harmony remotes) are starting to irritate me. What’s really stupid is how easily they could be fixed.
On the “Customize Buttons” pages, you can assign certain functions to the six buttons beside the LCD (three on each side). The functions can then (in theory) vary depending on the mode you’re in. This is great. I put a button to control the captioning display on my TV (my girlfriend likes to turn on subtitles, English is not her first language.) I also put on PgUp and PgDn buttons to use with the Rogers program guide on my digital terminal.
But, what really irritates me is that I can’t control which function is assigned to which button. PgUp and PgDn imply a particular positioning that one would expect… PgUp above PgDn. But I can’t get the remote to look like that no matter how I beg. It puts PgUp on the top right, and PgDn on the middle left. It’s stupid. The application has six slots for those buttons. All it has to do is program the buttons in the order I specified. How hard can that be?
And, another thing that bugs me: I tried to create a specially-customized “Watch TV” activity, specially designed to work with the Rogers Video-On-Demand service. Ie, Play, Stop, Rewind, etc should work. I wanted to start by just copying my existing “Watch TV” activity… but you can’t copy an activity, you have to start from scratch. So I did. But then, for some reason, when I changed the button assignments for the “Watch ROD” activity, they also changed for “Watch TV”. Sometimes the button assignments are independant for different activities (like, Watch a DVD seems to have independant button assignments), but other times they are shared. I’m not sure why.
And, why oh why does the backlight turn off 10 seconds after I turn it on, instead of 10 seconds after I stop using it? Always… ALWAYS, the backlight is turning off while I’m still using it. A constant irritation… but so easily fixed. Why won’t they fix it? I know I can’t be the only one who has ever complained about it.
Overall, I like the Harmony and the Harmony web-site… but even with my simple setup, I’m starting to feel it’s cramping my style. I know this nifty little gadget is capable of so much more.
Has anybody reverse-engineered this beast yet? Is there an open-source alternative firmware for the Harmony hardware that I could switch to?
I have set up the nested wireless network scheme I wrote about here. The “inner” network is maintained by a new LinkSys WRT54GS wireless router. But I’m seeing an occasional failure on the WRT54GS that displeases me mightily.
Sometimes, it gets into a state where the power LED blinks, and the wireless features become unavailable. The wired ports still work, but they appear to be connected directly to the WAN port. The internal DHCP server, web server, and firewall apparently stop working, instead my wired PC in effect seems to be directly connected to the WAN port. This is not a good failure mode for a product whose primarily purpose to me is security. If it has to fail, it should not fail insecure like this.
I usually don’t notice this state for days, because the only obvious symptom is the non-functioning wireless. But I don’t use the wireless all the time, just once in a while.
Fortunately, the situation isn’t as bad as it could be. The WAN port on my “inner” router is connected to a regular port on my “outer” router, which itself also has a firewall. So, this bug is not completely exposing me to the Internet. But it is exposing me to the outer network, which has the wireless running without WEP, only MAC address filtering. As long as no WiFi hackers come by, I guess I’m safe. And I do keep my PCs well patched.
Recent Comments