1
Cxx0 Hardware / repair advice for bad NAND blocks
« on: December 02, 2006, 06:31:25 pm »
I think someone earlier said only models later than the C700 can do that.
Select language:
News:
Show Posts
This section allows you to view all posts made by this member. Note that you can only see posts made in areas you currently have access to.
Pages: [1] 2
2
C1000/3x00 General discussions / Disable C3k's 4gb Md?
« on: January 03, 2005, 11:35:40 pm »
Sorry if this has already been answered, but... my SL-C700 with a 1 GB microdrive has this annoying habit of forcing the MD to eject and reinsert every time the device is suspended and resumed. It's impossible to run a swap file on the microdrive because your swap file would vanish after a suspend/resume. (So you'd have to remember to swapoff and unmount before suspending)
I assume the new device with the built-in MD doesn't have this problem. Is it OK to run swap off of a file on the HD with the C3000? If so, they've apparently solved those problems.
--Michael Spencer
I assume the new device with the built-in MD doesn't have this problem. Is it OK to run swap off of a file on the HD with the C3000? If so, they've apparently solved those problems.
--Michael Spencer
3
C1000/3x00 General discussions / Backup C3000 Microdrive?
« on: December 23, 2004, 09:59:09 pm »
Someone more familiar with the C3000 can adapt this so it'll actually work with the C3000's quirks (if any), but I'd create a network connection with a computer with 4 GB of storage available and some version of netcat.
On the desktop machine, type:
nc -l -p 1234 > C3000HD.bin
on the C3000 type:
dd if=/dev/hda bs=1024 | nc 10.1.2.3 -p 1234
Because it uses TCP it should recover from most errors.
To restore, just go the other way.
Desktop: nc 10.1.2.4 -p 1234 < C3000HD.bin
C3000: nc -l -p 1234 | dd of=/dev/hda bs=1024
I've used this method to mirror Windows NT installations. It's FAR less efficient than dedicated software, because it copies even empty parts of the disk, but it's a full backup alright. :-)
--Michael Spencer
On the desktop machine, type:
nc -l -p 1234 > C3000HD.bin
on the C3000 type:
dd if=/dev/hda bs=1024 | nc 10.1.2.3 -p 1234
Because it uses TCP it should recover from most errors.
To restore, just go the other way.
Desktop: nc 10.1.2.4 -p 1234 < C3000HD.bin
C3000: nc -l -p 1234 | dd of=/dev/hda bs=1024
I've used this method to mirror Windows NT installations. It's FAR less efficient than dedicated software, because it copies even empty parts of the disk, but it's a full backup alright. :-)
--Michael Spencer
4
Cxx0 Hardware / AC adapter doesn\'t give an SL-C760 a full charge in the USA
« on: December 05, 2004, 04:37:42 am »
If our machines' battery charge lights had different colors, wouldn't the D+M menu demonstrate that if you do a hardware test?
--Michael Spencer
--Michael Spencer
5
Cxx0 Hardware / Battery 3.7v Vs 4.2v
« on: December 04, 2004, 11:33:29 pm »
This is normal. 3.7v is the "nominal" voltage -- this is the voltage your battery will spend most of its life at. 4.2v for an unloaded 3.7v battery is perfectly normal. (Just as a 12v car battery can be measured at 13v or more soon after being charged)
--Michael Spencer
--Michael Spencer
6
Cxx0 Hardware / repair advice for bad NAND blocks
« on: December 04, 2004, 02:48:36 pm »
My C700, which I just got back from Japan a few months ago, is having the same problem again. I had programs lock up intermittently, and now it won't boot. Same flashing lights.
It's not going back to Japan again -- I'll buy a C3000 from conics.net first. But maybe I can salvage this device.
--Michael Spencer
It's not going back to Japan again -- I'll buy a C3000 from conics.net first. But maybe I can salvage this device.
--Michael Spencer
7
Cxx0 Hardware / Ratoc CFU1 USB Host
« on: October 20, 2004, 07:43:28 pm »
Digikey? I don't know if they make off-the-shelf DC-to-DC regulators -- when I tried to go that route (when I assumed I would only ever need to power the Z itself) I had to get a breadboard and build the reference circuit from the documentation.
--Spence
--Spence
8
Cxx0 Hardware / Ratoc CFU1 USB Host
« on: October 18, 2004, 11:02:50 pm »
That's two parts mentioned. One: http://www.batteriesamerica.com/newpage4.htm (My battery is something like the upper-left one. It's HEAVY, but I don't have to worry about running out of power. :-)
Two: http://www.walmart.com/catalog/product.gsp...duct_id=1818511 That's an inverter: it converts 12V DC into 110V AC. It's inefficient, but with such a huge battery I don't mind. I generally get about a week of heavy usage out of that system before needing to recharge the battery. The inverter also refuses to drain the battery down past about 40% of total capacity. That's bad because I need more battery capacity to compensate, but that's good because lead-acid batteries don't live very long if you always drain them down to flat.
Yeah, it's heavy, but what are you going to do? Because I'm using a USB hard disk enclosure, I need both 5V and 12V from a power supply. I can't use a plain old battery: I'm pretty sure a hard disk needs that voltage to hold steady, and even if I had a 5V battery it would probably give 5.5V when loaded very lightly and 4.5V when under heavy load.
So you need some kind of power regulation. My way is less efficient, but I figure I'll have fewer problems in the long run if I just convert my battery's output to 110 V AC and use the manufacturer's own AC adapters.
If you're sure you need a smaller battery, a 2.7 watt device needs a power supply that can deliver 2.7 watts for as many hours as you need. If you only need one hour of life, you could probably rig up 6 AAA batteries and a DC-DC switching regulator (which needs some circuitry you'll have to assemble). That would be somewhat compact, but wouldn't have a lot of battery life. Those 6 AAA batteries (750 mAh at 7.2 volts) provide 5.4 watt-hours if discharged slowly -- probably closer to 3 watt-hours at the speed you would need.
If you can tolerate bigger batteries, something like 6 D batteries, and don't mind spending nearly $100 for 6 top-of-the-line NiMH D cells, you can hold 9000 mAh at 7.2 volts, which is 64.8 watt-hours, which would give you about 16 hours of runtime. (remember, you need that regulator so the output voltage stays pegged at 5 volts exactly, even as the batteries die, so you're only about 60 to 75% efficient.
I hope you see my point: if you want to store a lot of power in a little volume you're talking about higher energy density. Future technologies might give us better batteries with better energy density, but what we have right now gets pretty bulky.
So just do what I do: carry around a massive tank of energy, and tolerate the increased weight. It's good exercise. :-) (not so healthy for laptop bags though. Put a towel in the bottom of your bag to avoid having holes worn through the bottom)
--Michael Spencer
Two: http://www.walmart.com/catalog/product.gsp...duct_id=1818511 That's an inverter: it converts 12V DC into 110V AC. It's inefficient, but with such a huge battery I don't mind. I generally get about a week of heavy usage out of that system before needing to recharge the battery. The inverter also refuses to drain the battery down past about 40% of total capacity. That's bad because I need more battery capacity to compensate, but that's good because lead-acid batteries don't live very long if you always drain them down to flat.
Yeah, it's heavy, but what are you going to do? Because I'm using a USB hard disk enclosure, I need both 5V and 12V from a power supply. I can't use a plain old battery: I'm pretty sure a hard disk needs that voltage to hold steady, and even if I had a 5V battery it would probably give 5.5V when loaded very lightly and 4.5V when under heavy load.
So you need some kind of power regulation. My way is less efficient, but I figure I'll have fewer problems in the long run if I just convert my battery's output to 110 V AC and use the manufacturer's own AC adapters.
If you're sure you need a smaller battery, a 2.7 watt device needs a power supply that can deliver 2.7 watts for as many hours as you need. If you only need one hour of life, you could probably rig up 6 AAA batteries and a DC-DC switching regulator (which needs some circuitry you'll have to assemble). That would be somewhat compact, but wouldn't have a lot of battery life. Those 6 AAA batteries (750 mAh at 7.2 volts) provide 5.4 watt-hours if discharged slowly -- probably closer to 3 watt-hours at the speed you would need.
If you can tolerate bigger batteries, something like 6 D batteries, and don't mind spending nearly $100 for 6 top-of-the-line NiMH D cells, you can hold 9000 mAh at 7.2 volts, which is 64.8 watt-hours, which would give you about 16 hours of runtime. (remember, you need that regulator so the output voltage stays pegged at 5 volts exactly, even as the batteries die, so you're only about 60 to 75% efficient.
I hope you see my point: if you want to store a lot of power in a little volume you're talking about higher energy density. Future technologies might give us better batteries with better energy density, but what we have right now gets pretty bulky.
So just do what I do: carry around a massive tank of energy, and tolerate the increased weight. It's good exercise. :-) (not so healthy for laptop bags though. Put a towel in the bottom of your bag to avoid having holes worn through the bottom)
--Michael Spencer
9
Accessories / wireless zaurus
« on: October 02, 2004, 04:33:17 pm »
I just called and talked to Sprint regarding this card ( http://www1.sprintpcs.com/explore/PhonesAc...D=1068954683464 ) and asked about their data plans.
They say the plan gets you a normal connection to the public Internet -- no filtered ports or whatnot.
Today (Saturday 10/2/2004) a sales rep quoted me three data plans:
20 MB per month for $40.00 per month
70 MB per month for $60.00 per month
300 MB per month for $100.00 per month
The sales rep said there are no unlimited-transfer plans.
You CAN switch plans in mid-month.
If you go over your plan's limit, you pay $0.02 per KILOBYTE, which is $20.48 per MB. So if you sign up for the 20 MB plan and use 70 MB, you pay $1064 instead of $60.
Quoted prices are for a one-year contract. You can go month-to-month for an additional $10.00 per month.
Hope this helps! (If anyone knows of better plans with other carriers out there, please let us know.)
--Michael Spencer
They say the plan gets you a normal connection to the public Internet -- no filtered ports or whatnot.
Today (Saturday 10/2/2004) a sales rep quoted me three data plans:
20 MB per month for $40.00 per month
70 MB per month for $60.00 per month
300 MB per month for $100.00 per month
The sales rep said there are no unlimited-transfer plans.
You CAN switch plans in mid-month.
If you go over your plan's limit, you pay $0.02 per KILOBYTE, which is $20.48 per MB. So if you sign up for the 20 MB plan and use 70 MB, you pay $1064 instead of $60.
Quoted prices are for a one-year contract. You can go month-to-month for an additional $10.00 per month.
Hope this helps! (If anyone knows of better plans with other carriers out there, please let us know.)
--Michael Spencer
10
Cxx0 Hardware / Ratoc CFU1 USB Host
« on: July 02, 2004, 11:38:12 pm »
Hmm...sorry I didn't see this reply sooner, but...if the device is just being slow because it's reading more than 5 MB of data to perform an 'ls -al', then maybe it wasn't such a good idea to put a 120 MB filesystem over that slow an interface.
I'll repartition and try to make a 1 MB filesystem if possible. I'll report if the speed changes. In theory, if the filesystem is only 1 MB in size, and the interface supports 1 MB/sec transfer rates, and the system uses cache for I/O, then an ls-al should take no more than 1 second. I'll confirm this.
--Michael Spencer
I'll repartition and try to make a 1 MB filesystem if possible. I'll report if the speed changes. In theory, if the filesystem is only 1 MB in size, and the interface supports 1 MB/sec transfer rates, and the system uses cache for I/O, then an ls-al should take no more than 1 second. I'll confirm this.
--Michael Spencer
11
Cxx0 Hardware / Ratoc CFU1 USB Host
« on: May 28, 2004, 10:46:29 pm »
Now that my C700 is back from Japan and working again, I can post my experiences with the Ratoc CFU1 USB host card.
I bought a generic USB hard disk enclosure and a 120 GB IDE hard disk for it. (The hard disk needs a separate power adapter, but I\'m already carrying a 12 volt 18 amp-hour SLA brick and inverter in my backpack.)
1) After fdisk\'ing you MUST reboot. mkfs.ext2 complained about being unable to determine the size of /dev/sda1 until after I rebooted.
2) Disk access is SLOW. It\'s apparently mounting the drive over loopback, and loopback filesystem accesses do something nasty to the system. (I think the whole system kinda pauses while disk accesses are running. Will a kernel upgrade fix this?)
mkfs.ext2 took a great deal of time (I think over 20 minutes) to create a filesystem on the whole drive.
I did a really crude benchmark, since it was so slow. From the root of that filesystem, I ran:
time dd if=/dev/zero of=/mnt/usbhd/testfile bs=1024 count=1000000
The total time came to around 29 1/2 minutes.
I did the same thing on a desktop-computer-type machine running Linux, and it completed the write in 30 seconds.
Both while creating the filesystem and while writing the test file, the Zaurus was unusably slow. It took nearly 5 seconds to respond to any keypress or pen click.
OK, so that\'s that.
Second, I tried hooking up a Creative Labs MuVo2 4 GB (via USB, nothing taken apart), which I haven\'t disassembled yet. (I\'m waiting until I clean off a nice photo-friendly workspace before I take the MuVo2 apart and install the 4 GB drive in the C700. Once I do that, though, the world will gain a nice instructional video for the procedure ) The C700 recognizes it as a hard disk with 3 partitions, (strangely) /dev/sda1, /dev/sda2, and /dev/sda4. fdisk can see the partition table, but believes the device is much bigger than it actually is, and complains about the physical and logical begin and end not being the same, and that the partition ends on other than a cylinder boundary. None of the partitions are mountable.
I think it\'s time someone put together a \'developers\' romdisk -- no handholding, lots of documentation, and abundant storage *required*. I\'d love to be able to compile my own kernel in-place and build my own tools. I want my C700 to be a *completely* self-sufficient Linux machine.
I plugged in a generic USB 802.11b adapter and it was recognized, but the syslog says no driver supports the device, or something similar.
Same messages from a USB webcam, scanner, and multi-format USB card reader (5-in-1).
A slightly larger camera ( http://externe.net/zaurus/forum/viewtopic.php?t=1123 ) is Right Out.
Plugging the C700 into itself using the sync cable didn\'t yield any amusing results. The syslog said the USB client side stuff was timing out waiting for some kind of response, and the USB host side said it didn\'t recognize the device.
More information to come.
--Michael Spencer
I bought a generic USB hard disk enclosure and a 120 GB IDE hard disk for it. (The hard disk needs a separate power adapter, but I\'m already carrying a 12 volt 18 amp-hour SLA brick and inverter in my backpack.)
1) After fdisk\'ing you MUST reboot. mkfs.ext2 complained about being unable to determine the size of /dev/sda1 until after I rebooted.
2) Disk access is SLOW. It\'s apparently mounting the drive over loopback, and loopback filesystem accesses do something nasty to the system. (I think the whole system kinda pauses while disk accesses are running. Will a kernel upgrade fix this?)
mkfs.ext2 took a great deal of time (I think over 20 minutes) to create a filesystem on the whole drive.
I did a really crude benchmark, since it was so slow. From the root of that filesystem, I ran:
time dd if=/dev/zero of=/mnt/usbhd/testfile bs=1024 count=1000000
The total time came to around 29 1/2 minutes.
I did the same thing on a desktop-computer-type machine running Linux, and it completed the write in 30 seconds.
Both while creating the filesystem and while writing the test file, the Zaurus was unusably slow. It took nearly 5 seconds to respond to any keypress or pen click.
OK, so that\'s that.
Second, I tried hooking up a Creative Labs MuVo2 4 GB (via USB, nothing taken apart), which I haven\'t disassembled yet. (I\'m waiting until I clean off a nice photo-friendly workspace before I take the MuVo2 apart and install the 4 GB drive in the C700. Once I do that, though, the world will gain a nice instructional video for the procedure ) The C700 recognizes it as a hard disk with 3 partitions, (strangely) /dev/sda1, /dev/sda2, and /dev/sda4. fdisk can see the partition table, but believes the device is much bigger than it actually is, and complains about the physical and logical begin and end not being the same, and that the partition ends on other than a cylinder boundary. None of the partitions are mountable.
I think it\'s time someone put together a \'developers\' romdisk -- no handholding, lots of documentation, and abundant storage *required*. I\'d love to be able to compile my own kernel in-place and build my own tools. I want my C700 to be a *completely* self-sufficient Linux machine.
I plugged in a generic USB 802.11b adapter and it was recognized, but the syslog says no driver supports the device, or something similar.
Same messages from a USB webcam, scanner, and multi-format USB card reader (5-in-1).
A slightly larger camera ( http://externe.net/zaurus/forum/viewtopic.php?t=1123 ) is Right Out.
Plugging the C700 into itself using the sync cable didn\'t yield any amusing results. The syslog said the USB client side stuff was timing out waiting for some kind of response, and the USB host side said it didn\'t recognize the device.
More information to come.
--Michael Spencer
12
Cxx0 Hardware / repair advice for bad NAND blocks
« on: May 22, 2004, 10:40:14 pm »Quote
mspencer,
I got the same as yours. I also find bad block in \"Zaurus Test\"-]\"Nand Flash (Full)\". I would like to know is that mean my Z is unrecoverable anymore. I can\'t restore aymore flash now.
Could you give me some advice?
Thanks in advance
You might consider trying to erase flash, and then run the bad blocks test again. Other than that, yeah, I\'d say you should send it in. I\'m no expert on the subject though.
--Michael Spencer
13
Cxx0 Hardware / repair advice for bad NAND blocks
« on: May 15, 2004, 05:10:32 pm »
I went ahead and sent my device to Dynamism. According to them, Sharp Japan charged the equivalent of $360 USD for the repair. It\'s being shipped back to me for $400 (including shipping).
Ouch. It\'ll be good to have the C700 back, though.
--Michael Spencer
Ouch. It\'ll be good to have the C700 back, though.
--Michael Spencer
14
Cxx0 Hardware / repair advice for bad NAND blocks
« on: April 11, 2004, 01:44:41 am »
Well, Take-sensei said she\'s going home to visit some family mid next month, so once I know when she\'ll be there I could send the Zaurus and have Sharp send it C.O.D. to her family\'s house. She also has an uncle in...did she say Chiba?...but she said her Uncle would probably pay for the repair and not allow anyone to pay him back, touching off some kind of gift-giving one-ups-manship as she tried to repay the favor. Japanese culture seems inscrutable that way, although still fascinating. (and listening to that phone conversation helped me see: even after four semesters of Japanese language study, normal adult usage of the language is complex both linguistically and socially, and I really only know enough Japanese to pass quizzes and tests -- not real actual communication yet.)
I\'m trying to get my old Agenda VR3 PDA working again, but it\'s not curbing the C700-withdrawl pains so well. Talk about bad timing, too -- I just received my MuVo2 4 GB last Monday, and have no Zaurus to use the cannibalized microdrive with. (Once I get my C700 back, I\'ll try to pay the community back by making digital camcorder video of the drive-extraction and Zaurus-preparation process. )
Thanks again guys!
--Michael Spencer
I\'m trying to get my old Agenda VR3 PDA working again, but it\'s not curbing the C700-withdrawl pains so well. Talk about bad timing, too -- I just received my MuVo2 4 GB last Monday, and have no Zaurus to use the cannibalized microdrive with. (Once I get my C700 back, I\'ll try to pay the community back by making digital camcorder video of the drive-extraction and Zaurus-preparation process. )
Thanks again guys!
--Michael Spencer
15
Cxx0 Hardware / repair advice for bad NAND blocks
« on: April 10, 2004, 12:58:18 pm »
I think I can understand why they might choose to only use that shipping method: if they take payment over the phone, whatever payment method the customer uses probably has chargeback rights for card-not-present sales. They\'re probably doing these repairs very close to at-cost, so they want to eliminate the risk of people charging back repair costs, claiming they didn\'t authorize the charge. If a delivery person is at their door, swiping a card through a machine in-person, the person can\'t claim they didn\'t authorize the sale. They also can\'t claim they never received the device.
It makes sense from a risk perspective, why they would work that way.
Doesn\'t make it any more convenient for me, though.
(I work for a major credit card processor, so these issues make sense to me at least )
Thanks again for your help, guys.
--Michael Spencer
It makes sense from a risk perspective, why they would work that way.
Doesn\'t make it any more convenient for me, though.
(I work for a major credit card processor, so these issues make sense to me at least )
Thanks again for your help, guys.
--Michael Spencer
Pages: [1] 2