startup ordeal with old 600x

[dcouzin]

My T42 must go to the shop so I tried to start my old 600x. It took about an hour of repeated power ons and offs. Sometimes the screen would be completely blank. Sometimes a cursor flash. Sometimes an error message. Then sometimes there'd be joyful beeps and the IBM error screen with the birdy pointer. That would give the opportunity to _again_ set date and time. Then sometimes a second error screen asking for off & on. What is interesting is that the more times one tries, the farther one generally gets. Eventually you get to Windows, and then it seems healed. From Windows I've restarted three times and each time it got back to Windows.
Now I will leave the 600x running for the week that T42 is away.
What's happening in my poor old 600x. It's as-if something inside needs warming-up. Could it be that simple? Should I put the 600x in a 100F oven before starting it?
Can this 600x startup characteristics be improved without major internal work?
I noticed that the screen colors with the 600x are better than with the T42.

[rkawakami]

With the description of the error screen (161 and 163 perhaps?) then the second screen indicating a power cycle of the laptop is required, I would say your CMOS (backup) battery is weak.

As far as the blank screen or just a cursor, I'm not sure if that would be indicative of a backup battery problem. That sounds more like a hard drive to me. You may want to try re-seating both the hard disk drive and memory modules and see if that improves the startup.

[The Spirit of X21]

What happens if you just pull the CMOS battery? I've heard that when the backup is going on PowerBook G3 (Pismos) that users can remove the CMOS battery and restart the machine in most cases.

[mlwhiten]

If the 600X has been sitting unused for quite a while, it may be that the main battery as well as the CMOS battery are just dying out.. That seemed to be the problem when I got mine. The person I got it from had let it sit unused for about 3 years or so, and it killed the CMOS battery. Also seemed to put a dead spot in the main battery, so that even with it claiming a 100% charge (after several days of charging) it would only last 5 minutes without the AC adapter.

[Stargate199]

Because the computer has been sitting around for a while, the CMOA is probably dead. That would not account for the irregular startups since the computer can function without the CMOS. My guess is that there might be some corrosion to the power circuitry. That would explain why the computer is having power up problems. That may in fact not be the case, but don't put it in the oven. Try powering one with out the battery on AC power only and see what happens. If it still isn't powering on properly, then take it apart and look for corrosion and/or burned circuit boards. If there is corrosions, carefully with a toothbrush and some rubbing alcohol clean off the affect parts, let it set for 30-60 minutes, put it back together and power on.

[dcouzin]

Thank you all for these convergent suggestions. I'll examine the CMOS battery after the week. I saw the following on a help site:

If your computer is losing its time or date settings, or you are receiving a message CMOS Read Error or CMOS Battery Failure, first attempt to leave the computer on for 24 hours. In some cases this can charge the battery and resolve your issue. This often resolves CMOS battery related issues when a computer has been left off for several months.

I'm surprised that a lithium battery can take a charge, but if something like this is going on it would explain why I became increasingly lucky with the startups.

The one description I found of the CMOS battery on the TP 600x says it is found under the RAM and has leads soldered (welded, brazed) to the battery shell.

[rkawakami]

It's been my experience that the CMOS battery does not get recharged, but I've never bothered to actually put a voltmeter on the battery connector to see if the laptop is forcing a voltage out there when powered up. The main reason I haven't done that is that the coin-type lithium backup batteries are not rechargeable in the first place. They differ from the lithium-ion batteries in the main battery pack, which are rechargeable.

The backup battery does have its leads welded onto the shell but there's a connector at the other end which simply plugs into a jack on the motherboard. No tools are needed to remove it (other than a screwdriver to open the memory panel). The battery is tucked up between the bottom of the laptop and the Ultrabay device.

[dcouzin]

It would be nice if we could use a generic lithium battery rather than having to buy one with welded leads. A pressure connection requires some sort of spring, e.g., rubber pads with unstretchable wrap, but is there enough space for this? There are silver-filled epoxies which never work. Standard soldering, I take it, doesn't work.

[Stargate199]

I have heard people just replacing the battery with a watch battery. This does require some soldering for the connectors. In my opinion, its best to go to eBay and look around for a replacement CMOS battery. There is a ton of them and go around $5.

[rkawakami]

Until recently I did not know that some of the older Thinkpads DO use a standard watch battery (i.e., no welded on leads). Some people here have pried off the thin metal strips that are welded onto the factory batteries and simply held them to an off-the-shelf lithium battery with electrical tape or heat-shink wrap. Soldering them is out of the question as they can explode with the application of heat. I generally find the cheapest seller on eBay, with a shipping discount on multiple quantities, and get the genuine article. That way I don't have to worry about intermittent contact on the battery causing weird CMOS errors.

[dcouzin]

As some of you suspected my problem is not just the CMOS battery.
I checked it. It measures 1.6 volts, so perhaps on the edge of functioning. But with the CMOS battery in or out there is still first the all black screen. With the marginal CMOS battery in, all black screen through 5-10 power off/on cycles. Then there's the IBM welcome, but nothing else. After about 20 power off/on cycles there's Windows! (This time there there no date errors reported.) What is going on with all this off/on action? Why does it almost always progress to Windows eventually?

[rkawakami]

Still sounds like a problem with the hard drive and/or the drive interface on the motherboard. Try this: remove the hard drive and any Ultrabay drive and power up. After the IBM splash screen, you should see (within a few seconds) a diagram on the screen which shows a floppy disk and an indication (with moving arrows) to insert said floppy disk into a drive labelled "0". At the bottom of the screen will be "buttons" representing keys F1 to F12. What this is saying, is that you should install a bootable disk and then press F1 to continue. Don't install anything and just press the F1 key. You should see an error message of "I9990305" displayed on the screen. That means the system has not found an operating system. This is a quick way of checking out your system to see if it is working normally. If this works as I've described, and more importantly, if it doesn't, then I would say your next step would be to get a copy of PC Doctor for DOS for your 600X system from either of these two places:

PC Doctor 600X - diskettes

or

PC Doctor 600X - .ISO (CD burnable) image

and start running some diagnostics. Start off with the Memory test, then CPU/Coprocessor, the Systemboard, then the Fixed Disk diags. The reason I suggest that order is that you need to be sure that most of the major sub-systems are working correctly before you run the disk diag.

[dcouzin]

Rkawakami: The first boot without hard drive stopped at the splash screen The second attempt went as you said, leading to error message I9990305.
Then with the hard drive reinstalled the first five boots stopped at the splash screen, and the sixth went to Windows.
The problem with this machine is that boots can be black, or can go to the splash screen, or can go beyond. Every time the machine must be turned off there is the danger that it won't run again.
I had already tried the DOS PC Doctor for 600X. It hung up while trying to load Test 46 of 65: Lucent Modem Loopback. I later ran this and the later tests separately: all passed. PC Doctor identified no problems.
Here is another little problem which might provide a clue. I'm using a 3Com 10/100 LAN card bus for this communication. Every few minutes it loses connection. (The DSL here is quite stable.) I pull the card out and push it in. It resets.

[rkawakami]

Hmm... then I would have this last (?) suggestion...

Pull or disconnect every bit of extra hardware out of the system (hard drive, Ultrabay, PCMCIA, USB, MiniPCI and memory) and see if you can get the system to power up reliably. Obviously if there is no operating system to be found, expect those kinds of errors. Just see if you can get past the IBM splash screen everytime. If that works, then put the MiniPCI modem back in first and try again. I have seen T2x system hang on boot from a bad MiniPCI modem but not ever with the 600X. Continue to re-install pieces (I'd do memory next, Ultrabay, then the hard drive) and repeat the power ons.

[dcouzin]

Phew! This time it took more than 30 repowerings to get moving. This could make a man religious, which isn't fair.
I did remove all externals as you suggested, but it didn't do the trick.
It did allow me to appreciate the pretty engineering of the 600X. My T42 is not in its league, finish wise.
Now the 600X must stay on all week while the T42 is at the shop.
Thanks.

[rkawakami]

Sorry that didn't do the trick. What you're left with then I suppose is a problem within the motherboard. It could be something in the power supply section or with the disk drive interface. In any event I think you have proved it's not any of the "extra" pieces you removed and re-installed.

[dcouzin]

I solved the problem with the 3Com 10/100 LAN card bus with a software setting -- "full duplex" -- whatever that means. So the old TP 600X is now perfectly functional provided it is left on. The brief depowering of a "restart" doesn't bother it. Modern equivalent of jalopy.

[Rob Mayercik]

As some of you suspected my problem is not just the CMOS battery.
I checked it. It measures 1.6 volts, so perhaps on the edge of functioning.

A new CMOS battery reads 3 volts. Yours isn't "on the edge," it's well over.

If you haven't replaced it yet, I'd say that ought to be your next step. It may or may not help, but it's a very good idea.

Rob

[dcouzin]

A new CMOS battery reads 3 volts. Yours isn't "on the edge," it's well over.

If you haven't replaced it yet, I'd say that ought to be your next step. It may or may not help, but it's a very good idea.
[/quote]

Rob,
Since I'm not having to reenter the date each time I restart, it appears that the 1.6 volts left from the battery [i]is[/i] sufficient for the BIOS, just now. A Panasonic CR2025 battery costs 6 Euros here. If as rkawakami suspects, the startup problem points to the system board, is the computer worth 6 Euros? Would you give $7.80 for it?
I wonder why I can't test your CMOS battery hypothesis with another 3 volt source. I'd hold two 1.5 volt button batteries in series and put this in parallel with the present battery just to see if this alters the course of startup. Why not? Why does the notice inside the hatchdoor insist "Danger of explosion if battery is incorrectly replaced. Replace only with the same or equivalent type."?

[farna]

For testing your idea of any other 3.0V source shold be fine. If the batteries are connected backwards it's possible to overheat them, which could lead to an explosion or the batteries "leaking", which would ruin the memory as well. As long as you're fairly certain you know what you're doing, and connecting the leads correctly, there's no danger.

I'll give you $7.80 for it when you get the new one back, and I don't need another laptop! It would cost $30 or more to ship from Europe though. It's worth $40 though...

[Rob Mayercik]

Since I'm not having to reenter the date each time I restart, it appears that the 1.6 volts left from the battery is sufficient for the BIOS, just now. A Panasonic CR2025 battery costs 6 Euros here. If as rkawakami suspects, the startup problem points to the system board, is the computer worth 6 Euros? Would you give $7.80 for it?

This "restarting" bit - do you mean you told Windows to restart the computer, or did you tell it to shut down, let it sit powered off for a minute or two, and then power back up?

I just replaced my CMOS batttery last month, and this time mine was at least as low as 1.6V, possibly lower. It wasn't even possible to get as far as CMOS to set the date and time until I actually disconnected the darn thing from the system board. The first time it went out on me(a couple years back), it was reading closer to 2V, and I was able to set the clock and continue the boot sequence with the dying cell still installed.

Remember, this little button cell also spins up the CPU fan briefly upon powerup (don't know about warm restart), so maybe the cell is only occasionally able to cough up enough power in its current state to run everything at once.

I wonder why I can't test your CMOS battery hypothesis with another 3 volt source. I'd hold two 1.5 volt button batteries in series and put this in parallel with the present battery just to see if this alters the course of startup. Why not? Why does the notice inside the hatchdoor insist "Danger of explosion if battery is incorrectly replaced. Replace only with the same or equivalent type."?

As Farna said, your proposed test method is fully acceptable - so long as you don't mix up positive and negative (in other words, take careful note of which way the connector plugs into the system board). If this shows that the CMOS battery is a factor, then put the real thing in, of course.

Once you have it hooked up, though, be sure that once you've set the date and time and booted, power it off for about 5 minutes and then see if it's persisting.

Honestly, I don't see 6 Euros as a significant cost for troubleshooting this - worst comes to worst, you can always take the leads off the new CMOS battery and use the cell in your car's keyless entry remote (most of the ones I've been inside have used 2032 cells, which is just a slightly thicker 2025).

[dcouzin]

There may be consternation here that one of us would present a problem to the rest which the former doesn't think worth 6 Euros to solve. Our struggles in these strands are not much about money but about making things right. rkawakami had already explained why the symptoms weren't from the CMOS battery. If the 1.6 volt battery managed to keep the time and date, then how could it be reponsible for the failure to get past the IBM splash screen? Well, I found a battery for 2.5 Euros and installed it and thisdidn't solve the problem. In fact I got all black screen (no Thinkpad splash, no flashing cursor) through 20 start attempts. (I wait between 5 and 10 seconds between.) The fan came on and the indicator lights went through a cycle, but no more. I then left it on for an hour. It literally warmed up. Next power cycle it went to time and date setting (since the CMOS battery had been removed), and then requested power off and on, whence it proceeded to Windows. This need for warmup is my 600X's problem.
I asked rhetorically whether such a machine is worth 6 Euros. This was not a sales pitch. Bill Morrow would have shuttled that to the Marketplace forum. Obviously the machine is worth more than 6 Euros as parts. But one doesn't add a new CMOS battery to a parts machine. The question was whether a laptop that one is never sure will start is worth having as a laptop. If the one hour warmup routine works consistently, then I'm glad I added the 2.5 Euro battery.
Incidentally, there is a nice technique for attaching the leads to the battery. The original battery's tabs are held by a few laser spot welds (soldering is out of the question). After prying up the tabs, hammer them very smooth. Then bend them with about 2 mm of bow and epoxy them to the battery just at their ends. The tab metal is springy, so their middles then make good contact with the battery.
One can own a Thinkpad without knowing what a "car's keyless entry remote" is. We're still searching for our common link to the bad black 'Pad.

[Rob Mayercik]

I see your point, and I understand where you're coming from.

All I was trying to get across is that for a cell that reads 3.0-3.2V new, 1.6V is very low. The first time mine failed, I got error codes for the motherboard and CPU fan, so I was thinking that it might be causing spurious other faults/fault codes on your machine.

Now, as to your comment about my keyless entry remote remark:

[quote=dcouzin]One can own a Thinkpad without knowing what a "car's keyless entry remote" is.[/quote]

I thought I was offering a practical use for a 2025 cell if you ended up buying one and it didn't help. Most cars these days have remotes for locking/unlocking, and many of them use the 2016/2025/2032 (same diameter, same voltage, but increasing thickness/capacity) cells. When I read the above remark, I got the distinct impression that you were offended by this. Despite the fact that I cannot fathom why this would be offensive to anyone, offense was not my intent.

Moving on to matters of more interest, your epoxy idea intrigues me - with the bow bent into the leads, what is your suggestion for pinning the lead ends down while the epoxy sets up? I would think that for any sort of contact tension, the ends of the bow would have to be pushed down almost to the cell surface. Do you advocate a "do one end, let it set up, and then do the other" approach? I have to admit, the idea of simply using some sort of conductive epoxy has crossed my mind in the past (got the idea from the car window defroster trace repair kits), but never bothered to look into it, but this seems like it might be just as effective with a more commonly available adhesive.

[/quote]

[dcouzin]

I used 5 minute epoxy, tacking one end first, then the other.
I've had poor luck with conductive epoxy. Perhaps the silver can migrate away from the interface -- I was attaching carbon fibers to aluminum. Conductive epoxy is also quite expensive, so you'd better have a use for the remainder.
The CMOS battery complete with leads and connector is just $5.60 (including shipping) on Ebay.