t21 power on problem. (with a twist!)

Message

dara212 · #1 Post by **dara212** » Thu Nov 30, 2006 10:50 pm

I am new to this forum and a bit of a novice so please bear with me. By reading through the other posts i see my problem is common. when i press the power button i get a quick flash of a led and a slight pop noise thru the speaker then nothing. I performed the bleed off procedure but it would only work for a day or two. Now here is the strange part. I connected a motherboard analyzer to the mini pci slot and the computer would power on every time. But when I removed it it was back to the same old "NO POWER ON". I also noticed that when the mini pci slot was empty if I gently touched pin #17 with a small paper clip it would power on every time with no problem. Maybe somebody with a lot more knowledge than me can explain this weird condition.

3-t22
2-t21
1-a30
2-a22m
2-a21m
1-t23
1-560z
2-600e

#2 Post by **rkawakami** » Thu Nov 30, 2006 10:57 pm

This needs futher explanation... What is the manufacturer/model number of the motherboard analyzer? Is the paper clip ONLY touching pin #17 and not anything else (i.e., adjacent pin(s))? And HOW did you determine it is/was only pin #17? I've got a T22 with this problem...

BTW, Welcome to thinkpads.com!!

dara212 · #3 Post by **dara212** » Thu Nov 30, 2006 11:29 pm

The PC analyzer is manufactured by Huatong Electronics Co. http://www.61131568.com/. I made sure the paper clip was only touching the pin and nothing else. I thought it might be using the resistance in my body to get to ground so i slipped on a nitrile glove. It still powered on with no problems. It took awhile but I was able to count the pins starting from the notch side. It was 17 pins (including the single pin before the notch. I did not push on the pin I just "breezed" against it. By the way I went down to my shop to get the computer and try it again. I'm using it right now.

vlyne · #4 Post by **vlyne** » Fri Dec 01, 2006 12:03 am

That's intriguing!! Assuming your pin 17 is the 17th pin along, by my calculation that's the ground pin. If so, you're basically discharging to ground. The thing to do next is to try and measure if possible the signal/voltage at that pin. It's a variant of the discharge trick I think. Talk about voodoo dolls and pins!!
Cheers

#5 Post by **rkawakami** » Fri Dec 01, 2006 12:58 am

17th pin on the top side of the 124 pin Mini-PCI (counting the one pin before the "notch") is AD31 (address/data 31) according to this spec:

Mini-PCI pinout document

I believe that the pin numbering system is odd on one side, even on the opposite side. This is based upon the number "123" appearing at the "right" end of the Mini-PCI connector on my T23. AD31 is the most significant address/data pin. Placing any metal on that pin would tend to increase the capacitance of that pin/signal.

(edit: Just realized this... If you are touching the socket pin closest to the motherboard, then that would make it pin PME#, which the spec defines as a PCI power bus management signal. Please clarify exactly which pin you are touching.)

vlyne · #6 Post by **vlyne** » Fri Dec 01, 2006 1:18 am

My error. I was looking at the 100 pin mini-pci. Ok, on the 124 pinout, the 17th pin is PME# - Power Management Event. Pin count starts from 2 which makes it pin 34. The pin next to it (the 16th is ground). The resistance between 17th pin and ground is 0.524K ohms. A bit more probing to do...
Cheers

vlyne · #7 Post by **vlyne** » Fri Dec 01, 2006 1:36 am

Ahha! So this pin is the one that controls the Wake-On-LAN function.

Edit: For reference, have you enabled the Wake-On-LAN BIOS setting?

#8 Post by **rkawakami** » Fri Dec 01, 2006 2:59 am

Pulled the old T22 out from under my desk where it has been living since the last time it decided to work for me for about a dozen turn-ons. Everything was intact except for the disk drive (both batteries were still in place). Tried to turn it on... still blinks, well that's good! Removed the modem/ethernet card and tried to power up; still blinks. Touched the 17th trace (PME#) with a fully metal screwdriver and tried to turn it on... blink. Touched PME# with just a paper clip (fully insulated from my hand)... blink again. It's still blinking after several attempts touching several pins on both sides of the PME# pin. I'm going to pull both batteries and let it sit again.

As far as the Wake-on-LAN function, I don't know if it is enabled or not on this T22. I remember diving into the BIOS during the last session to reset the date and time, but I can't be certain on the status of that feature. If and when it decides to power up again, I will go in and check it.

vlyne · #9 Post by **vlyne** » Fri Dec 01, 2006 4:46 am

Didn't work for me either.

I guess if the Wake-On-Lan is enabled we could jumper 3.3V from one of the pins (124?) and see if it does wake up. It's not earthed and with a resistance of around 0.5K ohm it should be OK but perhaps the current should be routed through a 1K ohm resistor just for safety. On a blinking A21m that pin has about 20K ohm resistance and there's a diode-like behaviour with earth - which is what I would expect as this pin is an input.

These symptoms are really odd. It's almost like there's a floating earth in the system.

Ah well, we've learnt something new in the process - at least I have.
Cheers

dara212 · #10 Post by **dara212** » Fri Dec 01, 2006 7:33 am

I don't have a schematic for the motherboard so I am calling it pin 17 based on its position. With the bottom of the unit facing you count the pins starting from the notch side. this pin is the most visible one attached to the board. That's the best way I can describe it without posting a pic.

vlyne · #11 Post by **vlyne** » Fri Dec 01, 2006 7:41 am

I'm not sure what you mean by visible. The ground pins "stand out" because they have a broad track that extends out from the pin. Your pin 17 is just to the left of one of these (looking at the slot with the notch to the right). I count 17 pins from the right (and proceeding to the left) starting at the designated "pin 2" to the right of the notch. Your pin 17 is actually not all that "visible" which makes me think we're not seeing the same pin?
Cheers

dara212 · #12 Post by **dara212** » Fri Dec 01, 2006 10:19 am

My explanation might be a bit misleading, sorry. When I get a chance I will post a pic with a arrow pointing to exactly the pin I am talking about. If somebody can provide me with a diagram as to the layout and pin designations of a t21 motherboard more so the mini pci slot I would greatly appreciate it. the row of pins that include this pin do extend and attach to board.

#13 Post by **rkawakami** » Fri Dec 01, 2006 12:01 pm

Okay, I've cropped a scan I made of a T23 motherboard to only show the Mini-PCI slot. The image is here:

Mini-PCI slot on T23 (315x600, 60KB)

Pin numbering is shown on the motherboard. The left side pins are odd, the right side, even. The orientation of this picture is if you have the front of the laptop at the "bottom" of the picture.

dara212 · #14 Post by **dara212** » Fri Dec 01, 2006 12:46 pm

According to the pic it would be pin #32 which is exactly in line with the (j22). By the way when I originally discovered this weird phenomenon the unit was stripped down, including key board with only a 64mb stick of memory and running on ac power. It also worked when fully assembled (less the Li-on batt.). I even had turn on by slowly running my finger over the terminals. I know this may sound far fetched but I couldn't believe it actually worked. I hope somebody a lot smarter than me can use this to come up with a "work around" for this all to common problem.

vlyne · #15 Post by **vlyne** » Fri Dec 01, 2006 5:38 pm

Ok, were on the right track (or pin). It's just that the T21 circuit tracks are different from the T23 shown by Ray. But, the pin is the same functional one. So, it's settled that is the PME# pin. So the only question now is are we talking about a T21 or T23? I'm assuming T21, and you've just used Ray's image to show us how you counted the pins?

dara212: can you please check what the status of your Wake-On-Lan setting is in the BIOS. I can't recall exactly where this is but I think it is the "Power" option.

What I'm guessing is happening is that your innocent clip-pin is actually causing a discharge that activates the Wake-On-Lan circuit. If you hold the clip-pin to ground first and also discharge yourself (use ground wrist strap) I suspect you won't be able to perform the trick successfully. The danger with this method is that the discharge is going to be way way over the 3.3V limit and you may end up killing the power controller chip.

I'm going to be away for a couple of days so unfortunately I won't get around to trying the 3.3V wake up trick. Trust someone else has fun with this

Cheers

Edit: pin #32 is ground I think but I'm ignoring this and going by pin counts.

#16 Post by **rkawakami** » Fri Dec 01, 2006 7:09 pm

dara212 wrote:According to the pic it would be pin #32 which is exactly in line with the (j22).

That would be the 16th pin counting down the "right" side and make it a ground pin, from the information in the Mini PCI document that I had referenced earlier. The 17th pin is pin 34 (PME#). As shown in my scan, the topmost pin is pin 2, the next one down (after the gap) is pin 4, and continues to count by twos until the last (124th) pin.

I'll try scanning the bottom of my T21 tonight and post that picture so we can compare apples-to-apples. I only posted my T23 scan since I had it handy.

vlyne · #17 Post by **vlyne** » Sun Dec 03, 2006 1:59 am

I'm back and now I'm confused. So, we're now talking about the 16th pin (pin #32) which is a ground pin. Ray: I can post you a copy of a T21 slot if you are willing to host it on your site. But, from data212's previous description about the pin being "visible" (the most visible pin on the T21 is pin #32 - ground) and counting the same way with the T23 it now appears as if it is the ground pin.
Cheers

#18 Post by **rkawakami** » Sun Dec 03, 2006 11:44 am

Okay, with vlyne's kind efforts, here's a scan of his T21 Mini-PCI slot, arranged the same way as my T23's:

T21 Mini-PCI (287x800 78KB)

@dara212: I've marked what I think are pins 32 and 34. Please confirm which pin you are touching.

dara212 · #19 Post by **dara212** » Sun Dec 03, 2006 3:11 pm

The pin is definitly # 34

vlyne · #20 Post by **vlyne** » Sun Dec 03, 2006 5:26 pm

Thanks Ray for hosting the image and labelling the pins.

data212: I'm glad that's sorted out!

So it is the PME# pin. If anyone else with a blinking T20/21/22 wants to chance the clip-pin or finger test go ahead and let us know what you find. I'll rig up a 3.3V test when I get the chance. It would be useful to know what the PME# fires off on the motherboard - the Maxim chip?? There should also be an adjoining pin on the port replicator connector - so dara212, your finger test may also work on the port replicator connector.
Cheers

furiousmonkey · #21 Post by **furiousmonkey** » Tue Feb 20, 2007 3:25 pm

I am going to give this a try, this evening, any thoughts on soldering a small ground wire from this to a ground elsewhere on the motherboard, in theory this would solve the problem once and for all.

Anyone else had success grounding out that pin?

#22 Post by **rkawakami** » Tue Feb 20, 2007 4:31 pm

I tried touching the pin in question several times on my T22 but never got it to power up. This past weekend I was prompted to pull that system out from under the desk it's been sitting under and now it turns on! Details here:

http://forum.thinkpads.com/viewtopic.ph ... 436#252436

At this point I don't know what to think about this problem. From the earlier description about the system powering up when the PME# pin is touched with a paper clip, I would assume that putting a small capacitor on that pin to system ground would more closely duplicate the human body effect. Looking at the spec for the MiniPCI interface:

Mini PCI Specification (previous link I posted in November is now dead)

appears to say that the pin is active low. In other words, grounding the pin triggers the function. However, leaving it permanently grounded may cause other issues.

I now have a T21 with this problem so I'll give this a try.

furiousmonkey · #23 Post by **furiousmonkey** » Tue Feb 20, 2007 5:00 pm

I have been using my T21 at home for the past month, I dont shut it down ever, just "sleep" I always leave it plugged in, and I never have the problem of "blink of death". The other day I went to town, I put the laptop to "sleep" for fear of the "blink of death" but it made no difference, the moment I used it, it was back.

So in response to your question, never turn it off, or leave it unplugged, basically you have one of the smallest Personal Computers!

If that pin makes a differents I will be tempted to add a sort of "switch" to the main power button, that would enable a sort of flash to the pin and boot the system.

EDIT:
My system did nothing this morning, no flashing lights no nothing, totally dead, then I removed the battery and plugged it all back in and it worked.

I was wondering if we could find a pattern to this power on problem, like which models do it.

If your laptop does this, please post what model # you have.

phool@round · #24 Post by **phool@round** » Sat Feb 24, 2007 5:35 pm

I don't have a system to submit but I do have a question. Wouldn't this problem be indicative of something further upstream, say a faulty component or a short? I know this sounds silly in review but I thought I'd ask.

Also if people post they should post their BIOS version, never know if it's a bug in the control software which IBM could correct if they are made aware of, maybe. T23's are current as of this year.

I would be tracing back from pin 34. If a cap is being discharged with the paperclip or finger then that tells me something isn't going to ground and releasing the signal when power is off. The signal is staying alive until all power is discharged either by time or finger....lol.

I've found many boards needing a little soldering iron touch up come back from oblivion. The older they are the more I touch up. I don't recommend that everyone should do so but..... I would be making sure my power components are reflowed at least.

The question is where does this trace from pin 34 lead to first and then second and so on? One component has to evidence the fault. If I had a T21 I'd be hot on the trail of this one....... sure would be nice to have a (free) schematic right now.......

Like my sig I'm not so sure I'd be leaving this system on all the time even in a sleep state............ unless it's on a glass desk and away from all combustables. Another short is all it might take to become a Tiki Torch Think Pad. Not a bad sounding host name.......

t21 power on problem. (with a twist!)

t21 power on problem. (with a twist!)

My t21 does the same thing

an interesting add

Who is online