T20 Blink Of Death Circuit Questions

Message

wireman · #1 Post by **wireman** » Sat Sep 01, 2007 12:13 pm

Hi All ! ,
New to the forum having just obtained a T20 with the common BOD problem. Symptoms are :
If the battery is removed then after a while it will boot successfully from the mains and is fine. However if the battery is inserted then no boot, only the BOD and it no longer boots with the mains either regardless of whether the battery is in or not. Once or twice it did boot from the battery solely and worked for an hour or so until it switched itself off. It was long enough to get a CPU temperature and battery monitoring program to run. The CPU temp was between 40-60degs so I don't think the CPU is overheating. Besides the FAN works ok anyway. Also I bought a new battery for it and repleced the bios battery and ran memory tests using PC doctor repeatedly and all passed even while shaking the laptop around in the air.

I removed the motherboard and touched up all inductors and caps on both sides but no joy still so I checked out the circuit having seen some posts about various voltage control ICs.

Anyway from what I can make out the MAX1632 is working ok as if I raise the RUN/ON3 pin to 5v via a resistor the 3.3v activates correctly. Also if the TIME/ON5 pin is raised the 5v rail activates also.
These voltages are sent to the ADP3421 and ADP3410 ICs which I think generate the core voltage. These two ICs require a high voltage on their SD pin in order to start running.
My question is does anybody know where these pins are fed from ?
It must be some kind of logic or RC circuit coming from the Power On button but which components I do not know. Can anybody help ?

wireman · #2 Post by **wireman** » Sat Sep 01, 2007 12:53 pm

Just to add on reflection. The MAX1632 is in Standby Mode as the SHDN and SEQ pins are high and at VRef levels respectively.

The MAX1632 must be told to raise the 5v and 3.3v from the Power On circuit also. So whatever IC or component raises these pins probably raises the SD pin also on the ADP3421 IC.

#3 Post by **rkawakami** » Sat Sep 01, 2007 1:55 pm

Welcome to thinkpads.com!

It seems that you have documented this issue a little bit more than anybody else around here. Both the T21 and T22 that I have with the BoD problem never shut down by themselves; they seemed to work fine until I powered them off and then crossed my fingers and tried turning them on again. Sometime they would work, mostly they did not. The T22 I have had without power for several weeks still does not power up no matter if I use a battery only, AC only or both.

Your circuit diagnostic appears to be sound. My thought on the root cause has always been a loose/badly soldered component on the motherboard. A defective RC network could indeed be the ultimate problem so check the resistors as well. The only advice I have would be to trace back from those two MAX1632 pins and see where they lead. Not an easy thing to do when the circuit schematics are not available.

SMA · #4 Post by **SMA** » Sat Sep 01, 2007 9:17 pm

First, I would like to say that I am glad that someone finaly starts talking about the real stuff behind this BOD problem.
So, thanks a lot for the above contribution, wireman. Especially, that about the temperatures releaves me for a lot of speculations.

For myselves, I have the problem with a T21 board. The last time I had the board out I got interrupted, but it is still on my 'to do' list.
I have already replaced the board with a working one, but somehow, I just cannot let the problem go - I have to know why.
I made my first notes on this board in march 2005, and I still does not know what's wrong - guess it is about time to get this issue closed.

Personally, I do not belive that all machines with these symptoms has the same root cause. But the varius causes are hidden
behinde the fact that the machines does turn off as soon as something abnormal is happening to any of the power lines coming out of the max1632.
Also, the symptoms differs slightly. Ray's machines does stay on when they do turn on. Mine and wireman's does not stay on.

The reason why none or so few of the BOD problems has been solved - I belive - is that it is impossible to measure what is going on just before the shutdown.
It takes approximately 18 milliseconds for a max1904 to shutdown - might well be the same for a max1632.
At that speed, a multimeter cannot visualize what is happening. Neither can an ordinary oscilloscope. A storage-scope might be able to, but then
there would be the issue of having the error condition reset. One cannot wait days between each measurement.

I hope that no one will disaggre, when I am stating it as a fact, that it is the max1632 chip that is detecting something as an error and initiating the shutdown.
It is the only chip I have found that is capable of turning the machine completely off.

By reading the datasheet for the max1632, I have discovered that maxim actually did anticipate, that it would be impossible for designers to design
boards with this chip, as designers would not be able to measure what was going on, if their experiments would not turn out as expected.
For this reason maxim actually made a chip similar to the max1632, but without the protection circuitry. It is named max1635.
One might wonder why maxim did not expected similar problems for repair people !
Maxim recommends using max1635 during development and max1632 for production - ibm seems to have followed that recommendation.

A side note here - I am currently working on a Toshiba made in 2004.
It also uses a maxim chip to control the powerup sequencing. It is similar to, but a newer generation of, the max1632/max1635.
It is from the max1901/max1904 series. The max1901 does have the protection circuitry, the max1904 does not.
The strange thing is that Toshiba designers have not followed the maxim recommendation of using the max1901 in the final product.
The chip actually found in this Toshiba machine is a max1904 - the one without the protection circuitry.
I belive that if Toshiba designers had chosen the max1901 for this model, then the machine I am working on, would now be showing the BOD symptoms.

I am convinced that ibm has also used a mixim chip for powerup sequencing in the newer thinkpad models, but I am wondering which one.
I just bought a defect T40 on ebay earlier today. I just have to know which mixim chip is in there.

What I think would be necessesary to do, to get further into this problem, is to find a way of disabling this shutdown.
Once the protection circuity has been disabled, the real error might reveal itselves by sending out a smoke signal or at least making it possible to do
some ordinary voltage measurements within the power circuitry.
It is even possible that the machines will run without any problems, if the shutdowns are only initiated due to a to restrictive protection circuitry.

My last experiment was to replace the adp3410. It did not help.
But I did notice one thing. While the adp3410 chip was removed, I could get power on the board. Both the 5V and the 3.3V was there steadily and the max1632 did not shutdown.
My next experiment will be to try to power up the board with the cpu removed. (The purpose of the adp3410 is to generate the core voltage for the cpu).
Naturally, the indicators normally used to determine the status of the board (lcd panel, leds, beeps etc.) will not be of any use then.
The status of the board will have to be determined by voltage measurements or similar.
I use a laboratory power supply to power the board. It has a display showing the amount of current drawn by the board. Enough for me to tell whether the board is on or off.
I have a strong feeling that it will be possible to keep the max1632 running by removing the cpu.

Last time I was doing measurements on this motherboard, I was using a setup, where I had attached the motherboard directly to a port-replicator.
No other thinkpad components involved. The motherboard directly on the port replicator.
Did not even have a fan on the cpu. Just kept an eye on the meter on the power supply and a finger on the cpu, so that I
would be warned, should the cpu temperature start rising - which it did not (the adp3410 was removed).

If only I could find a max1635, I would try to replace the max1632 in my board with it, just to see what will happen.

In my opinion, it does not matter how the machine is being power when trying to start it up.
Removing the power supply/battery only affectes resetting the error condition held by the max1632.
When the power adapter is attached or the battery is inserted the max1632 will always receive power and therefor it will not reset a previusly
detected error condition - even if the error condition has been releaved.
Also, the max1632 will not reset immediately when power sources are removed from the machine, as the capacitors on at least one of the 5V and 3.3V lines will hold
voltages for a long time (can be several days).
The datasheet for the max1632 says this clearly. All voltages must disappear before an error condition will be cleared.

Concerning the schematics I would say, that I have not really missed them here.
The datasheet's for the max1632 and the adp3421 does contain sample schematics on how these chips are intended to be implemented.
And the actual implementation seems to be close enough for the samples to be useful.

Speeder.dk · #5 Post by **Speeder.dk** » Sun Sep 02, 2007 5:51 am

Count me in as well... My T23 does the same as you guys mention...
Anyways, guess it wont help to replace the Mobo, since it seems like that the Maxim chip is causing the problem and not just a defect mobo? Or am I wrong ...

/Speeder

SMA · #6 Post by **SMA** » Sun Sep 02, 2007 8:36 am

Replacing the motherboard will help.
Even though there are a number of cases with boards that has this problem, it is still just a very small percentage of all boards ever made.
Finding a working board should not be a problem.

wireman · #7 Post by **wireman** » Sun Sep 02, 2007 10:25 am

Thanks for all the replies guys. Seems I'm not the only one who can't leave problems alone.
SMA seems to have spent quite a bit of time on this too and thank you for your information. I had to read it several times as you had so much done already on this.
I have done a bit more looking at this and from what I can see the schematic for the ADP3421 on the back page of its datasheet does not agree with the T20 circuit in that the SD pins are not tied together.
From my T20 It seems that the ADP3410 pin 7 ( Vcc Good) is supplying the pin 15 ( Shutdown ) of the ADP3421. The VCC good signal is generated when 5v is active and SD is high in the 3410. However I have tried to lift the SD pin on the 3410 to activate VCC Good but am unable to as the pads are very flimsy and am worried it will lift and break.
Anyway for me this comes back to this SD pin on the ADP3410 and to where its being activated from. I have buzzed out the board all over to try and find its source but have been unsuccessful so far. Any other ideas ?

aditya1956 · #8 Post by **aditya1956** » Sun Sep 02, 2007 1:26 pm

I had almost similar problems with my T20. I solved it by this.......

Go to BIOS
Press F9 to "Load the Default Settings"
Then go thru Config > Power > to "Intel Speedstep"
Disable it
Press F10 to "Save & Exit"

It may/should work for you.

Actually it is another way to manipulate with the PWR MNGMT isuues of the mobo without doing any hardware job, becoz amongst other things the Speed_Step thing also controls the CPU voltages.

As you have already dismantled the machine, take it as the oppurtunity to clean up all the mating points.

Pls leave a feedback.

Thanks

Speeder.dk · #9 Post by **Speeder.dk** » Mon Sep 03, 2007 10:34 am

Nice Aditya, will check if that works later... Just dissasembled my t23 completely and found a Max1718 chip underneath the mobo, which is a
"Notebook CPU Step-Down Controller for Intel
Mobile Voltage Positioning (IMVP-II)"
... Heh, just finds it a bit funny to know what the chips does:P

(and thanks for the info SMA)

/Speeder

wireman · #10 Post by **wireman** » Mon Sep 03, 2007 12:44 pm

well I think I just blew the motherboard !

Accidentally shorted the VL line to the V+ line and now the T20 just gives a high pitched squeak when I press the On button.

ebay here I come.

SMA · #11 Post by **SMA** » Mon Sep 03, 2007 1:20 pm

What a shame, that you have blown that board.

And that being now, where I think I might have found the answer to your original question.

Did some research on an A20 board (did not have a T board on hand) and found that adp3421 pin 15, the SD pin goes through a 10K resistor to pin 7 (VCCGD) on the adp3410. Might well be the same on a T2x board.

#12 Post by **rkawakami** » Mon Sep 03, 2007 2:07 pm

Speeder.dk wrote:... Heh, just finds it a bit funny to know what the chips does:P

If you need information on some of the parts on the T23 motherboard, try this: http://forum.thinkpads.com/viewtopic.php?t=37956

I have not updated that thread in some time.

I also started a thread for the T21 motherboard but have not had time to go through the board and look up the "interesting" pieces:

http://forum.thinkpads.com/viewtopic.php?t=47660

wireman · #13 Post by **wireman** » Mon Sep 03, 2007 3:12 pm

Hi SMA,

Thanks for that , It confirmed what I measured earlier.
Would you know where the SD pin on the ADP3410 originates from.
This is where I was getting stuck earlier. I would still like to know though.
If this raises high then the ADP3421 would activate also. Perhaps then the board would have powered up.

evo4jon · #14 Post by **evo4jon** » Fri Sep 21, 2007 7:49 pm

Hi,

I had the BOD on my T22 and i bought another mobo.... installed it and it was great for 2hrs on mains and then later on battery. Suddenly the laptop switched itself off and all i got on boot up was an error 175 which indicated the Bois chip was corrupted.

I mentioned this to my supplier who sent me out another mobo and after fitting it my laptop worked perfectly for 7 hrs....
....Next day I tried to switch it on and i got a BOD AGAIN !!!!!!!

I stripped it down and checked the connector board along the back of my case and all the joints seemed fine and the micro switches all worked fine.

THIS IS GETTING STUPID ... 3 Mother boards and all of them are faulty ??? This sounds very odd... surely they can't all be faulty ?

There is nothing else that it can be though ? right ?

QUESTION: Do I buy yet another motherboard or am I wasting my time ?

Jonathan

Robbyrobot · #15 Post by **Robbyrobot** » Sat Sep 22, 2007 1:42 am

QUESTION: Do I buy yet another motherboard or am I wasting my time ?

Unfortunately an unanswerable question. The basic dilemma is that we (I have two T21s with the same BOD symptoms) are experiencing what is apparently a common, mainboard-related problem whose root cause is unknown. And as long as it remains unknown, buying and installing a new (used - there are no really new ones to be had) mainboard may or may not solve it for an unknown/unspecified length of time. I also wish the situation were different.

Robbyrobot · #16 Post by **Robbyrobot** » Sun Sep 23, 2007 2:10 am

I hate to prove my own point, but unfortunately have done so. Bought a T22 base, including the mainboard, described as "in excellent condition and fully functional". My hope was to replace the mainboard in one of my "BOD" T21s.

Well you can probably already guess what happened...

I added the necessary components, started the computer and everything appeared OK. Updated the BIOS and did all the checks possible with PC-Doctor. Then, since I'm a suspicious person, I turned the system off but left the powered AC adapter in its plug in the base overnight.

The next morning, the computer wouldn't start and exhibited the "BOD" symptoms

.

Of course, this being a sporadic problem, I can't even assume the seller knew about it... sometimes "BOD" computers operate normally for quite a length of time (in my experience up to around 10 days). So now I have to contact the seller and see what can be done: a nasty situation for all involved.

It's not much satisfaction that the point I originally made, that a "new mainboard" is an uncertain cure for the "BOD" as long as the root cause is unknown, has been borne out here 100%. I would have liked to be wrong.

Website · #17 Post by **tfflivemb2** » Sun Sep 23, 2007 8:17 am

For these problems being exhibited on several systemboards being placed into the same base, I would begin to suspect something else... I haven't taken apart a T20-T22 in quite some time, but remember correctly, it is quite similar to the T23, which I have done many times...I would check the wires that connect to the systemboard, along with ANYTHING that attaches to the board, for any kind of potential shorting...

Robbyrobot · #18 Post by **Robbyrobot** » Sun Sep 23, 2007 10:50 am

For these problems being exhibited on several systemboards being placed into the same base, I would begin to suspect something else...

Good thought, but with one exception, all of the BOD mainboards I've experienced - including the one I last described - were bought together with the base. All I did was add a CPU, fan, one RAM stick and attach a working LCD display for testing.

One thing I might add is that in the last-described case, there is a soft "pop" noise and the power light flickers for an instant when the on/off button is pressed after removing the AC adapter (and always without the battery). This would certainly agree with the idea that a capacitor must have something to do with the problem, since the system has no outside power after the AC adapter is removed.

You know, I've never seen anyone describe "reflowing" of a T2x board with BOD similar to what some do with T4x boards with the "flexing" problem, but if everything else fails, that might be an interesting thing to try. Ray has always contended that he thought BOD was due to some part that became loose or separated from the mainboard, and such treatment might be effective in fixing such a problem.

evo4jon · #19 Post by **evo4jon** » Sun Sep 23, 2007 5:39 pm

OK...I also thought that maybe the chassis / mainboard connections could be a possible cause of BOD, so yesterday I stripped down the T22 containg the latest mobo which apparently died after 7 hours and looked at all connections and looked at the chassis board... put it all back together and it was still dead and exhibiting BOD signs.

THE BIG QUESTION IS....

Does anyone think that cause maybe related to the CPU ? Could it start to breakdown and then draw more current thru the Mobo that is expected - thus frying some component and render the Mobo dead ?

Jonathan

#20 Post by **rkawakami** » Sun Sep 23, 2007 6:50 pm

One of the symptoms of the BoD can be that the system will eventually power up normally. This points to an intermittent problem with a part or a connection to a component on the motherboard. If something got "fried", then the system would never boot at all. Most of the systems that I've heard of (and the two that I have) do or have come back to life for a couple of boots and then returned to the BoD state.

Rick the Green · #21 Post by **Rick the Green** » Sun Sep 23, 2007 7:28 pm

I don't have a t20 with the BOD syptom, but here's some thoughts that may be worth checking :

Electrolytic capacitor : i have fixed way back then 2 Toshiba 486's laptop with startup problems, Large electrolytic had dried out in the power circuitry or that had leaked on the board, electrolyte from from those capacitors will eat up traces on the board and/or will create a leakage between two trace. Very common problem in appliances that see a lot of heat (amplifier powered on 24/7) or a Laptop.

Large smd resistors : i've seen cracked solder on those because of heat or flexing. This are the first thing that i look for in Car Radio with surface mount component. This is kind of like the inductors trouble on T23's.

my two cent's

Rick

evo4jon · #22 Post by **evo4jon** » Mon Sep 24, 2007 7:12 pm

An Update on my situation...

I have just taken out the 2nd board I was sent (which failed after 7hrs use) in order to send it back to the supplier along with the first one which lasted 2hrs before dying and giving an error 175.

The original board that failed is now back in the machine after resoldering any power related sockets / componants - stuff like the battery connector and 4 pin mains connector + a few surface mount resistors etc...

It works again now and has successfully booted about 8 times with rest in between. Time will tell if the bits I soldered have had any effect long-term.

I can conclude from this that its DEFINITELY the Mobo and not the chassis and very likely to be dry joints that cause BOD IMO.

Having examined the other Mobo's i was sent, I could see that they were not new and had been resoldered in a few power related places and i think that this random failing is most probably not helped by the fact that companies are reselling reconditioned faulty boards. The fact that the solder has been dried out in those areas which see the most heat means that the symptom is alway power related but can be a variety of actual causes.

I don't think that there is a solution to BOD except to try and buy stock that has not seen long hours of use and therfore not been exposed to the heat to dry out the joints.

Anyone care to comment / blow my theory to bits etc ?

jonathan

#23 Post by **Harryc** » Mon Sep 24, 2007 7:27 pm

I like your theory. If it is indeed fixed it would be interesting to all if you could itemize which components that you resoldered. If you are correct the main issue will be to try to find a parts source for fresh T2X systemboards. Does anyone know of a decent, reliable source for new/er T2X boards? EBay is obviously not the right answer.

#24 Post by **rkawakami** » Mon Sep 24, 2007 8:54 pm

Rick the Green wrote:<snip> Electrolytic capacitor <snip>

A very common problem with the big "cans" a few years back when a Far East (Taiwan?) supplier shipped out a bunch of defective capacitors. They would work for a few days/weeks but would quickly leak or explode. I have not seen any evidence like this on the one T21 I stripped down.

Rick the Green wrote:Large smd resistors <snip>

This happened to me with the factory radio in my daughter's Jeep Grand Cheorkee. The previous owner said that although the radio worked, the fluorescent display did not. Before buying it I searched around the net and found that this was a common problem with those radios. It seems like a 10 ohm (I think) surface mount resistor near the back corner gets very hot and loosens up over time. The end result is a blank display. Bought the Jeep, took the radio out, popped off the lid and, yup, there was a half-on, half-off SMT resistor. Soldered it back on and the display worked perfectly.

tom_k · #25 Post by **tom_k** » Mon Sep 24, 2007 11:01 pm

Just some more info to solve the riddle, but no solution:

I met the bod several times, in one case from the start.
1st occasionally refusing to power up
until 1.5 years later it now lies in (nearly)total stasis
(sometimes if i do the hoobawooba ritual with the power button it will awake;)

During that time it did never shut down by itself, freeze or bsod,
because of that i had often let it run for weeks. Stable.
Also it did never refuse a warmboot !

Other one:
Got a T20 for repair und upgrade, the user assumed a bod.
Removing the first screws i thouht by myself "test it before you strip it down"
This testing did really last :D over hundred Attemps during several days showed no symptom of bod;)
Nearly convinced that is was ok i powered it up before writing a mail to the owner.
Should say i tried to power it up :D

I still have the boards, along with several other BoD´s
(but only a standard multimeter in case you`d like me to check something)

Faulty cpu can be nearly excluded, confirmed with several 100% working CPU`s (tested in other T2x)
Many of the bod boards are able to load a Battery.
They usually dont show error beeps witout ram.

tom_k

evo4jon · #26 Post by **evo4jon** » Tue Sep 25, 2007 7:01 pm

Since I posted that my soldering may have fixed the Mobo, I have had my laptop on test for over 24 hrs now.... it was switched off for 5 hrs and then back on for about 4... rebooted and randomly used with power and with battery... It seems 100%.

to see the parts of the board i resoldered, go here:

http://www.btinternet.com/~jcolling/lap ... ations.jpg

If anyone else could try to solder the areas and post there results it would help to establish whether I was just lucky / it was a fluke / it works etc...

Cheers

Jonathan

#27 Post by **Harryc** » Tue Sep 25, 2007 7:23 pm

Jonathan, thanks for taking the time to do this. Did you just heat up the solder points enough to reflow it or did you remove it (with a desolder braid) and resolder? If you were successful, I suppose the next step (for someone) would be to resolder each component one at a time ... a lot of work but if the component could be isolated it would be worth it to the community.

evo4jon · #28 Post by **evo4jon** » Tue Sep 25, 2007 7:32 pm

No Probs

I just reflowed using a hot soldering iron and added a little new solder if i thought that I could get away with it

(by this I mean, making sure that a big blob of solder did not land on other areas of the board to cause a short) .

The pins of the connecters looked like they could do with more solder and i noted that there were often connections on both sides of the board so a good hot soldering iron is a must inorder to transmit the heat to the flip side of the board.

Jonathan

#29 Post by **Harryc** » Tue Sep 25, 2007 7:36 pm

Well congrats, job well done. I think you're the first one to fix a BOD board. I can almost smell the soldering irons heating up across the world as we speak

Let us know (if it's not too much trouble ) how the machine operates over the next several days.

#30 Post by **rkawakami** » Tue Sep 25, 2007 8:32 pm

Has anybody been able to view evo4jon's image file? Numerous attempts to click on the link have led me to rtg.com (DNS problem with our company's ISP?)

T20 Blink Of Death Circuit Questions

T20 Blink Of Death Circuit Questions

Who is online