6800le 'Black Screen' - The Complete Guide: Identification, modification and reason

[KoR_Wraith]

Hello

- A MSI version 1 card, with "4 3 470 6E" capacitors
- Both earlier proposed capacitor modifications have been tested, but not worked for a long time
- Without any modification maximum usable (=no BS) clock frequency was 420 MHz !!!
- After replacing "4 3 470 6E" with Chemi-con's PXA-capacitors (820 uF/6.3V) BS was removed !!
- A memory voltage could be raised at least to 814 MHz
- Case stydy confirms: A CARD EVEN VERY BAD CONDITION CAN BE REPAIRED USING THIS MODIFICATION

I have exactly the same problem (also the same card) - I did the enhanced capacitor and it worked for a few months then BS again, so I replaced it and it worked for a few weeks. Have repeated this a few times - each time the time it works becomes increasingly short, now it only works for around a day!

I am in no way qualified to perform the ultimate capacitor modification but my uncle is very experienced in this area - having replaced capacitors and other components in his job.

((Maybe some free GOOD quality replacement capacitor for 'Ultimate capacitor modifiaction' still available (see Ch. 3.3))

I was just wondering if you had any left? Thanks

[SikaRippa]

Hello



I have exactly the same problem (also the same card) - I did the enhanced capacitor and it worked for a few months then BS again, so I replaced it and it worked for a few weeks. Have repeated this a few times - each time the time it works becomes increasingly short, now it only works for around a day!

I am in no way qualified to perform the ultimate capacitor modification but my uncle is very experienced in this area - having replaced capacitors and other components in his job.



I was just wondering if you had any left? Thanks

Yep. MSI-cards (very first models) will work only for a very short period with added capacitor - your case proofs that 'theory' once again !!

--

Have a few..already sent PM for you

[xan]

Hi, I was watching this thread for a quite long time recently, now I decided to post.
Here's my story ;|
I bought MSI 6800LE (v2 - with better cooling). It came with 4 6 470 6E capacitors.
It worked fine for a moment. After some hours spent playing DM3 i suddenly had a bs.
I thought that it was something wrong with other components in my PC, started playing again and everything was correct for some days.
When it came to overclocking - core overclocked very well (~370Mhz @ stock - 1.1v or 1.2) but when i rised memories over stock 700Mhz it gave me bs (immediately, in RivaTuner!)
I thought it was all about memory timings - so i flashed card with 16/6 @ +0.1v BIOS with somewhat looser timings. The problem was not resolved - card was bs'ing accidently. (core went up to 460Mhz )
I made some research again and found this very useful topic - which helped me a lot.
I put some 3300uF 6.3v cap into TP9-TP10. The problem was solved! Memory was going up to 900Mhz. I was very satisfied with this mod.

After a month the nightmare returned. I thought that underclocking card would help, but it changed nothing.

I RMA'd the card and after a month i got my card back with new caps. It worked for two weeks.
Here are the photos:


Here's my question... Can i do anything else to solve this problem? Maybe self-soldering Chemi-con's PXAs and combining it with TP9-TP10 trick would help?

[SikaRippa]

Hi, I was watching this thread for a quite long time recently, now I decided to post.
Here's my story ;|
I bought MSI 6800LE (v2 - with better cooling). It came with 4 6 470 6E capacitors.
It worked fine for a moment. After some hours spent playing DM3 i suddenly had a bs.
I thought that it was something wrong with other components in my PC, started playing again and everything was correct for some days.
When it came to overclocking - core overclocked very well (~370Mhz @ stock - 1.1v or 1.2) but when i rised memories over stock 700Mhz it gave me bs (immediately, in RivaTuner!)
I thought it was all about memory timings - so i flashed card with 16/6 @ +0.1v BIOS with somewhat looser timings. The problem was not resolved - card was bs'ing accidently. (core went up to 460Mhz )
I made some research again and found this very useful topic - which helped me a lot.
I put some 3300uF 6.3v cap into TP9-TP10. The problem was solved! Memory was going up to 900Mhz. I was very satisfied with this mod.

After a month the nightmare returned. I thought that underclocking card would help, but it changed nothing.

I RMA'd the card and after a month i got my card back with new caps. It worked for two weeks.
Here are the photos:
Here's my question... Can i do anything else to solve this problem? Maybe self-soldering Chemi-con's PXAs and combining it with TP9-TP10 trick would help?

Hi,

Thank you for interesting story, with non-happy end

--

I think everybody starting to be completely tired with MSI-cards:
THERE ARE STILL PROBLEMS AFTER CHANGING CAPACITORS

WHY ?? I have some theories, but before I'll post them in detail it would be extremely helpfull if you could provide following info/details about your card:

- You have now BS at stock frequencies ?
- There is a new heatsink above an inductor (grey metal besides purple capacitors), is it original or have you put it ??
- When looking photo extremely carefully one can seen that there have been something (like added capacitor?) between TP10-TP9, could you confirm that ?
- Is it possible to provide a good quality photos also bottom side of the card, I am especially interesting about this area under this power circuitry (and misaligned capacitors), see below:



--

At first point I do not suggest replacing capacitors, since if only two weeks old (?) they SHOULD be OK.

I think that this card has been modified afterwards (can been easily seen due to poor soldering quality of "1200 4V"-capacitors) and they have failed somehow repairing.. I wonder we could analyze why the failed, if you can provide enough info. BTW do you have a multimeter with capacitance measurement option ?

--

Many questions, I hope you could answer at least some

[xan]

- You have now BS at stock frequencies ?

yes, i got them playing games.

- There is a new heatsink above an inductor (grey metal besides purple capacitors), is it original or have you put it ??

I have put it, becouse it increased stability over 400Mhz @ core

- When looking photo extremely carefully one can seen that there have been something (like added capacitor?) between TP10-TP9, could you confirm that ?

yeah, it was 3300uF 6.3v cap which i used as a first cure for my card

- Is it possible to provide a good quality photos also bottom side of the card, I am especially interesting about this area under this power circuitry (and misaligned capacitors), see below:

[to be put later]

I think that this card has been modified afterwards (can been easily seen due to poor soldering quality of "1200 4V"-capacitors) and they have failed somehow repairing.. I wonder we could analyze why the failed, if you can provide enough info. BTW do you have a multimeter with capacitance measurement option ?

i'll visit my friend today to make the measurements.

Many questions, I hope you could answer at least some

I'll try to do my best.

[SikaRippa]

Thanks for comments.

yes, i got them playing games.

I have put it, becouse it increased stability over 400Mhz @ core

So it was running too hot (did you test temperature), have you seen those threads about that issue ?

i'll visit my friend today to make the measurements.

The important capacitors (AFAIK) are for those tiny feedback capacitors. It is quite tricky to measure those (take a look end of document, post #1), but if you are willing to help please comment and I will give more detailed instructions.

yeah, it was 3300uF 6.3v cap which i used as a first cure for my card

Did adding capacitor help at all ?

[xan]

So it was running too hot (did you test temperature), have you seen those threads about that issue ?

I observed some unstability, during some investigation I noticed that it's very hot. Then i put this small radiator and core become stable >400Mhz

Did adding capacitor help at all ?

Yep, it helped for a some time, with old caps. When a card came with new caps i haven't performed TP9-TP10 mod, untill it gave me bs. After some bss i put cap into TP9-TP10 and it's working good for some now... (put it just today)

[xan]

I think that's the best photo which I can provide for today.
[img=http://img72.imageshack.us/img72/8375/p10105519wg.th.jpg]

mportant capacitors (AFAIK) are for those tiny feedback capacitors. It is quite tricky to measure those (take a look end of document, post #1), but if you are willing to help please comment and I will give more detailed instructions.

Could you give me more detailed info?

[Sidolin]

Hmm, could it help to put a small fan on the capacitors? I dont want them to die as soon as the old capacitors

[robb-x]

so..solution to the problem (for very nasty msi cards ) is changing capacitors, put extra cap in TP10,9 and a fan in order to cool capacitors down?!!

if I have to do so i'll do...but...it would have been better if the card could work without any help

p.s:my card is still working (everything modded)

[SikaRippa]

I think that's the best photo which I can provide for today.
[img=http://img72.imageshack.us/img72/8375/p10105519wg.th.jpg]


Could you give me more detailed info?

Thanks for foto & info. The barcode is missing in foto, so I can not say whethet it is old (one of the very first cards) and modified afterwards. I suppose those tiny capacitors are not misaligned, but impossible to say 100 % sure with this foto..

If this card is one of the first MSI-cards (=had "4 3 470 6E"-capacitors) and MSI has replaced CV-EX-capacitors with SVP's (green->purple) WITHOUT altering feedback response (those small capacitors) the problems are similar to cases where users have replaced capacitors on they own (e.g. utilized PXA's). Since, excatly the same solution, adding extra capacitor between TP10-TP9, is working in both cases this makes sense.

What MSI did wrong ? They did not change to feedback response i.e. the card with 'MSI component values' gives BS at stock frequencies (some games) no matter how good replacement capacitors are used . Newer GW-card are working just fine, and those have different feedback components. Modifying those components have a HUGE effect as proofed by 'zxcv' :UP: Please take look post #396, page 16 for simulation results etc. about effect of feedback.

--

I will give later more detailed (and simplified) instrutions HOW to test what component values there are in feedback (a capacitance (multi)meter will be required)

[SikaRippa]

Hmm, could it help to put a small fan on the capacitors? I dont want them to die as soon as the old capacitors

A fan could have some effect, but in the case of new, just replaced, capacitors it propably does not solve the problem. If you are worried about temperatures, it is possible to 'measure' them gently by finger tip. I have earlier measured following values:



.. from BS affected MSI-card. As you can see there are MANY components running hotter that CV-EX-capacitors e.g. some capacitors, mosfets and 'R47'-coil (=the one 'xan' cooled with an extra heatsink). At least when having Vcore modded the heat is a severe, as considered in other threads in this forum and some heatsinks would be nice.

So, I think the replaced capacitors are not the FIRST the will fail due to excess heat.

[Firechicken]

Hi sikarippa, I hope yopu got the pm I sent you have not been able to post or pm(have also been back in hospital again ))

[Graversen]

Hey, I'm about to fix a graphic card for a friend.
Info:
MSI 6800LE
I don't know exactly which version it is, but it looks just like this one: http://img218.imageshack.us/my.php?i...10105345ey.jpg
Just with out the modding

I don't know exactly when the BS started, but i don't think it took Long, there are just a big different between games, in Zoo Tycoon 2 in a saved game, with details high and zoomed in, i can make it go black in less than 10 sec. ;D

But any way, i wanted to fix it, fair and square, so today i received 2 x CAPACITOR, 1200UF 4V from: http://dk.farnell.com/jsp/endecaSear...=4310263&N=401

And i should just change them, with the 2 green capacitors on the board, with the text: "4 6 470 6E" RIGHT?

My dad got a small firm, where he has some employees, who works with components like these every day, so changing them is not a problem.
What is a problem, is that i have read a lot, of threads where people have written that, it does not even help to change these components, i am seriously not keen about this.
But many of these people have also overclocked their cards, so maybe it could have something to do whit that. Well i also though i could be able to OC it, maybe soft mod it, so i could open up for some more pipe-lines. any suggestions?

--- Iknow my english is not that good ---

[Sidolin]

Hi,
today i put a 80mm fan on the capacitors and it seems it really helps much. I removed the additional capacitor between tp9 and tp10 and was able to play nfs:u2 for about 1,5h under windows, i'll try it with linux now (the last time i tried nfs with linux, i got bs even at the intro videos...)

Now the pc is even louder than before, but it seems to work

Sidolin

[SikaRippa]

Hi sikarippa, I hope yopu got the pm I sent you have not been able to post or pm(have also been back in hospital again ))

Hi,

I got your PM, and already replied

[SikaRippa]

Hey, I'm about to fix a graphic card for a friend.
Info:
MSI 6800LE
I don't know exactly which version it is, but it looks just like this one: http://img218.imageshack.us/my.php?i...10105345ey.jpg
Just with out the modding

I don't know exactly when the BS started, but i don't think it took Long, there are just a big different between games, in Zoo Tycoon 2 in a saved game, with details high and zoomed in, i can make it go black in less than 10 sec. ;D

But any way, i wanted to fix it, fair and square, so today i received 2 x CAPACITOR, 1200UF 4V from: http://dk.farnell.com/jsp/endecaSear...=4310263&N=401

And i should just change them, with the 2 green capacitors on the board, with the text: "4 6 470 6E" RIGHT?

My dad got a small firm, where he has some employees, who works with components like these every day, so changing them is not a problem.
What is a problem, is that i have read a lot, of threads where people have written that, it does not even help to change these components, i am seriously not keen about this.
But many of these people have also overclocked their cards, so maybe it could have something to do whit that. Well i also though i could be able to OC it, maybe soft mod it, so i could open up for some more pipe-lines. any suggestions?

--- Iknow my english is not that good ---

Hi Graversen (or Hej, I guess you are Danish )

So you have EXCATLY the same model on that foto or model like this:



..except your green capacitors have code "4 6 470 6E" ?? (=green capacitors, and NOT all purple)

Since, you mentioned code "4 6 470 6E" I suppose you have the same model as 'piaskoon' reported earlier. These models have indeed problems as you propably have noticed..

I am 99 % sure that your card will be BETTER condition after replacing originals with those SVP-capacitors you mentioned (when properly done). However, OC-capability might remain poor due to unoptimal feedback components. (What clock-frequencies you have now ? ) Therefore I suggest to also add extra capacitor between TP10-TP9 (get the best your daddy's 'firma' could provide = very low ERS ~ 50 mohm).

I assume the pipes remain the same condition when cap modded i.e. if you could not unlock them now, propably the situation will not be better afterwards - but no big deal to test

[SikaRippa]

Hi,
today i put a 80mm fan on the capacitors and it seems it really helps much. I removed the additional capacitor between tp9 and tp10 and was able to play nfs:u2 for about 1,5h under windows, i'll try it with linux now (the last time i tried nfs with linux, i got bs even at the intro videos...)

Now the pc is even louder than before, but it seems to work

Sidolin

interesting demo ! This could be good 'workaround' for those willing/capable to install extra fans. Please, post how your 'fan'-mod will work later/with Linux !

[SikaRippa]

Could you give me more detailed info?

Here comes information how to test what component values you have in the feedback of the regulator. Take a look figure below:



We are interested values of capacitors C1, C2, C3. I am almost sure the resistors are the same at all cards. I have measured (see document for details HOW-to-do) the values from two different cards MSI (BS-card) and GW (not-yet-BS). Also values for the Leadtek were provided by 'zxcv', those measured after removing i.e. 100 % correct values (thanks again 'zxcv')

1. C1=82n, C2=15n, C3=135n (Gainward 6800le version 1b)
2. C1=40n, C2=30n, C3=120n (MSI 6800le version 1, by 'money2')
3. C1=47n, C2=33n, C3=22n (Leadtek 6800le version "4 6 470 6E, by 'zxcv')

The procedure to measure values WITHOUT removing capacitors is *little bit* tricky, I agree. Anyway, it is easy to test whether the card has values of listed above. In order to test it, do as follows:

1. Take care of ESD somwhow (avoid static electricity when measuring)
2. Remove card from PC
3. Locate terminals 'N3' and N4' using figure above
4. Calibrate (=null) the capacitance meter if possible (there is usually option to adjust capacitance reading to zero somehow i.e. remove effect of cables)
5. Measure following capacitance
- N3 -GND (plus node to N3!!)
- N4 -GND (plus node to N4!!)

I have measured following values:

MSI (BS):
C(N3 -GND) = ~25 nF, C(N4 -GND) = ~63 nF

GW (non-BS):
C(N3 -GND) = ~56 nF, C(N4 -GND) = ~115 nF

.. are you getting values close to 'MSI' or GW ?

[Sidolin]

interesting demo ! This could be good 'workaround' for those willing/capable to install extra fans. Please, post how your 'fan'-mod will work later/with Linux !

Hi,
i tried it with linux today and it works! Maybe i'll try some overclocking later, but nfs works now without bs on the stock frequencies. Before i got bs 10 seconds after starting the game.
Sidolin

[SikaRippa]

Hi,
i tried it with linux today and it works! Maybe i'll try some overclocking later, but nfs works now without bs on the stock frequencies. Before i got bs 10 seconds after starting the game.
Sidolin

Very nice to heard

I summarize your experiments, please comment whether something missing/wrongly undestood.

- You replaced original capaciotors with PXA's -> BS 'propability' decreased, but still getting BS sometimes
- Adding extra capacitor between TP10-TP9 does not help too much (at all ?)
- Having extra fan cooling card (capacitors) prevent BS completely ?

Some questions:

- Where you have excatly the fan ? I mean whether is top of all those capacitors (cooling top side of the card) ?
- You got BS with ~770 MHz before extra fan, what is BS frequency now/fully working with 770 MHz ?

----

Temperature vs BS is a complex issue, I was thinking for a while WHAT temperature change causes: When temperature is increased:

- All IC-circuits are running slower
- Many resistances are increasing, like Ron of the MOSFET's
- ESR (resistance) of electrolytic capacitors DECREASES !!!

Since ESR increases when temperature decreases, more problems with capacitors occur lower temperatures (I proofed this myself with a BS-affected card, I cooled capacitors with 'cold spray' and got BS more easily). Another point is that high temperature all the time deteriorates components rapidly. BUT when a high temperature is reduced little bit why BS propability is reduced althought ESR is INCREASED !!

Therefore I propose that in this case (having good capacitors !!!) the decreasing of temperature improves performance of some other component than capacitors. Other components that are related to switching regulator operation are:

- Inductors, temperature decrease lowers resistance -> good !
- MOSFET's, temperature decrease lowers resistance -> very good !

Those resistance are in series in the high current path of the switching regulator, so small increase/deacrease of resistance have considerably effect on operation, like BS-propability.

---

BTW I tested Cold War demo1 with Window$, and did not get BS (at least soon/stock frequencies). Do you think it would be usefull to test also demo 2, was it harder test i.e. gave BS more easier ?

[Sidolin]

- You replaced original capaciotors with PXA's -> BS 'propability' decreased, but still getting BS sometimes

Exactly, i dont get bs as soon as before.

- Adding extra capacitor between TP10-TP9 does not help too much (at all ?)

It doesnt help at all, at least now, it helped half a year ago for about two hours, but never after this. I tried different capacitors.

- Having extra fan cooling card (capacitors) prevent BS completely ?

Unfortuanatly not. For nfs:u2 it helped (without the fan i get bs very soon) but in battlefield 2 i get bs after about one minute.

Some questions:

- Where you have excatly the fan ? I mean whether is top of all those capacitors (cooling top side of the card) ?
- You got BS with ~770 MHz before extra fan, what is BS frequency now/fully working with 770 MHz ?

1. On the same side as the normal fan is (when the card is plugged in the computer, on the bottom)
2. See above, i get bs in nfs:u2 with 770mhz, in battlefield even

BTW I tested Cold War demo1 with Window$, and did not get BS (at least soon/stock frequencies). Do you think it would be usefull to test also demo 2, was it harder test i.e. gave BS more easier ?

Hmm, maybe Cold War is only a game which more likely causes bs like battlefield2...

[xan]

I'll try to measure these values for you, SikaRippa.

I have some observations:

My PSU is rather not popular worldwide due it's manufactured by polish Ropla company.It's solid 400 Watt PSU. It's very good, it's running better than any of Ultra's (i got 2 before).
It has two cables with molex connectors, 2 per each cable.
When card is connected without any other devices on the same cable it works good, but when I connect it with DVD-RW or HDD, card is bs-affected!

[Sidolin]

xan, this is a good idea, i have a very cheap psu and many cards etc. in my computer. I'll remove some of them and test battlefield2 again

Edit: It didnt help, i even connected the card to a second psu (you can start a psu without a mainboard), but it didnt change anything. And it also doesnt matter if the fan is running or not (for battlefield, in nfs it really matters)
Very strange.

[SikaRippa]

Exactly, i dont get bs as soon as before.



It doesnt help at all, at least now, it helped half a year ago for about two hours, but never after this. I tried different capacitors.



Unfortuanatly not. For nfs:u2 it helped (without the fan i get bs very soon) but in battlefield 2 i get bs after about one minute.



1. On the same side as the normal fan is (when the card is plugged in the computer, on the bottom)
2. See above, i get bs in nfs:u2 with 770mhz, in battlefield even



Hmm, maybe Cold War is only a game which more likely causes bs like battlefield2...

OK. The added is fan not a final solution. After I was thinking 'the cooling problem' (my latest post), I was quite sure the effect of cooling has to be low or moderate i.e. no fits-for-solution. Anyway, it does not HURT to keep the card cooler