!! Version 2.0 of document posted !!
- Finally !
- 'Ultimate capacitor modification' succesfully tested and guide rewritten
- Many fotos of cards added e.g. good quality foto of GW6800 'dual fan'
--
I have plenty of more material, but I want to put this update NOW since many users have requested a better version for 'ultimate capacitor modification'. In next update I'll have e.g. more measurement results (oscilloscope fotos etc.)
Quote:
Originally Posted by SikaRippa
Hello, i thought i'd update you on my LE, working nearly 4 months BS free :woot:Quote:
Originally Posted by sunn
This time i included some proper size pictures
<a href="http://img89.imageshack.us/my.php?image=aopen6800le17gc.jpg" target="_blank"><img src="http://img89.imageshack.us/img89/7651/aopen6800le17gc.th.jpg" border="0" alt="Free Image Hosting at www.ImageShack.us" /></a>
<a href="http://img89.imageshack.us/my.php?image=aopen6800le22nv.jpg" target="_blank"><img src="http://img89.imageshack.us/img89/3430/aopen6800le22nv.th.jpg" border="0" alt="Free Image Hosting at www.ImageShack.us" /></a>
some news.
good: at 700 MHz no BS anymore
bad: at 800 - BS again... after 30 sec of GTA San Andreas...
i'm very disappointed now.
maybe 2x470uF is too low?
soon i'll recieve 2 caps of Sanyo SVP series (1200uF 4V).
maybe they will be my last chance.
:(
Nice fotos :thumbsup: I am planning to make a chapter '6800(le) cards not (yet?) affected to BS and these fotos would be good for that purpose.Quote:
Originally Posted by sunn
--
I found something special in those figures. Take a look this figure:
http://img388.imageshack.us/img388/3...ison1bp.th.jpg
.. where is zoomed area from Aopen and GW-cards. See the difference (size) of the ceramic filtering capacitors ?? The size corresponds to capacitance value in the case of the ceramics (in this case). This can be one good reason your card (model) is not giving BS. Moreover, the same kind of (larger) capacitors seem to be inserted many places at the bottom part of Aopen (all connected to Vmem!).
Aopen had also version, like this:
http://aopen.jp/products/vga/large/A...00-DV128lg.jpg
having 'normal' sized ceramics (bottom of green caps)
It is pity you can not get more 'power' :( . Maybe, your card has 2.8 ns memories (have you checked codes?). Larger caps might help, at least many users have get some (not too much in every case) increase to clocks with cap mod.Quote:
Originally Posted by zxcv
Quote:
Originally Posted by SikaRippa
yes. card has hynix 2.8 ns memories. but i can set memory clock to 933mhz.. there are no artifacts at all in some 3d applications..
but serious games and tests (like doom3, 3dmark05, GTA SA) gives BS even at 800 MHZ. without any artifacts.
I'm sure its nice to have 933Mhz working for 2d desktop but if you cant run any benchmark or games, then what's the point. How did you even get to 933? Disable internal clock test in rivatuner?
The cap mod doesn't really help overclocking of memory. I tried 2700uF and 4700uF, on both locations (TP1 and TP9), and at best only 5-10Mhz more before onset of artifacts.
However, I do believe that it helps people who are having problems with black screens which might be caused by insufficient power to the volt reg for memory. From the accounts SilkRippa mentions, it seems to be helping some people. Fortunately for me I have the 570uF caps, and dont have any black screen at all... if you do, please try to RMA card instead of toying around with it and possibly destroying it.
OK. Your memories ARE capable to high speed (=faster that 'regular' 2.8 ns). The problem is that the switching power supply design of this card is bad and usually (~allways) OC'ing is limited by BS. Some programs run very well, but when all memories are active in the same time card is not capable to provide enough current and shut down (BS).Quote:
Originally Posted by zxcv
In more detail, my theory says that the problem is (see document for deep details) failing overcurrent protection in such case. Added capacitor or better replacement capacitor provided more current sourcing capability and card survive 'short' condition for a longer time.
In some cards (like new GW's, version 1b) they have modified the feedback loop of the switching regulator (see document) to have a slower response to shorts. Therefore these new models are working better even having original capacitors. I am planning to examine this modification in detail, since I have now measured feedback component values for BS and non-BS cards. I wonder that this modification will help some users to have even more OC'ing capability or remove BS. This mod might (when invented in the first place ;) )be more hard to do, since soldering/desoldering of tiny SMD components will be required.
zxcv: Are you interesting about that kind of Extreme modding ? Do you have a multimeter with capacitance measurement function ?
I considered this little bit also previous post..Quote:
Originally Posted by ***Deimos***
--
My aim has been BS removing, but many users that have a BS-affected have reported all cases some increase in OC-capability (up to 50 MHz I think), because operation has been BS limited. Cards that are NOT BS-affected do not gain so much with cap modding. In your case (XFX version 2, Sanyo SVP 560uF ? capacitors) it is evident that improvent is only minimal: your have more capacitance for Vmem at the beginning and those caps are superior quality when compared to 'green capacitors' or CV-EX's.
SilkaRippa it sounds you know what you are doing.. but I think adding relatively large capacitor (like coupling/bypass cap) from vmem to ground is insufficient bandaid solution. Doesn't this adversely affect low-pass performance (cap to ground)? I'm also worried about the toll on the vreg of charging such large cap. Ultimately, I think you are right about improving the feedback loop. I'm not a practicing electronics so a little rusty, but if the problem is indeed spikes in current drawn, wouldn't an inductor be more appropriate to reduce the inductance in the circuit?
Yes, adding a large capacitor to Vmem node is a brute force method and very large value can cause a number of problems: card might not start up or feedback loop transfer characteristics changes too much (=>many problems).Quote:
Originally Posted by ***Deimos***
However, this mod (against BS) really works for the most users even when having cap as large as 4700 uF. I wonder, that usually quite large added capacitors do not cause problems because the long leads (=several mm at minimum) and moderate ESR value cause a higher impedance than due to original capacitors.
If the switching regulator circuit would re-designed I would suggest several changes. For instance the inductance value is far too low (1 uF), since lower inductance means higher current peaks (but faster response => trade-off). Also the input voltage -output voltage ratio is very low and output power level high - and varying extremely rapidly. (I have new measurement results (not yet fully analysed), that shows how large is the current level and how rapidly it is changing.)
In theory, using aqcuired data, an optimal set of components could be calculated. In practice following aspects make it a HARD task:
- Parasitics (extra resistance/capacitance/inductance) of components and PCB, has considerably effect to response
- Load due to memories is hard to model i.e. current min/max and risetime/falltime
- For a standard user it is too hard task to obtain a number different special components and solder those e.g. small SMD capacitors are extremely difficult to change
Therefore, I have mainly considered modifications that are simply to understand and easy to perform. Anyway, when I get some time I'll investigate the feedback modification more carefully - and report the results.
Yes, i'm very intrested.Quote:
Originally Posted by SikaRippa
Unfortunatly my home multimiter cannot measure capacity. But I'll try to find such multimeter.
Also i have good soldering equipment on my work. so... soldering of smd-components (capacitors) shouldn't cause any troubles =)
Fine ! I'll have to think/simulate/analyze about that very carefully - take some time, cannot promise anything yet.Quote:
Originally Posted by zxcv
raising R or lowering C values in FB2 RC-circuit should help avoid wrong shorts...
http://img121.exs.cx/my.php?loc=img1...ge=fig78fq.jpg
Yes.. but how about response to whole transfer function and what are the correct/optimal values.. . The desoldering/soldering SMD's is so nasty to perform that nobody want to do it several times. Moreover, tiny solder pads tend to 'burn' useless after few/some attempts.Quote:
Originally Posted by zxcv
--
I had nice ~ 2-3 hours while I analyzed component values (new ones got from a badly 'BS' card) and compared to 'GW version 1b. I use the same kind of method described in Ch 5.6. The components under interest can be located using a figure below:
http://img148.echo.cx/img148/2742/memfb17hp.th.jpg
I confirmed that *them* have CHANGED values for capacitances but AFAIK the resistor values are the same. Using an iterative trial-and-error method I found (!) values for those components with acceptable accuracy. I wonder (hope) they have used standard values so the obtained value can be rounded to nearest 'correct' one.
But, please be patient 'zxcv' (and everyone willing to perform this mod). I'll give those values after I have made electrical simulations and saw that those make sense. I want to really see what they have done, before giving detailed instructions. (I am too tired to do simulations tonight...)
I wish I knew one of the pcb designers.. it would be so much easier just asking them.. but probably not as rewarding.
I Sika, i received the card on saturday... its working fine!
i will reinstall my system now and make some benchs... then i will public the table with the oc results...
fyi money2 ( i had to answer public cause pm-systeme isnt working on my machine at work... ) :)
Yep, I would be great to heard ANY comments from those are responsible of designing/modifying this card. I am sure someone of those have read/see this document/thread, but the are not allowed (nor willing) to say anything.Quote:
Originally Posted by ***Deimos***
I was talking this issue with moderator of GW's forum. He confirmed that it is a 'design flaw' and GW has a fixed version version, but NO comments about identifying the faulty card (or what they have modified). You can not find it anymore, since the thread was deleled very soon after those comments, why ? Maybe he was not allowed to say anything about design flaw in public forum..
'zxcv' or somebody else interesting about 'Xtreme'-modding of GF6800 cards..
I have simulated the AC-response of switching regulator (memory part) with an iSim-simulator. I am not going to explain (in this post) detailed what this simulation electrically means, only show the results.
Component labels indicates those shown in this figure:
http://img148.imageshack.us/img148/2...mfb17hp.th.jpg
I acquired following values (see post #1 for details about HOW):
BS-affected card:
C1 = 40 nF (~ +/- 10 nF)
C2 = 33 nF (~ +/- 10 nF)
C3 = 120 nF (~ +/- 20 nF)
Non-BS-affected card:
C1 = 90 nF (~ +/- 10 nF)
C2 = 15 nF (~ +/- 5 nF)
C3 = 135 nF (~ +/- 20 nF)
.. not a big deal, does not affect on anything ???? Wrong, those little changes DO have an effect see the simulation below:
http://img138.imageshack.us/img138/4...sac05gi.th.jpg
So, the response of the non-BS card is faster i.e. the regulator has (more) gain at higher frequencies !! Propably this is enough to remove BS in certain conditions. BUT (a BIG BUT) this mod do not remove the high supply current variation (and ripple-current) and therefore BS can appear also with these cards - but it takes longer time.
--
How to do this mod ? Is it sensible to try it out ? For a highly qualified solderer this mod is piece of cake, but for standard user I do not recommend: soldering of tiny components IS difficult. If properly done it can be very usefull, BUT I CANNOT GUARANTEE ANYTHING SINCE THIS HAS NOT YET TESTED BY ANYONE.
--
Before you (zxcv?) try this out (I know you will anyway ;) ) Please do following test: Measure capacitance between 'n4' - GND and 'n3' - GND (GND to - terminal of meter !!!) and post the values (see figure above for those points) I will analyze values to find out if your card is similar to BS-card I have measured.
I have a GF Sparkle 6800 and I have problems with BS when playing HL2 and CSS. The suggestions I have got so far is to get a new PSU. The one I have is an AOpen 300 W which can deliver 15 A och the 12 V line. Should I get a better PSU or should I modify the card with the capacitor as described in this thread?
Hi,Quote:
Originally Posted by xotoy
Power supply has only a minor effect i.e. having very good supply BS will in some conditions disappear (becomes more rare). Usually the effect is so minimal, that new power supply cannot consider as a solution for the problem.
Easiest and most 'safe' mod is 'Enhanced capacitor mod', which in most cases removes BS. For some cards this mod does not work, and those 'green caps' should be changed (this mod has been given very good results with 'totally BS' card)
--
Is your card similar to this one:
http://img206.imageshack.us/img206/4...00x19rt.th.jpg
.. and it really is 6800 (not 6800le) ?. Could you specify any version code etc. for your card and the labels for 'green caps' (e.g "4 3 470 6E") :)
Thanks for your reply. It is a 6800 (NOT le). I will take closer look at the card and also supply a picture. I did change the cooling unit to an AC, do you want a picture of the original cooling unit as well?
Nice you are ready to provide info about Sparkle, it is rare and very few data available (so far! )Quote:
Originally Posted by xotoy
Fotos are welcome :) also about original cooler if not too hard task..
by the way sika, your thread reached already 15.000 Views :-D
Noticed :D, not bad... interesting issue (?)Quote:
Originally Posted by money2
--
BTW, I sent email for you, got it ?
Hi SikaRippa, I have found this thread and appreciate all the work you are doing :)
I just bought a Gainward 6800LE and it looks like it could be a version 1b card because the capacitors start with 47. I have managed to unlock it to 12 pipes and 6 vertex and overclocked it to 390 core and 850 memory. It has been running stable for two days. I have run the SkiAlpin many times and still no BS. Also I tried to stress the card running 3dmark05 and doom3 on highest settings and no BS. I have made some photos for you. on the back where you said the 1b card had some modification on the switching regulator mine does not appear to have it but you will be able to tell better.
Back of the card
http://img325.imageshack.us/img325/3...11413fp.th.jpg
Closeup
http://img325.imageshack.us/img325/6...11499ax.th.jpg
Front of card
http://img325.imageshack.us/img325/1...11442ia.th.jpg
Closeup
http://img325.imageshack.us/img325/1...11503lq.th.jpg
Hopefully it means this is one of the new cards that Gainward has "fixed" I will let you know if I encounter any problems
I have now taken a closer look at the card. It has the same text (SPAG40 REV P212A01) on it as the picture you posted. On the back side I found two stickers with some numbers on them. Are they of interest? Unfortunatly, the Arctic Cooler covers the capaciters, so I can't see the numbers on them. :(
Anyway, like you said, when I tested it in my friends computer with a PSU with more power, the result was the same (BS).
A also checked my receipt and the card is actually less than one year old, so perhaps I can change back to the original fan and RMA the card. Not sure what to do...
Hi !Quote:
Originally Posted by jabjab
Nice fotos, thank you :up: !!
--
I took a very close look your fotos and saw some pixels in wrong places :D:
http://img295.imageshack.us/img295/2...pmis6tt.th.jpg
.. are those capacitors misaligned, especially the one in the right side ? Could you check is those are 'not so nicely placed' that the other nearby ? Those two (marked with arrows) are changed at least in 'version 1b' -card I have investigated.
--
I also started to (again) think about version numbers. These codes were some reported cards (first in barcode, second at bottom of card)
a version 1 (#1 example): G040900553, xxx
a version 1 (#2 example): G041003120, xxx
a version 1 (#3 example): G040814005, xxx
a version 2 (#1 example): B041201502, GW6841101282
a version 2 (#2 example): B041134662, GW6841100549
a version 1b (#1 example): B041134230, GW5841100910
a version 1b (#2 example): B041201576, CW7041100629
a version 1b (#3 example): B050113539, xxx
a version 1b (#4 example): B050113522, xxx
a version 1b (jabjab's card): B050113475, GW5841100417
----
Logic ?!? Maybe the barcode number is one kind of identification or '41134230'th card of GW's production. That would make sense since your is new, and those two quite old ((one could calculate how much GW's is selling ;), the first 1b has got at April...)). The reason why '1b' card (modified) could have lower number is explained as follows (purely speculations): It was sold at first point as unmodified (had "4 3 470 6E" caps?) and after RMA'd *them* modified it and return - in theory makes sense (?). But how about 'GW***'-code ? Maybe GW6*** means unmodified and GW5*** (CW**) modified plus some extra info.
---
Please post if somebody has different codes on GW's card :) I think that those two codes ARE KEY ELEMENTS to distinguish easily GW's BS and non-BS-cards !!!
No bad feelings about capacitor codes ;) Maybe you can provide that info ehrn you (hopefully) RMA your card and get the original cooler back..Quote:
Originally Posted by xotoy
--
All codes are interesting (you saw my latest post ;) ), at least if there is a clear 'version x' etc marking which could help to distinguish different versions. I'll put that picture I attached earlier to next update, if you think it looks like identical to BS-affected model.
:DYou are absolutely right :) , the capacitors are misaligned, see mega zoomed photo below:Quote:
Originally Posted by SikaRippa
http://img381.imageshack.us/img381/3...11568mo.th.jpg
I suppose that means It could have been an ex-RMA card (either because of BS or because they could not get the full 16 pipes :D)
Do you think I am quite safe from having BS then? I think you mentioned there is one instance of a 1b card giving BS?
i have the xfx card, and i keep get the bs all the time, the prob is that i bought the
silencer and i cant find the location to the holes that the cap need to be inserted to them, can anyone show a pic of that holes on my xfx 6800le?
Heureka :woot: !! A single card repaired afterward could be explained as 'fault-in-production-line-but-fixed-ASAP) BUT two fixed excatly in the same manner - no chance by accident. And I think that CONFIRMS the theory how GW (and) others made some card working better than very first cards.Quote:
Originally Posted by jabjab
--
Are those working better enough ? As we all have seen those are working FAR better than old models. Yes, a single case has reported about fixed version (assumed using capacitor info) giving BS after ~ 5 months. The one I am investigating will be ~ 5 months old (got from RMA, can be older if really modified) in this month (!!!!!)
I cannot say for sure that modification is enough to remove BS a long time (>year). I have some unpublished measuremt data (acquired from modified version !!) and not fully convinced the current spikes are acceptable level - thus that single faulty case would be a typical case.
Maybe *them* count on (=hope) the fact users change cards so frequently, and after say 3 years nobody still using 6800le-cards (?).
---
EDITED: 'jabjab', Do you have a capacitance meter (multimeter with that feature) ? If 'yes', could you make an easy reference measurement (I'll give instructions later) :) ?
and another question:
if i replace my power supply to a better one(430W,480W) and flash the bios to 1.4 with nvflash
can it solve the problem?
Hi,Quote:
Originally Posted by inspire
At first power supply does not make big difference: good supply decreases BS probability little, but NOT a solution to problem.
--
About capacitors, you can use this figure to locate holes at the bottom side of the card:
http://img69.imageshack.us/img69/4061/capmod35gz.th.jpg
... your card should look like the same (?)
--
Could you see (under cooler?) what labels you have at top of those green capacitors ?
here some pic of my card:
http://img74.imageshack.us/img74/739...e0026gl.th.jpg
http://img74.imageshack.us/img74/766...e0016qe.th.jpg
http://img83.imageshack.us/img83/1327/amit1kg.th.jpg
At the third pic u can see where the tp10,tp9 on the card.
Yep, seems to be the same card i.e. you can use those TP's for cap mod.Quote:
Originally Posted by inspire
--
I noticed the version number (a sticker at bottom) is '2.0'. Could you see (under the coooler) whether those memory caps have green labels or purple ? This is important, since XFX has both versions, but I do not know (company) version numbers for those (green cap's should give BS, and purple not)
my caps are purple(u can see it on the pics if u save them and open them
on xp.
when i play hl2 and doom 3 for 20 min...the screen is suddenly shut down.
Sorry, I was not too accurate. There are allways 'purple' capacitors for Vcore etc. purposes in these cards. These capacitors (having values of 1500 or 330) can be seen through NV5. Memory capacitors can not be seen in this angle, since they are just below fan.Quote:
Originally Posted by inspire
Take a look this figure, where you can see difference between two version
(first with two green caps, second have only purple)
http://img194.echo.cx/my.php?image=xfx6800lev1v26ky.jpg
--
Shutting mysteriously down with HL2 & Doom3 is clear evidence having a BS-affected card. Have you tested 'SkiApine'-demo ((it is not so important, because seems to be so clear case)) ?
Sorry, I don't think my multimeter has capacitance measurement, its one of those cheap ones, looks exactly like this (same functions) http://www.montek.com/catalog/item1968.htmQuote:
Originally Posted by SikaRippa
Bad news :(
I decided to play Doom3 today with high settings and after only 5 minutes I got a BS! so in my case my version 1b card has lasted less than a week! So of course I am going to request an RMA ASAP as I'll be able to get a refund. I really like the 6800 cards though. I was thinking of getting the Leadtek 6800LE, is that much less likely to get me BS? Thanks for your advice anyway, If I manage to get a capacitance meter from work I'll let you know and maybe we can do those tests before I send it back :)
Quote:
Originally Posted by SikaRippa
sorry,but they all purple...
when i look behind the nv i saw that the caps are purple 2.
what does it mean?
i am going tomorrow to the store to buy the caps...
can u give me the exact name of the best caps that i need?
All purple and you still get BS ????? This confuses me *really*. You have (as far I understood/know) an XFX version 2 that should have Sanyo SVP-capacitors instead of those green CV-EX -caps. In detailed those capacitors are in other version card '10SVP560M' which should have VERY GOOD performance. Here is datasheet for those really interesting:Quote:
Originally Posted by inspire
http://www.secc.co.jp/pdf/os_e/2004/e28.pdf
Big question: why your card gives BS ? The other XFX version 2 cards have been reported to work OK and so far only version 1 has been known to be BS-affected. I'll have to think about this, but could you tell following details:
- How old your card is ?
- How soon the first BS came (weeks/months..)
- Your BS is allways the same: total shutdown, only power switch helps
---
About cap mod. Since the capacitors of your card are (should be) much better than in normal BS-case it makes things complicated. Because you get BS something in your card IS different compared to other XFX version 2 cards. If the capacitors are ruined (=should be unlike) the normal mod should work. ((Another option could be that another capacitor has poor contact (e.g. bent little bit))
Anyway, since your case seems to be somehow unique (AFAIK). I suggest to test cap mod carefully (not known how good it works this case)
- Start with low capacitance value e.g. 1000 uF or 2200 uF
- Ask for low ESR-capacitor voltage, rating > 6 V
- There are soooo MANY models, and every shop have different capacitors. Therefore, just go to shop with those specs
Please, post how your modding succeeded :)
Bad news :( indeed, I hope you get it RMA'd . Leadtek's have also problems, less that GW's but still some. I suggest to go for real 6800 (GT) if you can afford it, otherwise ASUS is the most 'safest' in the sense of BS.Quote:
Originally Posted by jabjab
Maybe you got unfamous record "How fast to get BS with card GW has modied to last for ever".
--
What does this mean ? The single case GW's version 1b (=modified by company) can give BS is not a single case i.e this modification does not give 100 % reliablity of working card (as I have presumed, since current levels are the same in all versions).
--
If you can measure capacitance, that would be nice :clap:
Here are the instructions (again):
Labels indicates those shown in this figure:
http://img148.imageshack.us/img148/2...mfb17hp.th.jpg
Measure capacitance between 'n4' - GND and 'n3' - GND (GND to - terminal of meter !!!).
DO NOT MEASURE WHILE POWER IS ON !! (much easier when card has removed from PC)
I see, ive looked at the other photos on this thread and there appears to be only 1 other card that have these larger ceramics [only one the front, no pic of the back] and thats nVidias reference card [1]Quote:
Originally Posted by SikaRippa
<a href="http://img355.imageshack.us/my.php?image=nv68000x19pr.jpg" target="_blank"><img src="http://img355.imageshack.us/img355/3599/nv68000x19pr.th.jpg" border="0" alt="Free Image Hosting at www.ImageShack.us" /></a>
nVidia and AOpen together
<a href="http://img384.imageshack.us/my.php?image=untitled18fd.jpg" target="_blank"><img src="http://img384.imageshack.us/img384/3089/untitled18fd.th.jpg" border="0" alt="Free Image Hosting at www.ImageShack.us" /></a>
AOpen and nVidia again split with Gainward
<a href="http://img293.imageshack.us/my.php?image=untitled21nz.jpg" target="_blank"><img src="http://img293.imageshack.us/img293/6084/untitled21nz.th.jpg" border="0" alt="Free Image Hosting at www.ImageShack.us" /></a>
???
:woot: :banana: :banana: :banana: ....it is working, i soldered the caps between t9 and t10 and now i was able
2 play hl2 on the best resolution with AA and i didnt get a single bloody black screen...
wow, sikaripa,when u can give me your icq or msn and i talk 2 u about all the details...now here some pics :toast:
http://img394.imageshack.us/img394/8...e0063oy.th.jpg http://img394.imageshack.us/img394/8...e0100su.th.jpg http://img394.imageshack.us/img394/4...e0128la.th.jpghttp://img394.imageshack.us/img394/4...e0140cy.th.jpg
Yep, many (most?) card have those 'tiny' ceramic capacitors, but some have larger. Eg. BGF OC 6800 has large ceramics as can be seen here:Quote:
Originally Posted by sunn
http://img75.imageshack.us/img75/2095/fig28dw.th.jpg
I made a little survey about that kind of capacitors. Here the physical size matters i.e. larger capacitor means higher capacitance value (in general capacitance value DO NOT correlate the size, but in this case a maximal capacitance vs. size is the most probable choice). The size of the solder pad is the same for all PCB's and only the height is varying. Very high capacitance value, say greater that 10 uF, in this size of package is EXPENSIVE and I guess the larger capacitors (like in BGF or Aopen) are 10 uF at maximum. The smaller capacitors (used by all other vendor GW, MSI, XFX..) are using the same logic 1 uF at maximum, propably much less (0.1 uF ?) because these vendors HAVE OPTIMIZED THE COST.
There are several ceramic capacitors bottom side of the card (at least 2 pcs for each memory chip + some extra). The electrical performance of the ceramic capacitor is excellent (e.g. ESR only few milliohms), so this 'modification' has a huge impact to the operation of the card (=no BS).
However, when having a BS-affected card (capacitors already ruined), changing ceramic capacitors does not help anymore. In the other hand, for a brand new (no BS yet !) this mod could give more life time - not too wise option with a new card :)
Fine, congratulations :up:Quote:
Originally Posted by inspire
--
What size of capacitor you used (I could not read the value..) ?
--
I will send a PM for you...
Its 10V 3300F.Quote:
Originally Posted by SikaRippa
I couldn't find the capacitance meter, sorry. I had the RMA accepted so I have sent the card back. One thing I noticed though is that as well as the audio issue (noises on my speakers that changed frequency on fps change) There was also this very annoying ticking noise coming from the card whenever I turned the wheel in my optical mouse (I made sure the noise was coming from the card and not the mouse)Quote:
Originally Posted by SikaRippa
(edit: typo)
Thanks for info ! It seems not to be too high value even when having 'purple' capacitors instead of 'green' ones..Quote:
Originally Posted by inspire
Please post if you will encounter any problems afterwards.
what about the modification?Quote:
Originally Posted by SikaRippa
How do i know who are the bad caps?
In general is very hard to say which are bad. Temperature raise over 5 C at capacitor surface is a good rule of thumb. This means the power (=ripple current) inside the capacitor is so high level that the life time is shortened quickly.Quote:
Originally Posted by inspire
In your case I was wondering if one (both?) capacitor is somehow mechanically damaged or loosen.
I sent an email for you including questions about card.
hello SikaRippaQuote:
Originally Posted by SikaRippa
i have much work these days. but spent some time soldering feedback caps!
1. i replaced C1 with 100nF one . mem clock decreased from 760 to 740 Mhz
2. then replaced C3 with 100nF. black screen on windows desktop again. (C1=100nF, C2-? nF)
3. replaced C2 with 10nF. no BS in 2d and 3d (not heavy app's yet) at 750 Mhz. (C1=C3=100nF)
(!)one strange moment(!). when i soldered C2 out (old cap) i used iron only from one side of cap. i saw that cap moved a little when i touched it only from one side with iron (320 C degrees). other side had very bad contact with PCB. and then cap was easy removed by tool. maybe it was bad soldered by manufacturer?
also I will try to find some time to go to service and measure old cap values.
are u sure u measured right
Non-BS-affected card:
C1 = 90 nF (~ +/- 10 nF)
C2 = 15 nF (~ +/- 5 nF)
C3 = 135 nF (~ +/- 20 nF)
values are close to image (http://img148.imageshack.us/img148/2742/memfb17hp.jpg)
were C1=C3=100nF C2=10nF.
and i got now these values on my card.
so! feedback caps due effect on BS
4. (!) replaced sanyo caps (blue from radeon) with... tantalum EPCOS caps...
http://www.epcos.com/web/generator/W...locale=en.html
just 2 caps 680uF 6.3V. the seller wasn't sure which series this caps from.
card booted up ok, 2d and 3d ok... but only 760 Mhz as before this modification. and then BS again :(
can anyone tell me if this card is affected be this black screen effect. it has 45 caps, which i see haven't been mentioned a lot in this topic. Oh and how do you tell if its a Version 2 or 1B, i might have missed it when reading through the very interesting topic.
SikaRippa your english isn't bad really, well i can read it and i'm english so i should know what i'm on about. You have also done some very exhaustive work here and i thank you for it.
EDIT. its a gainward 6800le with i think 2.8ns (i haven't checked yet) Memory
I have now changed back to the original cooler. The labels on the green capacitors are: 4 5 470 6E. Hope it is of interest. See also the attached pictures.Quote:
Originally Posted by SikaRippa
today i soldered C3 back... stock cap
on my work PC I was able to run 3dmark05 three times at 850Mhz. (pIV-2800 Mhz prescott, asus p4p800-vm, 512 dual channel).
i hope at home i will have same result at least(athlon 64 2800+ @ 2300 Mhz, EP-8KDA3j, 1 Gig Hynix D43)
now feedback cap config is C1=100nF C2=10nF C3=stock (120-135nF i think)
I am also VERY busy :slobber: , and give a short answer. I hope I'll have more time at weekend..Quote:
Originally Posted by zxcv
--
Very interesting results, I appreciate people making that kind mods/testing :toast:
At first you example corfirms the theory: feedback have considerably effect on BS-performance. Second point is that things can be go easily wrong i.e. to have even more poor performance. Key point is the correct values !
--
The values I gave HAVE large tolerance/uncertainty issues due to used indirect measurement setup. Summary of those results can be expressed as:
- values VARIES between different models
- the order (=rough value) for target value is obtained
- EVEN WHEN HAVING THESE CAPACITOR VALUES BS CAN APPEAR
.. as has been reported (=BS with 'modified version') by some users like 'jabjab' (see above).
--
'zxcv': If you could measure those unsoldered caps it would be very helpful :)
I'll try to perform (some day) some optimization with simulator to find out better values for the components.
Thank you for nice comments. I am sure there are *some* typos/bad language used in the document, but if an English(wo?)man does not complaint ... :)Quote:
Originally Posted by unrealuniverse
--
Yep, "4 5 470 6E" are rare. Card having those is 'version 1b' AFAIK at least it is not version 2'. If you could specify following details I can analyze more precisely:
- Codes of the two stickers bottom side of the card: First down/right (e.g. 'GW5841100910' ) and second in barcode (e.g. B041134230)
- Code in OK-sticker at connectors e.g. OK-64
- Made in China/Taiwan ?
- Misaligned tiny SMD-capacitors (see figure below) ?
http://img295.imageshack.us/img295/2...pmis6tt.th.jpg
OK, Thanks !!!Quote:
Originally Posted by xotoy
Now, I have good info to add to next update: Sparkle 6800 + "4 5 470 6E" = BS
That'll be an english man and everyone makes typo's and grammatical errors anyway :) . The codes are GW7041100880 and B041224758. the Ok sticker is OK11. it's made in china and has no misaligned capacitors where you have highlighted in your picture or anywhere else (that i can see, i might have missed it though if it is a very small difference :p: ).Quote:
Originally Posted by SikaRippa
Hope this info helps, i can post picutres if required. and if the capacitors are "faulty" then how long do i have before BS, Just like to know when to order some capacitors :D if i have to.
yes! 925 Mhz, no BS yet...
card modifications once again:
C1 changed to 100nF
C2 changed to 10nF
4 6 470 6E C136 and C143 changed to tantalum EPCOS caps 680uF 6.3 V
Thank you for codes. I listed so codes in post #328 (this data can not be yet found on document!)Quote:
Originally Posted by unrealuniverse
a version 1 (#1 example): G040900553, xxx
a version 1 (#2 example): G041003120, xxx
a version 1 (#3 example): G040814005, xxx
a version 2 (#1 example): B041201502, GW6841101282
a version 2 (#2 example): B041134662, GW6841100549
a version 1b (#1 example): B041134230, GW5841100910
a version 1b (#2 example): B041201576, CW7041100629
a version 1b (#3 example): B050113539, xxx
a version 1b (#4 example): B050113522, xxx
a version 1b (jabjab's card): B050113475, GW5841100417
your card: B041224758, GW7041100880 => 'version 1b' is the best match :)
The card "B041201576, CW7041100629" (C=G ??) was a brand new GW (~month) which has "4 9 470 6E" and not yet BS. Whether you will get BS in future, not yet known..
--
Those tiny capacitors might have changed also PROPERLY way i.e. NOT misaligned. As you can see in this foto (courtesy of 'jabjab' ;) ):
http://img381.imageshack.us/img381/3...11568mo.th.jpg
.. the 'degree of misaligment' can be LOW. The newest cards can of course be brand new (like you?) and *them* have been used altered values at the first point i.e. not modified from RMA'd cards. Maybe have different values for capacitors ?
IF (!!!!) this theory is correct we would have following list for codes (down/bottom sticker):
No sticker (v 1) -> BS ASAP
GW68*** (v 2)-> BS Later
GW58*** (v 1b, modified afterwards) -> can give BS later, usually working (?)
CW704** (v xx, new batch !) -> ????????
:up: :up:Quote:
Originally Posted by zxcv
To get some idea what you have done: Do you have different capacitors for C1 and C2 (C3 is original) now ? What is changed between condition you had earlier when you got BS at lower frequency ? You had those EPCOS caps then OR those from Radeon ???? EPCOS are very good, do you have excatly these models:
http://www.epcos.com/inf/20/40/ds/HighCap_DS_series.pdf
---
Did you check the values of those original capacitors ?
I wonder the original values (in version 1b -card) can be C1 = 100nF, C2 = 10nF (see the tolerances I put !!). E.g. 90 nF is not a strandard value and it is much more likely to find a 100 nF in a design than a 90 nF capacitor.
here is the sequence of modifications:
1. i replaced stock C1 with 100nF one. mem clock decreased from 760 to 740 Mhz
2. then replaced C3 with 100nF. black screen on windows desktop again. (C1=100nF, C2-? nF)
3. replaced C2 with 10nF. no BS in 2d and 3d at 730 Mhz. (C1=C3=100nF)
4. replaced C136 and C143 with EPCOS tantalum caps - max mem clock 760 MHz again (the seller wasn't sure about series, but told me, that usually caps are 197-A series, which is low ESR). there are marks on PCB for palcing tantalum size-E caps =))) instead of alluminium caps.
5. replaced C3 back to stock. now i am able to run 3dmark05 at 925 mhz
i always use 3dmark05 for BS testing three times
my card running always 16x6 @ 375Mhz(core)
now gonna measure old (manufacturers) C1 and C2 values when i'll have a little more free time.
Thanks for detailed sequence. Good point that there are markings for tantalum size-E caps !!! I'll check my fotographs of MSI-card (version 1) without those caps to see whether those markings were used also at beginning.Quote:
Originally Posted by zxcv
Some thoughts:
- C3 had major effect, most likely original value is C3 = 150 nF (=150 standard value)
- adding EPCOS tantalum caps had MAJOR effect when having C3 = stock
=> EPCOS tantalum caps or similar (e.g POSCAP) might be best replacement for C136&C143 !!!
I thought that earlier, but I was worried those have too low ESR-value which could cause stability issues. Your test showed it is not a problem :clap:
----
If I undestood you had following (major) combinations:
1. All original -> BS at too low frequence
2. C136&C143 changed to PXA's -> 'BS frequence' increased a little
3. Playing with C1/C2/C3 -> 'BS frequence' decreased
4. C136&C143 changed to EPCOS -> 'BS frequence' increased a little
5. C1 = 100nF C2 = 10 nF (both changed, but might have original values?) C3 = original (150 nF?) -> 'BS frequence' increased A LOT
Correct ? What kind of capacitors you used for C1-C3 ? Ceramic with leads, or SMD ? The inductance of leads (and wires have some effect on response.
If you can provide values for C1/C2 it will be fine no matter takes time, measure those when OK for you :)
you understood everything correct, SikaRippa!
before modification my card looked like this one
http://img355.imageshack.us/my.php?i...68000x19pr.jpg
there one missing cap on PCB. u can see marking for tantalum size-E cap. same markings i saw when soldered out C136 and C143.
about caps C1-C3. my new caps looks like prevous (manufacturers) caps. and we call them usually "Ceramic SMD caps".
I have no scanner at home but i will scan my card with good resolution and post images here.
P.S. tested card again. no BS at 925 Mhz yet.
SikaRippa thanks for the info. i'll put a post in if it does BS in the furure, it has the noisy audio problem and actually makes beeps and not just cracking to FPS although it does that as well. Strange?!?
all the problems have sort of taken a back seat as it does 16ps and 5vs, the sixth is faulty. haven't tried overclocking yet so that might give me some results.
I cannot see if the capacitors are the same as i don't have a setting on my multimeter to read them sorry. i think your final theory is a good logical one, but i have yet to figure out why mine "new" should have older 4 5 caps when the other example that is closest to it has 4 9 caps. Why would they use these older caps on a new board.
the card was manufactured roughly around week 17 in 05 as this is roughly what i get date wise from the lot numbers on the chips.Quote:
Originally Posted by SikaRippa
Exactly same label (or component frame) has been used by e.g. MSI and GW in that 'missing cap area' and beneath C136&C143. I though that marking is for SMD electrolytic capacitors, since two corners are non 90 degree just like the shape of the capacitor.Quote:
Originally Posted by zxcv
Since, it is clear EPCOS-capacitors gives better performance in your case, the exact series code would be *nice*. Can you see any code top of Epcos's ?
Here specifications of EPCOS-capacitors (B45197A-series, 'SpeedPower'):
[capacitance][case code][VR][Ileakage][ESR mohm][Iripple][code]
680 D 15 42.8 150 1.00 B45006D687*+157
680 E 6 42.8 45 1.91 B45006E687*+456
680 E 6 42.8 60 1.66 B45006E687*+606
680 E 15 42.8 100 1.28 B45006E687*+107
680 E 15 43.0 100 1.28 B45197A1687+50*
It is noteworthy that ESR is larger than with CV-EX/SVP/PXA -capacitors !! Also ripple-current is much lower. Maybe your capacitors are not this series ?
The noisy audio can be almost anything - I had beeps, cracles etc. e.g when playing Far Cry. It was very annoying at silient ambient :( . I will later post those oscillosscope fotos which show HOW wierd is the current consumption of this card. I think that every performance change (=memory operation) causes a current peak, which can be heard either from card or loadspeakers. Having one 'event' you get crackle, constantly varying changes gives whistle and random operation (normal program) generates everything you can imagine.Quote:
Originally Posted by unrealuniverse
About datecodes: what chip(s) you investigated ?
I think that theory after post and I am now quite sure it had to be correct or very near correct. The codes, manufacturing dates, misaligned caps all match to that model. I have also been thinking those capacitor codes (many months, in fact :) ), since it indeed does not make sense to put old capacitors (were you can get old lot !!) to brand new cards. One theory which could solve the problem is that they have used Hitano's capacitors in the new cards. Those look like exactly the same as CV-EX and have better performance (and probably different lot codes !!!)
http://www.globalsources.com/si/6008...GetProduct.htm
http://www.hitano.com.tw/PDF/EVS.pdf
I mainly went on the mosfets however there is another chip and i cannot remember where it is (as its now the following morning :p: ) that has a date code on as well, both turned out roughly the same. Would hitano's caps still have the same style of lot coding on it though (eg. 4 5 470 6E) as the CV-EX caps. it would however solve the problem of these newer cards not getting BS even when the caps are ment to be BS efffected, they've used a different manufacturer of caps.Quote:
Originally Posted by SikaRippa
Oh and you mentioned the audio and you can get pritty much anything, R2D2 sounds interesting in HL2. :)
hey great job SikaRippa :clap: I had a MSI 6800LE BS affected card untill 3 days ago but thanks to your guide I finally solved the problem (adding a capacitor between tp9 & tp10, 2200uF 6V). Two month ago I tryed with a capacitor in tp1(3300uF 25V), but only worked for 1 day and not with half life2. Now its all ok,nomore problems, its wonderful :banana:
The card has capacitor with numbers 4 3 470 6E. The first BS appeared only after a few month in HL2 all 3D marks GTA SA, Colin Mcrae Rally 05 but NEVER happened Playing DOOM 3(strange isn't it?)
I can also ear the "strange sound" whyle running 3Dmark
p.s this is my third MSI 6800LE card,The first 2 were RMAded (BS) but since this one could work in 16/6 configuration i wanted to keep it.
need other info?
--------------------------------------------------------------------------
AMD athlonXP 2500+@2400mhz
2X256mb CorsairTwinx @450mhz
AsusA7N8X-Deluxe
MSI 6800LE 16/6 370/700 :slapass:
I was studying mosfet's and regulator chip (when writing document) and found those were older with 'version 1b' than with 'version 2'. Chips on both card were relative old (2004) compared to your case.Quote:
Originally Posted by unrealuniverse
If the new card really have Hitano's capacitors, it gives proper explanation to senseless lot codes and better performace. Very hard to confirm which capacitors are used, take a look these fotos:
http://img357.imageshack.us/img357/3055/cvex1td.th.jpg
Sanyo CV-EX
http://img345.imageshack.us/img345/5...oevs0zi.th.jpg
Hitano EVS
Can you distinguish those ? The facts about 'green capacitors' are:
- I have seen in Chinese www-page (!!!) a review of GF6800le-card, and a string 'Sanyo CV-EX capacitors' was mentioned below figure (yes, I cannot read chinese but why to have that string ?)
- Lot codes are not very logical with NEWER cards
- I have bought (!!) Sanyo CV-EX capacitors (not the exactly same model) and THOSE have similar labels
- I have never seen Hitano's EVS capacitors 'in live' and NO OTHER foto except the www-page above mentioned (could it be wrong foto ???, I have seen wrong fotos of components earlier)
--
Have anyone ever seen Hitano's EVS capacitors in live ??
Hi,Quote:
Originally Posted by robb-x
Nice you get card working :up:
MSI-cards have been the hardest modify AFAIK. I do know the exact reason (yet), but maybe there are tiny differents at components values. At least MSI's cards have different inductors..
It is quite common NOT have BS while HL2 gives BS ASAP. It is strange indeed, since D3 is really heavy application. There have to be major differences inside programs how they used/allocate video memory etc.
The MSI-card (provided by 'money2' ) failed later on also with 'Enhanced capacitor mod', but now working again after changing capacitors :D. It was similar to yours ("4 3 470 6E"), and that might indicate you will have problems .. sorry to be so pessimistic. BUT in that case changing the capacitors can give final solution also to your problem !
--
Want to help, nice ! In Finish forum we started to collect detailed info about MSI-cards. I want to find out also differences between MSI's cards to have better identification list. Here are the questions I asked:
'Do you have MSI GF6800(le)-card, please give following details'
Real example (money2's card)
model: MSI 6800le (quadro)
age: ~ year, BS came very soon
mods: Adding capacitor to TP's did not help long, changing capacitors to Chemi-con's PXA-seriers (820 uF, 6.3 V) helped
Capacitors (C136&C143): "4 3 470 6E"
Bottom line of large barcode: 4719072010706, NX6800LE-TD128B0409148455
Small barcode: 4910958269'
Yellow sticker: 030
Other stickers/any specials: ?
I haven't seen hitano's caps (i might be staring at them right now), but i've found in the data sheets that the capacitor sizes are slightly different in height or length depending upon which way you look at them. the only problem i have is measuring accurately enough to distunguish between them.Quote:
Originally Posted by SikaRippa
1 hour later
i have just done the most accutate piece of measuring in the world (with a tape measure and straight pieces of wire) and i am nearly sure they're sanyo's caps as they corrispond exactly to the size stated in the data sheet. hitano's would be smaller in length and these are not..
hope this helps
Here are the codes... :( they are very similar to the ones of money2
..one day i'll have to change the caps...but for now I stick with the enhanced
cap mode. Maybe if I don't overclock the RAM It's going to work for a longer
time?!I'll let you know when i get BS back (better if not ;) )
model: MSI 6800le (quadro)
age: ~ year, BS came after ~2month
mods: Adding capacitor to TP1 did not help long especially with HL2,adding capacitor to tp9 & tp10 (2200uF,6V) is working at the moment
Capacitors (C136&C143): "4 3 470 6E"
Bottom line of large barcode: 4719072010786, NX6800LE-TD128B0409148450
Small barcode: 4910958042
Yellow sticker: 030
Other stickers/any specials: never BS in DOOM3
--------------------------------------------------------------------------
AMD athlonXP 2500+@2400mhz
2X256mb CorsairTwinx @450mhz
AsusA7N8X-Deluxe
MSI 6800LE 16/6 370/700
Hi
Nice document
I have a aopen Aeolus gefroce 6800le
I was try the cap mod but doesnt work on my aopen
I dont what can i do?
The capacitors on that card were C136 = 4 3 470 6E
I have a bs problem I was try to find a capacitor but that is really dicicult to find,
Someboy know where can I find the capacitors?
thanks
Good point: sizes can vary :) . I also investigated datasheets (once again..), but I not sure whether we can found relevant differences. The specified sizes vary BUT there are following tolerances e.g. for diameter:Quote:
Originally Posted by unrealuniverse
- CV-EX + 0.5 mm (maybe should be +/-)
- EVS +/- 0.5 mm
.. I will try anyway tomorrow (I have pair of "4 3 470 6E" and "4 7 470 6E" inside PC) with this apparatus (I do not remember english word):
http://www.savenmaa.fi//tuotekuvat/suuri_879.jpg
--
I have been thinking more about the labels. I have scanned fotos of three different cards and here comes very accurate figure of corresponding capacitors. It is interesting that in this figure are all (relevant) different capacitor models:
- '4 3 470 6E' (used by GW, MSI, BS ASAP)
- '4 8 470 6E' (used at least by GW BS later)
- '4 7 470 6E' (used at least by GW, MSI BS maybe..)
http://img98.imageshack.us/img98/385...ors02ai.th.jpg
I think that the font IS different with '4 7 470 6E', what you think ? Hitano ?
Thank you for codes :), I'll try to find some 'logic' of those codes later..Quote:
Originally Posted by robb-x
--
I hope your mod keeps going, it depends on used capaciotor (quality), how much you are using computer etc. I am sure OC'ing speeds up BS-process, so having stock values the card can last much longer.
Hi,Quote:
Originally Posted by jromon
Did you put the capacitor between TP9-TP10 or TP1 ? What kind of capacitor you used, low-ESR, capacitance value ? Long wires ?
About finding capacitors, where you live ?
Thanks for your replyQuote:
Originally Posted by SikaRippa
I live on Madrid ,Spain
I was try tp1 capacitor with 3300uf 6,3 v but doesnt work (long wires),
Only I can work if I do underclock 330 / 600 if I try 700 I can seen artefacts on screen in special doom3 and hl2
Sorry about my english
I hope that you can understand me
Give a try to the enhanced version since in many cases it gives better performance.Quote:
Originally Posted by jromon
http://img69.imageshack.us/img69/4061/capmod35gz.th.jpg
Please post how it succeeded :)
I wrote after last modifications I was able to run stable at 925 MHz. I was wrong. my card was overclocked later to 972 MHz.
today i soldered C3 out.
measured C1 C2 C3.
results are:
C1 = 47nF
C2 = 5 nF
C3 = 25(!) nF (22 or 33?)
just for experiment I soldered C3=10nF on my card. it is stable now at 972Mhz again.
more Mhz - artifacts. no BS!
current config C1=100nF,C2=C3=10nF.
you could have a point there, it does like like a similar but different font was used, but why was Hitano's caps just used for one version and then went back to sanyo's. I think another theory might be needed. also the base sizes and length of the capacitor would be slightly different in the two cases, hoping that average capacitors are exactly as specified in the data sheets then there should be a measureable difference.Quote:
Originally Posted by SikaRippa
Its a real pity that they don't put sanyo on the side or something.
HiQuote:
Originally Posted by SikaRippa
Just a couple minutes ago I did the cap mod tp9 tp10 but I haven´t any diferent
:(
I was try (3300uf at 6,3v)
I don´t know what can i do?
I think that the best solution is replace the bad capacitors
The problem is :
Where can I find a good capacitors?
Im not sure but, would you like a couple of pictures of my card?
Thank you VERY much about those values :toast: (I'll try to simulate also with those to see the difference).Quote:
Originally Posted by zxcv
Your experiments are so interesting, that I have to think/simulate more before to draw final conclusions. I also will check out (once again) my measurements/calculations.. Your card is Leadtek (and had originally "4 6 470 6E" caps?), and therefore newer than the very first cards. Maybe, Leadtek modified those values (after reported BS-feature) OR in fact many/most vendors have different feedback capacitor configurations.
Could you post the exact labels top of Epcos-capacitors (anything else except '197-A' ?
----
By the way, you have a nice OC-frequency, I wonder you are quite happy now ;)
I mean after they go for Hitano never back to Sanyo !!! The GW version 1 ('4 3 470 6E') and 2 ('4 8 470 6E') are older than version 1 b ('4 7 470 6E') in respect of when appeared to market. The version 1 b (I investigated) have older components than version 2 (I had earlier and RMA'd) and this version 1b have modified feedback capaciotors and Hitano's (?). So, after BS-affected cards were found *them* found a solution to overcome problem: feedback modifiaction (=TRUE) and Hitano's-capacitors (=still speculation). After that point all cards should have Hitano's (old RMA'd were modified), and the lot codes became totally meshed. Using this theory one can write following list:Quote:
Originally Posted by unrealuniverse
'4 3 470 6E' = CV-EX
'4 8 470 6E' = CV-EX
'4 X 470 6E' = Hitano's (?) X = 4,5,6,7,9 (2)
(I wonder some old MSI's having BS were using '4 9 470 6E' = CV-EX?)
--
Could anybody provide HIGH resolution (scanned e.g. 300dpi) foto of these green capacitors having labels '4 X 470 6E' X = 2,4,5,6,7,9 ???
If you not afraid to replace capacitors (have some soldering skills), that is the best advice I can give. I'll send for you a PM about caps ;)..Quote:
Originally Posted by jromon
--
Fotos are allways nice :) , please post if easy to arrange
Thanks for your replyQuote:
Originally Posted by SikaRippa
this is my geforce´s pcb
I have a solderind skills
<a href="http://img385.imageshack.us/my.php?image=pcb9ic.jpg" target="_blank"><img src="http://img385.imageshack.us/img385/2212/pcb9ic.th.jpg" border="0" alt="Free Image Hosting at www.ImageShack.us" /></a>
Good & Interesting fotos :up:Quote:
Originally Posted by jromon
--
When comparing to fotos posted by 'sunn' (newer version):
<a href="http://img89.imageshack.us/my.php?image=aopen6800le17gc.jpg" target="_blank"><img src="http://img89.imageshack.us/img89/7651/aopen6800le17gc.th.jpg" border="0" alt="Free Image Hosting at www.ImageShack.us" /></a>
<a href="http://img89.imageshack.us/my.php?image=aopen6800le22nv.jpg" target="_blank"><img src="http://img89.imageshack.us/img89/3430/aopen6800le22nv.th.jpg" border="0" alt="Free Image Hosting at www.ImageShack.us" /></a>
.. some differences have be noticed
- In old version (by 'jromon') some ceramic coupling capacitors are smaller, but in the bottom side of the card there are large ceramics (=for Vmem) in the both cards
- Only 'relevant' difference is "4 3 470 6E" vs "4 5 470 6E" capacitor (are the fonts between those different ??, I cannot say
--
'jromon' : I sent a private message for you about capacitors, did you get it ?
i could take a picture of my card with its 4 5 caps if you would like it. it might help to find the difference.Quote:
Originally Posted by SikaRippa
Fine :up: Yes, the more fotos are the easier is to analyze fonts/codes.Quote:
Originally Posted by unrealuniverse
--
I measure physical dimensions of "4 3 470 6E" and "4 7 470 6E" capacitors. The 'scales' I used has accuracy of 0.05 mm, BUT I could not measure any significant difference at any dimension. So, I wonder the physical size cannot be used to distinguish CV-EX and EVS capacitors (pity).
its a pity you cannot use size to find the difference between them. I'll post the picture at the weekend when i have more time to mess around.Quote:
Originally Posted by SikaRippa
I noticed one important detail related to differences between EVS vs CV-EX -capacitors. Take a look these captions of datasheets:Quote:
Originally Posted by unrealuniverse
http://img312.imageshack.us/img312/4...xds21bc.th.jpg
http://img312.imageshack.us/img312/9...sds20ey.th.jpg
.. the sizes of 470uF 6.3 V capacitors are completely different !! So, these capacitors (c136& c143) CAN NOT be Hitono' EVS-series (except there is an error in datasheet..)
i thought when i had a close look at it that the sizes were different. so that proves that we are on sanyo's CV-EX capaitors. It also now stick the question in the air of, why are the 4 4 and the 4 6 capacitors faulty and the 4 5 are supposidly not. its quite strange how the 4 5 is the only batch that has turned out alright, or is it because they are rare and infomation isn't availiable on wether they BS or not. I'll post my pics late today or early tomorrow with the 4 5 capsQuote:
Originally Posted by SikaRippa
The term faulty is little bit wrong or at least 'misleading'. Maybe better term would be 'there are major differences between capacitor lots'. I am going to refine so aspect to the document, using following guidelines:Quote:
Originally Posted by unrealuniverse
- Capacitor codes of c136 and c143 ( "4 x 470 6E" x indicating manufacturing lot) correspond highly to the version of card
- There are some (major) differences in electrical parameters of c136 and c143 between lots
- In addition to capacitor lots some other changes have been done to different versions e.g. feedback component values of switching regulator
- Feedback modification has a HUGE effect on BS-probability
- I propose that the cards having mofied feedback are working better at least when new(=no BS) even with poor c136 and c143 capacitors => The differences between lots might not been too big
- The capacitors c136 and c143 (Sanyo CV-EX) DO NOT HAVE PROPER ELECTRICAL PARAMETERS FOR THIS PURPOSE
- The ripple-current value (or charging/discharging) of memory voltage node is much higher than specified for CV-EX -capacitors (those are NOT proper for purpose where load is charged frequently, like here)
- The exceeded (ripple) current inside capacitors heats capacitors (because of internal resistance ESR) which REDUCE THE LIFE OF CAPACITORS AND ALTER ELECTRICAL PARAMETERS LIKE ESR
- Increase of ESR means even higher temperature -> wear out process of the capacitors accelerates even more !!!!
- FEEDBACK MODIFICATION DOES NOT REMOVE THE REASON TO THIS PROCESS, BUT ONLY DELAYS THE APPEAR OF 'BLACK SCREEN' (I have measured the currect levels are still too high with the modified card)
--
I'll try to update the document soon or 'near future' with those ideas and add all new measurement/test results I have.
I have been simulated more the effect of the feedback components of the switching regulator (memory part). The capacitor labels are related again following figure.
http://img148.imageshack.us/img148/2...mfb17hp.th.jpg
In the first simulation there are used feedback components of following cards (the values of first two have tolerances about +/- 10 nF):
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')
http://img259.imageshack.us/img259/1...imu08sr.th.jpg
It can be clearly seen that '1' has the highest gain at high frequencies i.e. faster response (=better performance against BS).
----
In the second simulation there are used feedback components of Leadtek card modified by 'zxcv' :toast: . Also I tested to alter value of C2 little bit to find out an optimal value
2. C1=100n, C2=10n, C3=100n (Leadtek 6800le version modified by 'zxcv')
2. C1=100n, C2=5n, C3=100n (Leadtek 6800le test version, more optimal?)
http://img259.imageshack.us/img259/3...imu11gf.th.jpg
I can be seen, that the values 'zxcv' utilized GIVE VERY GOOD PERFORMANCE :up: . I tried to play little bit with values, and find the optimization quite hard task: too many moving parts... However, I found that having C2 a little bit smaller value like 5n (=4.7n with practical components) would give even better performance. Having smaller value, say less that 1n, decreased the phase margin a lot and therefore I advice not to use so low values (can cause stability problems).
Hi againQuote:
Originally Posted by SikaRippa
Just a couple of minutes ago I did the replace capacitor
Well, the good Point is that I haven´t black screen but I can´t overclock the memory
I was try to do a black screen with doom3 hl2 far cry 3DMark03 and 3DMark05
All the programs work fine and no bs ( 385- 667) I can overclock the gpu to 400 but ussually Im working at 380 - 390
If I try to do a overclock the memory I can see artifacts on screen but no bs
I was try 3DMark03 with 385- 750 with artifacts on screen the test run fine at the midle of the firstone the test freeze for a second and then continue the test without counting frames.
Without overclock the memory at standar values (350.700) doom3 work with black points on screen in special on the front screen them you can see some times black points
On far cry the same like doom3 at this values, I can see bat geometry on screen
The estables values:
gpu 350 to 400
Memory 600- 671
the graphics card work fine but I haven´t improve the memory.
As you write on the board I have on my pcb a couple diferent ceramics capacitors (problably memory capacitors ??)
Conclusion
Iam happy about black screen but unhappy aswell about memory overclock
What do you think about ceramic capacitors??
<a href="http://img261.imageshack.us/my.php?image=im0006863em.jpg" target="_blank"><img src="http://img261.imageshack.us/img261/1425/im0006863em.th.jpg" border="0" alt="Free Image Hosting at www.ImageShack.us" /></a>
Good news you succeeded to mod the card :), but pity still having major problems :(Quote:
Originally Posted by jromon
--
Having artefacts etc. at low clock frequency (~700 MHz) is quite unusual. Could you answer following questions, so I can more easily analyze the problem:
- What is a 'BS-frequency' now i.e. card gives BS easily e.g. with 3dmark03/Skialpine ?
- How about before mod ?
- What drivers you have (give a try for re-install) ?
- Did you have similar artefacts before ?
- Did you checked what memories you have 2.2, 2.8 ns ... ?
- Can you seen ANY significant 'injuries' somewhere in your card, missing/burn components etc. ?
- Have you ever test 'ATITool' for artefact testing, what is result now ?
I do not believe changing ceramics would have any significant effect: the artefacts are totally different problem than BS.
Hello to the "BS" community lol.....
Here are some news of my card which I brought for RMA on the 8 of august 2005:
No news at all.... lol My reseller send it to Gainward I think, but I didn't see any graphic card within my postcards yet.... They told me that after the 30th of september, if no card is here, I would have to go to get one new or something like that.
What do you advice me to do then...... I expect to sell the retrieved card to someone I know after having checked it works well of course, cause I feel tired of this loooooooooooooooong problem.
I expect to buy a 6600GT or equivalent, should I keep the 6800LE if I have it (or a new) back? What is the most performant? What investment appears better right now, considering actual prices, the PCI EXPRESS technology etc..... ?
Thanks again, and again, and again......
Hi Godefroy !Quote:
Originally Posted by Godefroy
I hope they will keep the promise, and you will have brand new 'no-BS' card tomorrow :up: If you receive 6800le-card, please post details of card to see what kind of card they give now:
- Capacitor codes
- Number below barcode sticker
- Number at small sticker e.g GW684****
--
I am now quite sure that ALL card using the design (the same PCB) are potentially BS-affected. Some gives BS ASAP, other after few months and the rest will survive... longer. I could not say (yet) what 'longer' means, but I do not see any particular reason why ANY OF THESE CARDS COULD SURVIVE MANY YEARS. Another point of view: maybe the manufacturers are counting on, that we will not use them more that two years..
So, if you consider a new card, 6600GT can be very good candidate in that price range (ATI's out of question of course :D)