In fact we have a single case when having "4 6 470 6E" caps gives BS, if I remember correctly that was MSI 6800 card (not le).Quote:
Originally Posted by charlie
I wonder you have Leadtek-card.. Do you ever have got BS ? How old card you have? Do you know any www-page havong good quality foto of that model or could you provide one of your card ;) ? I am asking, because I have too little information about Leadtek's. I know that there are some wierd features with those cards, but no BS (?).
is it possible to change my 4 3 470 6e into a 4 6 470 6e??
or better said - would it help ?
Quote:
Originally Posted by money2
Yes you if you are working carefully ! Replacing the is more tricky that 'normal capacitor mod'. There is some advices in the document 'Ultimate capacitor modification', I will edit that for next update.
--
I propose that the ripple current value of the original CV-EX -capacitors "4 X 470 6E" (X = 2,3,4,6,8,9 especially 3) is too low to withstand the (abnormal) operation of the switching regulator. Therefore, better quality capacitors should be used. I have some samples Chemi-con PXA-series, having ~ 2 times higher maximum ripple current value. Also Sanyo SVP-series should be OK (that one is used in XFX version 2 !!!)
I do not know (yet) whether this mod will remove BS completely, but as fas as the electrical properties (except ripple current !!!) of the capacitors are ~ same I do not see particular risks when carefully done (!!!).
--
Interesting to get pair FREE ? See instructions and send private message..
There was a VERY informative discussion at GW's forum about BS-problem:
http://forum.gainward.de/viewtopic.php?t=2112
.. where member of GW' s staff CONFIRMS that BS-problem is 'A DESINGFLAW' !!! Moreover, he says problem is now FIXED and the new and RMA'd card would be OK (??)
It is pity, he was not allowed (or did not want to other reason) tell how to identify old/new cards (BS/NON-BS). As explained in the document, i have been thinking that 'Gainward version 1b' has been modified afterwards. in that case they have modified (changed) feedback components of the switching regulator. Properly performed, that could remove oscillation (change loop characteristics) and propability to get BS will go down (disappear?). The modification of those tiny components is VERY difficult task to perform at home (=without professional equipment). Therefore, I'll try to invent methods that can be used even not professionals.
I doubt GW's modification will be working 100 % confidence. The main reason to be sceptic is 'noisy audio' issue, which seems to be exactly the same level in all version.
--
Gaiward MUST have a method/marking to detect old/new cards. Since barcodes, etc are excatly the same in all versions those little OK? stickers at connector could be indication for different versions.
If someone has a GW6800le which does not have BS (e.g. new one/got from RMA) and is has a sticker number different from those shown in #1 (e.g. OK64 for version 1b), please post ! In that case please give also other details of your card: age, labels of 'green' caps..
--
In order to find out differences between old/new version 'noisy audio' recordnings could be used. I would be extremely gratefull, if somebody would perform following test:
- A BS affected card required
- CD-audio cable connected to soundcard (integrated is also OK)
- CD-audio level to 'maximum' at volume control
- Record a noisy audio generated by 6800le-card e.g. this program
http://www.snapfiles.com/get/wavepad.html
- Audio source for this recordning is either internal CD-line or by connecting headphone-line to mic externally
- Following recording are requested
1. PC is 'ON', but no other processes (applications)
2. Running a 3d-benchmarks e.g. aquamark3 (audio of the benchmark should be OFF !!)
- Send this (these) audiofiles (mp3) for me 'sikarippa*yahoo.co.uk' (please note here * = @ )
Quote:
Originally Posted by SikaRippa
sure i want a free pair but i am a really noob in soldering or any electrical work ^^ So i won´t do it by myself. I ll go to a good friend of mine - he´s works in a tv-shop and can do those things easily - i hope :p:
Perhaps u can tell me some things which are really to attend while changing the capciators.
A little Checklist ( for a Soldering specialist ;) elseif it would be never ending list ) what should not happen during the change would be nice.
I hope u understood my freakin english... :D
hehe - looking forward to hear from you soon :-)
---------------
another option:
i ll send my card to you and you´ll do it. In good hope that it wont be destroyed ;) But i am more than sure that u work carefully :-)
If anything bad happens to the card - doesn´t matter, there´s no guarantee anymore... :D
kind regards money2
I have written more detailed version of 'capacitor replacement guide' including some hints&warnings, and it will be added to next update (when?, soon I hope..)Quote:
Originally Posted by money2
--
I sent a private message for you considering capacitor replacement ;)
i answered or?!
??? .. You sent private message (?) At least I have not received yet...Quote:
Originally Posted by money2
Hmm, my card started BSing again...around 3 months after I did the enhanced capacitor mod. I had to take out the capacitor I put in first, and put a new one on it :S
(XS is up again :D)Quote:
Originally Posted by KoR_Wraith
--
Yep. It seems that some cards gives BS again, even with 'enhanced capacitor mod'.
In that case the changing capaciotors should give a solution. If your considered it please use PROPER capacitors for that purpose. Also I suggest to change BOTH capacitors in the same time, since having one old parallel in good one might be risk (the ripple current is mainly over one capacitor in that case).
--
I still have some capacitor samples left, send private message if interested ;).
--
We will hopefully get test results about the capacitors change in near future (~ week). So If you want to heard results first, please wait a little bit.
For the enhanced capacitor mod, if my card doesn't have BS, would it have any effect (like overclocks) to do it?
And if I use a 470uF capacitor would it still be useful? The capacitors are high quality ones that I have in my CMOY amp (which I don't use anymore) so I thought that I could cannablize parts if it would do something.
BTW, the caps on my card (Leadtek 6800LE) are:
4 5 470 6E
4 9 470 16A
Hi,Quote:
Originally Posted by ChronoReverse
Thank you for capacitor information !
--
Many users who have made (enhanced) capacitor mod have reported increace of memory OC capability up to 50 MHz. And this do not require a BS card,only important thing is to have exactly same PCB or design for the card (like in the case of all (???) Leadtek 6800le-cards).
The cap value of 470 uF is not too high, but I wonder it helps (at least in the case of good quality caps). Cannablizing of old ones is OK (I do it all the time :) ), but be sure those are still working (no short..) !!
hello
1st of all^
SikaRippa - outstanding work!!
I have Leadtek 6800LE card with BS "feature". my caps are 4 6 470 6E.
BS appears usually in first 3-5 minutes of heavy graphics load.
I don't wanna RMA my card, its 16x6 =)
now I'm waitnig 2 caps Sanyo SVP series (4SVP1200M). wanna do extreme mod =) I hope it will solve BS problem. additional cap (Hitano EXR series 1000 uf 6.3v) helped avoid BS at 700MHz).
also one my friend found smth intresting... he helped me to find caps for replacement. maybe this will help a little... =)
http://www.globalsources.com/gsol/I/...1000858291.htm
Thank you for comments & information !! Especially 'Hitano'-capacitors are very interesting, I'll propably add those to 'good' caps list.Quote:
Originally Posted by zxcv
I think that those Sanyo caps would be just perfect (= high ripple current value), if you want some hint&tips please wait few days for update.
--
I have done MANY things considering BS-issue during a few days. I am now documenting those, but there are soooo many items: maybe I'll do a minor update first mainly for 'Ultimate capacitor modification'.
--
For instance I have successfully tested this mod, greetings to 'money2' who provided card :toast:
Here comes a preview figure (after):
http://img331.imageshack.us/img331/8...p3003rd.th.jpg
.. completely BS-affected card turn live again ! (no other mod worked for this card)
:D
I ll vet it ;)
and i made Sika immortal - look at my signature :banana: :banana: :banana:
I already have solder two caps, like these... on this picture of 6800. taken from ATI Radeon dead cards.Quote:
Originally Posted by SikaRippa
but they were only 470uF 6.3V.
http://www.ixbt.com/video2/images/re...scan-front.jpg
so.
BlackScreen came right after 3 seconds showing windows desktop...
then soldered back green 4 6 470 6E (Hitano?) caps... and one additional Hitano EXR between TP9 & TP10... now waiting Sanyo SVP...
It's possible, that smth else should be modified... so.. i'm already waiting for BS FAQ update =)
I don't own a 6800LE but I read through your guide, you've done alot of research there. Good job :thumbsup:
Quote:
Originally Posted by money2
Thanks !! For me you comments is a signal that work has not been useless !!Quote:
Originally Posted by masterofpuppets
OK.. I'll try summarize what you have done:Quote:
Originally Posted by zxcv
1. BS with original caps
2. You removed 'green caps' and replaced those with got from radeon (light blue, might be Chemi-con's PXA-capacitors 470 uF 6.3 V) -> BS ASAP, even without 3d-load
3. Soldered back ORIGINAL caps and one leaded cap between TP's ?? Or did you use those Hitano's shown in your link ? You do not get BS with this setup ?
If I understood correctly (?) it is very wierd the card was not working in the case '2'. Since the capacitance value (470 vs 820) should not make difference, more important is ESR (and max ripple current after ageing). And if those were PXA-capacitors (?) ESR should be low enough (maybe caps were burn/got bad when desoldered from radeon?) . Are you sure you had both capacitors soldered properly ?
yes.. everything summarized correct. #1,#2 and #3
about #2 and #3.
yes, replaced those with got from radeon (light blue, might be Chemi-con's PXA-capacitors 470 uF 6.3 V) -> BS ASAP, even without 3d-load... 3, maybe 5 seconds showing windows desktop.
I soldered back original 4 6 470 6E caps... tested card. it restored to #1. then did enhaced capacitor mod, using Hitano EXR-series cap (1000uF 6.3v) between Tp9 Tp10. now it's no BS at 700 MHz. i haven't tested max memory clock yet.
but.. i still want to replace 4 6 470 6e caps.
about caps from radeon. i'm not sure both capacitors were soldered properly from radeon. one man gave them to me. i only soldered them on 6800le. at least my card still working with 4 6 470 6e caps after my soldering as earlier.
u think I should try another pair of those caps?? (blue, from radeon)
should i solder anything else on my card, except C136 C143 capacitors. maybe some little SMD caps?
forgot one moment. memory voltage on my videocard is 2.80 V.
also i think, possibly 4 6 470 6e - also Hitano's caps...
maybe i'm wrong, but havent seen before, that two manufacturers used same marks for their alluminium SMD-caps.
Quote:
Originally Posted by zxcv
OK.. I think those 'light blue' cap were bad/badly soldered. These caps are sensitive to high temperature and those might be ruined when soldering iron has touched capacitor case ! Having good contact beneath capacitors is also hard to obtain i.e. solder does not melt properly. Since you got the card working with old caps, the reason is most likely bad 'light blue' caps.
It is a little bit risky to use old caps (especially SMD caps in high performance applications), but if you rip those off VERY carefully you can try again (make it yourself and DO NOT touch caps with iron). However, note that PCB of the card (copper areas) is getting worse every time you solder/desolder..
--
It is interesting question, indeeed : Are ALL/NONE those capacitors Sanyo CV-EX ? Or Hitano's ? The same markings are very wierd... I have some brand NEW CV-EX capacitors are those HAVE that kind of markings. Also I have seen in some www-page review where said these are CV-EX -series. Sometimes figures in www-pages are not correct, anybody seen Hitano's capacitors (100 % sure) in real life ?
IF cards really have capacitors from different series Sanyo CV-EX <-> Hitano EVS, that would explain many things :) (Hitano EVS are better..)
ok..
today at morning i've token this dead radeon. it isn't 9700. it is 9800. on this card were 4 more blue caps (470uF 6.3V).
few minutes before i soldered another two blue caps. by myself. this time i soldered them from radeon and i soldered them on 6800le.
somth changed =) at least windows booted-up correct and installed the driver. next boot-up was correct too.
card works in 2d now without problems. the only thing is - this PC is on my work. i have only quake3arena here =) so... no BS appeared yet.
later i'll test 6800LE at home. i have a little hope now...
about caps.
i'll try to order some Hitano EVS series caps. if they will be the same as those green 4 6 470 6e... it will be really intresting. but it will take some time. maybe 4 weeks maybe even more.
thank u SikaRippa for ure investigations and support.
i'll post here later about my card. also i'll do a photo (scan) of my video.
Big thanx for your work on this prob SikaRippa
did the Cap mod on my gainward 6800le, which was a very bad BS
Now everything working solid,
Np overclocking np at 400/800 although fitted an artic cooler just keep those temps down
UKasassin
Nice :clap: , likely those you first tried were bad.Quote:
Originally Posted by zxcv
--
If you get Hitano's capacitors, could you provide a foto of those to make a comparison whether they look like exactly the same as CV-EX. Here is a foto of CV-EX -capacitors:
http://img357.imageshack.us/img357/3055/cvex1td.th.jpg
I both caps look like the same, it would be in theory possible that models not having BS are using Hitano's capacitors and BS-affected have CV-EX !!!