K7SEM capacitor identification

[deconstructor]

Hi all, I've been going through my old motherboards getting rid of the obsolete ones and I came across a K7SEM mobo. It was one of the first broken computers I got when I started getting interested in computer repair. I couldn't fix it at the time, so I did what I usually do in that situation... I desoldered all the components off it to use for future repairs. Well.....I'm pretty sure it was leaky capacitors (which I didn't know about at the time) and I'm looking to put the components back on that I took off. If anyone has this motherboard, I'd greatly appreciate them telling me what the capacitor ratings of the following labeled components are: CE32, CE33, CE38, CE37, CE34, CE45, CE9, CE10, CE15, CE25, CE16, CE26, CE6, CE7, C219 and L15. I've tried to look for pinouts of the mobo, but can't find anything. I'm just glad I didn't tear anything more off....Many thanks!!!

[pfrcom]

Re: K7SEM capacitor identification

I believe ECS K7SEM is the same as PC Chips M810

Here's what's on my M810

CE32 Up against AGP slot - probably same as CE34
CE33 Up against AGP slot - probably same as CE34
CE38 1000uf 10v
CE37 1000uf 10v
CE34 1500uf 6.3v
CE45 1800uf 6.3v
CE9 1800uf 6.3v
CE10 1800uf 6.3v
CE15 2200uf 6.3v
CE25 2200uf 6.3v
CE16 2200uf 6.3v
CE26 2200uf 6.3v
CE6 1000uf 10v
CE7 1000uf 10v
C219 100uf 16v
L15 Where is it ?

[PCBONEZ]

Re: K7SEM capacitor identification

L15 is an inductor, not a cap.

[Pyr0Beast]

Re: K7SEM capacitor identification

K7SEM is junk. Throw it away.

[deconstructor]

Re: K7SEM capacitor identification

Thank you pfrcon and pcbonez. Just google imaged the m810 and it is exactly like the k7sem (with the exception of an interchangeable processor socket). Based on the image, L15 is definitely an inductor. I'll throw em on the mobo and test. I do realize that it is a piece of junk, but it maxes out at 1 gb of ram (got nothing else to do with the pc133 512 sticks I have lying around) and the processor beats the socket 7's that are more useful on my computer board collage. I appreciate the speedy responses very much. Thanks again!

[xrror]

Re: K7SEM capacitor identification

The K7SEM may be junk, but it's still looks like a fun project. I just happen to have one I'm trying to refurb for a kid's computer, and it indeed has some bulging caps. It's a revision 1.0, so if you need any info on what caps it has i'll be more than happy to pull it out and list them off.

Also... these old PCChips/ECS mobos. While some are hopeless, many are pretty darn entertaining to abuse for "retro-overclocking"... The M863G v5.1 was the board that got me back into serious soldering iron work on motherboards. While stock it's a joke, the lax SiS timings and chipset power overprovisioning (for the onboard video) let it do some ridiculous overclocks. The max "supported" speed was 166. It would run 250. And the Newegg bundle that was literally $99 for the mobo with an Athlon XP Mobile. Bahahaha...

At 2.5Ghz, the "junk" board was/is some of my best memories. And yes I still have it, and it still runs. And ironically if this K7SEM is too hard for me to find 512mb SDRAM guess what's going in that case? hehh

Anyhow enough of my rambling, I'm hoping you haven't thrown away the K7SEM yet, i've read rumors that there is a way to biosmod or clockgen the FSB up (okay, so I just need to find a datasheet for the PLL) and this would just be fun to beat up an old board. Also you can add multiplier adjustment to any Socket A board, http://fab51.com/cpu/barton/athlon-e20.html


While I don't work there anymore, the page is still up with pics. Be afraid of a franken-ECS http://ux1.eiu.edu/~jckoontz/personal/m863g51mod/

[Pyr0Beast]

Re: K7SEM capacitor identification

Oh sweet mod.

250MHz is nice, even nForce2 boards have troubles coping with such speeds.

[xrror]

Re: K7SEM capacitor identification

Well to be honest, the M863G wasn't exactly happy about 250. Later on when I was running Rosetta@Home on it, around 240 was needed to keep it from throwing computation errors occasionally. I think the biggest "limitation" was the 2-phase power. I had the processor bridge modded for 2.0v (yes, nuke mode) and it would sag to 1.8v under load, so heh... So quality wise that it didn't blow the board during it's eternal abuse - they're not as bad as people make them out to be. I seriously doubt ECS/PCChips was targeting this board to be paired with a volt-modded Geforce 6800 that was on it too. Did I mention this system was a furnace? =p

Back on topic with the K7SEM. I swapped out the Duron (Morgan 1Ghz, 100fsb) with an Athlon XP (Palomino 1.6Ghz, 133fsb) and can't run it stable at 133fsb in the board. 112fsb works fine, so I'm thinking it's probably those capacitors being toast, plus I really need to check inside the stumpy 100w power supply in that case too.

I need to round up all the junk I have sitting around that needs capacitors, and bulk it up into an order, so when I do that this will be on the re-cap list.

[Pyr0Beast]

Re: K7SEM capacitor identification

2V isn't that much. Max. what you could do safely on air tho. But water. hhhheeeeheee
2.2V and more )

You had lots of luck that CPU liked such high FSB. Many stop even before they reach 200.

High voltage on NB/SB could work wonders. I would love to see that board peak 280MHz, it really seems doable

Palominos are friggin cooking stoves. Caps might be and probably are cooked. I am surprised it even works at 133MHz. I had trouble getting it to 124MHz on this board with duron 1.3GHz.

[xrror]

Re: K7SEM capacitor identification

well, the lack of knowing what/if dividers on the M863G past 166 kinda started being a problem over 252 ;D While the bios maxed out at 255, you could run clockgen on it to retardation ... but pretty much you couldn't keep anything running on the board, but you didn't really care since your OS image was data rot central by that point anyway. Plus it didn't really have any memory ratios downward... so 250fsb was truly 250 (500DDR). The hard drive kept corrupting over 252, and after trying three different drives (different models/brands) it was pretty much as suspected the southbridge. The scary part never was sure about... just how fast was the AGP running at that point? The downer was I still couldn't catch nForce2 boards in benchmarks even at that speed, but it was still hella fun. I mean, $99 for a board with Athlon XP mobile (barton)... you bought the processor, and got a board free.

So back to the K7SEM, I pulled the board out since I was cutting on the case anyway. I had some Sanyo caps in the salvage bucket (not the best I know, but it's hopefully better than the bulgers).

The board silkscreen: K7SEM REV:1.0 15-A01-010100
judging from the names given to the capacitor locations, this isn't the same version deconstructor has.

There are a TON of tiny caps on this board, where they use an abbreviated code for them. Which I'm too lazy at the moment to hunt down - hopefully these don't go bad too much, or the stuff doesn't care. G-Luxon seems to be the brand for majority for these little guys, 203(M), 204A(M), also many "little fatties" 16v 100uF luxons.

Anyway, here are the major caps.

purple OST I.Q 6.3v 3300uF RLX 105c
Four tall ones, directly beind IO sheild
TC8, TC12, TC15, TC24

green G-LUXON 1000uF 6.3v
TC31 - directly under TC24
TC38 - directly below ATX power connector
TC5, TC7, TC21, TC33 - around memory slots

dark green OST I.Q 6.3v 2200uF
Eight, paired by twos with a coil. Power phases?
TC3 TC4, TC9 TC10, TC13 TC14, TC17 TC18

green G-LUXON 1500uF 6.3v
Two, one above and one below end of AGP slot
TC44, TC55

To indicate just how many caps are on this thing, the highest cap number I've found so far is TC60 !

So the caps that were obviously bulged were three of the eight OST I.Q 6.3v 2200uF. Now what's really interesting is there seem to be two "styles" of these. They both look identical, except the vent "seams". Five of them have a "tri-wing" vent, while three have a "cross or X" pattern vent. All of the cross pattern ones are bulged, and none of the tri-wing ones. Just odd.

I didn't have eight sanyos, so I redistributed the caps into pairs, so there is one OST and one Sanyo per pair. Yea, I pulled an extra OST out (that wasn't bulged). This probably wasn't needed, but since the sanyos I have are 3300uF instead of the 2200uF they replaced this was just a hedge in case the board didn't like the phases (if that's what they are) having different capacitances.

So now, after typing this small novel I'll go put the board back in and see if it actually works!

[xrror]

Re: K7SEM capacitor identification

yeay, it still works.

The inability to boot at 133fsb in retrospect was due to me having Dram Background Command and LD-Off Dram RD/WR Cycles set to "Normal" instead of "Delay 1T"

yea... so errr, don't use "Normal" when running 133fsb on this board. Totally makes sense . That or possibly it works with fewer sticks of memory and/or smaller ones.

Wish this board had 124 and anything over 133. Maybe that'll be the next part, time to look up the clock PLL

[Scenic]

Re: K7SEM capacitor identification

i would take a look at the bios for hidden settings and stuff, but the ECS website is f*cked up as usual..

how the hell is one supposed to know if he has a K7SEM 1.0 or a K7SEM 1.0+B1 !?

edit: apparently, Socket 423 is now an AMD socket (at least for ECS)
-> http://www.ecs.com.tw/ECSWebSite/Pro...uID=24&LanID=0

[Uranium-235]

Re: K7SEM capacitor identification

I saw soooooo many SEM's at goodwill that were flat out dead with bad caps. I probably came across like 15

[xrror]

Re: K7SEM capacitor identification

Thanks for offering to look at the BIOS. I think this is the "non-B" one. When I get a chance I'll see if I can confirm it. If all else fails I can dump the bios with awdflash or whatever this board uses.

I need to check to see if it has the latest BIOS anyway, so hopefully later today I'll get that info.

@Uranium-235
I wish our goodwill even had fun stuff like that. Then again you can't find any old video game stuff for anything in thrift stores around here which is really annoying. And the pawn shops around here are all "ebay highest price" = ripoff.

[xrror]

Re: K7SEM capacitor identification

Alright, so I flashed the board with the bios listed for the generic K7SEM rev1.0 and it works. So I'm guessing that's the right one.
http://download.ecs.com.tw/dlfileecs...m/k7sem630.exe

Hah, and in true ECS humor, when downloading awdflash from the site:
You download a self-extracting zip named awd832B.exe which then attempts to extract the program named awd832B.exe. Think about that for a moment then laugh

So the clockgen for this board is an ICS 94208EF. Which seems the same as the PCChips M810D rev7.5. If I can get some motherboard pictures and/or other confirmation of which PCChips boards are equivalent of this one that opens up more potential bios's to mod from the PCChips side. Back in the day there were many modded bios's for PCChips so I just need to find the equivalent for this ECS board.

From looking at this thread http://www.rebelshavenforum.com/sis-...;f=19;t=000260 it appears that at least the clockgen supports more speeds above 133.

140/35.0/140 <-this would be nice to try
150/30.0/150 <-no vcore adjustment, and it's a Athlon pally... probably not gonna go
166/33.3/166 !!! <- some of the 256mb sticks will do 170'ish, but i dunno about the 512mb
200/33.3/200 !!!! lol SDRAM BURN

But they also talk about the need for adjustments to PCI latency. But that's a hurdle for when/if I get there.

I did do some tinkering with the award bios tools (i think?) way back in the day, where you could add/remove modules and enable/disable menu options. But I never got to the competence of actually adding NEW options or hexing in assembly jumps to hack in new functionality. Soo... I'd be more than happy for you to take a shot at it Scenic. At least the bios chip on this is removable LOL... unfortunately all my K7S5A's (literally had five of those from random places) are all dead now. Like... volt mod dead so I don't have an easy hotflash board anymore.

[Pyr0Beast]

Re: K7SEM capacitor identification

Repair them !

I doubt 150 is even possible on such crappy board. Cas3 perhaps yes. Cas 2 ... I really doubt it.

[Uranium-235]

Re: K7SEM capacitor identification

you coulden't pay me to get one of those POS's. ooold K7 + SDR bleh, and an ecs at that

[PCBONEZ]

Re: K7SEM capacitor identification

You can pay me to give me parts all you want.

[xrror]

Re: K7SEM capacitor identification

Heh... HoneyX modded BIOS's... much damage potential here
http://www.lejabeach.com/M810/M810.htm

It looks like many of the PCChips M810 versions use the same bios. I just need to make sure the board I have is one of them.

I need a good free image hosting service, I'm pretty sure after digging around this forum that somewhere a person has already taken a pic of this board. If it helps any this version of the K7SEM doesn't have the integrated LAN jack on the I/O shield. It has a pin header on the motherboard and a small breakout cable (can't remember what you call them ATM). The physical jack on the end of the cable is a backplate in an empty case slot. It's not a CNR card, just a header cable.

So basically if there is a PCChips M810 variant that has a "LAN header" that's probably this board. Now to dig. Probably a google search for someone going "I lost the %&#* proprietary ECS cable" The SuperIO chip is an ITE IT8705F, and the LAN is a Realtek RTL8139C if that helps any. Random thought... with a CNR card, you would have dual NIC's (the 2nd from the SiS 730 chipset)... I digress.. again.

@Pyr0Beast lol, now I MUST try 150 haha. I'm more just wanting to get 140fsb, and even that's more for the 37mhz PCI and a ram speed over 133. I can bridgemod higher mult if I want CPU clock. 150 seems doable, I mean SiS chipsets are unified (which I forgot about in my M863G rantings above) which I think is why the traditional southbridge functions are so resistant to FSB overage, that and some active cooling on the chipset. But based on the rebelshaven post on the clockgen, 150fsb is using a 1/5 divider (!) for a "PCI speed" of 30 which kinda defeats what I want. The onboard video (which I'm not using) should hopefully imply some power overhead built into the board for the chipset. And as SiS chipsets usually have kinda slow default memory timings actually works in favor of pulling higher clocks - granted you'll be slower per clock, but hey at least you might actually be able to reach them. And lol no, I seriously doubt any of my "mongrel" collection of memory is badass enough to hit 150+ at CAS2. I mean ugh the only 512mb SDRAM stick I have I think came out of some G3 Macintosh that was surplussed at the university. Be amazed it was even PC133...

Restraining all this enthusiasm is the 100w gutless wonder power supply in this case. The case is so cramped a normal size PSU pretty much would be where the 5 1/4 bays are. Why not just use another case? Well this is a kid's machine... and it happens to be painted up and done as a Barbie case (no, not the official Mattel Barbie PC... which is ridiculous cramped). Lol so whatever I do, It's gotta fit in this thing in the end.

@Uranium-235 lol yea I'd never buy one. Unless they were like $2 a piece.. and even then the requirement for 512mb SDRAM to make them useful for anything modern outside of some linux appliance pretty much makes them useless. I just happen to have all this stuff laying around, so it's all for fun. The only reason I'm still posting on it is hopefully it gives some amusement for everyone else with a similar morbid curiosity as me... Basically, if someone WAS dumb enough to "pimp their K7SEM" would it still suck? Probably, but hey maybe some surprise will come of it. Or blue smoke. Regardless, it's for fun =D

I mean for perspective, the M863G that I keep going off topic with - I could just throw that in right now, with 2GB of DDR and be done with it. Except for the power supply space issue. And I'd need to find another high AGP speed tolerant card - the Geforce 6800 paired with that board died long ago, plus ugh it'd be WAY too hot for this. But even with those issues that would still be a way better solution than torturing the K7SEM.

But what fun is that? Yes, I am a hardware "gidiot" ... I enjoy this stuff, even if it's of absolutely no interest or relevance. I still want to put a Cyrix 5x86 under a peltier and see if it's possible to hit 200mhz. Pointless yes... but hah. I have no life.

@PCBONEZ lol yea sign me up for that payment program too.

What's weird is in a roundabout way, if deconstructor ever revisits this thread... is if this board turns out to be compatible with the other M810's then whatever BIOS carnage I do would actually be relevant info for him. You know, assuming he also has a sadistic hardware streak

[xrror]

Re: K7SEM capacitor identification

This post just for off topic fun, plus me being happy the links still worked. I used to work at a university but I got laid-off. Well technically someone who had the same job classification as me was laid-off, and had seniority so they bumped me out of my job =(

Anyhow, they haven't killed the account, so you get pictures of other ECS abuse. This is the Geforce 6800 non-ultra AGP that was in that M863G
http://ux1.eiu.edu/~jckoontz/personal/Leadtek6800nu/

And this is me trying to beat my buddy who had a DFI LanParty nForce2. I couldn't catch him, he was consistently 200points ahead with a 6800GS 512mb. Bah. So while CPU wise I had him easily, the fact his video card was DDR and mine SDRAM he still had me. Not for lack of trying, I bios modded that card to death down to individual memory timings (some went down to zero, how does that work?) and minimum latencies on top of the volt mod, pixel/vertex unlock ridiculousness . The nail I couldn't fix though was my card was only 128mb, which would hitch sometimes in 3dMark05.
http://ux1.eiu.edu/~jckoontz/personal/BenchHighs/

Hehe, but in the end? My cpu and proc only cost $99 combined. And the vid-card was like $180. Soo... 200points. No sweat. Also later his DFI died from guess what? Bad caps.