About Soltek 75FRN2-L boards...

A couple people asked about these boards here recently, one person looking
to replace one due to blown caps. Well, just thought I'd mention that I run
a few of these and I have two 'revisions', E3 and E4. (Silk-screeded by the
edge of the board near the last two PCI slots) The E3 uses mainly
purple/dark blue caps made by "GSC" and these are extremely failure-prone,
especially the five large 3,300uF caps in a row next to the CPU as well as
three small 1,000uF caps near the RAM slots. The E4 board uses different
caps (can't discern name but they're black) that don't have anywhere near
the same problems.

This is not to say that the E4 is fine. I've had to either slightly raise
the vcore or lower the clock-speed of these over the last year to keep them
stable. However, they haven't failed catastrophically like the E3 did. The
caps on the E4 still look fine whereas the caps I mentioned on the E3 are
visibly blown and leaking electrolyte. If I can source them I'll replace the
caps on the E4s as well as the E3s.

Just FYI. I really like these boards, one of the better Socket A boards that
were available IMO, and am dissappointed that this problem is showing up in
boards made as late as 2003. The 'bad caps' problem was well-known around
the world by then and there's no excuse for them to have been used on new
boards. Perhaps that's why Soltek no longer exist.

I just wish I could source some low ESR, 3,300uF, 6.3v, 10mm diameter caps
in New Zealand. (Or even in the world, at the right price as postage is a
killer) Preferably a good make such as Rubycon or Nichicon. However it seems
that the agents here have decided that there isn't a market for them. The
Nichicon NZ agent told me they didn't have any ex-stock and minimum order
was 2,000 units with a 90-day turn-around and I'm having similar problems
everywhere I look. The guy at www.badcaps.com won't ship capacitors outside
of the US/Canada. Meanwhile I have what is a fast and capable machine
sitting in parts for want of a handful of caps. I hope that it's not just
the first of several. :-(

Cheers,
--
Shaun.

"~misfit~" <> wrote in message
news:44f375b1$...
>A couple people asked about these boards here recently, one person looking
>to replace one due to blown caps. Well, just thought I'd mention that I run
>a few of these and I have two 'revisions', E3 and E4. (Silk-screeded by the
>edge of the board near the last two PCI slots) The E3 uses mainly
>purple/dark blue caps made by "GSC" and these are extremely failure-prone,
>especially the five large 3,300uF caps in a row next to the CPU as well as
>three small 1,000uF caps near the RAM slots. The E4 board uses different
>caps (can't discern name but they're black) that don't have anywhere near
>the same problems.

I have an E4 with green caps... running 2.5 years now without any major
problems.

> This is not to say that the E4 is fine. I've had to either slightly raise
> the vcore or lower the clock-speed of these over the last year to keep
> them stable. However, they haven't failed catastrophically like the E3
> did. The caps on the E4 still look fine whereas the caps I mentioned on
> the E3 are visibly blown and leaking electrolyte. If I can source them
> I'll replace the caps on the E4s as well as the E3s.

A while back I got a bit panicky when I thought I saw the caps bulging a
bit. I sent an email to Soltek and they sent me replacement caps (black
ones). They're sitting in a box in my colset, as the original green caps are
still doing well.

(They still seem to be bulging, if you shine a light on them at the right
angle, and tilt your head, and squint really hard... but in any case,
they're not getting any worse.)

No problems with stability, in fact I'm running a 2500+ Barton at 2.2GHz
(3200+) and I got the thing *undervolted* at 1.5V. I did find out, however,
that the board for whatever reason doesn't like any BIOS other than 1.7L

> I just wish I could source some low ESR, 3,300uF, 6.3v, 10mm diameter caps
> in New Zealand.

That's what you get for living at the end of the world. Move to a normal
country. :-D

Mercury wrote:
> "~misfit~" <> wrote in message
> news:44f375b1$...
> > A couple people asked about these boards here recently, one person
> > looking to replace one due to blown caps. Well, just thought I'd
> > mention that I run a few of these and I have two 'revisions', E3
> > and E4. (Silk-screeded by the edge of the board near the last two
> > PCI slots) The E3 uses mainly purple/dark blue caps made by "GSC"
> > and these are extremely failure-prone, especially the five large
> > 3,300uF caps in a row next to the CPU as well as three small
> > 1,000uF caps near the RAM slots. The E4 board uses different caps
> > (can't discern name but they're black) that don't have anywhere
> > near the same problems.
>
> I have an E4 with green caps... running 2.5 years now without any
> major problems.

Ok. My E4 that I have here has brown caps. Running semi-fine. So that's
brown, blue and green for the 3,300uF ones. I heard after the fact that
Soltek was grabbing any components they could get cheaply down at the
markets towards the end there, and slapping them in their mobos. A shame
they went bust as, other than the caps, this is a mighty-fine board. I think
that they were a small company which tried to expand too quickly with
insufficient capital.

> > This is not to say that the E4 is fine. I've had to either slightly
> > raise the vcore or lower the clock-speed of these over the last
> > year to keep them stable. However, they haven't failed
> > catastrophically like the E3 did. The caps on the E4 still look
> > fine whereas the caps I mentioned on the E3 are visibly blown and
> > leaking electrolyte. If I can source them I'll replace the caps on
> > the E4s as well as the E3s.
>
> A while back I got a bit panicky when I thought I saw the caps
> bulging a bit. I sent an email to Soltek and they sent me replacement
> caps (black ones). They're sitting in a box in my colset, as the
> original green caps are still doing well.

Yeah, I know of one other guy who did the same thing and he has the caps
just sitting there too. However, his will no longer run at 200MHz FSB, he's
had to drop it to 180 to get it stable. I tried emailing them quite a while
ago and get returned email errors, "that domain no longer exists". I missed
out on Soltek sending me caps.

> (They still seem to be bulging, if you shine a light on them at the
> right angle, and tilt your head, and squint really hard... but in any
> case, they're not getting any worse.)

Good to hear it.

> No problems with stability, in fact I'm running a 2500+ Barton at
> 2.2GHz (3200+) and I got the thing *undervolted* at 1.5V. I did find
> out, however, that the board for whatever reason doesn't like any
> BIOS other than 1.7L

I'm running all mine on 1.7L as well. I never upgraded further as the
further BIOS upgrades are only to show the Sempron range's names correctly.

My main E4 used to run an XP2500+ at 2.2GHz at default voltage, I never
tried lowering vcore at first, it didn't occur to me, I was just happy that
my 2500 would run at 3200 speed without *raising* vcore. However, after a
couple years or so it would keep re-booting just as Windows was loading.
Sometimes it would re-boot 6 or 7 times before it'd actually load Windows.

Because the machine would run Prime95 all night and Memtest all day I didn't
suspect the CPU/RAM/Mobo. At first I thought it might be the IDE cable, then
the HDD so I used it as an excuse to upgrade. When it kept happening I
suspected PSU and checked the 12v and 5v rails while the machine was
booting/re-booting to see if they sagged. They were fine, as I thought they
would be, I have a good 520W PSU.

Next I suspected the IDE controller might be flakey. I swapped the HDD onto
IDE2 but it still did the same thing. Finally, convinced that it must be
something effecting both controllers I bought a PCI - ATA/SATA controller
and ran the HDD off that. Still the same. As the machine ran fine and passed
all stress-tests I could throw at it I didn't suspect that it could be
easilly fixed in 2 minutes by raising vcore! I spent lots of money and a
month trying everything else first.

I'm assuming that the caps went out-of-spec over the time it was running and
the load on the 12v rail as the HDD was thrashing and the CPU running at
100% while the OS was loading was too much for the failing caps and they
couldn't hold the voltage. (The 12v rail from the PSU was fine, it must have
been after it hit the board's circuitry) It was only by accident that I
discovered the cause of the problem. I decided to down-clock my machine as I
was leaving it running 24/7 and I wasn't using anything like it's full power
anyway.

As soon as I reduced clock speed (multi) it booted clean every time. If I
put it back up to 11 x I had to take the vcore to 1.725v before it would
boot cleanly. It was a bit warm at that voltage (60°C on-die after 30
minutes at 100% CPU) so I decided to run it a bit more slowly. It's now
doing 10 x 200 at 1.55v. At first, 3 months ago, it did 2GHz fine at 1.50v
so it's definitly not getting better. I raised it to 1.55v last month after
a few false starts at booting.

My 3,300uF caps don't show any signs of failure but I'm dead sure that they
are the problem. It wasn't until after I had this problem that I had a PC I
built using an E3 board come back and I saw it had bulging *and* leaking
3,300 caps and 1,000uF caps near the RAM. Probably why some boards won't run
stably at 200MHz, the big 3,300 caps are just power filters/smoothers for
the CPU, I suspect the 1,000uF ones have more to do with the FSB.

I've got a mate in the US who has ordered the caps I need from
www.badcaps.com (who won't ship outside of the US/Canada) and he'll post
them on to me. I'm a lucky guy.

> > I just wish I could source some low ESR, 3,300uF, 6.3v, 10mm
> > diameter caps in New Zealand.
>
> That's what you get for living at the end of the world. Move to a
> normal country. :-D

Heh! It has it's pros and cons. All-in-all I'm happy here. "Normal"
countries scare me, they have terrorists bombing them for perceived wrongs.
NZ never hurt anyone and the scenery and the weather are just fine. :-)

Cheers,
--
Shaun.

"~misfit~" <> wrote in message
news:...
> I've got a mate in the US who has ordered the caps I need from
> www.badcaps.com (who won't ship outside of the US/Canada) and he'll post
> them on to me. I'm a lucky guy.
>
> Heh! It has it's pros and cons. All-in-all I'm happy here. "Normal"
> countries scare me, they have terrorists bombing them for perceived
> wrongs. NZ never hurt anyone and the scenery and the weather are just
> fine. :-)

Except you're too far from the US Midwest to fix mine once you have yours
figured out. Still got my acid core solder and propane torch at the
ready. LOL

Mark
--
Four boxes protect our freedom: the soap box, the
ballot box, the jury box, and the cartridge box.

pheasant wrote:
> "~misfit~" <> wrote in message
> news:...
> > I've got a mate in the US who has ordered the caps I need from
> > www.badcaps.com (who won't ship outside of the US/Canada) and he'll
> > post them on to me. I'm a lucky guy.
> >
> > Heh! It has it's pros and cons. All-in-all I'm happy here. "Normal"
> > countries scare me, they have terrorists bombing them for perceived
> > wrongs. NZ never hurt anyone and the scenery and the weather are
> > just fine. :-)
>
> Except you're too far from the US Midwest to fix mine once you have
> yours figured out. Still got my acid core solder and propane
> torch at the ready. LOL

Hi Mark,

Yeah, sorry about that. ;-)

I've replaced the 3 x 1,000uF caps around the top corner by the RAM slots
that were bulging tonight. The original caps were 8mm diameter but even the
guy at Badcaps has trouble getting those. His site says that he has them in
stock again but check to see if you can use 10mm caps as the 8mm ones are
*very* expensive. Only one of the caps on the 75FRN2-L is situated between
two other caps so is a bit tight. I used some Sanyo caps that I took off a
nVidia FX5200 that prematurely failed as replacements. The video card was
only a year old, just out of warranty. The caps look perfect so I used those
on this first board.

Just waiting for the 3,300uF ones now. That'll take a while though.

Cheers,
--
Shaun.