Shuttle K45 recap

[Half-Saint]

Hi guys,

I'd like to recap the motherboard as there are 8 bulging 1800uF 6.3V OST RLZ caps. The RLZ series is not listed on the OST website: http://www.ost.com.tw/products_ec_list.asp so I have no clue what the specs are. I imagine they must be low ESR. There are also four 1500uF 10V RLX caps which are supposedly Ultra Low ESR but they look fine.

The dimensions of the 1800uF caps are 8 x 20 mm with 3.5mm lead spacing.

The only suitable replacements that actually fit the dimensions are Nichion HN and HM series, HM being about 50% cheaper. The question is, would it be safe to use HM caps?

I'm also considering doing a poly mod but can't figure out which caps would be alright to use.

Cheers,
SainT

[Half-Saint]

Re: Shuttle K45 recap

Well I went ahead and bought ten Nichion HM caps and I'll do the recap when I get them.

[c_hegge]

Re: Shuttle K45 recap

Yup. Nichicon HM will work fine there.

[momaka]

Re: Shuttle K45 recap

I would suggest replacing most of the OST caps. Those Shuttle cases run rather hot, and OST does not like the heat very well. At a minimum, also do the Northbridge and RAM power rails.

I have an older Shuttle PC with a dead Northbridge thanks to bad OST caps - NB VRM commited suicide when the caps went bad. Took out the upper MOSFET and the PWM controller.

[Half-Saint]

Re: Shuttle K45 recap

I replaced all 1800uF 6.3V caps but it still won't boot. There's an SMD LED onboard that blinks once when you flip the PSU power switch. The motherboard comes without a speaker and there's no way of connecting one. When I power it up, fans spin and that's it. Should I also replace 1500uF 10V caps? The other large caps are Sanyo and should be okay

I tried three diff. processors, three diff. sticks of RAM and two known power supplies so I'm pretty sure it's the motherboard.

EDIT: I replaced all four 1500uF 10V caps with 1500uF 16V Rubycon MBZ caps and it hasn't fixed it. There are two or three small OST caps left on board but I seriously doubt replacing them would solve the problem.

EDIT #2: I measured voltages over the caps and they are as follows:

- the group of six 1800uF 6.3V - 1.28V each
- 1500uF 16V - 1.5V
- 1800uF 6.3V between ATX connecto and NB - 5.11V
- two 1500uF 16V next to the RAM slot - around 30mV (shouldn't that be 1.8V?)
- 1800uF 6.3V closest to the NB - 1.5V
- both Sanyo 1800uF 16V - 11.92V each
- 1500uF 16V next to the PCI slot - 1.2V

I borrowed the photo for illustration purposes - my board has the exact same layout of caps.

[Half-Saint]

Re: Shuttle K45 recap

Anyone have any idea why there's no voltage on the caps next to the RAM slots?

[momaka]

Re: Shuttle K45 recap

EDIT #2: I measured voltages over the caps and they are as follows:

- the group of six 1800uF 6.3V - 1.28V each

That seems okay. What CPU did you have in the board when you tested that voltage?

- 1500uF 16V - 1.5V

and

- 1800uF 6.3V closest to the NB - 1.5V

I am guessing that 1500uF 16V cap is next to the group of 1800uF 6.3V caps above the CPU?
If so, then that cap may be connected together with the 1800uF cap closest to the NB. Check with a multimeter, though (i.e. see if their positive leads show <1 Ohm resistance). You have a i945 chipset, so this is probably the Northbridge voltage. Looks right. I'm guessing you replaced the two caps above? Or at least one of them? (just to rule out that ripple isn't getting to the chipset)

- 1800uF 6.3V between ATX connecto and NB - 5.11V

Interesting...
Check whether that cap is connected to 5VSB or 5V rails. I'm curious why it has that voltage. Doesn't seem like there are any things nearby that need 5V. Perhaps Q34 or Q38 - the two D^2PAK / TO-252 transistors under the CPU socket... Actually, speaking of those, does one of them have 3.3V on any of its pins, even when the PC is turned OFF but plugged in? I think one of them might be a 3.3V rail for the Southbridge. At least that's what I remember it was for my Shuttle SB83G5A (FB83 motherboard). Let me check...

- two 1500uF 16V next to the RAM slot - around 30mV (shouldn't that be 1.8V?)

Yes, 1.8V sounds more like it.
On my Shuttle mentioned above, there are 3 caps like that close to the DDR slots. They are connected to the RAM Vdd rail, and I have 2.5V on them (DDR memory uses 2.5V).
By the way, my thread can be found on the link below. But I don't know if it will be of any useful info, so save yourself the long read if you want
https://www.badcaps.net/forum/showthread.php?t=17062

Anyways, there are two TO-252 MOSFETs under the ATX connector and two of the same close to the 1500uF 16V cap that has 1.5V. Each set of two likely forms a buck converter rail (making that two converter rails total, as suggested by the "1R2" square inductors). That means one of these is for the 1.5V rail for the NB and the other should be for the RAM Vdd. See which MOSFET's Drain connects to the caps next to the RAM slots (if any). Let me know what you find here.

- both Sanyo 1800uF 16V - 11.92V each

Those are the CPU VRM high side caps. Voltage looks normal. They should be connected to the 4-pin 12V PSU connector.

- 1500uF 16V next to the PCI slot - 1.2V

That looks like CPU secondary / CPU Vtt rail. I get 1.2V on mine as well, but only when I turn ON the computer.

There's an SMD LED onboard that blinks once when you flip the PSU power switch.

Sounds like something is not right there. I think the LED should stay ON - at least for the first time you flip the PSU switch. Again, referencing info from my Shuttle PC, my LED stays ON when I flip the PSU switch (but keep the motherboard OFF). When I turn ON the motherboard, the LED stays lit. Only after I turn the motherboard OFF, then the LED turns OFF.
So basically, the LED ON/OFF state seems to coincide with the RAM Vdd (2.5V) rail. When I flip the PSU switch, my RAM Vdd comes up immediately. Not all motherboards are like this, but a few are. After I cycle the motherboard ON and OFF, the RAM Vdd no longer stays ON all of the time. I think this is normal, because I have seen it on other non-Shuttle motherboards as well.

Also, if you look all the way on the right of your picture, you will probably see some small caps and resistors next to the rightmost DDR2 memory slot - these should be connected to RAM Vtt. RAM Vtt is usually about 1/2 the voltage of the RAM Vdd... so about 0.9V to 1V for DDR2 if I remember correctly. Most likely, that 8-pin MSOP chip at the top-right of the board is responsible for that. When you turn ON the motherboard, both RAM Vdd and Vtt must turn ON. There is also a high chance that 8-pin MSOP chip has at least one of its pins connected to the PSU 5VSB rail - if it does, then that chip is very likely the RAM Vtt generator.

There are also some other regulators/MOSFETs on the left side of the board, like Q14, Q19, Q26 (all below the CPU socket), and U22. These probably generate 3.3V, 2.5V, or 1.8V for the smaller IO chips on the board. Make sure none of them are shorted input-to-output or Drain-to-Source. Some of these may be always ON as long as the PSU is plugged in and its switch flipped.

[mike-s]

Re: Shuttle K45 recap

I'll be honest, I have replaced all the elecrolytics on my k45 (using low esr caps of the same value), and never had this level of problem with my board.

Then again this is a slight threadsurrection.