JetWay N2PAP Ultra recap

[trodas]

As seen in the bad-caps pic thread, these are the Vcore caps on N2PAP Ultra board from JetWay - Evercon caps, BTW:



The machine has problem starting up (sometimes more that six tries to sucesfull boot-up was need) and also the max. clock was about 2100Mhz for 1.650Vcore and about 2300Mhz semi-stable (but so was the board in general) for 1.575Vcore in Mobile AXP...

The board lack the additional 12V connector, tough my intention is to make it stable and fold for me on more that 2300Mhz. To do it, I intent to replace all the caps, and up the capacity of caps used for the Vcore recap + bridge them with 100nF ceramic caps from their bottoms to make things work better.

The five final Vcore caps are 1500uF 6.3V. The source-filter ones are 2200uF 16V ones. I think about 2200uF and 4700uF ones, but I think the 16V is bogus there and example of bad design, as there is not 12V power connector and therefore the Vcore get powered from 5V (or 3.3V) rail and therefore 6.3V caps are more that enought, while 16V is example of BAD DESIGN, IMHO.

I want to change that and see what the results will be

So, what do you guys think - should I try?

[yanz]

Re: JetWay N2PAP Ultra recap

your site must be exceed the bandwidth limit, i can't see any picture.

trodas, on atx connector there are a (or some) 12v line (yelow) iirc.. are you sure the vrm doesn get the input from 12v? many p4 motherboards didn't have 12v p4 connector but relay on those 12v on atx connector. cmiiw.

[MD Willington]

Re: JetWay N2PAP Ultra recap

ooh sexy...

...not...

Well if you get it fixed those boards can be screamers...

[trodas]

Re: JetWay N2PAP Ultra recap

yanz - slibe.com is slow, just wait, works for me, no bandwithc limit And yes, you are right, the main 20min ATX power connector feature 12V too. But I will of course measure first, what voltages are on the caps... But mostly these are powered from 5V rail, so I bet I can use 6.3V caps there for great benefit - 16V ones are also damn big...!


MD Willington - not sexy for you? Damn, you are picky! And true enought - there is some pitching noise - when the board is not loaded. 100% load by folding = 100% quiet
So yea, it CAN fold. It can't stop, tough

We see if I could manage more that 2300Mhz semi-stable out of this thing

[MD Willington]

Re: JetWay N2PAP Ultra recap

No.. I meant the bloated caps were not sexy... LOL

Now curves on other things... thats a different story..

(I live in a University town and the sun is out if you catch my meaning)

[trodas]

Re: JetWay N2PAP Ultra recap

Time for a little update :
First I measured the Vcore with oscilloscope and the results are pretty ughly:





As I suspected, the big (huge!) 2200uF 16V caps are sitting on 5V rail, and therefore I replaced them with 4700uF 6.3V Pannasonic ones. I recapped the rest of the mainboard as well, except the five last 1500uF 6.3V caps, witch I want to replace with 2200uF ones to get the chance for higher O/C.
I also bridged the 4700uF caps with 100nF ceramics SMD caps for better filtering - the current Vcore fluctuation was terrible, eeeeh...

And results ATM? Board is stable - hoooray. I did not tried higher OC, as the last Vcore caps are still not changed (I did not have such caps, so I have to get another batch of them...), but I think I made good progress :


MD Willington - I think you mean curwes on womans, did not you?

[willawake]

Re: JetWay N2PAP Ultra recap

what is the model of that scope trodas? looks nice

[trodas]

Re: JetWay N2PAP Ultra recap

Well, ATM I don't know how to play much with it, and I starting missing the compensation thing - eg. to be able enlarge the spikes on the display... But hey, it is this suxxka:
http://www.professionalequipment.com...qx/default.htm

[PeteS in CA]

Re: JetWay N2PAP Ultra recap

Spec Summary:
Oscilloscope
Bandwidth: 1 MHz
Sample Rate: 20 Ms/s
Channels: Single

With only 1MHz BW, you aren't going to see the full amplitude of spikes, and if the switch frequency is much above 200KHz, even the switch frequency component of the ripple will be attenuated. P/S PARD measurement is usually specified as with a 20MHz BW. In handhelds: the Fluke 123 is 20MHz, ~$1200; the Fluke 124 is 40MHz, ~$1500; the Fluke 192B is 60MHz, ~$2000. All three are 2-channel; the Fluke 192B (and the higher BW//$$ 196B, 199B, and 196C) channels are isolated from each other.

[willawake]

Re: JetWay N2PAP Ultra recap

i saw that as well, also the max capacitance that can be measured is very low. but the units you quoted are just too expensive. having even just the the max/min values is quite useful for trodas's tweaks.

i think the computer based oscilloscopes are probably best cost/performance ratio. also good if you want to publish the results. what do you think?

[Per Hansson]

Re: JetWay N2PAP Ultra recap

Yea, I was looking at the Velleman PCS100A myself, 12Mhz BW, but it's only 150$...

[Spacedye69]

Re: JetWay N2PAP Ultra recap

Any input on what to look for in buying a scope, especially under $500, would be awsome.

[willawake]

Re: JetWay N2PAP Ultra recap

i bet kc8 can find you one cheap

[arneson]

Re: JetWay N2PAP Ultra recap

I have a very small ProbeScope, connects to serial port for display on screen.
Very slow, very cheap, but only the size of a probe, matching vrm is also small.
I work in very confined spaces

[willawake]

Re: JetWay N2PAP Ultra recap

occasionally i get called to ships to go into confined spaces because some of our guys are too fat.