Original Xbox Replacement Caps

[stj]

no, the total esr would be too high

[jayjr1105]

Originally Posted by stj

no, the total esr would be too high

The ESR goes down as capacitance goes up though. So three 3300uf caps might have an ESR of .018 each and two 4700uf caps would be around .015 ESR each. Does that make sense? Am I misunderstanding ESR across multiple capacitors?

[momaka]

Originally Posted by jayjr1105

Any harm in using fewer caps around the CPU? The XBOX either has five 1500uf caps or three 3300uf caps. Could you simply populate two 4700uf caps on the left and right most sides?

It depends.
If the caps are in parallel, then... MAYBE

I say "maybe" because of ESR.

For a number of _identical_ caps in parallel, the total equivalent ESR/impedance in the circuit, Z, will be:

Z = (ESR of 1 capacitor) / number of identical capacitors in parallel

So with the example of the three 3300 uF caps with ESR of 0.018 and two 4700 uF caps with ESR of 0.015

Z (of 3x3300 uF caps) = 0.018 / 3 = 0.006 mOhms
Z (of 2x4700 uF caps) = 0.015 / 2 = 0.0075 mOhms

So as you can see, the ESR of the two 4700 uF caps is higher than the three 3300 uF caps.

Now, would 0.0015 mOhms of difference in ESR matter in this circuit?
- Probably not. The two ESR values above are still within 25% of each other, so that's not too radical of a change. Most circuits can work OK with that difference, if designed properly. Usually, it takes a lot more deviation from the standard values to break the circuit function (10-fold is typical, hence why some devices can still work, even when their caps have already started failing.)

But generally, you want to match ESR to be the same or better, and then if possible match the capacitance too. If you had to sacrifice a bit of capacitance to get the right ESR, that's OK. But of course, don't go too far away from the original capacitance too. Again, I'd say try to stay within 25% of original.

Originally Posted by Per Hansson

The direct links to imgur are still there for the inline pictures.

Ah.
Maybe because I haven't cleaned my history/cache in a while, so they probably finally all pre-loaded.

[jayjr1105]

Originally Posted by momaka

It depends.
If the caps are in parallel, then... MAYBE

I say "maybe" because of ESR.

For a number of _identical_ caps in parallel, the total equivalent ESR/impedance in the circuit, Z, will be:

Z = (ESR of 1 capacitor) / number of identical capacitors in parallel

So with the example of the three 3300 uF caps with ESR of 0.018 and two 4700 uF caps with ESR of 0.015

Z (of 3x3300 uF caps) = 0.018 / 3 = 0.006 mOhms
Z (of 2x4700 uF caps) = 0.015 / 3 = 0.0075 mOhms

So as you can see, the ESR of the two 4700 uF caps is higher than the three 3300 uF caps.

Now, would 0.0015 mOhms of difference in ESR matter in this circuit?
- Probably not. The two ESR values above are still within 25% of each other, so that's not too radical of a change. Most circuits can work OK with that difference, if designed properly. Usually, it takes a lot more deviation from the standard values to break the circuit function (10-fold is typical, hence why some devices can still work, even when their caps have already started failing.)

But generally, you want to match ESR to be the same or better, and then if possible match the capacitance too. If you had to sacrifice a bit of capacitance to get the right ESR, that's OK. But of course, don't go too far away from the original capacitance too. Again, I'd say try to stay within 25% of original.


Ah.
Maybe because I haven't cleaned my history/cache in a while, so they probably finally all pre-loaded.

Thanks for that explanation. I'll keep that math in mind. BTW I think you meant 0.015 / 2 = 0.0075 but your math was still correct and valid. But yes, these would be a cluster of caps in parallel.

[momaka]

Originally Posted by jayjr1105

Thanks for that explanation. I'll keep that math in mind. BTW I think you meant 0.015 / 2 = 0.0075 but your math was still correct and valid.

Ah... yeah.... that would be my copypasta.
Thanks for catching that! Fixed it now.

Originally Posted by jayjr1105

But yes, these would be a cluster of caps in parallel.

In that case, likely OK to use them.

[jayjr1105]

I assume the original Rubycon MBZ's are also good to leave alone if not bulging?

[stj]

i have seen in a lot of xboxes and never seen a rubycon MBZ in one - maybe they are replacements.
if MBZ's fail you will usually see it - so probably o.k.

[jayjr1105]

Originally Posted by stj

i have seen in a lot of xboxes and never seen a rubycon MBZ in one - maybe they are replacements.
if MBZ's fail you will usually see it - so probably o.k.

99% sure they are original. Solder joints on the bottom are too good to be done by hand. This is a 1.6 board btw.

[stj]

could be original then - 1.6 had a number of updates relating to the caps.

[momaka]

Yeah, MBZ is pretty stable.

MCZ and MFZ, on the other hand, can be a bit more unstable - usually when placed in hot spots. But those two are newer series, so you won't see them on the original Xbox. The Xbox 360 used MFZ and MCZ quite often, though. I've seen a few failures (again, mostly related to being overheated), but they are very reliable otherwise.

[jayjr1105]

Kudo's to Microsoft overall. Really the XBOX had one consistently bad capacitor in the Nichicon HM. In all the XBOX's I've refurbed, I've never seen anything else really bad other than those Nichicon's. Even the power supply's with Teapo's, Ltec, Samxon, and Taicon. I've recapped a lot of power supply's just because but none of those 2nd tier caps were ever out of spec. This all aside from the supercap obviously.

[momaka]

Nichicon had a manufacturing flaw/problem between 2001 and 2004 with the HM, HN, and HZ series. In 2005, they got most of that squared away - at least with the 16V and higher voltage -rated caps. Really 2006 (caps with date codes H06xx) and forward is when they got things under control with those series.

As far as the 2nd tier caps... I don't know, I still feel uneasy about leaving them in a device that I don't want to come back for repair again. Teapo and Ltec, in particular, seem to do poorly over time. I think it's the fact that the original Xbox PSU runs rather cool and doesn't stress the caps a whole lot is why we don't see those caps fail there as much. Once they get stressed a little more and/or are placed in a hotter environment, then they start to get a lot more problematic. Case in point with many of the post-2005 OEM ATX PSUs, when PC power consumption on the CPU alone jumped from a mere 30-50 Watts from the Pentium 3 / Athlon era to 60-100+ Watts in the Pentium 4 era. PSU design also matters, of course. But between two identical PSUs - one with Japanese caps and another with 2nd tier caps - the PSU with the Japanese caps would not have problems with cap failures. So it's clear that 2nd tier caps aren't quite on par with the good Japanese brands, even when datasheets suggest nearly identical specs and performance.

[jayjr1105]

Pulled out three more Nichicon HM's on someone else's machine. All bloated and reading really high like 9700uf again (each). ESR is still 0.01 which is crazy since these were bulging.

I measured the voltage around the CPU in an empty capacitor spot and it's 1.7v. Why did they spec 6.3v caps when 4 or 2.5 would have sufficed?

[stj]

because 6.3v was as low as you can go on large electrolytics, only polymers and small smd fo down to 4v or 2.5v
and the ripple handling will drop like a stone on smd electrolytics.
.

[jayjr1105]

Originally Posted by stj

because 6.3v was as low as you can go on large electrolytics, only polymers and small smd fo down to 4v or 2.5v
and the ripple handling will drop like a stone on smd electrolytics.
.

There are 4v 2700uf polys with excellent ripple handling. Three of those and you'd be golden

[stj]

yes, but we arent talking about poly's now - we are talking about std caps

[jayjr1105]

Originally Posted by stj

yes, but we arent talking about poly's now - we are talking about std caps

2700uf 4v poly's are almost $2 a piece. Pass. I'll stick with my Rubycon ZLQ/ZLH and Nichicon UHV for the CPU replacement caps.

[momaka]

Originally Posted by jayjr1105

2700uf 4v poly's are almost $2 a piece. Pass. I'll stick with my Rubycon ZLQ/ZLH and Nichicon UHV for the CPU replacement caps.

Yup, that's why MS didn't fit this board with polymers. They're just more expensive - especially even more so back in the day. Not only that, but the circuit doesn't really need capacitors with that kind of spec. After all, the original Xbox is nothing more than a Pentium 3 CPU with a GeForce 3 GPU. Neither of those can be considered as "power hungry", and hence they don't really need a beefy VRM or beefy caps to filter power. So that's why on a lot of older stuff, you can get away with regular low-ESR caps.

[stj]

Originally Posted by momaka

Yup, that's why MS didn't fit this board with polymers.

your forgetting the age of the xbox.
they couldnt fit polymers even if they wanted to.

[jayjr1105]

Saved another 1.6 from pissing Nichicons