XBOX One Power Brick Repair

[jayjr1105]

Pulled the old Samxon and Elite caps in favor of some Kemet Poly's. Used 2200uf in every spot even though the Elite cap was 1500uf. Working well so far.

[momaka]

Re: XBOX One Power Brick Repair

Interesting repair.
Were the Samxon and Elite caps failed? If yes, what capacitance and ESR did they show approximately. At least visually they don't appear to have any signs of failure.

Also, use of polymer capacitors in power supplies is generally a no-no. However, this is a flyback converter design (as evident by the lack of output toroid filter), which generally are less prone to oscillating / output filter ringing. So the chances of the recap not working from ultra-low ESR caps are much much smaller compared to a standard PSUs. Certainly avoid poly-modding regular ATX PSUs, though... unless the caps to be replaced were already rated for ultra low ESR variety.

Also, is that really all there is in an Xbox One power adapter? Looks nothing like the old Xbox 360 adapters, which were a lot more packed with components... not to mention they weren't rated for that much power either (~200 Watts.) My guess, however, is that the Xbox One doesn't use anywhere near as much power (at least average power - peak may be another discussion), so probably that's why the One adapter can get away with being smaller. It looks like a synchronous rectifier adapter too, so probably quite efficient too.

[jayjr1105]

Re: XBOX One Power Brick Repair

They were not failed (yet) but have seen a lot with failed Samxon caps so I did this a preventative. What makes a Poly a no no on the output side of a smps?

[stj]

Re: XBOX One Power Brick Repair

psu's can sometimes mis-interpret the lower esr as a short circuit on the output.
obviously not in this case

i wonder if they will work in the internal psu of the "s" version?

[Wester547]

Re: XBOX One Power Brick Repair

Also, use of polymer capacitors in power supplies is generally a no-no. However, this is a flyback converter design (as evident by the lack of output toroid filter), which generally are less prone to oscillating / output filter ringing. So the chances of the recap not working from ultra-low ESR caps are much much smaller compared to a standard PSUs. Certainly avoid poly-modding regular ATX PSUs, though... unless the caps to be replaced were already rated for ultra low ESR variety.

Do the Xbox 360 and Xbox One power bricks use flyback or push-pull topology on the secondary side? I ask because japlytic posted this Hipro Xbox 360 power brick almost nine years ago in the current power supply build quality thread and mentioned the primary side MOSFETs were configured in a “push-pull” manner - is this referencing the topology or just the MOSFETs?

Regarding the thread itself, as far as I know the highest failure rate cap wise in the Xbox One power bricks is LTEC (by far, from the launch day consoles, in the Chicony units). It was my impression the Elite and Samxon (and of course Chemi-con) held up better but it would be a good preemptive recap regarding lower tier brands.

[jayjr1105]

Re: XBOX One Power Brick Repair

Here is the XBOX 360 Chicony


And a second one I have with mostly Ltec caps

[momaka]

Re: XBOX One Power Brick Repair

Do the Xbox 360 and Xbox One power bricks use flyback or push-pull topology on the secondary side? I ask because japlytic posted this Hipro Xbox 360 power brick almost nine years ago in the current power supply build quality thread and mentioned the primary side MOSFETs were configured in a “push-pull” manner - is this referencing the topology or just the MOSFETs?

So, the topology (configuration of the switch devices on the primary) and power transfer design (continuous vs. discontinuous... a.k.a. flyback for the latter) are two different things.

The topology really refers to how the switch devices (MOSFETs, BJTs, IGBT(s)) and primary-side winding configuration on the transformer are set up. With forward configurations (single transistor, double forward, and active clamp), you typically have power transfer through the transformer only for max of 50% of the duty cycle and the pulse is only in one direction. The other 50% (or longer), the transformer is let into the reset state. On the other hand, push-pull, bridge (half and full), and LLC resonant, can push pulses through the transformer for nearly 100% of the duty cycle, with pulses of current going back and forth. Now the secondary side rectifier configuration is another whole discussion... but let's just say, depending on the topology of the primary side, this will also determine how the secondary side is set up. (For example, a center-tapped secondary is not practical for forward-configurations.)

The power transfer design, on the other hand, is how power is transferred through the "transformer". I put transformer in quotes, because it is such a device only for a continuous design. That is, in continuous mode, a pulse of current on the primary winding of the transformer directly results into a pulse of current(s) on the secondary winding(s), and there is (ideally) no power held in the transformer's core. With discontinuous (flyback) mode, that's not the case: a pulse of current on the primary winding stores energy into the core of the "transformer" (technically speaking, coupled inductor would be the more appropriate term here **). Once the pulse "passes" / ends, only then through the collapse of the magnetic field in the core, there is a pulse of current generated on the secondary winding(s). So in flyback / discontinuous mode, the transformer really behaves more like a coupled inductor (** but they are still called transformers due to providing primary-secondary isolation, in addition to some transformer action.) Moreover, with flyback / discontinuous mode, the primary-to-secondary windings ration isn't necessarily what determines the output voltages. This is why flyback designs are regularly used for "universal input" power adapters (even without APFC), because can work with a wide range of voltages on their primary side and still produce the desired voltages on the secondary. With continuous designs, this is much more difficult, because the output voltages are determined by the ratio of windings between primary and secondary, as well as primary-side input voltage.

So what does all of this have to do with the discussion about using polymer capacitors on the output? - A lot.

What makes a Poly a no no on the output side of a smps?

What stj mentioned, but also more than that.

Again, it depends if the power supply uses flyback design or normal transformer design.

With continuous mode, a square pulse of current on the primary of the transformer will also produce a square pulse of current on the secondary winding(s). If you put a rectifier and caps behind this, you will surely blow/damage/overheat your caps (from excessive ripple current) because of extremely high dI/dt from those square pulses. Therefore, continuous mode designs must always have an inductor between the rectifier(s) and the output filter caps. Since inductors resist sudden changes in current (and can also store energy in their cores in the form of a magnetic field), they essentially turn the square current pulses into triangular waves... which greatly reduces the dI/dt (and ripple current) that the output caps see. Also, the combination of inductor followed by capacitors forms a low-pass LC filter that essentially greatly filters out the switching frequency noise on the output of the power supply.

In contrast, with discontinuous / flyback design, the transformer acts more like the output inductor itself and the pulses on the primary aren't going to look too square. Therefore, as the magnetic field collapses and produces a pulse on the secondary winding(s), this pulse won't be very square either. For this reason, flyback design don't actually need an output inductor, so the output capacitors can be placed right after the output rectifier. (In fact, an output inductor after the flyback can mess with the operation, so flybacks just don't have output inductors.)

The presence or lack of an output inductor is what makes most of the difference as to whether ultra-low ESR capacitors (or polymers) may or may not be used. As I mentioned, in continuous mode, the output inductor and output capacitors form an LC low-pass filter. But like all LC circuits, this configuration will have its own resonance frequency. In practice, of course, such a circuit is actually an LCR circuit, with R being a resistance component from the inductor's windings, any trace resistances, and capacitors' ESR. The first two usually have somewhat negligible resistances, so most of the resistance really comes down from the ESR of the capacitor. In LCR-tuned circuits, the resistance tends to diminish self-resonant noise/spikes/waves. Therefore, capacitors with extremely low ESR may not provide enough resistance in such a circuit (this is called an "under-damped" LCR filter), and thus the power supply will end up producing "ringing" noise from its output filters. This is undesirable. On the other hand, if R is too high, then self-resonant noise from the LCR filter will be minimized (this represents an "over-damped" LCR circuit.) This is good... but then the problem becomes that the output capacitor(s) won't be able to filter as much of the noise and ripple coming out from the inductor, so this noise and ripple will end up on the output of the power supply. So in short, both ESR that is too high and too low may present a problem for continuous-mode designs. To further complicate issues, the feedback controller/IC also has its own filter when it "samples" the output voltages. This filter on the feedback controller is called compensation and can also be used to fine-tune feedback performance at various loads. The problem is, if the compensation circuit experiences excessive ripple/noise (either from output capacitors going bad with really high ESR or output capacitors used with too low of an ESR, producing tons of self-resonance noise), this might make the IC/feedback controller ramp up/down pulses erroneously and end up amplifying the noise even further in a positive feedback fashion... which can lead to the PSU to do unpredictable things on the output and even possible blow up a component.

So for this reason, large changes/deviations in the ESR on the output capacitors of continuous mode PSUs are discouraged. If you seen an output toroid inductor, the PSU is more than likely continuous design... and in such cases, it is best to stay close to the ESR of the original capacitors. (Note: "PI" coils don't really count as output inductors. Although these also form a LC low-pass filter when there are more output capacitors behind them, they have relatively low inductance, so the self-resonant frequency noise tends to be insignificant.)

That said, these Xbox power bricks / adapters look like flyback/discontinuous designs, as evident by the lack of a toroidal inductor on the secondary after the transformer. Discontinuous/flyback designs aren't nearly affected as much by the ESR of the output caps as continuous designs. In fact, although flyback designs won't produce sharp square current pulses from their transformers, the current pulses will still have higher dI/dt than that after the inductor of a continuous mode PSU. Therefore, capacitors with lower ESR and higher ripple current ratings are often beneficial at the output filter of a flyback design. This is the sole reason why ultra-low (and polymer) capacitors typically work fine on a flyback / discontinuous design.

Alright kids, PSU class dismissed. Time for a break (or so says my keyboard. )

[Wester547]

Re: XBOX One Power Brick Repair

Wow, thanks for the exceptionally detailed explanation!

I inquired as to whether the power transfer design was flyback or not (which it was, I should have observed, based on the lack of an output toroid on the secondary side) because discontinuous flyback designs puts great stress (in terms of heat output and ripple current) on the output caps, much more than continuous designs IIRC. The same goes for higher frequency VRM buck regulators on motherboards - the high ripple current, especially when it's close to the rated voltage of the cap, is certain to be a quick death for crap caps. I guess a wide range flyback design (or APFC with flyback) is ultimately simpler to implement for universal input adapters, especially given their more compact size, it means less components to use.

[WayOfTheRoad]

Re: XBOX One Power Brick Repair

Yeah, caps failing in those bricks is probably the most common failure of the OG/Phat models. So many came to me with the console being bad and it was just the power brick. Honestly, over half the time it was just the brick and almost every time it was the main or out caps.

Nice work and write up.

[sam_sam_sam]

Re: XBOX One Power Brick Repair

So, the topology (configuration of the switch devices on the primary) and power transfer design (continuous vs. discontinuous... a.k.a. flyback for the latter) **** and the hole post

I always read your post for explanation of how stuff works thanks for sharing this information

[Leeroy89]

Re: XBOX One Power Brick Repair

hey guys, yes i know old post, i have the xbox one power brick, replaced the ltec cap as was swelled, plugs in wouldnt go past the green xbox screen a few mins later didnt power upat all on a closer look one on the MOSFET looks like its burnt, i will post a pic, but i cannot see the spec on it, its Q205 and cannot find anythink online either, so i came to here
thank you in advance

[stj]

Re: XBOX One Power Brick Repair

i think it's threadlock,
you should have changed the other caps - specially the small ones

[Unspun01]

Re: XBOX One Power Brick Repair

@Leeroy89: I think those 2 mosfets are same, and operate in push-pull configuration. If you can read the one at left side, then I think you can find datasheet and replacement for the one on the right. In my power supply, I could not sourse exact replacement, so I replaced both as matched pair with new MOSFETS of equal or better ratings. If I recall, they are low voltage, high amperage MOSFETs and I may have ended up pulling some off an old P4 PC motherboard.

Usually power supply won't do anything if those transistors are shorted, so if your xbox turned on, then the transistors may not be the issue. Or if you are saying it turned on, and now it no longer turns on, then try my advice above to source replacements. Sorry I don't have one in front of me to read the markings on them