Guide Wii Undervolting "guide" and stats. (incomplete)

[Wesk]

Here is the first dirty release of the Wii undervolting "guide".

This isn't an official release yet because:

1. I'm not really sure how to structure this information in a forum post yet.​
2. I need a larger sample size of systems and tests to really dial everything in.​

Brief synopsis/overview:

The Wii (4 layer) can be undervolted by a significant margin, on average reducing power draw by 30% and reducing heat generated by 15%.


What we still need:

We need more samples of successful undervolts from an array of boards. Any additional information you can provide will be helpful in building a definitive undervolting guide for the Wii.

Here is the current compiled information spreadsheet that you can use as an outline for performing your own undervolts:

Wii Undervolting "Guide"




Please post your undervolt successes/failures so that it can be added to the sheet.

​

Thanks to @thedrew and @YveltalGriffin for initial testing and data compilation assistance

 

[YveltalGriffin]

I designed an adjustable Wii undervolting board with current sense resistors. The reg board isn't really for portables, but instead to help characterize mobos' undervolting potential and take accurate current measurements. It's open source and available on my GitHub: LINK

• 3.5 - 5.5V input
• x4 MUN3CAD03-SE 0.6-3.3V 3A buck regulators
• Trimpots for output voltage adjustment (feedback resistors pre-tuned for Wii undervolting)
• 4-terminal current sense resistors for input and outputs
• Mouser project with BOM

Vin and each regulator output have dedicated 100-milliohm 4-terminal sense resistors. When powering a load from the reg board, you can measure the voltage across the resistors by poking your multimeter probes into the plated test pads. Multiply the voltage by 10 to get the load current in amps.

Each regulator has a trimpot for adjusting the voltage. Turn it clockwise to increase Vout, and counterclockwise to decrease Vout. You can turn the pots with tweezers or a tiny, blunt Phillips screwdriver. The Vout ranges are bounded by resistors so that the entire pot range is usable (and to avoid pumping 5v into any of the rails.)

Note: the trimpots are fragile mechanical devices. If you turn them too far in one direction, they'll 'wrap around' to the other side. When overturned like this, there's a small region that won't produce a reasonable resistance, causing Vout to collapse. Just keep turning and Vout will jump back to the intended range.

Schematic
View attachment undervolt.png

Happy undervolting!

 

[loopj]

LET THE UNDERVOLTING COMMENCE

 

[thedrew]

After many months of testing on 2 portables that I undervolted, I have found the 1V line doesn't like to be undervolted. My advice is to leave the 1V line as is and it's only a 0.3W saving anyway and doesn't seem to run any coller undervolted anyway, so it's not worth risking stability. These are for Hollywood-1 boards, but I suggest the same for Hollywood-2. It worked fine at first but a couple months later, couldn't go longer than 10 minutes without freezing and crashing on both portables. I had them set to 0.92V under load, but raising both back to stock 1V fixed the freezing. Haven't had issues undervolting the 1.15V line to 1.05V under load and the 1.8V line to 1.45V under load yet. I specify under load as voltage levels drop under load, so I have the voltages set a little high to compensate for the under load voltage drop.

 

[Stitches]

Long term usage data hype!

 

[Shank]

Haven't had issues undervolting the 1.15V line to 1.05V under load and the 1.8V line to 1.45V under load yet.

Makes me wonder if running both the 1v and 1.15v off a shared, single 1-1.05V rail would have any benefits? Would love for someone smarter to chine in on this idea.

 

[GingerOfOz]

Makes me wonder if running both the 1v and 1.15v off a shared, single 1-1.05V rail would have any benefits? Would love for someone smarter to chine in on this idea.

I've got a Wii board that runs very stably (so far) with the 1.15v line pulled down to 1.02v. I've undervolted a PMS-2, connected the 1v and 1.15v lines to a single 1.02v reg, and put it inside of the WiiX, so I'll report on how it does

 

[thedrew]

Makes me wonder if running both the 1v and 1.15v off a shared, single 1-1.05V rail would have any benefits? Would love for someone smarter to chine in on this idea.

I have thought about this too, but getting too close to 1V has caused instability from my testing on the CPU and if we tie both lines together, we would be slightly overvolting the GPU as the most stable undervolt voltage for the CPU has been 1.04V.I'm sure the GPU can handle the slight overvolt, but would probably end up drawing a little more power. I'll have to test that.

I would think there is no benefit connecting both lines together other than a smaller overall footprint of a custom reg board. Splitting the load in two different regs would lighten the workload of the regs and run cooler than 1 reg doing double the work than the usually setup, however, there are some pretty efficient regs that can easily handle 3W workload of the two combined lines though... so might not make much of a difference. Time to get to work with this!

 

[thedrew]

It has been a month since the last post, so quick little update. I have played through Metroid Prime and other light games here and there with the 1V and 1.15V line tied together running at 1.04v under load without any issues. All other voltages are stock, only the 1v and 1.15v were undervolted. Has not froze once. I thought we would be overvolting the 1V line a little by bringing it up to 1.04v (which is the stable undervolt voltage for the 1.15v line on most 4 layer motherboards it seems) but turns out we have been undervolting the 1V line this whole time lol. Stock 1v line is actually 1.06V-1.07v from my testing between 3 different boards! All these years we have been running it as close to 1v as possible, or at least I have.

 

[loopj]

This is a really cool finding.

It still sounds we'll save some power consumption by undervolting both 1V and 1V15 seperately BUT it does mean we have the option of simplifying builds. On non-portable builds, we can drop a regulator completely!

 

[Shank]

I wonder if I could squeeze out a few extra seconds of battery life on the Kill Mii

 

[Wesk]

I wonder if I could squeeze out a few extra seconds of battery life on the Kill Mii

Do it

 

[VoxelTek]

I've got a Wii board that runs very stably (so far) with the 1.15v line pulled down to 1.02v. I've undervolted a PMS-2, connected the 1v and 1.15v lines to a single 1.02v reg, and put it inside of the WiiX, so I'll report on how it does

@GingerOfOz I saw you undervolted the WiiX in your video, and that you mentioned the stable point of that Wii "changed over time". Is this what you had before the Wii began disliking it, or is this what you switched to afterwards? And if the former, what did you then change your values to?

 

[m0thm4n]

I have a few motherboard revision trims I'm interested in experimenting with undervolting on. I just have a quick question. So after finding stable voltages, I could just add in some fixed voltage dividers (with resistors) on my PMS power lines in the context of an actual build, right? Just want to make sure I'm not missing anything important. Thank you, and looking forward to adding to the data pool in the future!

 

[Shank]

I have a few motherboard revision trims I'm interested in experimenting with undervolting on. I just have a quick question. So after finding stable voltages, I could just add in some fixed voltage dividers (with resistors) on my PMS power lines in the context of an actual build, right? Just want to make sure I'm not missing anything important. Thank you, and looking forward to adding to the data pool in the future!

Yes, you should be able to just replace the feedback resistors on the PMS boards to change the output voltage.

 

[VoxelTek]

I have a few motherboard revision trims I'm interested in experimenting with undervolting on. I just have a quick question. So after finding stable voltages, I could just add in some fixed voltage dividers (with resistors) on my PMS power lines in the context of an actual build, right? Just want to make sure I'm not missing anything important. Thank you, and looking forward to adding to the data pool in the future!

@Shank I don't think they were talking about feedback resistors.


@m0thm4n Do NOT do what you're suggesting. Resistor dividers are absolutely not a good way to reduce an output voltage.

The RVL-PMS line of products all use resistors to configure the output voltage, HOWEVER, the actual voltage outputs are NOT channeled through resistors. Instead, a resistor divider is used to provide a feedback voltage line to the regulator. It's important to note that NO OUTPUT POWER is actually going THROUGH these resistors. They simply divide the voltage down, and provide it to the regulator, providing a closed-loop feedback system.

I'd definitely suggest researching into buck DC regulators, before attempting any of this.


Adjusting the output voltage on any of the RVL-PMS products would require replacing the tiny SMD capacitors on the board with your own. You likely wouldn't be able to use standard through-hole resistors, their large size would make it practically impossible. If you're not familiar with working with SMD components, this is definitely not where you should start off.

 

[Shank]

Oh jeez yeah I misunderstood. Do not attempt to reduce any power supply with a voltage divider. That is a very fast way to burn things up.

 

[Stitches]

Pls don't Mumble your PMS, it's still too soon

 

[m0thm4n]

@m0thm4n Do NOT do what you're suggesting. Resistor dividers are absolutely not a good way to reduce an output voltage.

The RVL-PMS line of products all use resistors to configure the output voltage, HOWEVER, the actual voltage outputs are NOT channeled through resistors. Instead, a resistor divider is used to provide a feedback voltage line to the regulator.

This makes a lot more sense My background is mostly in audio stuff where the the signal current I’m working with is waaaayyy lower, but yeah I definitely get why that’d burn up in a power situation like this. I should’ve engaged my critical thinking a bit more before asking that question hahah, my bad. Thanks for the info/tips on the correct way to go about it lol