It's So Bad! Modding the Nintendo Power Glove (picture intensive)


You probably already know what the Nintendo power glove is. It was a critical commercial failure that died with the 80’s. Even with its dubious reputation, I have often wondered if the power glove was really as bad as people made it out to be. Was it the glove’s fault or just how it was implemented in gaming? Could the power glove’s reputation been saved with better tech? Is it possible to science the crapola out of it and bring an idea that was seemingly ahead of it’s time to true glory?

I intend to find out.

But It just arrived, so today I’m just going to share my first impressions with you.

I aquired my power glove on Ebay for under 40$. The going rate is usually closer to 65$, even more if it’s still in the box or at least includes the sensor bar. However, the family dog looks like he decided to nibble a bit on the cable.

Bad day for Fido, but it works out for me. I don’t have a nintendo, and now I don’t feel so bad about hacking into it.

Other than a few scuff marks the power glove is in near mint condition. I originally wanted a power glove for the flex sensors in the fingers, so I’m happy that they’re probably not damaged.

Here you can see two microphones located on the back of the hand. The sensor bar would have emitted an ultrasonic pulse that these microphones could pick up on. The powerglove could determine where it was in 3D space by the time it took for each pulse to reach the microphones. But I have other ideas that involve an IMU.

Moving onto the ergonomics, the glove is held on by a Velcro strap around the arm and a tabat the base of the palm.

Holes in the tips of the finger and the stretchy outer fabric makes the glove conform to nearly any hand size. While the black outer fabric is breathable, the fabric that rests on top of the hand is backed by the power glove’s plastic shell. If you wear it for even a modest amount of time your hand will start sweating.

The plastic cover doesn’t just make the glove hot, but it also sacrifices some of your flexibility. When I make a fist like in the picture above my knuckles actually “bottom out” on the plastic cover, preventing me from making a tight fist. It is still possible to flex my fingers individually and grab objects but it does make more demanding activities like typing or using an arcade stick harder.

But I can forgive all of that because of how awesome it makes all my hand gesture’s look.

Seriously, I love the way this thing looks.

Power glove can be metal too!

I eventually got tired of posing with the glove and decided to take a peek inside it. Modding it was my original plan after all. There are a suprising number of screws holding this thing together. Here I just took off the eight screws holding the covers on, but it gets so much more intricate.

Here’s a look inside the hand cover. The mass of wires to the lower left are all for the flex sensors. I’ll be testing those later to confirm all my flex sensors are working. The rest of the wiring is totaly alien to me, including that metal disk on the cover. I have to assume it’s for the microphones. I also know there’s a rotation sensor on here somewhere, but I won’t be using it. From what I read it was pretty bad…

The underside. Now you can see the wires going into the hand so that they can read the flex sensors.

This is the button cluster on the arm. Look at the size of that processor! It’s as thick as two quarters. I’m not suprised though. There won’t be another controller with this many buttons, let alone VR capable, for a very long time.

The reverse side that actually faces the button cluster. not unlike modern controllers, only there are waaaay more of them.

And lastly, the back side of the buttons. Nothing special to note.

That’s it for now.Since the shipping time on a lot of this stuff is really slow, I probably won’t mod it too soon. However, I’ll be testing the flex sensors while I’m waiting for everything else to get here.


Tact switch mod that shit. Put a new RGB LED on it instead of the red one, and put a mini Kaimana in it for cool RGB effects. You could probably put some RGB LEDs behind the microphones so that when a certain gesture is done, they light up, like you’re firing lasers or something. Maybe even some fiber optic cable next to the flex sensor wires with LEDs at the base to shine light through them. If you poke them out of the finger tips slightly, you could glow color through them.

Man there is a lot of cool stuff you could do. Depends on what you want to do though. You could probably get a Wii Motion+ board and get that in there for motion control. Or, you could probably hack in a PS Move, and use that LED controller board for the LEDs behind the microphones and play PS Move games or something similar.


That metal disk on the hand cover is actually a speaker.
The original power glove can be programmed with various settings, hence all the extra buttons.
And when you properly entered in a program setting the glove would beep.


I believe these are the same flex sensors:


Yea, I’ve seen a lot of people do some really cool stuff with the power glove. Lots of indestructibles and blogs on the process.

Primary goal is to wire the flex sensors to an arduino and have all the buttons perform certain keyboard functions and adjust flex sensor and IMU sensitivity on the fly. I don’t really want to have a bunch of wires sticking out of it, but I do like the idea of using fiber optic wires and RGB LEDs to enhance the appearance. I’ll see how much room I have once I get the flex sensors and buttons working. The IMU is fairly small, and I plan on running the wires for the flex sensors into the wrist. So there might be room for a light show in there.

Makes sense. I was aware of the programmable buttons, but didn’t realize there was a speaker in there. I should probably look at an owners manual some time. Maybe I’ll have to take advantage of that speaker if it’s capable of more than beeping.

you are correct, 4.5 inch flex sensors. They’re about 13$ everywhere, That’s what made this power glove such a good deal!


I wonder if the impedance curve changed over 20 years?


Since it’s going to be some time before my “stuff” arrives, I decided that testing the speaker took priority over the flex sensors. It came out pretty easy, too easy actually. My soldering iron didn’t even reach full heat before one side came out, and the ground side just pulled out of the PCB without heat.

The speaker works though! I hooked it up to my arduino uno to test it out. I ran a built in chip tune and the speaker played it perfectly. While i’d like to test actual sound files, I think the hardware might be too bulky for the power glove. I did come up with the idea of playing something when you power it on, like the Metroid “item get” tune. I’d like to show you that today, but the sample program can only handle 8 notes before it starts to glitch. I’ll probably work on that over the week.

Electronics certainly can degrade with time, especially with repeated use. But the fact that the power glove was such a catastrophic failure, it’s in near mint condition, and the fact that Fido effectively prevented his previous owner from further using it makes me think I don’t have much to worry about.

The biggest boon for me in that regard is that I’m using my own PCB with my own code. If there is a change in the flex sensors’ resistance, I can adjust for it. It’s only a problem if while flexing I get an erratic increase/decrease in resistance. I only see that happening if one of the sensors are broken.


The original settings for the Power Glove did not have that much precision, it only measured if the fingers are bent or not (a change of resistance) and not the actual resistance from the sensors.
It is also why the original power Glove has a reset, to re-calibrate for all the flex sensors.


My MUX board arrived, so I decided to start installing it today.

Yea, the work bench needs a little cleaning, but I really felt like making some progress on this before I did so. There is actually nothing on the PCB that I can really use, so I decided to entirely remove it instead of cutting it up. So I started to remove the solder from all the tiny wires and the PCB. Once I did this on both sides, I was left with…

…This wiring harness. To preserve the appearance of the power glove, I will be reusing this. I might also remove the microphones from the palm PCB later, but I’ll have a better idea what I’m going to do with that when my IMU and micro switches arrive.

That red PCB is the MUX. In layman’s terms it gives me more “pins” for my micro controller. I currently have all 4 of the flex sensors wired to the MUX . The red wires are ground, so they will go into a large resister in the wrist PCB once I get everything in the palm wired up. I think I have a week or two to wait before that shows up in the mail. It also looks like there will be plenty of room in the palm and on the MUX for some bonus features, but I’ll have to wait and see how big the IMU is before I decide if I want to add lights to anything or not.

I wanted to do more, but it IS father’s day. It’ll progress smoother once I get the rest of my parts in.


Yay! I wired up the flex sensors!

First order of business was to finish wiring up the multiplexer. I’m not 100% happy with how my solder joints turned out, but I have to desolder some of that stuff when the IMU arrives anyway.

Here’s the pro micro. I got it because it was thin and came with a micro USB slot. If I thought ahead enough though, I’d have wired my own micro USB slot onto something with more than 4 analog slots. FML.

I figured this would help explain what’s going on, since there was a teansy weansy bit of misinformation in the last post. I confess, I didn’t know how some of this stuff worked. So I educated myself and it became less frustrating.

Okay, so the left side of the schematic is the 4 flex sensors. In the last post I said the red wires were going to ground, but they’re actually signal. It does need to be grounded to a resister to prevent micro-voltage from screwing up the readings.

The multiplexor is hooked up to the flex sensors via the brown, white, black, and blue wires. Because the MUX is a giant switch board, it can only select one wire at a time to send a signal through. However, it’s set up to flip through these bad boys so fast that it checks a finger about 100 times a second. Typically the signal sent trough the wires is whatever’s on the SIG pin. However, I want to measure resistance. So the SIG pin is hooked to power so that I can measure the resistance in the sensors.

Pins S0 through S3 are how the MUX selects which C pin to send the SIG voltage through. It’s binary code, where each of the S pins represents a digit, with ground being 0 and power being 1. so 0000 would be C0, and 1100 would be C3. At the moment I’m only using 4 pins, so I grounded s2 and s3 so that the last 2 digits would always be 0.

Fritzing didn’t have a pro micro in its database, so I’ve opted to use a pro mini to represent it. The difference being the mini has extra analog slots, but no built-in micro USB port, and the micro-controller is an Atmega328 instead of the Atmega32u4, which can write commands to the computer like a keyboard/mouse.

Having tested the flex sensors I can confirm that at least three of them work. the middle finger is not giving much in the way of readings. I’m not prepared to rule out my wiring yet. I will check it with a multi-meter later. Even if it needs to be replaced, it was still cheaper to buy the glove than 4 new sensors.

I realize I never announced concrete goals I’d like to accomplish with this project. I guess I’ll share that with you all now.

My primary goal is to transform the power glove into a “Modern Virtual Reality Device”. I don’t plan on using it on Steam VR games, but I’d like to make it function as a buttonless, stickless arcade stick and a 3D mouse for my CAD programs. I would like to switch between these two modes from the power glove’s wrist pad, as well as adjust the sensitivity of the flex sensors. While doing this, I don’t want to have to cut any openings in the power glove or do anything that would compromise its retro appearance.

Secondary is the cool factor. I’d like to get the speaker to play a tune when you first plug it in, and maybe add lights behind the microphones, etc. These features are not nearly as important to me as the above, but I will try to incorporate them if it won’t compromise the main goal.

…And now begins the 2-3 week wait for the IMU. My micro buttons for the wrist controller should be arriving any week now.


Really looking forward to seeing what you finish off with this in the end.

Your soldering does need a bit of work, though; I can see what looks like strands from a GND wire almost touching the VCC solder point…


So my IMU arrived 2 weeks early, and has actually been in my room for 2-3 days without me knowing it. Apparently my brother decided to sign for it when the FedEx guy arrived and just quietly placed it on my nightstand without telling me. The fricking thing is smaller than a postage stamp. FML.

At least I will be able to toy with it this weekend.

One thing that makes these Arduino boards, and even some copy-cat boards, so awesome is all the cool little fail safes they have built into them. You can short ground directly to the power rail and the worst thing that will happen is it will momentarily kill all communication until it’s resolved. Speaking from experience!

The strands of wire are hard to see with the naked eye, and i didn’t notice it until after I uploaded the picture. I have since taken care of them.


Guessing you saw this?

If not, it may help you in your project a bit. Go for it!


Didn’t do as much with the power glove this week as I wanted to. Got caught up in 4th of July festivities and the steam summer sale.

So here’s a picture of the IMU beside a quarter. Hard to believe it can do so much.

I have to do some reading up, but it has eight holes, six presumably analog outputs for the 3-axis gyro and accelerometer. I get the sneaky suspicion that I’m going to need another mux to cut down on the need for so many analog wires. Problem with the existing mux is that it’s set up with power on the signal wire. This second mux is going to need a signal wire that goes straight into the arduino. But that’s worse case scenario. I have plenty of wires left on the harness, and plenty of room left in the palm enclosure.

I’ve been working so diligently with the palm enclosure, I haven’t had much time to explain my plans for the wrist controller. I’m not a huge fan of the old style buttons, the parts that aid with the contacts wear out pretty quickly and they aren’t easy to repair/replace. So under every button is going to be a tactile button just like the one above.

I had the foresight to buy some solderable breadboards for my next controller project. It never happened, but now I have something to attach the buttons to.

They fit so snugly they almost don’t need to be soldered. Eventually I’ll cut it to fit the wrist controller. Due to the large number of buttons, I’m almost 100% possitive this will require a mux as well. However, I am unsure if I can fit a mux inside with the bread board in place. It’ll be a tight fit for sure!

Here’s a picture of the faulty flex sensor. I put my multi meter on it and I get good resistance, and it even changes as the sensor flexes.

The problem is actually with the tabs themselves. They’ve pulled away from the rest of the flex sensor, and they seem to be connected to each other, creating a detour for current that I want to run through the flex sensor. Because electricity is inherently lazy, it will always take the path of least resistance. I’m toying with the idea of repairing the flex sensor VS replacing it. I mean, anything to avoid a lengthy wait for parts right?

But how did I get to the flex sensor? Being that the power glove is sealed together and isn’t really designed to be taken apart?

I dissected it.

My cuts were not as clean as they could have been, but I think I can patch it back together well enough.

With the glove removed, you can see the flex sensors are all packed into little clear stockings. The one on the thumb is most interesting, as it actually has two 45 degree bends in it.

The glove, even without the plastic shell, still makes my hand sweaty. I’m toying with the idea of removing the rubber top layer of the glove and adding some breathable membrane. However, I’m not sure if that won’t be just as bad once it gets attached to the rest of the shell.

…And no, I’m not using gym chalk.

Underside of the glove still looks good.

I was so conflicted, but how else was I going to get to the flex sensor?

That plastic tab is actually what the palm housing bolts to. beneath it is all of the wire work going to the flex sensors. With an exception to the whole “cut the glove apart” thing, these gloves are actually very well put together. desoldering and removing the faulty flex sensor took seconds once the glove was cut apart.

So checklist for this week is…

-Determine if the flex sensor can be saved, then fix or replace
-Read up on how this particular IMU works, then put it in a prototype circuit.
-Determine the exact number of mux units I need
-Place an order for all required parts

It might take more than a week seeing as I have a BBQ to work off and steam games to play am busy with other projects. This is going to start moving along pretty fast once I know how I’m going to finish it. 'Till next time!

Nope. While I read a few different instructables, this one was the most useful for helping me decide how to go about modding the power glove.

The largest deviations from the power glove in the above link and mine are:

  • I chose a wired connection for an infinite supply of regulated power.
  • I chose to completely remove the old PCBs, because I didn’t want to destroy them. (I’m weird like that)
  • Instead of just an accelerometer, I’m going to put a full IMU in mine.


If it was glued you could have used a hair dryer to heat the glue up to get it soft enough to pull the halves of the glove apart. :frowning:


Is there actually continuity between the two pins? If not you should be fine resoldering them to the traces and securing them in your final project so the sensor doesn’t bend near them.


I know it looks bad : C

Directly under the plastic shell was a very soft glue that pulled apart fairly easily.

The seams were fused together. The plastic itself was melted at the edges to mesh with the glove’s fibers. At the base of that it was stitched together for good measure. So even if heat did work, I’d have to cut the stitches and entirely unravel the glove.

I think this will be easier to repair than re-stitching the glove. But I’ve repaired plastic/rubber in the past, and I’ve never stitched, so I’ll confess slight bias towards my approach.

Yea, it doesn’t matter if I touch both pins or just one pin, the multimeter reads 0.2 ohms. I’m hoping if I remove the pins from the sensor I’ll be able to see where the connectivity is happening, address it, then position the pins so that they are crimping the traces to the non-conductive paper. That’s how the other ones are set up.


I figured out how to fix the flex sensor!

So what I was struggling with (aside from the continuity between the metal tabs) was that solder didn’t stick to the conductive paper/plastic strip. In fact, it almost melted it. So I needed some way to press metal into it without destroying anything. Turns out the answer was as close as my desk stapler!

It’s not pretty, but I saved 12$ + shipping, so that’s okay.

The staples also appeared to be coated in something. They were plenty conductive, but the solder didn’t really want to stick. It tried to jump back onto my iron every chance it got. I eventully burned most of it off, hence the black scorching on the staples.

I used a wee dab of hot glue to make sure the staples didn’t move.

After that I just slid it back into the finger and ran tests on it. It’s not as sensitive as the other fingers, but it’s nothing that I can’t overcome with code.

I also re-soldered a few wires on the MUX and micro controller. it looks a little better and now I freed up enough of the wiring harness to know I have room for one more MUX to assist with the IMU. I’ll also have a few wires left over to use on the speaker and any lights I may wish to run. Nothing really worth taking a picture of at this time.

Looks like my shopping list consists of a few more Mux. It’s the only way I can get all 20 buttons to work with the remaining 14 digital pins. The last two big hurdles for this project are going to be getting everything to fit in that wafer-thin wrist enclosure and writing up the code. I’ll talk about the code more once I get all the hardware in place.


You could of used conductive glue


I knew it was going to be EXTREMELY hard to set up, but I really wanted to get the IMU working so that the palm enclosure was done. The good news is that it actually only needs 5 pins, and none of them are analog. So this means that I won’t need that extra multiplexer and can still have 3 analog pins for lights and sound. The bad news is that this is because it uses I2C pins, which are very technical and hard for a novice like myself to get working. The very bad news is that the particular IMU I got, while extremely powerful and cheap, (it can actually detect the earth’s rotation and the pull of gravity) The company that made it released virtually nothing about how it’s suppose to work. Thankfully, a guy called Jeff Rowberg has been backwards engineering the thing for years. Without him and his code, i probably wouldn’t be able to use this IMU in my power glove.

Something this advanced was going to be a chore to get working, so I tested it out on a bread board with my uno before installing it. There’s a lot of second hand information out there, and a lot of it is bad. It’s a simple enough setup, but false information really made it hard to get this setup to work.

What made this set up difficult was that the board is obscure enough that the code needed to be modified for it to work. I actually had the IMU working this morning, but it kept crashing after just a few minutes. After playing around both with the code and the wires, I finally got a setup that works. Let’s look at a wiring diagram so that it’s easier to explain.

The green block on the right side is the IMU. Just like last time, VCC is the power rail and GND is the ground. the Inter-integrated Circuit protocol (I2C) only requires two lines, SCL (clock) and SDA (Data). The INT wire is Interrupt, and is used to assist with timing. So I spent most of the day doing trial and error, editing the code and shifting around wire pins until I got a setup that works.

If you want the nitty gritty, I think sparkfun does a good job of explaining it. I won’t lie, I haven’t gotten all the way through the article yet.

It doesn’t look like much, but that was my weekend. The code for the IMU eats up about 70% of my Pro Micro’s memory, so I’ll need to start trimming away features that I don’t need to fit in my flex sensor program and all the input data for the buttons. I feel very successful in that, aside from the speakers and lights, I can close the palm enclosure. The only things left to get this working is the wrist enclosure and the code.