Hacking DEAD [Shutdown]DragonInjector - Game Cart Payload Injector (Trinket M0 Clone)

[MatinatorX]

just wondering for the test batch r u going to have a cfw preloaded on them?

The actual CFW goes on your MicroSD card if you're using one, so I'll assume you mean a payload. I'd like to include one, but I need to check up on the ones currently available to see if I'd be allowed to include one and what accreditations I'd need to provide. I definitely won't be including the SX OS payload though as that's someone else's commercial product, but it's easy to install it on your own.

do you happen to still have it? I've resurrected a few 3d printers before and would love to take a crack at an SLA.

I don't, unfortunately. I'm fairly handy myself and would have loved to get it back up and running, but it took a tumble down a set of concrete stairs. Every conceivable piece of metal was bent, the acrylic cover was in shambles, both laser galvos broke and the screen cracked. By now it's probably part of a few hundred soda cans

Oh yeah also interested in purchasing a bare board and making my own case

Of course! Bare boards will be available with the seoncd batch, I'll probably put up 30 of them alongside the 100 full packages. I won't be able to offer cases separately until I get my own SLA printers, though.

Glad to hear you got the price down again. I'm kinda tempted to try building one of these myself in the future (i have too many things on the todo list right now). This is a great project and very well designed. I like it.

Thanks! I spent a lot of time on the mundane things like prettying up the traces and making all the edges on the model line up and stuff so I'm glad it shows! If enough people want, I can even put together a kit with all the parts + board or just the parts or whatever people want. Just need to let me know!

This is excellent. I would perhaps buy this on day 1.

My only complaint about that are jig pins. They don't seem very Joy-Con slot friendly (could damage Joy-Con slot).

IMHO the pins should look like something like the pins on picture bellow. The black one is TX's jig.

I see where you're coming from, and it's certainly a valid concern, but I spent a great deal of time taking measurements and testing the jig to make sure it wouldn't damage the Switch. I've used it a good 50 times now on my own Switch and checked the pins under a microscope and haven't noticed any difference. The feel of sliding in my joycon is exactly the same as before as well. The grooves you see in my jig are actually divets and not ridges, and just give the pins somewhere to go while keeping them aligned properly. They do not stick up above where the flat areas of existing jigs go. The pins themselves are minuscule, even smaller than on the SX jig. They're just the leads from a 1/8w resistor and only stick up 0.2mm from the jig surface. I might be more concerned if other jigs were using spring pins or something to alleviate some pressure, but from what I can see, despite the "peak" shape of the wire bend, there is still no room for the pin to depress as it's sitting right against the case (like the red example you provided) or fixed in place (like the black TX jig).

I went with my design because SLA printing is very precise and I thought I might take advantage of that fact to make the alignment divets. With the design above, if you happen to have a chunk of dirt in your joycon rail or something else happens where it raises the jig up too much, the pins just fall into the divets. With the flat faces on the other jigs, the pins might become under pressure, and with nothing guiding them properly they might bend sideways. It's a very low risk (maybe not even that), but one I didn't want to take if I could avoid it.

That said, if it does somehow turn out to be a problem it's an easy one to fix. I'd like to hear a bit more about why you think it would damage the joycon rail if you have the time! The last thing I want it to damage anyone's Switch.

1st batch colors don't effect me. really don't care what it is. CHEAPER IS better but quality is always something to keep in mind.
GREY IS FINE
question does this come with a battery ? or do we need to pick one up. would like to know before getting it so i can have them on hand

Good to know! It's hard to see in the pictures but the grey actually looks really nice. It's quite dark with a slight blue tint.

Also, there's no battery needed - the device runs off of a supercapacitor that charges from the Type C port or from your cartridge slot. I designed it to pull less power from the card slot than a standard game cart. It should outlive the battery in your Switch by a country mile.

Oh man thats a wondeful project please dont forget me in the next batch!!!

Thanks! I'll be posting here as soon as the next batch is ready!

 

[munky-head]

I'm interested in at least one, possibly even 4 units at once!
Keep me in the loop please.

 

[mss1988]

I see where you're coming from, and it's certainly a valid concern, but I spent a great deal of time taking measurements and testing the jig to make sure it wouldn't damage the Switch. I've used it a good 50 times now on my own Switch and checked the pins under a microscope and haven't noticed any difference. The feel of sliding in my joycon is exactly the same as before as well. The grooves you see in my jig are actually divets and not ridges, and just give the pins somewhere to go while keeping them aligned properly. They do not stick up above where the flat areas of existing jigs go. The pins themselves are minuscule, even smaller than on the SX jig. They're just the leads from a 1/8w resistor and only stick up 0.2mm from the jig surface. I might be more concerned if other jigs were using spring pins or something to alleviate some pressure, but from what I can see, despite the "peak" shape of the wire bend, there is still no room for the pin to depress as it's sitting right against the case (like the red example you provided) or fixed in place (like the black TX jig).

I went with my design because SLA printing is very precise and I thought I might take advantage of that fact to make the alignment divets. With the design above, if you happen to have a chunk of dirt in your joycon rail or something else happens where it raises the jig up too much, the pins just fall into the divets. With the flat faces on the other jigs, the pins might become under pressure, and with nothing guiding them properly they might bend sideways. It's a very low risk (maybe not even that), but one I didn't want to take if I could avoid it.

That said, if it does somehow turn out to be a problem it's an easy one to fix. I'd like to hear a bit more about why you think it would damage the joycon rail if you have the time! The last thing I want it to damage anyone's Switch.

It seems that your jig is even more reliable than the jigs I posted. I'm looking forward for this product Good luck bro

 

[MatinatorX]

It seems that your jig is even more reliable than the jigs I posted. I'm looking forward for this product Good luck bro

Thank you! We'll know for sure before too long

 

[SavagePR]

This looks very pro and convinient, i want two.

 

[MatinatorX]

Alright, time for another update!

I was messing around with some testing this weekend on a breadboard to see if I could get some stuff out of the way before the actual boards arrive, and I found out some good stuff as well as one bad thing. The latter meant that I had to redesign some things and order a new batch of boards, so everything has been pushed back about 3 days. This means that rather than the end of this week, the 10 tester units should be going out early next week instead (sorry!). I decided to take this opportunity to eschew sleep for a few days add in some features I'd been working on but didn't think I'd be able to get working as quickly as I did. I also ordered enough parts and PCB's for the second batch of 100 since 10 PCB's somehow cost the same as 100 and the components won't be changing again. Read on for the details.

So first thing, I am super happy with the Type C connectors I chose. They may be a bit expensive, but they're very easy to solder and because of the slotted design, you can solder the casing to the PCB for much more mechanical strength. I found I could snap a 0.6mm PCB that was 25mm long clean in half without bending or damaging the connector once it was fully soldered. So, whatever little worry I had about that is now long gone.

Secondly, I managed to get my Trinket M0 powered from both my PC and my Switch's USB port. A small triumph, yes, but important nonetheless. Which leads me into the one bad thing...

I was previously unaware that - starting with USB Type C - VCC is always cold until something tells it how much power it should deliver. Anything older then Type C always has VCC hot, so whatever you hook up to it can get power right away. With Type C, even if you're only using USB 2.0 speeds and power levels, at minimum you need to at least tell it you want to run in "legacy mode". This is accomplished by putting a 5.1KΩ resistor between CC1 and GND. This is how I got my Trinket M0 powered off of my PC's native Type C port, as well as from the Switch. You can see my super elegant testing process in the photos below. It should be noted that when operating in USB slave mode (like the Switch is when in RCM mode) USB Type C won't deliver VCC even with said resistor.

What this means is that the current crop of boards in the mail can only charge from the game card slot or when plugged into a USB Type A to Type C adapter. It's likely 90% of people will be charging this way anyways, but I wasn't really happy with that, so I went back to the drawing board and added in the offending resistor as well as a bunch of other stuff.

So let's get to what's changed:

1) The dongle will actually charge from native Type C ports now (oops).

2) The power switch is GONE! Thanks to @jcrorxp cluing me in to how the SX dongle wakes up from standby, I was able to do some testing and get it working myself. There is now an N-Channel MOSFET that powers up the ATSAMD21. Whenever external 5V is applied through the Type C port, in addition to charging the dongle, it also goes to the gate of the MOSFET. Any USB device in host mode will send this 5V and the dongle will wake up. Now the cool part - it turns out that while USB devices in slave mode won't send us anything on VCC, and so we can't use it to turn on our ATSAMD21, they do pull D+ up to 3.3v as a way to advertise their speed capabilities. We can use this to our advantage to wake up the dongle when connected to a USB slave device - for example a Switch running in RCM mode - by also sending that 3.3v to the gate of our MOSFET. A few safety diodes later and a pull-down resistor to drain the MOSFET and we have a superior solution for powering the dongle. In summary, when you plug the dongle into something, it'll turn on. When you unplug it, it'll turn off. Thus, the power switch is redundant now and has been retired.

3) It turns out the Type C connectors I'm using have an indent for alignment purposes, so I added it to the board to make soldering much easier.

4) A few people didn't like the idea of charging via the game card slot, so I've routed the 3.3v from the card slot pin via a single trace through a fairly empty area of the board and marked it. This means that if you don't want to charge the dongle via the card slot for some reason, you can just slit the trace with a knife where marked and it will only charge from USB afterwards.

5) I ditched the DotStar LED again. The thing is expensive and inefficient and it was driving me crazy. 2mA quiescent current draw when it's completely off? What is this, 1980? No, not in my dongle. I'm now using a 1515 size common anode RGB LED that uses less board space, costs less than a tenth the price, is actually designed to run at 3.3v and doesn't gorge on power while sitting pretty. Of course, this means I'll have to fork the code again, but I'm actually kind of looking forward to that after seeing some of the cool work @mattytrog @electronrancher and @Canna are doing right now. Go check out their stuff if you haven't already! As long as I'm going to make my own firmware, I might as well implement one more change...

6) Payload selection! With the PCB space gained from removing the power switch, I was able to fit a second push button in its place and connect it to D0. You will be able to store multiple payloads and switch between them by pushing this button. I'll put up the code on my GitHub when it's ready along with instructions on how to get it working. I don't think I'll be able to include any actual payloads as they belong to others, but I'm still looking into it. Definitely don't count on the SX payload being included though. You will have to add it yourself using the provided instructions.

So, that's where we stand right now! It's a shame it's going to be a few more days before the tester dongles go out, but I'm excited about the things I was able to include in the latest (hopefully final) revision. The best part is that due to the high per-part cost of the power switch, the BOM is actually marginally less than it was for the last version, even with all the new stuff. This means the price will remain $30 CAD/$23 USD (unless our dollar takes a nose dive again). I should mention that the enclosure remains unchanged, as the new board has the same layout for all the dimensionally important bits.

I haven't yet had the time to update the renders or the poster, but I'll get to it in the next day or so. For now, I've included snaps from the JLCPCB gerber viewer below as well as the latest schematics. Again, the full source files will be released once I have a fully working board in hand.

Before I go, just wanted to thank you all again for all the support! You're all awesome!

 

[XiorgON]

@MatinatorX

Wow that is amazing news.
same price but better product.

What do you mean by the new push button ? it will let us switch between payload on the tricket memory ?

I can PM you but as i already say reserve me 2x please.
If you want i can transfert cash with paypal to your account

Thanks for all the energy, passion, quality and brain you put on your project

XiorgON

 

[MatinatorX]

Forgot to add this to my last post - I dove into the power stuff a little more and I have some numbers now. (WARNING! MATH BELOW!)

- Charging the dongle for the first time (0v) to a full charge (3.2v) will take 90 seconds, but will be usable (2v) after 25 seconds.
- Charging the dongle from low battery (1.7v) to a full charge (3.2v) will take 70 seconds, but will be usable (2v) after 5 seconds.
- Peak charging current for first charge (0v) to full charge (3.2v) is 27.5mA, with current decreasing logarithmically as capacitor voltage increases.
- Peak charging current from low battery (1.7v) to full charge (3.2v) is 13.4mA, with current decreasing logarithmically as capacitor voltage increases.

These numbers only apply when charging the dongle from the game card slot. See a few paragraphs down for USB charging numbers. When charging via USB, the ATSAMD21 is powered up and will be consuming a portion of the 27.5mA current we have to play with so charging takes longer.

For those who choose to charge via the game card slot, thanks to the way a capacitor charges, battery drain will not be an issue. If we were pulling the maximum charging current of 27.5mA from the Switch whenever the game card slot was energized, we would see a loss of around 0.5% battery per hour that the screen was on (or the Switch was docked and not in standby). However, the capacitor will pull exponentially less current as it charges, eventually pulling practically nothing. Since a full charge will take a maximum of around 90 seconds, after which hardly any current will be used, effect on battery life will be pretty much insubstantial. The dongle will not charge when the Switch is sleeping, so it will not sap your standby power at all, nor will the dongle lose any charge beyond what is normal for an idle capacitor.

As far as the number of uses you can get on a full charge, that's going to depend heavily on how much current the ATSAMD21 uses. I've seen reports suggesting about 12mA is consumed while injecting the payload, but that seems rather high to me, so I'm hoping I can to decrease it. However, if we assume this isn't possible and that the required current will be 12mA, we can get a pretty good idea how many uses we can get. Keep in mind that 12mA of current draw is probably a worst case scenario.

A capacitor's energy is measured in coulombs. We can figure out how many coulombs we have to play with by multiplying the capacitance by the usable voltage.

Our ATSAMD21 works down to 1.7 volts (1.65 according to the spec) and our supply voltage/fully charged capacitor voltage will be 3.2v, so our usable voltage is 3.2 - 1.7 = 1.5 volts. Multiplied by our supercapacitor's capacitance of 0.22F, we should have about 0.330 coulombs to play with.

A single coulomb is exactly enough energy to power a one amp load for one second. It can also be expressed as powering a 1000mA load for one second, a 16.7mA load for one minute or 0.28mA load for one hour (mAh).

With our 0.330 coulombs, we power a 330mA load for one second, a 5.5mA load for one minute or a 0.0916mA load for one hour (mAh).

If we want to convert that into a power on time for our dongle that has a 12mA load, we can take our "330mA for one second" result and divide it by our 12mA load current. This gives us about 27.5 seconds of power-on time.

This might not sound like much, but remember that the dongle will only turn on and use power from the supercapacitor when connected to a device in SLAVE mode. If it's connected to a device in HOST mode, it will be getting 5V and will be charging the battery instead. The Switch is in SLAVE mode for less than a second for payload injection before powering the port completely off, then turning it back on in HOST mode a few seconds later, charging the dongle back up. Real world results will likely vary a bit, but you can expect at least a dozen payload injections between charges, perhaps even two dozen. Not bad for a power source half the size of a tic-tac!

We can also take that 12mA estimated current usage and use it to calculate our charging times using USB charging. We need only subtract that load from our 27.5mA of total possible current. This gives us a charging current of 15.5mA:

- Charging the dongle for the first time (0v) to a full charge (3.2v) will take 165 seconds, but will be usable (2v) after 45 seconds.
- Charging the dongle from low battery (1.7v) to a full charge (3.2v) will take 130 seconds, but will be usable (2v) after 10 seconds.
- Peak charging current for first charge (0v) to full charge (3.2v) is 15.5mA (not including ATSAMD21 or other loads like LED), with current decreasing logarithmically as capacitor voltage increases.
- Peak charging current from low battery (1.7v) to full charge (3.2v) is 7.5mA (not including ATSAMD21 or other loads like LED), with current decreasing logarithmically as capacitor voltage increases.

Running the numbers, what I consider a "usable" charge level of 2v should give us a power on time of about 5.5 seconds, which should be good for a few injections, so if you're in a hurry you'll never have to wait more than 10 seconds to get your payload injection on. Woo!

Here's some nerd links for nerds like me if you want to calculate stuff yourself:

http://www.ohmslawcalculator.com/ohms-law-calculator

http://mustcalculate.com/electronics/capacitorchargeanddischarge.php

https://www.electronics2000.co.uk/calc/capacitor-charge-calculator.php

https://www.rapidtables.com/convert/charge/coulomb-to-ah.html

--------------------- MERGED ---------------------------

@MatinatorX

Wow that is amazing news.
same price but better product.

What do you mean by the new push button ? it will let us switch between payload on the tricket memory ?

I can PM you but as i already say reserve me 2x please.
If you want i can transfert cash with paypal to your account

Thanks for all the energy, passion, quality and brain you put on your project

XiorgON

Yup! I'm going to program it to flash the green LED quickly for the number of the payload. (1 fast blink for 1st payload option, 2 fast blinks for second, etc.) I'll make sure it won't do this in the middle of sending a payload obviously and I'll commit the change to EEPROM each time so it will remember what you picked forever.

Really appreciate the support! I'll be posting here as soon as the second batch of 100 is ready, and if I can figure out how to do preorders/backorders on tindie I'll probably do that after those are gone while I scramble to make more.

 

[345742D]

Okay, I'm ready to throw my money at you. How can I reserve one?

 

[Cobalt`]

Thank you for this well documented and kick-ass Canadian project I res'ed my account just to commend your efforts

 

[Riky_Xerez]

I want one or two too

 

[Andalitez]

My body is ready

Edit : Though Seriously,excellent work.
will you be limiting orders per customer?

I figured batch orders would be easier to ship,but i'd like to get 4 of them tbh but if I can only get one from the first order I'd be happy with that.

Thanks again for the exceptional work

 

[clancy94]

Hey mate, are you shipping internationally? (I'm in Melbourne, Australia)

Keep up the amazing work!

 

[fragged]

My body is ready

Edit : Though Seriously,excellent work.
will you be limiting orders per customer?

I figured batch orders would be easier to ship,but i'd like to get 4 of them tbh but if I can only get one from the first order I'd be happy with that.

Thanks again for the exceptional work

I think diehard supporters should be limited to 1 each for the first 100 batch, then open it up to more bulk orders after that.

 

[MatinatorX]

Thanks for all the awesome support everyone! New boards are in production and should ship Wednesday. If I'm lucky, I'll get them by Friday. I'll see what I can do about a demo video once I have a working prototype in hand.

I think diehard supporters should be limited to 1 each for the first 100 batch, then open it up to more bulk orders after that.

I don't know if tindie allows for purchase count limiting, but if they do I'll be limiting to 2 per order for the first batch of 100. After that, there won't be any limits. Once the first batch of 100 is out I should be in good shape to make more fairly quickly, so those who want to place bulk orders or those who missed out on the first round won't have to wait long.

--------------------- MERGED ---------------------------

Hey mate, are you shipping internationally? (I'm in Melbourne, Australia)

Keep up the amazing work!

I don't see why not! I'm not some big corporation with a shipping department with weird rules and exclusions and stuff.

Another nice thing about using a supercapacitor is that I won't have to deal with lithium battery import/export regulations or anything funny like that, so it should make shipping pretty painless.

 

[Thetoto]

Looks sooooo good, well done!
I really want one

 

[Lumince]

Okay, I'm ready to throw my money at you. How can I reserve one?

Yea same, I need to reserve one for the first 100!

@MatinatorX How do I do?

 

[Andalitez]

Yea same, I need to reserve one for the first 100!

@MatinatorX How do I do?

fall in line soldier..lol

 

[0000ff]

Wahooo! Definitely want to reserve 1

 

[linuxares]

So is it working good for you?

 

[MatinatorX]

So is it working good for you?

Still waiting on the newest boards I designed on the weekend, hopefully they'll ship tomorrow and be here Friday or Monday. Pretty much at a standstill hardware-wise until they get here.

Until then, I'll be updating the poster in the main post and will fork the payload code this weekend for my non-dotstar LED, payload selection button and some power optimizations.