Hi guys! This guide is for absolute beginners and is a step by step guide on creating an internal Soarer's Converter. For reference, here's the end product:
1.jpg (1.42 MiB) Viewed 10535 times
This adaptor is designed for IBM's Model F but can definitely be re-purposed for other keyboards. It's powered by a Pro Micro (Teensy clone) with 4 Dupont male pins leading out of it allowing you to plug and play into an IBM Model F. It can then be followed by a MicroUSB Cable of your liking.
Please PM me if you have queries. I'm active on geekhack as well.
I also hand make these adaptors on eBay and I ship them worldwide. PM me if you're interested
For the Dupont cables, we just need 4. Alternatively, you could get any wires you want, but it'll make it much harder to plug into the Model F. For the Heat shrink tubing and black tape, it's not necessary, but it'll make things more sturdy and prettier.
Obviously this requires a soldering iron and some other equipment like scissors and a pair of steady hands. This is also written for Windows.
Change the COM5 into the COM Port that shows up on your computer.
7.PNG (31.65 KiB) Viewed 10535 times
9) If it times out (you have a buffer of about 5s before the Pro Micro resets itself), short the pins again and type in the command again. I personally type the command before hand and press enter when I short the teensy.
10) If the firmware is successfully flashed, you should see this:
8.PNG (36.64 KiB) Viewed 10535 times
9.PNG (30.3 KiB) Viewed 10535 times
11) Under device manager, there should be no more COM Ports and you should see a Soarer's converter in the devices panel in control panel.
10.PNG (29.47 KiB) Viewed 10535 times
Yay! You've flashed the firmware successfully!
Now onto the next part, soldering and making the converter:
1) You'll need 4 Dupont male to female cables. I've got 4 and they're about 5 inches long.
11.jpg (1.06 MiB) Viewed 10535 times
2) I used a plier to bend the male pin into a right angle as it makes soldering much easier.
12.jpg (856.98 KiB) Viewed 10535 times
13.jpg (893.01 KiB) Viewed 10535 times
14.jpg (973.06 KiB) Viewed 10535 times
3) Now get to soldering! I have a helping hand so I can clamp the cable down on the board and the right angle sticks the pin into the hole.
The pins to solder is: 2, 3, GND and VCC. For GND, there are a few that you could use but I recommend using the one along the side of the VCC.
15.jpg (911.25 KiB) Viewed 10535 times
16.jpg (1.25 MiB) Viewed 10535 times
4) Snip off any excess solder which might affect soldering:
17.jpg (1.26 MiB) Viewed 10535 times
5) Solder the rest of the cables!
18.jpg (1.13 MiB) Viewed 10535 times
21.jpg (1.24 MiB) Viewed 10535 times
22.jpg (1.34 MiB) Viewed 10535 times
6) The end result:
23.jpg (1.52 MiB) Viewed 10535 times
7) Time to test the Pro Micro! I'm using a Model F XT here:
to find out the pinout. You'll need a multimeter for this. I made a simple diagram to follow:
25.png (1.85 MiB) Viewed 10535 times
It's not the best illustration, so here's a written explanation:
Use a multimeter to figure out which pins of the DIN connector (Model F XT for example) leads to which pin on the actual board connector and use the keyboard pinout documentation as reference. I'll leave the Model F XT and 122 pinouts here:
F 122 Connector view:
C G D
X N V
C - CLOCK
G - GROUND
D - DATA
N - NO PIN
V - VCC
9) Time to wrap the adaptor up! At this point, the adaptor is completely functional so if you wanna stop here and start using, it's done. I'm doing this part to insulate the adaptor. I'll be wrapping it up in insulating black electrical tape followed by heat shrink wrappings.
26.jpg (1.29 MiB) Viewed 10535 times
10) Electrical tape:
27.jpg (1.36 MiB) Viewed 10535 times
28.jpg (1.37 MiB) Viewed 10535 times
11) Heat shrink wrapping, you can use any heat source to shrink it:
30.jpg (1.01 MiB) Viewed 10535 times
31.jpg (1.03 MiB) Viewed 10535 times
12) All done!
32.jpg (553.22 KiB) Viewed 10535 times
It's done! This section is showing off some pigtails and female USB mounts.
1) Female pigtail:
33.JPG (930.95 KiB) Viewed 10535 times
35.jpg (1.83 MiB) Viewed 10535 times
36.jpg (1.41 MiB) Viewed 10535 times
37.jpg (331.16 KiB) Viewed 10535 times
38.jpg (1.49 MiB) Viewed 10535 times
2) Mounting the female USB onto the board:
39.JPG (1.8 MiB) Viewed 10535 times
This is a magnetic microUSB cable I sourced, allowing me to attach and detach magnetically.
40.JPG (1.93 MiB) Viewed 10535 times
MicroUSB female mounted into a Model F AT.
41.jpg (1.75 MiB) Viewed 10535 times
42.jpg (1.48 MiB) Viewed 10535 times
Model M50 with a USB-B female mounted into the case.
Also, the Pro Micro is tiny. It's about half the length and less than half the width of a sticky note. If you're old and have bad eyes, like me, you'll really want a magnifying lens. If you don't have one, you should probably get your "Helping Hands" with one.
The guide talks about the "2 & 3" spots on the Pro Micro. These are for "Clock" and "Data." If you connect it to your keyboard incorrectly, the keyboard just won't work. Just swap the wires on your keyboard.
I tested the Pro Micro every time I made any change. A couple of times, I accidentally pulled a cable from the keyboard side. Another time, I found that I put on the electrical tape too tight and that shorted a connection. Just take it slow!
Hi, how would you deal with the lack of grounding for this mod? I have to do an internal converter for my F XT and it just inputs random characters to the pc, someone pointed out the lack of grounding for my F XT but i don't really know how to incorporate that with this mod.
I just wanted to add what fixed the random characters issue for me because it might help others. I've been plagued by this problem for years (it would usually appear in the morning when I started my computer up and then eventually go away after some jiggling/unplugging), and I eventually had had enough of dealing with it and started researching a real solution.
What finally worked for me was to solder the j1 jumper on the pro micro. Afterwards I plugged my Model F in and it started working 100% flawlessly. I couldn't tell you exactly what the problem was, something to do with 3.3v vs 5v power, idk, but it worked.
Normally all Sparkfun 5V Pro Micros ship with the J1 jumper shorted.
And all Sparkfun 3.3V Pro Micros ship with the J1 jumper not shorted.
That's not the only difference between them, also:
1) The 3.3V variant has a 8MHz crystal, while the 5V variant has a 16 MHz crystal
2) The 5v variant has an 5V voltage regulator (this is used when powering it from a battery of a higher voltage), While the 3.3V one has a 3.3V regulator which is always used.
So it's not as simple as opening up the J1 jumper to turn a 5V pro micro into a 3.3V one.
Normally in a 5V Pro Micro the J1 jumper is shorted which bypasses the regulator, and the D2 protective diode.
See full schematics here: http://cdn.sparkfun.com/datasheets/Dev/ ... o_v13b.pdf
What happens if you have a 5V Pro Micro, without J1 shorted, is that you just get a significant voltage drop from the input, through the diode, and the regulator. This is very undesireable. On one non-shorted Pro Micro I have here I'm measuring 4.62V on the output (with the input being 4.99V)
At this voltage the pro micro itself is expected to work fine most of the time, since the atmega32u4 can operate at 16MHz down to 4.5V supply voltage:
atmega32u4_voltages.png (12.54 KiB) Viewed 5643 times
The cheaper 5V Pro Micro clone manufacturers from ebay/aliexpress/etc pretty much all of them forget to to short the J1 jumper. And they are absolutely wrong about this. And the only reason they have not been forced to short it, is that the pro micro itself kinda works at that voltage!
However trouble comes from many directions:
1) the voltage will not always be above 4.5V. The diode to be used is not specified on Sparkfun's schematics, and someone making a Pro Micro clone could easily choose a diode with a higher forward voltage drop, and bring the voltage below 4.5V. And at that point it won't fail immediately, it will just operate out of spec, and only some devices will fail, rarely.
2) trouble comes when other stuff is being powered off of the bad voltage-dropped 5V line. The diode's forward voltage drop is current dependent, here's a typical diodes forward voltage curve:
diode.png (14.32 KiB) Viewed 5643 times
Also, the regulators drop-out voltage is current dependent:
regulator.png (26.89 KiB) Viewed 5643 times
If you're drawing 500mA, which would be USB maximum, and also depending on diode selection you could drop even 1.7V total across diode and regulator, leading to an output voltage of 5V - 1.7V = 3.3V.
The atmega32u4 chip works at 3.3V but not with a 16MHz crystal, at that voltage it's only specified up to 8MHz, so it will likely fail in this condition.
Also trouble can come from whatever else you are powering not having enough voltage (i.e. your keyboard).
And also if current consumption keeps changing, then the supply voltage will vary as well!
So I hope I convinced everyone reading this, that: You should short J1 on ALL 5V Pro Micros*
* The only exception I have heard of, is that some split keyboards, using QMK, are kinda relying on the wrongly-built ebay/aliexpress pro micros, because the way they are determining which side of the keyboard is slave, and which one is master, is by looking for the presence of voltage on the VBUS pin of the atmega32u4. And if J1 is not shorted, then only the half-keyboard that has been plugged into USB has any voltage on VBUS. If I had one of these keyboards I would try to change the firmware to not use VBUS detection, or if that is not an option, then I would mod the Pro Micros in the following way: 1) change D2 diode to a very low forward voltage schottky diode (it might already be a schottky on some Pro Micros, just not the lowest possible) 2) short RAW pin to VCC pin, so you bypass the drop on the voltage regulator.
* What I have written above only applies to Pro Micros, and ebay/aliexpress clones that are direct copies of the original Pro Micro schematic. I have not studied in detail other boards, that are promicro-inspired, or the footprint is Pro Micro -compatible. They might not even have a J1 jumper after all. Those cases should be studied independently.