DescentBB

Built mine tonight and it worked great!

Attached a picture of the board. I went with a cheaper PCB board and soldered everything. Since the Teensy 2.0's are out of stock right now from Pjrc, I socketed mine until I can get another couple.

Going this route keeps the cost down pretty cheap. I had everything needed except the 15 pin connector. Ordered that from Jameco.

Here is a breakdown of the parts if doing my method.

Teensy $18.00

Protoboard = $5.49

2.2K Resistor 100 pack $2.00

D-sub 15pin Female = $0.79

.001 uF Capacitor 10 pack = $0.60

Connector pins quantity of 2 $0.59 each total $1.18

Total before shipping is $28.06

I put a quantity of two of the header pins so the Jameco order would be over $10 (their minimum)

If you don't mind soldering this is a pretty cheap way to make one. The 15pin d-sub take a little \"pin tweaking\" to get the pins to line up with the standard board spacing, but it isn't too hard.

I didn't look, but Jameco might have a project box to house this board, or they do have many other boards that would go with a project box.

You can also save your self some time and get the teensy with pins. You save a dollar or so doing it yourself, but if you want to minimize your soldering, get the teensy with pins for $3 more, and remove the connector pins from your order. Course your Jameco order will be below $10 so you might want to get extra 15 pin d-subs, or anything else you might need.

Nice work!

I know a few Descent Rangers that are using that Manhattan usb converter and seem to be happy with it and the sidewinder.

This is a dead thread, time to unsticky it.

KoolBear wrote:This is a dead thread, time to unsticky it.

KB,

With all due respect, this thread should stay stickied. Grendel's design is the only one which provides full functionality (all other adapters are limited to emulation mode), so this is a valuable resource.

Since this is pretty much the hub of the converters documentation it shall stay sticky. If you disagree you can PM me or discuss it in the mod section.

Just snagged a New In The Box (Sealed) 3d pro off ebay.

Grendel, the teensy that you refer to http://www.pjrc.com/store/teensy_pins.html is out of stock. Can the teensy++2.0 work?

Yes, but the firmware needs to be adapted to the ++ 2.0. I think I put provisions for that chip into it, needs testing tho. I do have a v1 ++ board, have to check that tomorrow at work. I could try to get it going on that board, if time permits.

I've ordered the ++2.0 and hope to have it by next week.

Is there any advantage to using the Teensy over a 3dpvert r2 other than the ease of programming? Is there a poll rate, power consumption advantage of any sort?

The Teensy based converter does have a higher poll rate and less latency (USB 2.0 vs. 1.1). The Teensy ++ board I have is v1, will take it home and poke at it.

Ok. Did a quick jab at the code w/ my Teensy++ v1.0 -- seems to work fine, needs some more serious testing tho.

For the hardware, I just took the Teensy 2.0 prototype, pulled the board, added a 2nd small breadboard, added the ++, and connected the two. The firmware uses the same ports -- B4/5 and D0-3.

I should be receiving the 3d pro tomorrow and the Teensy should arrive on Friday.

Considering the prospective availability of the Teensy 2.0, it appears as though the future of the 3dpvert will be reliant on the ++2.0

Did you have to alter the firmware any to get it to load on the ++? Is there any sacrifice to using the AT90USB1286 instead of the ATMEGA32U4?

Wasp wrote:Did you have to alter the firmware any to get it to load on the ++?

Yes, the firmware needs code specific to the MCU (mostly initialisation). The current version (not publicly available yet) does support all four Teensy boards (1.0, 2.0, ++ 1.0, ++ 2.0) but needs to be recompiled for each.

Wasp wrote:Is there any sacrifice to using the AT90USB1286 instead of the ATMEGA32U4?

Just the price & size of the board. The AVR core and the USB module are equal in capabilities from the firmware point of view. Technically the 32U4 is already a lot more MCU than really necessary, the 1286 even more so. Many things could be done w/ these... (WingMan support anyone ?)

If you PM me your email address I'll send you the current firmware image for the Teensy++ 2.0

Edit: actually, I may be able to send you a PM w/ the code attached. I'll try that in the evening.

Ok, can't add attachments to PMs. Let's see if I can attach it here. ... Nope, ".hex extensions are not allowed". Wonder what is allowed... Fallback to plan A -- you'll need to send me your email address via PM.

PM with email addy sent.

I received the 3d pro yesterday, I expect to have the teensy on Friday. I'll get working on it the moment it arrives and post an update.

It appears as though the teensy is programming ok, but afterwards, nothing detected. Not even the "USB Input Device" which was seen when programming. Windoze 7 (64)

Wasp wrote:It appears as though the teensy is programming ok, but afterwards, nothing detected. Not even the "USB Input Device" which was seen when programming. Windoze 7 (64)

Correct -- as long you don't have a supported joystick connected. The software will not enumerate as a device w/o first detecting and initializing a joystick. While it's waiting for a stick the LED should flash at about 2Hz.

The package that the teensy came in has a label "Teensy 2.0 ATMEGA32U4 Chip". It's definately a ++, but no way to tell what chip is on it as the writing on the chip itself is too dark to read.... After further examination, I can see that it is a 90USB1286.

When the stick is not connected, the Teensy blinks constantly...When I connect the JS, the Teensy stops blinking after about 3 blinks. If I disconnect the USB from the computer and reconnect, it blinks a few times and stops. If I disconnect the JS, the Teensy does not resume blinking like it did prior to connecting. I'm going to try it on a XP 32bit ma chine and see what happens.

Tried programming and testing on xp 32bit with same results. I programmed the sample fast / slow flash hex and Teensy seems to respond correctly to that so i'm guessing that the Chip is ok. btw, are those 1/8 watt resistors on your breadboard?

Right, the Teensy 2.0 is based on the 32U4, the larger Teensy++ 2.0 on the 90USB1286 -- probably just a packaging issue..

Hm, your description indicates that the firmware does detect the joystick and now should attach to USB. That's a bit confusing since the 646 and 1286 chips have the same USB module (actually the only difference betw. these chips should be the amount of memory) and I would expect that part to work just fine. I'll go double check the code and the chip manual.

Edit: I think I found the problem, the USB clock prescaler value in the PLLCSR register needs to be different for the 1286 chip (thanks Atmel...) Sent a new hex image.

Its now recognized in control panel. So far, it looks like everything is functioning normally. The Teensy 2.0++ has a very rapid and very dim blink while the JS is connected. Disconnecting the JS results in a bright blinking Teensy...reconnecting the JS results in the fast and dim blink. Appears very normal to this layman's eyes.

I've tested multiple button presses with simultaneous hat and stick movements and can see no problem whatsoever. I'll do some gaming with it shortly. I'm curious if that increased poll rate is noticable.

I'm very grateful for all your effort and continued support. Having this adapter along with a brand new 3DPro is like heaven to this old gamer.

I'll chime in with some photos when I get it stuffed into a project box.

Thanks again, you guys really rock!

Glad it works, enjoy !

Yes, the dim LED flashing is normal. The code will very briefly flash the LED every time a data packet is send out. Since the data is only send if the values are different from the last set sent, the LED behavior can be quite erratic.

Hi

Doing some tidy up at my place I found my old MS FFP and I was glad to find your project to make it work.
Unfortunatly I have the same issue than Wasp, I've got a Tensy 2.0++ and it doesn't work with the .hex file I've got.
Is there's a way to get the correct file for it? That would be much appreciated.

Thanks for your great job.

So i started reading through this thread but its getting late and I'm lazy. Did anyone end up making some of these they are selling or am I gonna have to attempt to make one myself? Lemme know cause I'm fiending for some Descent right now!

You can try your luck here but you might want to PM him about them. As far as I know, he's done making them.
Though you never know, you might get lucky.

Grendel,

Can you tell us what USB speed this device uses? I experimented with a joy to mouse software and ran mousespeed which revealed 105Hz, but I wasn't sure if the joy to mouse software was influencing the outcome.

What does that program measure/report ? The USB speed is 2.0 full speed. The poll interval is set to 5ms (200Hz poll rate on the INT channel.) The joystick may deliver only up to 100 different values / s tho (never checked that.)

Grendel wrote:What does that program measure/report ?

Poll interval

Using the mousespeed software and hovering the mouse over the test pad, you can detect the poll interval (125Hz) of the mouse. This allows me to test the changes I make to the USB port (rate). By using a joystick to mouse software, I was hovering the mouse cursor over the software test pad, via the joystick, hoping to detect the poll rate of the joystick and got (105Hz).

I'm pretty confident that's not the hardware poll rate. As said, the device is requesting 5ms and I did confirm w/ a USB protocol analyzer that the rate actually is that value (at least on XP.)

I think you're right, using a different joystick to mouse software, mousespeed returned 20Hz.

Looking in device manager under "Standard OpenHCD USB Host Controller" I see my mouse and the joystick, is that normal? Another high speed device shows up under "Standard Enhanced PCI to USB Controller" which made me wonder why the Joystick wasn't in there instead.

Just a heads-up, I updated the source code.

Great job! Thank you!

Great work! I just finished building my converter with the Teensy ++2.0, and it works great. Thanks for the updated code for the 1286 chip. I built the converter on a 2 1/2 inch square pc board and mounted it with a pc bracket just like a normal pc card in the computer. I ran the teensy's usb port to an internal usb connector on my motherboard. I just plug my joystick into the game port and it works.

Excellent work Grendel. I just built my converter using the Radio Shack breadboard assembly and it works perfectly. The only small issue that arose is that Winford added a 16th pin for a shield to their DB15 breadboard adapter. I just trimmed it down flush with the board and it was good to go.

Thanks again. It's great to use the old Sidewinder once more.

I own the Microsoft SideWinder 3D Pro Plus #97462 and the Microsoft SideWinder Force Feedback Pro. I haven't been able to use them for years and they're in mint condition. Can anyone direct me to a couple of these USB adapters? I wouldn't be able to build them so I'd need to purchase them.

Thanks in advance!

Anyone?

I found the breadboard design quite simple to assemble. I spent about $60 on the parts listed and the illustrations were easy to follow. Event the wires were color coded by length.

Someone correct me if I'm mistaken, but I recall seeing a note by Grendel somwhere that the converter did not include support for the force feedback option.

Yes, correct -- from the 1st post:

Grendel wrote:Just to clear this -- the converter in the current state does not support force feedback effects for the FFP and probably never will. Here's why:

Most gameports back in the day were located on a sound card sharing some pins with the cards MIDI interface. M$ added pin 12 to the FFP (MIDI TxD) and uses MIDI channel 6 to issue FF commands. It's too much work trying to reverse engineer the commands, add a MIDI interface to the 3DP-Vert and modify its USB code to have a fully featured FF device.

Yo Grendal, Found a PPro with a gameport for 2 bucks . I tried plugging it into the 1st version of your converter and although it does initialize (weird) all the inputs are skewed. Before I setup an environment to flash with your latest firmware for Ver.1, will that firmware in fact support this PPro or do I need to consider the teensy one? I have parts for 2 more Ver.1 Converters and had rather just utilize them.

Another thing. What speed crystal did you finally decide on being the most stable. I've been running the 16 MHZ for awhile now (too lazy ) and 90% of the time it works fine, but sometimes it will disconnect in the game and send me spinning. I've checked the connections numerous times and can only think it has something to do with the USB timing somehow. Do you suggest the 15 MHZ over the 16 ?

flip wrote:Yo Grendal, Found a PPro with a gameport for 2 bucks . I tried plugging it into the 1st version of your converter and although it does initialize (weird) all the inputs are skewed. Before I setup an environment to flash with your latest firmware for Ver.1, will that firmware in fact support this PPro or do I need to consider the teensy one? I have parts for 2 more Ver.1 Converters and had rather just utilize them.

In general yes, the current code does support the PPro but C8 & C9 need to be smaller than the originally designed 10nF. I successfully used 2.2nF and 1nF caps (these eg.)

flip wrote:Another thing. What speed crystal did you finally decide on being the most stable. I've been running the 16 MHZ for awhile now (too lazy ) and 90% of the time it works fine, but sometimes it will disconnect in the game and send me spinning. I've checked the connections numerous times and can only think it has something to do with the USB timing somehow. Do you suggest the 15 MHZ over the 16 ?

16 MHz should be fine if you flash a more recent binary. The 16MHz module in the AVR-USB software stack had a few bugs fixed since '08, including one that could cause that behavior. Drop me an email describing the converter hardware (mega8 or 461, if the former -- 556 or not ?) and I will send you a freshly compiled binary.

Message sent.