Linus Torvalds | 6d0b842 | 2007-09-20 11:33:45 -0700 | [diff] [blame] | 1 | ** Introduction |
| 2 | This document describes what I managed to discover about the protocol used to |
| 3 | specify force effects to I-Force 2.0 devices. None of this information comes |
| 4 | from Immerse. That's why you should not trust what is written in this |
| 5 | document. This document is intended to help understanding the protocol. |
| 6 | This is not a reference. Comments and corrections are welcome. To contact me, |
Johann Deneux | 118e78d | 2007-10-20 00:47:32 +0200 | [diff] [blame] | 7 | send an email to: johann.deneux@gmail.com |
Linus Torvalds | 6d0b842 | 2007-09-20 11:33:45 -0700 | [diff] [blame] | 8 | |
| 9 | ** WARNING ** |
Matt LaPlante | 01dd2fb | 2007-10-20 01:34:40 +0200 | [diff] [blame] | 10 | I shall not be held responsible for any damage or harm caused if you try to |
Linus Torvalds | 6d0b842 | 2007-09-20 11:33:45 -0700 | [diff] [blame] | 11 | send data to your I-Force device based on what you read in this document. |
| 12 | |
| 13 | ** Preliminary Notes: |
| 14 | All values are hexadecimal with big-endian encoding (msb on the left). Beware, |
| 15 | values inside packets are encoded using little-endian. Bytes whose roles are |
| 16 | unknown are marked ??? Information that needs deeper inspection is marked (?) |
| 17 | |
| 18 | ** General form of a packet ** |
| 19 | This is how packets look when the device uses the rs232 to communicate. |
| 20 | 2B OP LEN DATA CS |
| 21 | CS is the checksum. It is equal to the exclusive or of all bytes. |
| 22 | |
| 23 | When using USB: |
| 24 | OP DATA |
| 25 | The 2B, LEN and CS fields have disappeared, probably because USB handles frames and |
| 26 | data corruption is handled or unsignificant. |
| 27 | |
| 28 | First, I describe effects that are sent by the device to the computer |
| 29 | |
| 30 | ** Device input state |
| 31 | This packet is used to indicate the state of each button and the value of each |
| 32 | axis |
| 33 | OP= 01 for a joystick, 03 for a wheel |
| 34 | LEN= Varies from device to device |
| 35 | 00 X-Axis lsb |
| 36 | 01 X-Axis msb |
| 37 | 02 Y-Axis lsb, or gas pedal for a wheel |
| 38 | 03 Y-Axis msb, or brake pedal for a wheel |
| 39 | 04 Throttle |
| 40 | 05 Buttons |
| 41 | 06 Lower 4 bits: Buttons |
| 42 | Upper 4 bits: Hat |
| 43 | 07 Rudder |
| 44 | |
| 45 | ** Device effects states |
| 46 | OP= 02 |
| 47 | LEN= Varies |
| 48 | 00 ? Bit 1 (Value 2) is the value of the deadman switch |
| 49 | 01 Bit 8 is set if the effect is playing. Bits 0 to 7 are the effect id. |
| 50 | 02 ?? |
| 51 | 03 Address of parameter block changed (lsb) |
| 52 | 04 Address of parameter block changed (msb) |
| 53 | 05 Address of second parameter block changed (lsb) |
| 54 | ... depending on the number of parameter blocks updated |
| 55 | |
| 56 | ** Force effect ** |
| 57 | OP= 01 |
| 58 | LEN= 0e |
| 59 | 00 Channel (when playing several effects at the same time, each must be assigned a channel) |
| 60 | 01 Wave form |
| 61 | Val 00 Constant |
| 62 | Val 20 Square |
| 63 | Val 21 Triangle |
| 64 | Val 22 Sine |
| 65 | Val 23 Sawtooth up |
| 66 | Val 24 Sawtooth down |
| 67 | Val 40 Spring (Force = f(pos)) |
| 68 | Val 41 Friction (Force = f(velocity)) and Inertia (Force = f(acceleration)) |
| 69 | |
| 70 | |
| 71 | 02 Axes affected and trigger |
| 72 | Bits 4-7: Val 2 = effect along one axis. Byte 05 indicates direction |
| 73 | Val 4 = X axis only. Byte 05 must contain 5a |
| 74 | Val 8 = Y axis only. Byte 05 must contain b4 |
| 75 | Val c = X and Y axes. Bytes 05 must contain 60 |
| 76 | Bits 0-3: Val 0 = No trigger |
| 77 | Val x+1 = Button x triggers the effect |
| 78 | When the whole byte is 0, cancel the previously set trigger |
| 79 | |
| 80 | 03-04 Duration of effect (little endian encoding, in ms) |
| 81 | |
| 82 | 05 Direction of effect, if applicable. Else, see 02 for value to assign. |
| 83 | |
| 84 | 06-07 Minimum time between triggering. |
| 85 | |
| 86 | 08-09 Address of periodicity or magnitude parameters |
| 87 | 0a-0b Address of attack and fade parameters, or ffff if none. |
| 88 | *or* |
| 89 | 08-09 Address of interactive parameters for X-axis, or ffff if not applicable |
| 90 | 0a-0b Address of interactive parameters for Y-axis, or ffff if not applicable |
| 91 | |
| 92 | 0c-0d Delay before execution of effect (little endian encoding, in ms) |
| 93 | |
| 94 | |
| 95 | ** Time based parameters ** |
| 96 | |
| 97 | *** Attack and fade *** |
| 98 | OP= 02 |
| 99 | LEN= 08 |
| 100 | 00-01 Address where to store the parameteres |
| 101 | 02-03 Duration of attack (little endian encoding, in ms) |
| 102 | 04 Level at end of attack. Signed byte. |
| 103 | 05-06 Duration of fade. |
| 104 | 07 Level at end of fade. |
| 105 | |
| 106 | *** Magnitude *** |
| 107 | OP= 03 |
| 108 | LEN= 03 |
| 109 | 00-01 Address |
| 110 | 02 Level. Signed byte. |
| 111 | |
| 112 | *** Periodicity *** |
| 113 | OP= 04 |
| 114 | LEN= 07 |
| 115 | 00-01 Address |
| 116 | 02 Magnitude. Signed byte. |
| 117 | 03 Offset. Signed byte. |
| 118 | 04 Phase. Val 00 = 0 deg, Val 40 = 90 degs. |
| 119 | 05-06 Period (little endian encoding, in ms) |
| 120 | |
| 121 | ** Interactive parameters ** |
| 122 | OP= 05 |
| 123 | LEN= 0a |
| 124 | 00-01 Address |
| 125 | 02 Positive Coeff |
| 126 | 03 Negative Coeff |
| 127 | 04+05 Offset (center) |
| 128 | 06+07 Dead band (Val 01F4 = 5000 (decimal)) |
| 129 | 08 Positive saturation (Val 0a = 1000 (decimal) Val 64 = 10000 (decimal)) |
| 130 | 09 Negative saturation |
| 131 | |
| 132 | The encoding is a bit funny here: For coeffs, these are signed values. The |
| 133 | maximum value is 64 (100 decimal), the min is 9c. |
| 134 | For the offset, the minimum value is FE0C, the maximum value is 01F4. |
| 135 | For the deadband, the minimum value is 0, the max is 03E8. |
| 136 | |
| 137 | ** Controls ** |
| 138 | OP= 41 |
| 139 | LEN= 03 |
| 140 | 00 Channel |
| 141 | 01 Start/Stop |
| 142 | Val 00: Stop |
| 143 | Val 01: Start and play once. |
| 144 | Val 41: Start and play n times (See byte 02 below) |
| 145 | 02 Number of iterations n. |
| 146 | |
| 147 | ** Init ** |
| 148 | |
| 149 | *** Querying features *** |
| 150 | OP= ff |
| 151 | Query command. Length varies according to the query type. |
| 152 | The general format of this packet is: |
| 153 | ff 01 QUERY [INDEX] CHECKSUM |
Matt LaPlante | 01dd2fb | 2007-10-20 01:34:40 +0200 | [diff] [blame] | 154 | responses are of the same form: |
Linus Torvalds | 6d0b842 | 2007-09-20 11:33:45 -0700 | [diff] [blame] | 155 | FF LEN QUERY VALUE_QUERIED CHECKSUM2 |
| 156 | where LEN = 1 + length(VALUE_QUERIED) |
| 157 | |
| 158 | **** Query ram size **** |
| 159 | QUERY = 42 ('B'uffer size) |
Matt LaPlante | 01dd2fb | 2007-10-20 01:34:40 +0200 | [diff] [blame] | 160 | The device should reply with the same packet plus two additional bytes |
Linus Torvalds | 6d0b842 | 2007-09-20 11:33:45 -0700 | [diff] [blame] | 161 | containing the size of the memory: |
| 162 | ff 03 42 03 e8 CS would mean that the device has 1000 bytes of ram available. |
| 163 | |
| 164 | **** Query number of effects **** |
| 165 | QUERY = 4e ('N'umber of effects) |
| 166 | The device should respond by sending the number of effects that can be played |
| 167 | at the same time (one byte) |
| 168 | ff 02 4e 14 CS would stand for 20 effects. |
| 169 | |
| 170 | **** Vendor's id **** |
| 171 | QUERY = 4d ('M'anufacturer) |
| 172 | Query the vendors'id (2 bytes) |
| 173 | |
| 174 | **** Product id ***** |
| 175 | QUERY = 50 ('P'roduct) |
| 176 | Query the product id (2 bytes) |
| 177 | |
| 178 | **** Open device **** |
| 179 | QUERY = 4f ('O'pen) |
| 180 | No data returned. |
| 181 | |
| 182 | **** Close device ***** |
| 183 | QUERY = 43 ('C')lose |
| 184 | No data returned. |
| 185 | |
| 186 | **** Query effect **** |
| 187 | QUERY = 45 ('E') |
| 188 | Send effect type. |
| 189 | Returns nonzero if supported (2 bytes) |
| 190 | |
| 191 | **** Firmware Version **** |
| 192 | QUERY = 56 ('V'ersion) |
| 193 | Sends back 3 bytes - major, minor, subminor |
| 194 | |
| 195 | *** Initialisation of the device *** |
| 196 | |
| 197 | **** Set Control **** |
| 198 | !!! Device dependent, can be different on different models !!! |
| 199 | OP= 40 <idx> <val> [<val>] |
| 200 | LEN= 2 or 3 |
| 201 | 00 Idx |
| 202 | Idx 00 Set dead zone (0..2048) |
| 203 | Idx 01 Ignore Deadman sensor (0..1) |
| 204 | Idx 02 Enable comm watchdog (0..1) |
| 205 | Idx 03 Set the strength of the spring (0..100) |
| 206 | Idx 04 Enable or disable the spring (0/1) |
| 207 | Idx 05 Set axis saturation threshold (0..2048) |
| 208 | |
| 209 | **** Set Effect State **** |
| 210 | OP= 42 <val> |
| 211 | LEN= 1 |
| 212 | 00 State |
| 213 | Bit 3 Pause force feedback |
| 214 | Bit 2 Enable force feedback |
| 215 | Bit 0 Stop all effects |
| 216 | |
| 217 | **** Set overall gain **** |
| 218 | OP= 43 <val> |
| 219 | LEN= 1 |
| 220 | 00 Gain |
| 221 | Val 00 = 0% |
| 222 | Val 40 = 50% |
| 223 | Val 80 = 100% |
| 224 | |
| 225 | ** Parameter memory ** |
| 226 | |
| 227 | Each device has a certain amount of memory to store parameters of effects. |
| 228 | The amount of RAM may vary, I encountered values from 200 to 1000 bytes. Below |
| 229 | is the amount of memory apparently needed for every set of parameters: |
| 230 | - period : 0c |
| 231 | - magnitude : 02 |
| 232 | - attack and fade : 0e |
| 233 | - interactive : 08 |
| 234 | |
| 235 | ** Appendix: How to study the protocol ? ** |
| 236 | |
Matt LaPlante | 01dd2fb | 2007-10-20 01:34:40 +0200 | [diff] [blame] | 237 | 1. Generate effects using the force editor provided with the DirectX SDK, or |
| 238 | use Immersion Studio (freely available at their web site in the developer section: |
| 239 | www.immersion.com) |
| 240 | 2. Start a soft spying RS232 or USB (depending on where you connected your |
| 241 | joystick/wheel). I used ComPortSpy from fCoder (alpha version!) |
Linus Torvalds | 6d0b842 | 2007-09-20 11:33:45 -0700 | [diff] [blame] | 242 | 3. Play the effect, and watch what happens on the spy screen. |
| 243 | |
| 244 | A few words about ComPortSpy: |
Matt LaPlante | 01dd2fb | 2007-10-20 01:34:40 +0200 | [diff] [blame] | 245 | At first glance, this software seems, hum, well... buggy. In fact, data appear with a |
| 246 | few seconds latency. Personally, I restart it every time I play an effect. |
Linus Torvalds | 6d0b842 | 2007-09-20 11:33:45 -0700 | [diff] [blame] | 247 | Remember it's free (as in free beer) and alpha! |
| 248 | |
| 249 | ** URLS ** |
| 250 | Check www.immerse.com for Immersion Studio, and www.fcoder.com for ComPortSpy. |
| 251 | |
| 252 | ** Author of this document ** |
Johann Deneux | 118e78d | 2007-10-20 00:47:32 +0200 | [diff] [blame] | 253 | Johann Deneux <johann.deneux@gmail.com> |
Justin P. Mattock | 0ea6e61 | 2010-07-23 20:51:24 -0700 | [diff] [blame] | 254 | Home page at http://web.archive.org/web/*/http://www.esil.univ-mrs.fr |
Linus Torvalds | 6d0b842 | 2007-09-20 11:33:45 -0700 | [diff] [blame] | 255 | |
| 256 | Additions by Vojtech Pavlik. |
| 257 | |
| 258 | I-Force is trademark of Immersion Corp. |