very interesting. I think it would be a good resouce to take a look at. (provided we don’t break any copyright rules by posting it for download of course.)
I did a bit of hunting and found the website I must’ve gotten the PhraseALator from. They’re using frames on the site, so here’s the direct link to the downloads page: magnevation.com/software.htm
I think the only thing I added to the dictionary I have is Lynxmotion. 
lynxmotion= \LE \FAST \IH \NGE \KO \FAST \SO \FAST \MM \OW \SH \AX \NE
You can download the software, users manual, or just the dictionary. Hope this helps!
Thanks for the link Beth!
I remember now why I never used that software. My SJ sensor board doesn’t have a DB9 connector to program the SJ. I think I need to add the SJ developmnet board on my “need to get” list
If you do a new revision of your board, maybe you can find space to add a header (not DB9) that you can attach a cable to. A header should not take as much space as a DB9 attached to your board. The cable would have the DB9 on the other end. The cable would be similar to those older motherboard cables that go from the board to the back panel or to a bracket that fits one of the expansion slot bracket spaces.
This might allow you to do in circuit programming.
8-Dale
I do have another board design that I might do later. I want to experimant with a band pass filter or maybe even an pot to adjust sensitivity of the audio portion. A DB9 connector might be possible on the board if I remove the PING sensor header to make more room.
No, don’t remove anything unless you really have to or really want to for some reason. Just put a header on the board to attach a cable with a DB9 on it. A small single row header should not take up too much board space.
8-Dale
I hear what you are saying but the PING connector has no real functionality. The PING header is only a header connected to the 10pin microcontroller header via traces. The reason I put that there was to allow easy remove and replacement of the PING sensor without have to mess with anything else. When boards are stacked on one another, access to pins become limited, and therefore I decided to have the PING “ported” through the SJ board along with all other sensors and have everything feed through one 10pin header that remains connected to the Bot board.
With my design, if you ever needed to remove my bothead for example, you simply “unplug” the ears, and remove the head and servo while the board remains mounted still connected to the bot board. Same with the PING sensor, just unplug and remove, or the tilt sensor, just unplug it and use it for another project.
The Tilt sensor socket does have built in resistors of the proper values so that it’s plug -n-play. The board also functions with or with out the Memsic sensor installed as with the SJ chip as well without any problems with the board function.
Ah, OK, I understand now. The PING header is just a convenience rather then necessity.
It’s nice to not have to take everything apart to remove or change out a sensor. However, everyone doesn’t want sensors mounted the same way, including me. 
I’d probably have to stack your board using longer standoffs to be sure of having access to all the ABB connections.
I wonder if there would be interference from motor noise if all the electronics were mounted on the third deck of my robot. Spacing between decks is 2 5/8" now and I have the motors mounted for maximum ground clearance.
Which tilt sensor are you using? I have an Analog Devices ADXL330 3-Axis accelerometer module to eventually put on W.A.L.T.E.R.
I am going to have to check your board out again soon. I would like to see how this would work on W.A.L.T.E.R. I’d love to have him be able to “hear” and talk. 
8-Dale
I designed the board to use the Parallax Memsic tilt sensor, not sure of the model number though.
YES! That be the one! 8)
So I received Jim’s SpeakJet (SJ) board on Friday and messed with writing some Basic Atom Pro (BAP) code to make it work. Here is the way it goes.
The SJ has 3 outputs that can be configured to report with an active high level “ready”, “speaking”, and “buffer half full”. I point out the “can be” part because even though the SJ manual says this is the default the chip on the board Jim sent was not defaulting to this configuration. For this reason I recommend initializing the chip with a "reset’ command and a “portcfg 7” to enable these lines. Note that because handshaking may or may not be enabled you can’t use handshaking on the serial output of the initialization string.
The serial mode needs to be inverted and defaults to 9600 baud. For the BAP this is I9600. According to a post in the other BAP bug thread by AcidTech the serout hardware handshaking feature is active high, that is the input must be in the non-zero state for the data to be transmitted. Previously on the BA we were wiring the Buffer Half Full (BHF) signal as the handshake. This is the wrong active state for the BAP. I noted however that the active state of the ready output seems to be inverted to the BHF output once the chip is initialized. So with the BAP we can use the READY signal as our handshake.
The other signal of interest is the speaking (SPK) output. The SJ manual does show the duration of each phoneme in ms but unless really pinched for I/O lines it makes good sense to bring SPK in as a higher order flow control signal. Without it you would have to estimate or hand tune how long each phrase takes to say to avoid just mashing stuff together in one big long run on sentence. This line is active high when the SJ is speaking so it can be used to block additional phrases from bring sent until the last one has completed. This is different then the READY or BHF signals in that they are preventing you from over-running the serial buffer in the SJ and loosing data.
Here is a link to some video I took of the BAP sending speech to the SJ. On the last long sentence, after “mission completed”, I purposely made a sentence that would over-run the SJ buffer if the handshaking was not working. youtube.com/watch?v=DKqwrGThoyQ It’s not the greatest video as I just used my wife’s digital camera to record it but it shows the thing working at least.
I took the code SN96 posted in the BAP bug thread and used all his constant declarations for the phonemes. I have only included a short sample of the serout statements since most of it is wash, rinse, repeat the same thing over and over.
[code]; sjbap.bas
; Driving the SpeakJet IC with a Basic Atom Pro module demo code
; P4 is serial line to SJ
; P7 is SPK line from SJ
; P8 is RDY line from SJ
; the BHF signal from the SJ is not used
; serout mode I9600 is used
; SpeakJet phoneme codes follow
PA0 CON 0 ; pauses
PA1 CON 1
PA2 CON 2
PA3 CON 3
PA4 CON 4
PA5 CON 5
PA6 CON 6
Fast CON 7
Slow CON 8
Stress CON 14
Relax CON 15
_Wait CON 16 ; note underscore
Soft CON 18
Volume CON 20
Speed CON 21
Pitch CON 22
Bend CON 23
PortCtr CON 24
Port CON 25
_Repeat CON 26 ; note underscore
CallPhr CON 28
GotoPhr CON 29
Delay CON 30
Reset CON 31
IY CON 128
IH CON 129
EY CON 130
EH_ CON 131
AY CON 132
AX CON 133
UX CON 134
OH CON 135
AW CON 136
OW CON 137
UH CON 138
UW CON 139
MM CON 140
NE CON 141
NO CON 142
NGE CON 143
NGO CON 144
LE CON 145
LO CON 146
WW CON 147
RR CON 148
IYRR CON 149
EYRR CON 150
AXRR CON 151
AWRR CON 152
OWRR CON 153
EYIY CON 154
OHIY CON 155
OWIY CON 156
OHIH CON 157
IYEH CON 158
EHLL CON 159
IYUW CON 160
AXUW CON 161
IHWW CON 162
AYWW CON 163
OWWW CON 164
JH CON 165
VV CON 166
ZZ CON 167
ZH CON 168
DH_ CON 169
BE CON 170
BO CON 171
EB CON 172
OB CON 173
DE CON 174
DO CON 175 ; note underscore
ED CON 176
OD CON 177
GE CON 178
GO CON 179
EG CON 180
OG CON 181
CH CON 182
HE CON 183
HO CON 184
WH CON 185
FF CON 186
SE CON 187
SO CON 188
SH CON 189
TH CON 190
TT CON 191
TU CON 192
TS CON 193
KE CON 194
KO CON 195
EK CON 196
OK CON 197
PE CON 198
PO CON 199
RO CON 200
R1 CON 201
R2 CON 202
R3 CON 203
R4 CON 204
R5 CON 205
R6 CON 206
R7 CON 207
R8 CON 208
R9 CON 209
A0 CON 210
A1 CON 211
A2 CON 212
A3 CON 213
A4 CON 214
A5 CON 215
A6 CON 216
A7 CON 217
A8 CON 218
A9 CON 219
_B0 CON 220 ; note underscores
_B1 CON 221
_B2 CON 222
_B3 CON 223
_B4 CON 224
_B5 CON 225
_B6 CON 226
_B7 CON 227
_B8 CON 228
_B9 CON 229
C0 CON 230
C1 CON 231
C2 CON 232
C3 CON 233
C4 CON 234
C5 CON 235
C6 CON 236
C7 CON 237
C8 CON 238
C9 CON 239
D0 CON 240 ; DTMF tones
D1 CON 241
D2 CON 242
D3 CON 243
D4 CON 244
D5 CON 245
D6 CON 246
D7 CON 247
D8 CON 248
D9 CON 249
D10 CON 250
D11 CON 251
M0 CON 252 ; sonar ping
M1 CON 253 ; pistol shot
M2 CON 254 ; WOW
EOS CON 255
; nice names for I/O pins and some program constants
SJTXD CON P4 ; this is the serout transmit pin
SJRDY CON P8 ; this is the active high ready pin
SJSPK CON P7 ; we actually need to define this differently
; for now it is hard coded in WaitForSilence
; as IN7, but assigning IN7 as a constant doesn’t
; seem to work right.
MyPin var bit ; variable to read the SJSPK input pin in WaitForSilence
MyPin = IN7 ; access these to make sure they are configured as inputs
MyPin = IN8
HIGH SJTXD ; set TXD pin to marking high so we don’t confuse the SJ
PAUSE 1000 ; just a nice pause from hitting RESET button
; init commands don’t use handshake since we can’t be sure OUT[2:0] are configured yet
SEROUT SJTXD, I9600, [Reset, PortCtr, 7, Volume, 127, Pitch, 120]
;‘Initializing Speakjet’
SEROUT SJTXD\SJRDY, I9600, [FAST,IH,NE,IH,SH,SH,UX,LE,OHIH,ZZ,ZZ,NGE,PA2]
SEROUT SJTXD\SJRDY, I9600, [SE,PE,SLOW,IY,EK,JH,EH_,TT,PA2]
GOSUB WaitForSilence ; wait for the SJ to stop talking
PAUSE 1000 ; put a break before next phrase is sent
;‘Lynxmotion Mini SJB’
SEROUT SJTXD\SJRDY, I9600, [LE,FAST,IH,NGE,KO,FAST,SO,FAST,MM,OW,SH,AX,NE,PA3]
SEROUT SJTXD\SJRDY, I9600, [MM,FAST,IH,NE,IY,PA2, EH_,SE,SE,PA1, JH,EYIY,PA1, BE,IY,IY]
GOSUB WaitForSilence ; wait for the SJ to stop talking
PAUSE 1500 ; put a break before next phrase is sent
;‘Status OK’
SEROUT SJTXD\SJRDY, I9600, [SE,SE,TT,AY,TT,UX,SO,SO,PA2]
SEROUT SJTXD\SJRDY, I9600, [OW,KE,EYIY,PA2]
GOSUB WaitForSilence ; wait for the SJ to stop talking
PAUSE 2500 ; put a break before next phrase is sent
;‘Mission Completed’
SEROUT SJTXD\SJRDY, I9600, [MM,IH,FAST,SH,SH,FAST,IYEH,Slow,NE,PA2]
SEROUT SJTXD\SJRDY, I9600, [KO,MM,PO,LE,IY,TT,EH_,ED,PA2]
GOTO complete
; subroutine to wait for the SJ to stop talking
WaitForSilence:
MyPin = IN7
If MyPin <> 0 Then WaitForSilence ; wait for the phrase to complete
RETURN
; all done, go to low power mode until reset is pressed.
complete:
END
[/code]
Note that I could not figure out how to assign P7 to a constant in such a way to get the value of the pin into the MyPin variable using that constant. I tried what I thought were the obvious ways but I got tired of messing with what seemingly should have been something simple and straight forward. As a result the SJSPK input is hard coded in the WaitForSilence subroutine as an IN7. Maybe someone can enlighten me on the proper way to do this. Someone should be able to assign SJSPK as a constant up above and not have to find every instance of IN7 in the code body to port it to their application.
EB
:mrgreen:
That is pretty cool. It is a bit difficult to make out, but it seems to me that if one would probably have to spend a lot of time to get pronounciation just right. Thats real good skills you have Eddie!
P.S. The Arnold impression at the end was dead on 
The links Beth posted earlier include a text file with several hundred common words broken out to phonemes.
Eddie,
Thanks for all the help you have provided on this. I really appreciate all your efforts. Thanks for posting that code also!
I’m a little confused as to the handshake methods though. I understand how you used the SPK, BHF, and the RDY lines as flow control factors, but per the SJ manual, only BHF is required for handshaking. The Basic Stamp II worked fantastic using the BHF line as the flow control. I have never purposely tried to over run the buffer so I don’t know if the SJ could handle it with only the BHF line alone, but I usually broke a phrase in to chunks if it was long winded. This method, however, would use up more programming space to achieve the same results I suppose, but not enough to any significant degree.
Thinking about it, using SPK, RDY, and BHF lines I guess makes for a much better system and I never thought about using these lines as another way of using flow control.
I am going to need to replace my SJ chip before I can resume trying this stuff out. 
By the way, what version of the IDE were you using to make the SJ work?
Also, can you PM me your mailing address? I would like to send you my SJ Chip and see if you can get the sucker to work because I’m almost certain mine is toast. If you don’t want to bother that’s ok, I don’t want to make you feel obligated.
You really only “need” the RDY line for the BAP, just like you only used the BHF line on the BA. The entire purpose of this handshake is to prevent your overflowing the buffer in the SJ chip. Once the SJ chip is initialized inside and ready to take commands the RDY line appears to take the roll of BHF inverted, that is if BHF is high then RDY is low. Since it appears the hardware handshakes work with different active levels for the BA vs. the BAP this means you can use the RDY line.
The advantage of also including the SPK line as a secondary handshake is it allows you to pace phrases or sentences with natural breaks in between. You might send a phrase, add a 1 second delay, and then send a second phrase with the expectation of there being 1 second between the two phrases when you hear them. However if the particular first phrase has several long syllables it is possible to absorb most of your 1 second break and have the two phrases mashed together because you did not wait for the SJ to stop speaking before starting the delay. Using the SPK line you can know the SJ has stopped speaking, start your delay, and then send the next phrase so you actually get two distinct phrases instead of one big one. You don’t “need” to do this but the end results will be more natural (as much as that applies with this thing) sounding than not using it.
I d/l and installed the most recent IDE that AcidTech had provided a link to in the orriginal thread for this subject. I think 8.0.1.3, perhaps?
I completely understand this and it’s an excellent solution.
PM me your address. 