Some details on how I have wired this up.
 
I took a standard servo wire, and removed the black wire from the plug side, and plugged the red/white into the pins 7-8 of the blackfin 32pin bus. Black pin went to pin2 GND. These were connected to drive S1 (white, pin7 of bf, left side motor), S2 (red, pin8, right side motor). Black wire went to GND of sabertooth.
 
Blackfin____________________Saber Tooth 2x10
Pin2   <--------Black-------------> GND
Pin7   <--------White-------------> S1
Pin8   <--------Red---------------> S2
 
 
I also confirmed that switch 4 of the Sabertooth needs to be in OFF position to support differential (non mixed drive). I configured a "elevon" mix on my RC radio side, and things are happy.

I would leave em as is.. I can modify the srv.config to use the new commands and that should be fine..

Okay - I've posted a test version of code at  
 
    http://www.surveyor.com/blackfin/srv-blackfin-servo-020108.zip
 
I have added the 'S' and 's' commands to the protocol.  At the moment, the range of supported values is 0-100 (0x00 - 0x64) rather than 0-255, but we can change that later.  So in srv.config, the command
 
     533232
 
would be the equivalent of 'S' 50 50, which generates 1.5ms pulses at 50Hz on timer channels 2 & 3.  An 's' (0x73) command has an equivalent effect on channels 6 & 7.
 
The 'M' commands still work as well.  If the processor first sees an 'M' command, it will initialize the timer channels 2 & 3 in PWM mode for the robot H-bridge, and the keypad commands are activated in this mode after the first 'M' command is issued.  
 
If the processor first sees an 'S' command, it will initialize timer channels 2 & 3 in PPM (servo) mode, and when it first sees the 's' command, it will initialize timer channels 6 & 7 in PPM (servo) mode as well.
 
I tested the same code on an SRV-1 robot and on the coaxial rotor flyer using all 4 channels, and everything seems to be working, but please let me know how things go for you.
 
By the way, you can still drive this from a terminal - here's a link to an ASCII chart
    http://www.cdrummond.qc.ca/cegep/informat/Professeurs/Alain/images/ASCII1.GIF
 
so the sequence "S22" would be the same as 533232 or 'S' 50 50
 

I was able to map out a few controls for srv1.config tonight
 
Stop point S22
 
Forward Medium S##
 
Reverse Medium SAA
 
Left Rotate SA#
 
Right Rotate S#A
 
I plan to make a new srv.config and post this tomorrow with the new controls..  
 
I did have an issue with my wifi antenna, I think it was related to the pin, it appears the new pin material has pushed the metal sleeve inside the plug in too far, and doesnt make a good connection now, for now I have butchered an old wireless router that had the same connectors and I am using its antenna and it seems to be better now..  
 
Now I need to map the s channel and make some servo movements for the camera p/t..

Thanks for the update - I'm very glad to hear that the new commands are working for you and look forward to the video, which we'll post in a link from the Surveyor Robotics Journal. Also, thanks for noting the antenna issue - I hadn't heard of problems with the replacement pin before.  
 
My goal today is to get the coaxial flyer airborne with the new firmware.  I will have to modify the keypad controls to produce a useful command set, since I need to be able to continuously adjust throttle and yaw trim as well as the pitch/roll servo settings.  However, since there's now logic to branch depending on whether we're in PWM or PPM mode, I can add this code without breaking anything.  
 
My thought is to use the keypad hits to incrementally adjust the last values sent by the 'S' or 's' functions.  For my application, the '+' and '-' keys would be used to incrementally adjust throttle.  '0' and '.' would be used for yaw trim.  The remaining numbers would be used to incrementally adjust pitch/roll settings.  

I have the 4-channel servo interface working as well.  
 

 
You can find a description and a couple of video clips here -  
 
    http://diydrones.com/profiles/blog/show?id=705844%3ABlogPost%3A23537
 
There's still work to be done on the flight controls, but the Blackfin interface is solid.
 
Here's the updated code -  
 
    http://www.surveyor.com/blackfin/srv-blackfin-coax-020308.zip
 
In the PPM (servo) mode, the '+' and '-' keys are intended as throttle controls, increasing or decreasing both speed control channels.  The '<' and '>' keys are used to trim the two channels for yaw or steering control.

Very nice !
 
I just posted a brief description of your project here -  
    http://www.surveyor.com/cgi-bin/robot_journal2.cgi/2008/02/04#141
 
We will release the code changes supporting servos with the mainstream SVN code base after a bit more testing, as the same firmware works with the existing SRV-1 base, your Lynxmotion base, and the coaxial rotor flyer described above.

Quote from osmogr on 02/05/08 at 5:32pm:

 
At some point, I need to document the S, s, +, -, <, >.  I think once more of the sensor code is integrated, we'll freeze code, update the SRV_protocol document, and update the SVN archive.
 
There were requests in the past to enable the processor to jump into the C interpreter with a switch or jumper.  Again, this is something we might want to assign to a GPIO line, though we'd need to add a pullup resistor and jumper to the sensed line.  Maybe that's something to add to the next expansion board revision.