Taijigamer2
Gutsy Member
So on my adventures trying to look into the elusive hardware of the Panasonic 3DO, I decided to try and figure out the lesser known BT encoders. The later FZ-10 mainly came with the VA536 video encoder which fortunately has a well documented datasheet telling us which pins do what. Unfortunately the BT (BrookTree) encoders have no datasheets available. I have searched high and low for these elusive datasheets but no luck. The BT910x encoders have been seen in both 3DO and Philips CDi hardware.
So without a datasheet I decided to try and use trial and elimination to work out as much of these encoders as I could. Fortunately my FZ-1 came with the BT9103 encoder. This encoder has the same package as the VP536 (68 pin QFP) and we know from Airs' thread that 2 pins are responsible for selecting between PAL/NTSE and Progressive/Interlace. I was hoping the BT encoders followed similar principals.
So eliminating pins we could already work out (24 RGB pins, VS, HS, Y, C, CVBS) The next step would be to lift remaining pins and tie them high or low and observe the result. Obviously the chance of damaging the chip was probably quite high. Fortunately after a little searching, I came across a CDi thread which mentioned the BT856 encoder. http://mjpeg.sourceforge.net/driver-zoran/datasheets/bt856.pdf
This encoder has the same pinout as the BT9103 encoder. As we can see, 2 pins control the region and interpolation (pins 59, 60).
I checked my board and found as expected pins 59 and 60 held high (5V) which corresponded to my region (PAL) and interlaced. When pin 59 was lifted and tied low to GND, the console was put into Progressive mode.
I couldn't get a good picture of the result as I don't have an RGB mod or capture card but there was a definite improvement on picture with CVBS, pixel crawling stopped and the pixel edges became much more defined. S-Video, the result wasn't as profound. So if anyone has the BT9103 encoder and wants to try this mod to confirm it with RGB and capture card, please do.
I soldered the wires to a DPDT switch with the middle wire going to the original pin 59 (BLUE). Another wire going to 5V and the final wire going to GND. I chose 5V from a nearby pad (RED) but any 5V point will do, same with GND (GREEN).
I found unreliable behaviour with the pin still soldered to the board. The datasheet has the maximum voltage for a low input at 0.8V. When I measured the voltage with both 5V and GND tied to the pin, it measured 0.2V. This could be causing erratic behaviour so I recommend lifting the pin from the board for best results. Also, the FZ-1 had no problems booting in 240p in this configuration. Switching between 240p and 480i with the BT9103 causes no issues mid operation as the chip switches during ODD fields to prevent syncronisation disturbance.
You could also tie the pin to GND without the switch to give permanent 240p but I have read that some games behave weirdly in 240p.
Hopefully this mod is useful to people with the BT9103 encoder. After experimenting thoroughly with the pins, unfortunately the RGB pin seems to be redundant in these chips. This has also been confirmed for the BT9106 encoder used in some Philips CDi units as per this thread . The BT9107 may have the RGB pin active so there is a possibility to swap the chips to potentially mod the console for native RGB but I would need a BT9107 to test this. Until then, I will be also looking into dedicated RGB encoding with modern ICs.
The next task is tackling the BT9101 encoder. This encoder has 160 pins and no similar datasheet so it will be a bigger challenge. We now know the 'style' of the BT chips so once we eliminate the known pin functions, it shouldn't take too long to work out and may even retain the RGB function.
If anyone has any 3DO with BT9101 encoders and fancies a challenge
Special thanks to:
Superg
Airs
and RetroRGB
for their generous sharing of knowledge in the field of RGB modding of legacy consoles.