Digital Ballast Bypass Expirimenting

[jan.didden]

Originally Posted by crazifunguy

The purpurse of removing the ballast is so I can retrofit an HID or LED lamp and keep using the projector insted of buying a lamp for ~$300. The projector will run just fine with it removed as long as the projector thinks it is still talking with the ballast. This also cuts way down on power consumption as the projector isnt burning 250W into the lamp

OK, so if I recall your scope pics, it seems that the ballast returns the exact same commanded data but inverted, is that correct? So then, is it an option to just connect whatever comes out of the command micro back into the ballast input? Just looping the serial link, inverted?

jd

[crazifunguy]

Here is a signal from the tx and rx of the microprocessor. It looks to have alot of noise but this is the signal going to the opto-couplers



Here is a picture of the ballast and input wires.
Wires are as follows:

[crazifunguy]

Here is a signal from the tx and rx of the microprocessor. It looks to have alot of noise but this is the signal going to the opto-couplers



Here is a picture of the ballast and input wires.
Wires are as follows:
RED - Signal Back to Projector From Ballast
WHITE - Ground
BLUE - +5v
YELLOW - Square Wave CLOCK
BLACK - Signal From Projector To Ballast



This is where I get confused.....
The bottom 2 optoisolaters are for input. The middle one being the clock signal and the lower one being the serial information to the ballast.

I was looking at the data sheet for the isolaters and tracing the circuits.
http://www.datasheetcatalog.org/data...shay/83666.pdf

The top pins get +5V from the blue wire
The signal lines go through 100 ohm resistors and to the bottom pins(Most likley current limiting)

The other side of the the lower optocouplers feed into the microprocessor on the rx and clock lines

On the return the microprocessor feeds ithe tx output into a transistor then to the optocoupler and back to the board

The signals I meaured from the pins of the micro processor are not at all what I expected. The signal looks really noisy

I am thinking maybe I could cut out the micro processor and just bridge the TX and RX traces but I'm not sure it would work

[crazifunguy]

I decided to desolder the ballast controller from the actual ballast. I got some high rez scans of both sides. Unfortunatley it is a 4 layer circuit board. I am going to try to trace the circuit paths and create a block diagram from input to output. This might take a few days to complete.

Front Side


Back Side

[crazifunguy]

I did some circuit tracing. This thing is pretty high tech for just a ballast. Theres a whole lot going on with this little pain in the butt.

Front

Back


I am wondering if it woudl be easier to buy a programmer for the microchip and try to redo the internal program eliminating ballast functions.

[crazifunguy]

Well I managed to get a copy of the "Holy Grail" for this projector....AKA the service manual. Its funny that you were talking about using a PIC to communicate with the ballast. This projector uses a PIC to control all ballast functions.

After studying the schematics and trouble shooting. I figured that the Lamp_Err signal coming out of the PIC was the place to start. This signal is around 70mV and jumps to 5V when a lamp error is detected. This signal is sent through a buffer IC then to the main processor.

http://www.fairchildsemi.com/ds/74%2F74LCX541.pdf

By cutting the trace from the pic to the buffer and connecting the buffer pin to the buffer ground the projector stays on....

SUCCESS!! So far anyways. I havent tested for more than 2 minutes. I dont know what repeated errors will do to the pic or any other part of the circuit.

[jan.didden]

Originally Posted by crazifunguy

[snip]I am wondering if it woudl be easier to buy a programmer for the microchip and try to redo the internal program eliminating ballast functions.

Don't if your are new to ucontroller programming. UYou can do it but the need to learn a whole new trade costs you a lot of time.

Better to find someone who is good at it already.

And don't forget you need to figure out the whole control logic structure first!

jd

[jan.didden]

Originally Posted by crazifunguy

Well I managed to get a copy of the "Holy Grail" for this projector....AKA the service manual. Its funny that you were talking about using a PIC to communicate with the ballast. This projector uses a PIC to control all ballast functions.

After studying the schematics and trouble shooting. I figured that the Lamp_Err signal coming out of the PIC was the place to start. This signal is around 70mV and jumps to 5V when a lamp error is detected. This signal is sent through a buffer IC then to the main processor.

http://www.fairchildsemi.com/ds/74%2F74LCX541.pdf

By cutting the trace from the pic to the buffer and connecting the buffer pin to the buffer ground the projector stays on....

SUCCESS!! So far anyways. I havent tested for more than 2 minutes. I dont know what repeated errors will do to the pic or any other part of the circuit.

Well that seems a good solution!

jd

[Andrew Eckhardt]

I just wanted to mention, if you're not careful, with that much of the plate through hole stuck to the ballast pins and a four layer board you could easily wind up with some discrepancies between what you see in the drawing and what's actually wired on the main board..

[crazifunguy]

Originally Posted by Andrew Eckhardt

I just wanted to mention, if you're not careful, with that much of the plate through hole stuck to the ballast pins and a four layer board you could easily wind up with some discrepancies between what you see in the drawing and what's actually wired on the main board..

Luckily the only connections made are on the bottom of the ballast board. That board is only 2 layer.

Oddly enough these ballasts do not have a set wattage rating. The wattage is variable down to 0% now if through the service menu if the power output is lowered down to say 15% would the ballast still function and would it power a 35w HID lamp and not explode?