Digital Ballast Bypass Expirimenting

[crazifunguy]

Tonight I made some headway on bypassing newer digital ballasts that require a signal rather than a simple short.

I got out my cheap osciliscope and took some measurements. It was interesting to find that on the projector sends a signal to the ballast then it seems like it waits for a reply. There is only 1 optocoupler sending back to the projector so I assume that this is the needed signal. The other 2 optocouplers i'm assuming are for the lamp enable and the input signal.

This projector has 5 wires going to the ballast.
1 - Signal Back to Projector
2 - Ground
3 - +5v
4 - Square Wave
5 - Signal From Projector

I measured the out and in signals and took some snap shots. The signal going to the ballast starts at 5v and drops to around 3v. The signal returning from the ballast starts at 0V and goes to 5V. The signals appear to be the same pulse width just offset and inverted.

Low Power Mode


Med Power Mode


Full Power Mode






The microcontroller used in the ballast is this 1 here
http://www.nxp.com/documents/data_sheet/P87LPC769.pdf

The signal input to the ballast goes into pin 11 which is a serial input

The signal from the ballast goes to pin 12 which is serial output

The square wave goes into pin 8 External Interupt.

This ballast design for sure uses some sort of serial communication. I am positive that this complicates the bypass process greatly. For now I will do some reading and see what I can come up with.

[jan.didden]

Indeed looks like a serial comms. Assuming a standard 8-bit control or command word, the last shots looks like a 10101011 data word.
The forward direction would be the command, the return would be the actual setting, (which of course when all went well would be identical to the command).
Educated guessing.
You can easily figure out the data values for the other modes.

jd

[crazifunguy]

Can this communication be emulated or would a similar microcontroller with serial com be needed? Its a challenge now. I need to prove that we are smarter than the ballast...hahaha


I dont know much about serial communication. I am reading that a start bit is sent then the 8 bits then a stop bit or bits. This would explain the return signal waiting for the complete input signal.

Looking closer at the data sheet, I would say its a very accurate guess that the Serial Port is running in mode 0 with external timing.

My guesses for the other modes are
Low
10 00 01 00 00
Med
10 01 01 11 or 10 00 10 11

[jan.didden]

Originally Posted by crazifunguy

Can this communication be emulated or would a similar microcontroller with serial com be needed? Its a challenge now. I need to prove that we are smarter than the ballast...hahaha


I dont know much about serial communication. I am reading that a start bit is sent then the 8 bits then a stop bit or bits. This would explain the return signal waiting for the complete input signal.

Looking closer at the data sheet, I would say its a very accurate guess that the Serial Port is running in mode 0 with external timing.

My guesses for the other modes are
Low
10 00 01 00 00
Med
10 01 01 11 or 10 00 10 11

You can hang your scope's 2nd probe on a clock signal so you see the bit transistions clearly. You value for low has a few bits too many

It's simple to build your own controller and play with the values and resulting brightness etc, using for instance a PIC microcontroller. You don't need to worry about start-stop bits and the like, that's all low level stuff handled by the micro and the development system (which, for the PIC is free). BUT you need to be somewhat experienced in this.....

jd

[crazifunguy]

The brightness is actually controlled by the remote and the software on the projector. The different signals are sent for each mode to the ballast. I am trying to eliminate the ballast completly and just feed the proper message back to the projector so it thinks the ballast is installed and the lamp is working.

[crazifunguy]

Heres a question....Would a PIC be able to accept the incomming serial message then send the correct message back to the projector main board allowing the ballast to be removed? I know it would take some programming just wondering if the device is capable

[jan.didden]

Originally Posted by crazifunguy

Heres a question....Would a PIC be able to accept the incomming serial message then send the correct message back to the projector main board allowing the ballast to be removed? I know it would take some programming just wondering if the device is capable

Of course it is. This is 2-fingers-in-your-nose stuff. ANY PIC even the $ 0.85 12-series in a DIL08 can do that. Take a look at the microchip site.

I'm not familiar with this projector stuff. How does this work? You send a 'light level' command to the projector, then what happens? How does the lamp and supply and ballast work together?

jd

[crazifunguy]

The projector has a main controller board and the ballast for the lamp. The main board controls all functions of the projector. In most projectors the main board sends a signal to enable the ballast and power on. Then the ballast sends a signal back to the projector that the lamp is lit and working. Usually by pulling the lamp lit signal low (grounding) most projectors are happy and will stay running with no errors. This is not the case with newer projectors that use serial communication.

This Eiki projector essentially works the same as most but uses serial communication.

The brightness modes are selected with the remote. The main board then alters the signal going to the ballast and the ballast responds accordingly.

The main goal is to remove the ballast and have the projector run giving no errors thinking that the ballast is still installed and the lamp is functioning. This requires requires:
1 - Recieving the signal from the main board.
2 - Processing the signal
3 - Returning the correct signal

I will post some schematics with more detailed scope readings. All signals to and from the ballast are isolated with opto-couplers. I measured the signal at the TX and RX pins of the microchip. I will post some signals before and after the opto-couplers for comparison on both input and output

Also...Please note that I appreciate you accepting the challenge to understand this and try to help figure this out. It will help me and many others continue to modify newer projectors.

[jan.didden]

OK.... But I assume the ballast is there for a purpose. So removing it would that not give any problems with the lamp lifetime or startup or brilliance stability or whatever? Just trying to wrap my head around it.

jd

[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