TSL1401 Linescan Imaging Daughterboard

John Abshier

I have has some success today using the TSL1401 with a ADC0831 all mounted on a Propeller DemoBoard.· The program waits for a keypress from Parallax Serial Terminal(PST) and then spits the 128 values to PST.· The camera is operating in one shot mode.· I then copied the values and pasted them into Excell to visually see the one line image.· This code is almost a literal translation of the BS2 code in the TSL1401 documentation.· The program is 100% Spin.· My first value was always 255.· An extra clock pulse to the TSL1401 solved that problem.· But I don't know why I had to do that.· The last six values are flaky, too low.· I don't know if it is my code or the camera.· I considered continuous acquisition, but today I don't have a way to visualize the results so I did the one shot.

John Abshier

Phil Pilgrim (PhiPi)

John,

I haven't tried this yet, but I think the reason your first pixel is always 255 is because you haven't clocked all the pixels out during the exposure. Your repeat 250 should be repeat 256 just to be on the safe side.

As to the droop at the end, this is probalby due to the long readout time: serial ADC combined with Spin's slowness. There's an explanatioon in the docs about pixel droop. Basically it's due to the internal charge storage getting discharged before it has a chance to output.

-Phil

John Abshier

I thought I saw pixel droop when I had the ADC as a function. I moved the ADC inline and that made everything better except the last 6. I know that this needs to be implemented in assembly. I still suffer from having to program a RCA mainframe in machine language using absolute addresses--no assemblier; I avoid assembly whenever I can. I will work on it some more when I get back from UPENE.

John Abshier

Phil Pilgrim (PhiPi)

I really need to write an object for this. There's a Prop board coming out soon that the TSL1401-DB plugs into, but it uses the Prop's delta-sigma ADC, rather than a dedicated chip. Stay tuned...

-Phil