Having a 4 channel oscilloscope makes tinkering so much easier. Now I can look at four incomprehensible and unexplainable voltages at the same time!

I have a Sun/Sony GDM-5410 monitor at work. It's a 21-inch Trinitron boat anchor, a solid performer that throws me images at a resolution of 1504x1128 and more than 82 times a second. I'm very happy with it, it helps making work bearable.

Unfortunately that monitor has recently developed a problem with the brightness settings: it's way too bright, even with the controls down at zero.

As it turns out, drift of this parameter is a common problem with this series of (microcontroller-controlled) monitor. There's two fixes: soldering in different resistors (yuck, inelegant, nasty and not exactly cool to do, working close to Quite High Voltage), or using Sony's adjustment program called Windas (for Digital Alignment System). Method two it must be!

Windas is available out there, and all one needs is an RS-232 to TTL converter cable. Which I promptly soldered up: some breadboard, caps, connectors and a MAX232 are all you need. Cool chip, that 232.


The cable/converter setup looks like nothing and it didn't work. First, I couldn't convince Wine to run all of the sodding Windas thing (some of it seems to be cooked up with a hot needle and a bad dose of crack). Ok, thinks I, I have an old lapdog, I can run up a Vmware with some Windows for Windas.

Only it still didn't work...dammit! Obviously it was time to Have A Closer Look, setting up the debugging-ICU: TEK 2246, multimeter, solderless breadboard and lots of jumpers, some source of 5V, lapdog.


The initial test confirmed my suspicions that RS-232 sucks particularly badly on laptops: mine makes -13V nicely, but the "positive" voltage reaches only -0.38V. The standard says +12V, and the MAX wants to see at least +1.5V or so. Something like that a multimeter can't tell you, because even at 300bps the transitions are way too fast...and the negative state is the idle state.

After some head-scratching I realized that fixing this isn't as hard as I had thought initially (think: op-amp, voltage booster and so on). All that is required is some positive voltage shift: with -13V there is lots of headspace until the MAX would stop detecting the negative state and all I need would be a volt or two for the positive state.

I am not well-versed in electronics. It took me a while to figure out that all I needed to do was to shunt some little current from my 5V source into the RS-232 signal to offset it upwards. A bit more testing with a pot, watching two channels simultaneously and with the multimeter showing the current draw, I found out that 400ohm between 5V and the signal gives me peaks of -8 and +1.8V, at the cost of an extra 15mA. The MAX is happy with those voltages, and I'm happy and hopeful: tomorrow I'll lug the lapdog to work and give fixing the monitor a go.

If it still doesn't work, I'll go poke it with a stick; if that fails I'll smack it with the stick: I believe electronics work best when they have reason to fear you. (But don't start anthropomorphizing computers: they hate that.)

[ published on Mon 09.07.2007 21:15 | filed in interests/tinkering | ]
Debian Silver Server
© Alexander Zangerl