Would you look at that, he's got one of the mysterious K's.

...But it looks to me like the version K is what is to blame here. K's just don't follow the rules when it comes to making calculators.

Little did I know that some BE's are more basic than others.  It seems I have been cursed with the elusive 'K'.  Darn.  Here are some other funky things about my calc:

• I need ALCDFIX to play some games.  Something to do with a slow driver
• Omnicalc's Ram Recovery function doesn't work.  Something to do with missing memory
• I got it a Wal-Mart for 80 bucks.  Either it was priced wrong, or Wal-Mart knew something I didn't.

Personally, I don't think that the last letter of K's even means anything. Like I've seen version O, which isn't even out yet.

I need ALCDFIX to play some games.  Something to do with a slow driver

zStart does this too :).  Mine has a bad screen and the delay is 44. (44/4 = 11 clock cycles added). Whereas good screens, like wabbitemu have delays of 12.

Yeah, the Ks follow no pattern.  For example, my friend that has a K has the serial number 1142038377_K-0906B (Yes, that's an underscore) yet calcs bought at the same time are like 2365063662 S-0207G.  It's completely different!  As you can see, K's are weird.

Weird, if you see those calcs again be sure to note the serial.

Well since this topic was originally about 16 level grayscale, let me share some facts that I just figured out.

I thought, well maybe I can get 64 levels out of it by putting one pixels on in the first column, two in the second, three in the third, and so on until the 64th column is completely full. I then changed the z-address to scroll the screen 100,000 times per second. (For those who don't know, the z-address is where the lcd driver starts to display at, so if you set it to 1, the very top row of pixels appears at the bottom and all others scroll up one, a 2 does the same except by two pixels.) The z-addressing is nice because it only takes one command to update the entire lcd. But this didn't work at all and I got these weird scrolling patterns. So then I thought, I'll figure out the refresh rate.

I made a refresh rate program. Through careful tweaking I found that at 3,811 z-addresses per second made a perfectly still image. So every 1/3,811 of a second, the driver displays another line. Divide this by 64 and the screen refreshes at 59.5 hz.

What does this mean, this means that if your grayscale program updates the lcd faster than 59.5 hz, it won't make a difference. My 16 level just happens to update it 120 times per second. I suppose this means that within a few hours I will have an 8 level grayscale viewer instead.

I included an example to show you how perfect the grayscale looks when you update the lcd at the right speed. This is a weird version of tetrisM I am working on. If your driver updates at 59.5 hz, you will see 100% flickerless grayscale with a little slant the scrolls through the screen.

I'm not sure it hold true for all screens, I would assume that it is different for the good and bad drivers. And maybe it's even different within the driver types.

If it's the first case, it's really easy to figure it out, just a quick lcd delay test will tell you, but if it's the second, to be used in games you would have to have it user adjusted.

All I know is that the way I have it adjusted, it looks perfect on my calculator.

I'm not sure, I'm currently writing an 8 level grayscaler. Doing it that way, there wouldn't be a diagonal line, I just wonder what the cross over point would look like then. Maybe there wouldn't be one if you're lucky.

Also, I think you would have to start at the bottom and work up, because I feel that's how the lcd is updated. But I'm not positive.