Atari 2600 Alarm Clock Calendar

[SeaGtGruff]

I had posted a (horribly unfinished and somewhat buggy) clock program in the Atari 2600 forum, but I figured I should move any further postings to this forum. For anyone who's interested, the original thread is here:

 

http://www.atariage.com/forums/index.php?s...c=99473&hl=

 

I'm not quite ready to post the next (still horribly unfinished but less buggy) version of my program, but in the meantime I have a few comments and questions that I'm hoping someone can help with.

 

(1) (This is really a comment rather than a question, but any replies relevant to this issue would be greatly appreciated.) I know the PAL Atari 2600 has a different clock rate than the NTSC Atari 2600. From my readings about TV and computer clock rates (especially NTSC Atari computers versus PAL Atari computers), I expect that the PAL Atari 2600's color clock rate is equal to four-fifths of the PAL color frequency (i.e., 4433618.75 Hertz times 4/5), or 3546895 Hertz. I think the spec sheets show it as 3546894 Hertz, but I suspect that the difference may be due to dropping the fractional portion of the numbers before and after the calculation, since it seems to be very common to round or truncate Hertz values when discussing TV and computer frequencies (i.e., 4433618.75 truncates to 4433618; 4433618 * 4 / 5 is 3546894.4; and 3546894.4 truncates to 3546894). On the other hand, 3546895 is *not* a multiple of 3, whereas 3546894 *is* a multiple of 3; and since the color clock rate is equal to the CPU clock rate times 3, perhaps 3546894 is actually correct after all. But on the *other* other hand, since crystals don't *have* to generate integer Hertz frequencies, there's no reason why the 3546895 value can't be correct. This may seem rather unimportant, but that tiny difference matters to my clock program, because it changes the exact mixed number which represents the PAL Atari 2600's precise frame rate. However, I'm not going to worry about that issue right now, except to say that it will be interesting to let the final program run on a PAL Atari 2600 for an extended period of time, and see whether it keeps perfect time or gets off by a second after a while.

 

(2) For the time being, I've decided to stick to a 312-line display for PAL Atari 2600s; but since I plan to let the user switch between 262 lines and 312 lines using the "TV TYPE" console switch, there's no reason why a PAL user couldn't run the program with a 262-line display (i.e., PAL/60). But that raises an interesting question about PAL/60. Since PAL/60 carts are run on PAL Atari 2600s, and since PAL Atari 2600s have a different clock rate than NTSC Atari 2600s, it follows that a PAL/60 display must have a slightly different frame rate than an NTSC/60 display. Presumably, this difference won't be great enough to matter for most programs, but it *will* matter to my clock program. I could easily add support for a PAL/60 display-- and I may eventually do that (because I think it would be pretty cool! )-- but I'm wondering if anyone has ever pondered or wrangled with this PAL/60 frame rate issue before?

 

(3) One of the things I'm hoping to eventually do is include the option of displaying a dial clock face, rather than just a digital time display. I'd be using the players to draw the clock dial, so it's going to depend on whether I can draw the hours, minutes, and seconds hands. Anyway, assuming that I do eventually include a dial clock face, I want the dial to be as close to a circle as possible, which raises the issue of the aspect ratio of the pixels on the different Atari 2600s. For an NTSC Atari 2600, I've been assuming that a 160x200 NTSC Atari 2600 screen has a 4:3 ratio, the same as a 640x480 or 800x600 computer screen. Thus, whenever I design a mockup of an Atari 2600 screen in a graphics program, I use a picture size of 800x576, and assume that the aspect ratio of the NTSC Atari 2600's pixels is 5x3 (i.e., 800 / 160 = 5, and 600 / 200 = 3, and 192 x 3 = 576). But is that correct? What is the actual aspect ratio of the pixels on an NTSC Atari 2600? I can create a square test display and try it on a real Atari, and measure to see whether the square is actually square, but presumably the results might vary a bit depending on the TV set? And if that's the case, then I suppose I could just go with a presumed 5x3 pixel aspect ratio and not worry about it. But does anyone know the actual aspect ratio of the pixels on an NTSC Atari 2600?

 

(4) What about the aspect ratio of the pixels on a PAL Atari 2600? I think the standard PAL TV picture has 575 active scan lines, so presumably 575 interlaced PAL/50 scan lines have the same height as 483 interlaced NTSC/60 scan lines? Is that correct? Does anyone know the actual aspect ratio of the pixels on a PAL Atari 2600? I should note that this is *not* the same as the aspect ratio of PAL/50 TV pixels, since the PAL Atari 2600 has 228 color clocks per scan line, whereas a standard PAL TV signal has 283.7516 color subcarrier cycles per scan line.

 

(5) What about the aspect ratio of the pixels on a PAL/60 display? Are they the same as the aspect ratio of the pixels on an NTSC Atari 2600, or does the slight difference in the frame rates have any kind of impact on the aspect ratio of the pixels?

 

(6) If I remember correctly, I once read in the AtariAge forums that a SECAM Atari 2600 is basically a PAL Atari 2600 with the "TV TYPE" console switch permanently wired to the "B-W" setting. Is that correct? Or more to the point, does the SECAM Atari 2600 have the same color clock rate and CPU clock rate as the PAL Atari 2600?

 

I know there's no need to worry about the aspect ratio of the pixels on the different Atari 2600s, but I'm interested in those sorts of details, so I'd like to know just for the sake of knowing. Of greater importance for my clock program is the issue of the exact frame rates on the PAL and SECAM Atari 2600s. I can calculate the exact PAL frame rates for either 3546895 or 3546894 Hertz (depending on which of those values is the correct one), but I'm not sure if either of those values is correct for the SECAM Atari 2600's color clock frequency.

 

Thanks in advance to anyone who can provide answers or comments.

 

Michael

[Cybergoth]

Chris Cracknell (sp?) released a 2600 video clock some 5 or 6 years ago, which was pretty cool and sold through Hozer. There should be severeral threads in the [stella] archive discussing accurate timing of such a clock.

[alex_79]

These are the two oscillators on my PAL light sixer board. I hope this can help you.

 

3546.894 KHz (Y200)

 

 

4433.618 KC (Y201)

[SeaGtGruff]

Chris Cracknell (sp?) released a 2600 video clock some 5 or 6 years ago, which was pretty cool and sold through Hozer. There should be severeral threads in the [stella] archive discussing accurate timing of such a clock.

I downloaded the clock-v3.bin from the Stella archive, and I read the posts in that thread, but the program runs with 272 lines per frame, and the thread didn't really reveal very much, other than a brief reference to adding 1 second after a certain period of time. I'll search to see if there are other threads wih more information.

 

Michael

[Cybergoth]

Hm... I thought there was more... but it's been a while

[SeaGtGruff]

These are the two oscillators on my PAL light sixer board. I hope this can help you.

 

3546.894 KHz (Y200)

 

 

4433.618 KC (Y201)

Thank you for posting these! They confirm the numbers I saw on the spec sheets-- which I can't seem to find now. Can anyone post a link to the spec sheets? Are there spec sheets for the SECAM Atari 2600?

 

Michael