Quote:

Does it really need access to the video chip to manage more than just doubling the refresh rate?

Yes, this flickerfixer requires access to the Lisa chip, as it's "listening" to the inter-chip communication between Alice and Lisa. This is using knowledge that we have gained during the (still ongoing) development of Clone-A, and it's going a little bit beyond what other flickerfixers have done before. There are a few parameters that the CPU can set inside the FPGA, and that is also going through the chip registers: We've added a few bits in unused places of the AGA chipset and protected them from accidential writing with a fairly complicated activation procedure.

Setting parameters, accessing the flash chip (which contains the hardware behaviour data of the FPGA) and detecting all those different screenmodes would not be possible if we didn't have access to the (otherwise mostly mysterious) chip-register bus.

Quote:

Wouldn't it be possible to have a "scan tripler" that just plugged into a video output?

Triple frequency would be OK for PAL, but if you triple an NTSC screen, you end up with 90Hz. For broad compatibility, you should stay in the 60-85Hz range, where most affordable monitors are working. We're particularly aiming at those people who either still have a CRT somewhere in the basement, or who need to replace a 1084 monitor with something of "about the same size" due to space constraints on the table.

In order to stay within a "most compatible range", we're multiplying with a fractional factor of 2.5. Yes, this is a bit more complicated than multiplying with an integer, but it makes things a lot easier for the customer. You get the idea: The keyword is "customer value": I want to provide something that continues to be valueable (as opposed to most other computer hardware).

CRTs are traded for about a buck an inch these days, and this is what Indivision AGA wants to be compatible with: Spend money on something that does not wear out (the flickerfixer!), and save big time on the thing that needs to be replaced every other year.

For an external version, it would be hard to maintain good picture quality. AGA has 24 bits colour resolution (three channels of 8 bits each). If this is converted to analog, there's noise in the output that's hard to compensate for. The only possibility would be "brute force": Nothing less than a 10-bit AD-converter would come anywhere close to the picture quality of Indivision AGA. Now multiply this by three, because you need to sample red, green and blue. The whole design would require full 30-bit operation from input to framebuffer to output. It would be *a*lot* more expensive just because of the electrical requirements and the required additional logic. In addition to that, you need to pay for a case, cabling (23-pin Sub-D's anyone?), and it'll be quite hard to achieve the "no adjustments necessary" feature.

As you can see, I have given an external unit quite some thought, but I have dropped it after making the first calculations. There might be a few people out there willing to pay 950 USD for an external flickerfixer, but the broad average would either hesitate or skip on a product for 200,- EUR (look at the exchange rate: that's 311 USD today!). An internal unit makes the thing affordable for a lot more people.

Jens

Add me to your list of customers for an A4000 model when it is available.

_________________
AmigaOne X1000, A4000(T), A3000, A2000, A1200(T), A1200, A500, CD32, Minimig+ARM, FPGA Arcade, Chameleon64, C-One, C128, C128D, C64C, C64, VIC-20, CBM 8032, CBM4032, Efika, Ultimate64

I only need to know one thing.

I need to have an composite video output, because my TFT monitor does not support low resolutions, I explain, my monitor accept 15 khz frecuencies, but if I select 640x256 it is not showed, if I select 640x512 at 15 khz, it is showed correctly. So... Will this flicker fixer have a composite output too?

Thank you.