[Coco] Cococompatible monitors...

[John Kent]

Hi Gene,

On 3/07/2011 11:37 PM, gene heskett wrote:
> On Sunday, July 03, 2011 09:13:16 AM John Kent did opine:
>
> The gime's pallette is somehow extracted from a 2 bit d/a on each color,
> for a 4 level per color lashup.  Now if this FPGA in a Spartan 3 could just
> drop into the gime socket, I could get interested.  Real interested.
> Running a 63C09, and a 2 meg disto memory kit, all on an old AT power
> supply, I have nearly zero heat, the gime is the warmest chip in it, and
> the power use in that socket could go up to several watts with a suitable
> heat sink.
>

The Spartan 3 would not be a direct drop in. It would need to be mounted 
on a daughter board with some sort of pin header arrangement for the 
GIME chip. It would also need on board 5v to 3.3V bus switches to make 
it 5V compatible as well as some regulators, bypass caps and so on and 
perhaps an oscillator chip if the GIME has a crystal oscillator circuit 
built in.

Power consumption and heat dissipation shouldn't be a problem at the 
sort of clock rates the GIME chip is running at. Most low cost FPGA 
boards running with a 50MHz clock don't require cooling, in fact the 
FPGAs in the systems I have designed barely get warm. The entire DE1 
board can be powered by a USB socket, i.e. uses less than 500mA or less 
than 2.5W and that includes all the peripheral chips such as SDRAM, 
SRAM, Flash, RS232 and audio codecs and so on. Most of this is running 
off 3.3V so about 1/3 of the power is dissipated in the regulators.

I'm not sure about the GIME chip, but the 6809 I think was originally NMOS.

There is a Spartan 3 power calculator here:

http://www.origin.xilinx.com/cgi-bin/power_tool/power_Spartan3

The Spartan 3 runs off 1.2V, 2.5V and 3.3V by the looks of it. Current 
consumption at a guess would be under 100mA depending on what clock 
rates the logic was working at. Not all of it would be working at 25MHz. 
You need to synthesize the design to get an accurate estimate of the 
power consumption.

John.

> I am well aware of the limitations imposed by the coco's OEM power supply
> and regulator, which could not support that power usage increase without
> active cooling.  That however is also adjustable, I have done it a couple
> of times now.  That very inefficient analog design is 90% of the emitted
> heat in a box stock coco.  It can drive a small 12 volt dc fan from the
> output of the rectifier, and this is a huge help in maintaining reasonable
> internal temps.  Whether that additional 150 ma is still within the design
> envelope of that transformer I don't know, but I have a coco2 and a coco3
> equipt with fans, and the coco2 ran 24/7 for 13 years, with 2 drive
> controller failures.  The lack of a 3 pin power plug on the drive boxes and
> the coco2 meant that when the studio tower was being used as a lightning
> rod by mother nature, the resultant EMP surges were pretty hard on the
> controller.  I eventually married them all together with copper braid,
> grounded to the rack.
>
>
> Cheers, gene

-- 
http://www.johnkent.com.au
http://members.optusnet.com.au/jekent