Searching \ for '[PIC][SX] Video overlay replacements' in subject line. ()
Make payments with PayPal - it's fast, free and secure! Help us get a faster server
FAQ page: techref.massmind.org/techref/ubicom/lib/io/index.htm?key=video
Search entire site for: '[SX] Video overlay replacements'.

Exact match. Not showing close matches.
PICList Thread
'[PIC][SX] Video overlay replacements'
2006\08\18@023516 by William Chops Westfield

face picon face

On Aug 17, 2006, at 5:55 PM, James Newtons Massmind wrote:

>> replace the ST Video overlay chips that are used in
>> police cars everywhere. ST has dropped them. They have to
>> paint standard 8 x 10 pixel characters, 26 characters per
>> line, 12 lines. Only BOBIII can do it at the moment.
>
> 26 x 12 is 312 so you would need external RAM. The SX's
> have 262 bytes at most.

If I understand Bob's application, he doesn't need 312 characters,
he needs smaller numbers of characters in the corners of a video
screen, sized and placed in spots at the corners of the screen,
in spaces corresponding to the 26x12 matrix.

What about a chip/circuit that does that explicitly? 8x1 in each
of the four corners or similar?  Speed limitations are different
than memory limitations...  Or does it need to be a drop-in
replacement for the old ST part?

BillW

2006\08\18@055841 by William Chops Westfield

face picon face

On Aug 17, 2006, at 11:35 PM, William Chops Westfield wrote:

> What about a chip/circuit that does that explicitly? 8x1 in each
> of the four corners or similar?

Or maybe up to 4 lines of 26 chars, but with each line having a
settable scanline to start at?

BillW

2006\08\18@111941 by Bob Axtell

face picon face
William Chops Westfield wrote:
{Quote hidden}

Unfortunately, every police setup is unique. Been down that path before.
They want to
put it ANYWHERE. Not only that, but full-screen text is used to setup
the car video
system.

--Bob

2006\08\18@120427 by Alan B. Pearce

face picon face
>>> 26 x 12 is 312 so you would need external RAM. The SX's
>>> have 262 bytes at most.
>>>
>>
>> If I understand Bob's application, he doesn't need 312 characters,
>> he needs smaller numbers of characters in the corners of a video
>> screen, sized and placed in spots at the corners of the screen,
>> in spaces corresponding to the 26x12 matrix.
>>
>> What about a chip/circuit that does that explicitly? 8x1 in each
>> of the four corners or similar?  Speed limitations are different
>> than memory limitations...  Or does it need to be a drop-in
>> replacement for the old ST part?
>>
>> BillW
>>
>Unfortunately, every police setup is unique. Been down that path
>before. They want to put it ANYWHERE. Not only that, but full-screen
>text is used to setup the car video system.

So, I have suggested before using a dsPic as a high speed PIC to generate
video. Use a colour burst crystal (or preferably 4x colour burst freq.) as
the chip oscillator and use the PLL multiplier to step up as fast as you can
go - you are in a car so current draw isn't a problem (relatively), and the
dsPic should have enough RAM for a character buffer, and the chip is
probably fast enough internally to get the dots from a character table on
the fly. I suspect that judicious use of the CCP input could probably
genlock the data stream to an external video source, and would be looking at
using the MSSP to send the dot stream, clocked from one of the timers at the
dot clock rate.

There you are, all designed without looking at the data sheet - now probably
find the MSSP won't clock fast enough for the dot clock. ;)) One day I'll
get to try it out.

As an alternative a PIC24 or PIC33 so you are not paying for the DSP part of
a dsPIC may be a cheaper alternative for production once you have the
original up and running.

2006\08\18@122319 by David VanHorn

picon face
This sort of thing used to be done by 1 MHz 6502 systems, using a rom, and
only doing non-video stuff during the sync pulses.

2006\08\18@124031 by Maarten Hofman

face picon face
2006/8/18, David VanHorn <spam_OUTdvanhornTakeThisOuTspammicrobrix.com>:
>
> This sort of thing used to be done by 1 MHz 6502 systems, using a rom, and
> only doing non-video stuff during the sync pulses.


I believe the 1 MHz video system you might be referring to actually had a so
called "VIC" which handled the video part. The only thing the processor did
was share the memory with it occasionally. The only system old system that I
recall that did video stuff on the main processor used a Z80A at 4.77 MHz
and had a special logic chip that aided in generating the (black and white)
video output. It had a "FAST" option, that disabled the video processing,
causing the system to run 5x faster than normal, but without showing
anything, obviously.

Greetings,
Maarten Hofman.

2006\08\18@131014 by David VanHorn

picon face
> I believe the 1 MHz video system you might be referring to actually had a
> so
> called "VIC" which handled the video part. The only thing the processor
> did
> was share the memory with it occasionally.



My OSI-C1P had no such chip.



The only system old system that I
> recall that did video stuff on the main processor used a Z80A at 4.77 MHz
> and had a special logic chip that aided in generating the (black and
> white)
> video output. It had a "FAST" option, that disabled the video processing,
> causing the system to run 5x faster than normal, but without showing
> anything, obviously.


Sounds like a ZX-81



Don Lancaster's TV Typewriter books have the basics.
Overlaying is a bit more complicated, as you have to sync up with the
incoming video, but not THAT much worse.

2006\08\18@154222 by William Chops Westfield

face picon face

>> I believe the 1 MHz video system you might be referring to
>> actually had a so called "VIC" which handled the video part.

The 1MHz video systems had a shift-register external to the
CPU.  The XX MHz scenix chips and cute hacks for PICs are
doing even the serialization in software...

BillW

2006\08\18@160029 by Maarten Hofman

face picon face
>
>
> The 1MHz video systems had a shift-register external to the
> CPU.  The XX MHz scenix chips and cute hacks for PICs are
> doing even the serialization in software...


Well, the specifications that I read from the OSI-CxP seemed to indicate
that they had an external character ROM and a video RAM. It would count
through the video RAM and look up the corresponding values in ROM and then
shift the data out to video. I'm not sure whether the 6502 did more than
just put the correct values in the video RAM, but I might of course also be
looking at the wrong schematics. The so-called "VDU" of the OSI-CxP seems to
have up to 15 chips associated with it in the schematics I saw. A similar
system could indeed be done using PICmicro, but I enjoy the one chip
solution I made myself, despite its limited abilities.

Greetings,
Maarten Hofman.

2006\08\19@165516 by Mike Harrison

flavicon
face
On Fri, 18 Aug 2006 17:04:11 +0100, you wrote:

>>>> 26 x 12 is 312 so you would need external RAM. The SX's
>>>> have 262 bytes at most.
>>>>
>>>
>>> If I understand Bob's application, he doesn't need 312 characters,
>>> he needs smaller numbers of characters in the corners of a video
>>> screen, sized and placed in spots at the corners of the screen,
>>> in spaces corresponding to the 26x12 matrix.
>>>
>>> What about a chip/circuit that does that explicitly? 8x1 in each
>>> of the four corners or similar?  Speed limitations are different
>>> than memory limitations...  Or does it need to be a drop-in
>>> replacement for the old ST part?
>>>
>>> BillW
>>>

I'd suggest taking a hard look at the Philips LPC2101 ARM -  $2 (100x), 70MIPs, has 2 SPI/SSP ports,
one of which has a fifo, and also variable word length, so you could for example use 10 bits to do
8x8 chars with spacing, and as long as you keep the fifo full, it ought to come out in a solid line
of pixels.
IAR do a free version of their C/Asm devtool with 32K limit, which should be plenty.

2006\08\19@204700 by Bob Axtell

face picon face
Mike Harrison wrote:
>
> I'd suggest taking a hard look at the Philips LPC2101 ARM -  $2 (100x), 70MIPs, has 2 SPI/SSP ports,
> one of which has a fifo, and also variable word length, so you could for example use 10 bits to do
> 8x8 chars with spacing, and as long as you keep the fifo full, it ought to come out in a solid line
> of pixels.
> IAR do a free version of their C/Asm devtool with 32K limit, which should be plenty.
>
>  
Thanks, Mike. I will.

--Bob

2006\08\19@204949 by John Chung

picon face
Does Philips LPC have similar devtools like Microchip?
How are their pricing like?

John

--- Mike Harrison <.....mikeKILLspamspam@spam@whitewing.co.uk> wrote:

{Quote hidden}

> --

2006\08\20@011616 by William Chops Westfield

face picon face

>> I'd suggest taking a hard look at the Philips
>> LPC2101 ARM -  $2 (100x), 70MIPs,

Dual voltage supplies :-(

They do some sort of weird interleaving on the flash to get
"full speed"; that may make it tricky to get deterministic
timing for something like video...

BillW

2006\08\20@023458 by Wouter van Ooijen

face picon face
> They do some sort of weird interleaving on the flash to get
> "full speed"; that may make it tricky to get deterministic
> timing for something like video...

Timing-critical code could be run from RAM. Or use the thumb
instructions, I think that halves the amount of bits that must be
fetched from FLASH (but I have not verified this).

Wouter van Ooijen

-- -------------------------------------------
Van Ooijen Technische Informatica: http://www.voti.nl
consultancy, development, PICmicro products
docent Hogeschool van Utrecht: http://www.voti.nl/hvu


2006\08\20@051342 by Mike Harrison

flavicon
face
On Sat, 19 Aug 2006 22:16:08 -0700, you wrote:

>
>>> I'd suggest taking a hard look at the Philips
>>> LPC2101 ARM -  $2 (100x), 70MIPs,
>
>Dual voltage supplies :-(

1.8v regulator - 20-30c  I think - not exactly a major problem. 1.8v is only for the core, IO is
3.3v

>They do some sort of weird interleaving on the flash to get
>"full speed"; that may make it tricky to get deterministic
>timing for something like video...

I've not played with it myself yet but I think it's basically just 128 bits wide & muxed down using
a few tricks - I'm fairly sure it's deterministic.
However you can also run code from RAM if you want.


2006\08\20@051751 by Mike Harrison

flavicon
face
On Sat, 19 Aug 2006 17:49:48 -0700 (PDT), you wrote:

>Does Philips LPC have similar devtools like Microchip?
>How are their pricing like?

The chip has on-chip debug facilities that are fairly close to a proper ICE, with trace etc. i.e.
closer to mplab-ice2000  than ICD. This is accessed via JTAG and there are a number of JTAG
interfaces out there. The IAR software includes support for this BTW.

There is a fairly active list called LPC2000 at yahoogroups for the Philips range.  

{Quote hidden}

>> --

2006\08\20@095935 by Byron A Jeff

face picon face
On Sat, Aug 19, 2006 at 09:59:55PM +0100, Mike Harrison wrote:

> I'd suggest taking a hard look at the Philips LPC2101 ARM -  $2 (100x), 70MIPs, has 2 SPI/SSP ports,
> one of which has a fifo, and also variable word length, so you could for example use 10 bits to do
> 8x8 chars with spacing, and as long as you keep the fifo full, it ought to come out in a solid line
> of pixels.
> IAR do a free version of their C/Asm devtool with 32K limit, which should be plenty.

Wow! I need to get out more often.

Is there a source of 44 pin PLCC parts anywhere?

I'll have to give this part a look.

BAJ

More... (looser matching)
- Last day of these posts
- In 2006 , 2007 only
- Today
- New search...