Welcome, Guest. Please login or register.

Author Topic: MiniMig with AGA  (Read 347559 times)

Description:

0 Members and 1 Guest are viewing this topic.

Offline freqmax

  • Hero Member
  • *****
  • Join Date: Mar 2006
  • Posts: 2179
    • Show all replies
Re: MiniMig with AGA
« Reply #59 from previous page: November 26, 2010, 01:24:12 PM »
What's the status on the tests so far?
 

Offline freqmax

  • Hero Member
  • *****
  • Join Date: Mar 2006
  • Posts: 2179
    • Show all replies
Re: MiniMig with AGA
« Reply #60 on: November 26, 2010, 04:17:38 PM »
What do you mean with reverse leakage ..?
Is that the only issue? ie the LDO has too much reverese leakage wich prevent the board from turning power off?
 

Offline freqmax

  • Hero Member
  • *****
  • Join Date: Mar 2006
  • Posts: 2179
    • Show all replies
Re: MiniMig with AGA
« Reply #61 on: November 26, 2010, 04:56:00 PM »
Maybe a Zener diode could solve this in combination with some resistor current limiter?
Or a plain scottchy diode in series with the 3,3 V regulator or somewhere in the power supply chain?
Diodes have a constant voltage drop which makes it easier to calculate with in a constant voltage circuit.

I suppose the DAC problem is only affecting the analog part of the DVI connection? (no analog provision in HDMI)

Btw, an auto power off from software would be neat ;)
 

Offline freqmax

  • Hero Member
  • *****
  • Join Date: Mar 2006
  • Posts: 2179
    • Show all replies
Re: MiniMig with AGA
« Reply #62 on: November 26, 2010, 08:59:32 PM »
xyzzy, Why do you feel the need to divert from the topic in such trolling way rather than focus on what's useful ?
 

Offline freqmax

  • Hero Member
  • *****
  • Join Date: Mar 2006
  • Posts: 2179
    • Show all replies
Re: MiniMig with AGA
« Reply #63 on: November 26, 2010, 11:12:32 PM »
* VHDL, tough!, you have to eat shit initially but you can do the bit-flank-bang any way you want in the end. Get used to Cobol fingers(tm). Modeled after ADA thanks to DoD .. Doh!..

* Verilog, Modeled on the C language, Shoot from the hip! Fast way to bit-flank-bang heaven with limitations in advanced level.

Then there's the C++/C whatever to VHDL/Verilog translator. BUT interpreting programming languages is the wrong tool for the job. A shortcut that can make you suffer the brick wall.
 

Offline freqmax

  • Hero Member
  • *****
  • Join Date: Mar 2006
  • Posts: 2179
    • Show all replies
 

Offline freqmax

  • Hero Member
  • *****
  • Join Date: Mar 2006
  • Posts: 2179
    • Show all replies
Re: MiniMig with AGA
« Reply #65 on: November 28, 2010, 02:02:59 AM »
I'm not suprised that a 1-2 persons in their home can beat a commercial development department. Usualy those institutions are hampered with red-tape, clueless managers that doesn't get the task at hand etc.. Ofcourse the original creators had to do things in hardware without any good simulation tools. But the former still applies, even now.

lou_dias, Cache helps, but the 16-byte cache is way smaller than the original 68030 256-byte cache. Increasing from 16 to 256 bytes gives cirka 7% more performance. But cost lots of matrix estate. So not worth it. So any performance from cache on the TG68 comes really from another section. The only noteworthy is that the data cache gave TG68 a 2x boost. I suspect there are other parts that can be optimized. But one has to stay compatible too. As soon as some software uses tight I/O loops with external hardware or multimedia chips there might be a problem otherwise.

I see a common problem with both VHDL and Verilog. They both want to be imperative language look-alike. Personally I don't like ADA, but like the ability to do serious flank setups in VHDL. The problem with imperative and even functional language is that they are designed with a sequential processing in mind.
A language which is centered around a logic array configuration would be far superior.

(wonder when those new linear regulators show up at mikej ;) )
 

Offline freqmax

  • Hero Member
  • *****
  • Join Date: Mar 2006
  • Posts: 2179
    • Show all replies
Re: MiniMig with AGA
« Reply #66 on: December 06, 2010, 12:48:00 PM »
@yaqube, volts?
 

Offline freqmax

  • Hero Member
  • *****
  • Join Date: Mar 2006
  • Posts: 2179
    • Show all replies
Re: MiniMig with AGA
« Reply #67 on: December 06, 2010, 01:55:25 PM »
Just wanted to be sure. The board uses regulators. So the input voltage can vary with the same power.
 

Offline freqmax

  • Hero Member
  • *****
  • Join Date: Mar 2006
  • Posts: 2179
    • Show all replies
Re: MiniMig with AGA
« Reply #68 on: December 06, 2010, 10:15:51 PM »
Assuming a cable area of 0.205 mm² (cat.5) which gives 0,188 ohm/meter and 5 V with a 5% tolerance. Ohms law U=R*I gives the maximum cable length to be 2,2 meters. Quite tight.

I hope the next board at least has space to populate it with onboard 5 V regulation.
 

Offline freqmax

  • Hero Member
  • *****
  • Join Date: Mar 2006
  • Posts: 2179
    • Show all replies
Re: MiniMig with AGA
« Reply #69 on: December 07, 2010, 11:56:36 AM »
I just used the cat.5 as an example to demonstrate that onboard regulators have an advantage as all wires inherently will loose some voltage. The important factors is the conductor diameter and voltage tolerance.
 

Offline freqmax

  • Hero Member
  • *****
  • Join Date: Mar 2006
  • Posts: 2179
    • Show all replies
Re: MiniMig with AGA
« Reply #70 on: December 17, 2010, 01:00:28 AM »
Didn't think of that (tm) ;-)
 

Offline freqmax

  • Hero Member
  • *****
  • Join Date: Mar 2006
  • Posts: 2179
    • Show all replies
Re: MiniMig with AGA
« Reply #71 on: December 17, 2010, 08:11:04 PM »
There are hundreds of things to get correct, and check in the project. Some things are bound to screw up, that's project-life ;)
On the next board revision the plug(s) may be relocated provided its possibe route and space to put a new connector.
 

Offline freqmax

  • Hero Member
  • *****
  • Join Date: Mar 2006
  • Posts: 2179
    • Show all replies
Re: MiniMig with AGA
« Reply #72 on: December 17, 2010, 09:00:27 PM »
Cabling out DVI and other high frequency connections might be hard to accomplish without signal degradation.
 

Offline freqmax

  • Hero Member
  • *****
  • Join Date: Mar 2006
  • Posts: 2179
    • Show all replies
Re: MiniMig with AGA
« Reply #73 on: December 17, 2010, 09:16:00 PM »
Every impedance alteration or significant cable length causes a bump in the signal path and thus deteriorates the signal. The less, the better.
 

Offline freqmax

  • Hero Member
  • *****
  • Join Date: Mar 2006
  • Posts: 2179
    • Show all replies
Re: MiniMig with AGA
« Reply #74 on: December 18, 2010, 12:49:25 AM »
Impedance alteration translates to a road bump. So not only length matters.