Shop OBEX P1 Docs P2 Docs Learn Events
Overclocking - is it possible — Parallax Forums

Overclocking - is it possible

soshimososhimo Posts: 215
edited 2011-04-05 19:54 in Propeller 1
I've had mixed results overclocking AVR. On the ATMega168, for example, it's possible to get NTSC by overclocking the chip at a frequency divisible by the NTSC frequency (I think it was 26Mhz or something like that - can't remember off the top of my head). The point is that you could overclock the AVR and as long as you didn't drive it too fast you could get some improvement in execution speed. Is this possible with the propeller? And is it even advisable?

TIA

Comments

  • ctwardellctwardell Posts: 1,716
    edited 2010-07-17 00:55
    soshimo,

    Yes it is possible, see:

    http://mikronauts.com/products/mikronauts-625mhz-crystal/

    C.W.
  • KyeKye Posts: 2,200
    edited 2010-07-17 02:41
    Its not even hard. Just pop in a 6.25Mhz crystal instead of the 5Mhz one and you're done.

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    Nyamekye,
  • JamesRonaldJamesRonald Posts: 39
    edited 2010-07-17 03:35
    Perhaps worth mentioning is that over clocking is not necessary. As we all know, the propeller can already generate video signals (It's built in).

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    James Ronald
    My Robots
  • soshimososhimo Posts: 215
    edited 2010-07-17 03:59
    JamesRonald said...
    Perhaps worth mentioning is that over clocking is not necessary. As we all know, the propeller can already generate video signals (It's built in).

    I only mentioned NTSC as an example of why I needed to overclock the AVR, which is why prop > AVR [noparse]:)[/noparse]. I don't really have a need, at this point, to overclock the prop, but I would like to know if the option is available which it sounds like it is.
  • potatoheadpotatohead Posts: 10,261
    edited 2010-07-17 05:33
    Yeah, run 96mhz with a 6Mhz xtal, or 100Mhz with a 6.25Mhz xtal, and life is good. The chips run great at those speeds, and it's a nice boost over the reference 80Mhz.

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    Propeller Wiki: Share the coolness!
    8x8 color 80 Column NTSC Text Object
    Wondering how to set tile colors in the graphics_demo.spin?
    Safety Tip: Life is as good as YOU think it is!
  • soshimososhimo Posts: 215
    edited 2010-07-17 06:28
    potatohead said...
    Yeah, run 96mhz with a 6Mhz xtal, or 100Mhz with a 6.25Mhz xtal, and life is good. The chips run great at those speeds, and it's a nice boost over the reference 80Mhz.

    25% speed increase is not too shabby. I'll have to order some 6.25Mhz xtals.
  • LeonLeon Posts: 7,620
    edited 2010-07-17 06:39
    Bear in mind that the Propeller hasn't actually been properly characterised by Parallax at out of spec. speeds, a limited number of samples was tested. I don't think that anyone has reported any problems at 100 MHz, though.

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    Leon Heller
    Amateur radio callsign: G1HSM
  • potatoheadpotatohead Posts: 10,261
    edited 2010-07-17 06:48
    Packs a real punch on the video too. When working at TV frequencies, that 25 percent translates into a lot. At higher frequencies, it means something being possible, or taking one less COG [noparse]:)[/noparse]

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    Propeller Wiki: Share the coolness!
    8x8 color 80 Column NTSC Text Object
    Wondering how to set tile colors in the graphics_demo.spin?
    Safety Tip: Life is as good as YOU think it is!
  • Cluso99Cluso99 Posts: 18,069
    edited 2010-07-17 11:16
    I run the RamBlade at 104MHz because the 6.5MHz xtals were readily available. I also supply 13.5MHz (108MHz) although I have not tested thoroughly. Spaieha has my TriBlade running at 1120MHz using a 15MHz xtal but RamBlade will not run at this speed. 96MHz (6MHz xtal) is also common.

    However, not every pcb will run overclocked, at least to this extend, because special layout and decoupling is required. There is a few threads that discuss this. The DIP chip appears to run better at overclocking than the QFP chip. I am not sure why as I am sure it is not heat. By increasing the voltage Coley achieved the highest frequency by using a signal generator.

    So 100MHz = 25%, 104MHz = 30%, 108MHz = 35%, 120MHz = 50%. Not bad hey.

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    Links to other interesting threads:

    · Home of the MultiBladeProps: TriBlade,·RamBlade,·SixBlade, website
    · Single Board Computer:·3 Propeller ICs·and a·TriBladeProp board (ZiCog Z80 Emulator)
    · Prop Tools under Development or Completed (Index)
    · Emulators: CPUs Z80 etc; Micros Altair etc;· Terminals·VT100 etc; (Index) ZiCog (Z80) , MoCog (6809)·
    · Prop OS: SphinxOS·, PropDos , PropCmd··· Search the Propeller forums·(uses advanced Google search)
    My cruising website is: ·www.bluemagic.biz·· MultiBlade Props: www.cluso.bluemagic.biz
  • davidsaundersdavidsaunders Posts: 1,559
    edited 2011-03-27 13:56
    Has any one attempted 128MHz, using a 8MHz crystal and PLL x16 or 16Mhz crystal PLL x8???
  • Mike GreenMike Green Posts: 23,101
    edited 2011-03-27 14:12
    Look at the graphs on the Propeller Datasheet. Beau Schwabe did some testing in an environmental chamber characterizing the maximum reliable speed over a range of supply voltages and ambient temperatures. It was not a large number of chips, but should be representative.
  • Cluso99Cluso99 Posts: 18,069
    edited 2011-03-27 14:33
    Sapieha has had a TriBlade running at 120MHz (15MHz x8). However, I was unable to get a RamBlade running at anything beyond 14.31818MHz x8. Note that this was not fully tested. Coley did get a prop running the fastest, but again not fully tested.

    I supply RamBlades with 104MHz (6.5MHz x16) and include an untested 108MHz (13.5MHz x8).

    For USB you require 96MHz (6MHz x16).

    However, when overclocking, it is very specific to your pcb layout and decoupling. There are a number of threads where this is discussed.
  • davidsaundersdavidsaunders Posts: 1,559
    edited 2011-03-27 14:45
    Thank you. By the graph, it appears that at least on some chips it should be possible to reach 128MHz @ 27 degrees C. I just hope this can be made stable :) . Just in the mood to live a bit dangerously. 4Ms/s poor mans oscilloscope at 4 bit res, or 8Ms/s at 3 bit any one? I guess it is time to dig out the heat sink and fan. And maybe try to run stable at 2.8V.
  • jazzedjazzed Posts: 11,803
    edited 2011-03-27 15:35
    104MHz is very easy to get on a 2 layer PCB.

    PPLA is a Propeller based logic analyzer that samples at 10ns intervals.

    Click here for the PPLA Hackaday.com article.
  • davidsaundersdavidsaunders Posts: 1,559
    edited 2011-04-05 16:27
    WOW! I have been running video driver tests, for about 5 hours with one of the VDD connections not hooked up. I am running it at 96MHz (I do not have any 5MHz crystals so I use 6MHz), and it has not had any problems, and this is on a breadboard. After reading about some people frying there Props PLL running at 80MHz by doing this for a couple minutes, I must consider myself lucky.

    There was a wire that must have got pulled loose before I finished the circuit.
  • Oldbitcollector (Jeff)Oldbitcollector (Jeff) Posts: 8,091
    edited 2011-04-05 17:21
    6.00 and 6.25 Mhz crystals have been verified by many as being stable with the Propeller.


    IIRC, I even managed 108mhz from a crystal pulled from a VGA card. Can't speak for stability, but it did work. The only reason my boards run at 80mhz is to abide by the "official' standard.

    OBC
  • davidsaundersdavidsaunders Posts: 1,559
    edited 2011-04-05 19:09
    Jeff:
    Yes I am aware that it is possible to overclock stably to up to 120MHz, My surprise comes in that I was running at 96Mhz with one of the VDD connections NOT connected, for over 5 hours, while some have reported frying the PLL doing the same at 80MHz for a couple of minutes.
  • Oldbitcollector (Jeff)Oldbitcollector (Jeff) Posts: 8,091
    edited 2011-04-05 19:54
    I have done that here without damaging the Propeller, then again, I haven't manage to mangle one yet given all the crazy stuff I done. Extremely robust little chip. Hoping that PropII is just as tolerant.

    OBC
Sign In or Register to comment.