Shop OBEX P1 Docs P2 Docs Learn Events
Overclocking — Parallax Forums

Overclocking

BradCBradC Posts: 2,601
edited 2008-06-14 12:28 in Propeller 1
G'day all,

I'm in the process of re-working the USB stack low level code and I'm incorporating automatic clock speed compensation.
Can those of you who are overclocking (I'm looking at YOU hippy) drop a note to let me know the crystal/clock configurations you are using?
I'm going to place an order for some crystals and I'd like to be able to cover a large proportion of the current usage base for testing.

(I've only tested with a 5MHZ Crystal and a 6MHZ resonator thus far)

Ta [noparse]:)[/noparse]

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Pull my finger!

Comments

  • Oldbitcollector (Jeff)Oldbitcollector (Jeff) Posts: 8,091
    edited 2008-06-10 13:41
    There is good information here...

    propeller.wikispaces.com/Oscillator

    OBC

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    New to the Propeller?

    Getting started with the Protoboard? - Propeller Cookbook 1.4
    Updates to the Cookbook are now posted to: Propeller.warrantyvoid.us
    Got an SD card? - PropDOS
    Need a part? Got spare electronics? - The Electronics Exchange
  • hippyhippy Posts: 1,981
    edited 2008-06-10 14:42
    I use 5MHz (80MHz), 7.3728MHz (~118MHz) and 14.31818MHz ( ~114MHz PLL8x ).

    7.3728MHz has proven reliable for me, even now we've hit the higher temperatures of summer.
  • Phil Pilgrim (PhiPi)Phil Pilgrim (PhiPi) Posts: 23,514
    edited 2008-06-10 17:21
    hippy said...
    7.3728MHz has proven reliable for me, even now we've hit the higher temperatures of summer.
    Did you mean to imply, by omission, that 14.31818MHz is unreliable?

    -Phil

    O, how I yearn for the "higher temperatures of summer". It's 50°F (10°C) here and raining. 'Has been virtually since last October. And I thought England had crappy weather!
  • BradCBradC Posts: 2,601
    edited 2008-06-10 17:27
    Phil Pilgrim (PhiPi) said...

    O, how I yearn for the "higher temperatures of summer". It's 50°F (10°C) here and raining. 'Has been virtually since last October. And I thought England had crappy weather!

    Pop over here for a visit (UAE).. it's just cracking low to mid 40's in the day preparing for 50's late next month.. (C not F)

    Thanks for the info Hippy, I've added them to my order..

    Brad

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    Pull my finger!
  • Ken PetersonKen Peterson Posts: 806
    edited 2008-06-10 17:37
    @BradC: I was in Las Vegas several years ago when it got up to 50C for a few days. There was a strong breeze, but it felt like I was standing in front of a giant hair dryer. What's it like there at night?

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
  • Oldbitcollector (Jeff)Oldbitcollector (Jeff) Posts: 8,091
    edited 2008-06-10 18:27
    I have used the 14.318 here without issue, however I've migrated to 6mhz
    due to the fact it's a little more common here.. [noparse]:)[/noparse]

    Phil Pilgrim (PhiPi) said...
    hippy said...
    7.3728MHz has proven reliable for me, even now we've hit the higher temperatures of summer.
    Did you mean to imply, by omission, that 14.31818MHz is unreliable?

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    New to the Propeller?

    Getting started with the Protoboard? - Propeller Cookbook 1.4
    Updates to the Cookbook are now posted to: Propeller.warrantyvoid.us
    Got an SD card? - PropDOS
    Need a part? Got spare electronics? - The Electronics Exchange
  • hippyhippy Posts: 1,981
    edited 2008-06-10 21:11
    Phil Pilgrim (PhiPi) said...
    Did you mean to imply, by omission, that 14.31818MHz is unreliable?

    No, I just don't have any analysis for that one. Since plugging the 7.3728MHz into the board that's all I've used.


    My personal preference is for 7.3728MHz over 14.31818MHz because it is 'less strain' on the PLL and gives a slightly faster system clock.
  • BradCBradC Posts: 2,601
    edited 2008-06-11 04:14
    Ken Peterson said...
    @BradC: I was in Las Vegas several years ago when it got up to 50C for a few days. There was a strong breeze, but it felt like I was standing in front of a giant hair dryer. What's it like there at night?

    Wet!
    After about 8pm the humidity rockets up past the high 80's. If you have the AC on in the car cold enough the outside of your windscreen frosts up.

    Hot and Dry during the day, Hot and Wet at night.. just how we like it [noparse]:)[/noparse]

    Give it another 6 weeks and we should have a couple of 55 days.. that gets too hot, even for me.

    Brad

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    Pull my finger!
  • Lord SteveLord Steve Posts: 206
    edited 2008-06-11 04:31
    Not to get back on topic, but I use a 6.25 MHz crystal at 16x.
  • David E.David E. Posts: 13
    edited 2008-06-11 17:23
    Lord Steve said...
    Not to get back on topic, but I use a 6.25 MHz crystal at 16x.

    I like that combo as it's an easy-peasy 100mhz. Easier to work with the math in my head that way.
  • Ken PetersonKen Peterson Posts: 806
    edited 2008-06-11 17:33
    Lord Steve: Sorry for the OT comments. BTW - where do you get 6.25MHz crystals? I didn't see anything that value at Digikey.

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
  • David E.David E. Posts: 13
    edited 2008-06-11 21:20
    Ken Peterson said...
    Lord Steve: Sorry for the OT comments. BTW - where do you get 6.25MHz crystals? I didn't see anything that value at Digikey.

    Couldn't I just use a 100mhz crystal and not use the PLL? Or a 50mhz and 2xPLL?

    Sorry but I haven't played with the PLL/clkfreq at all other than just running it at 80mhz and don't know the limits on the crystals I can use.
  • Phil Pilgrim (PhiPi)Phil Pilgrim (PhiPi) Posts: 23,514
    edited 2008-06-11 22:29
    The specs say that, to use the PLL, the crystal frequency has to be between 4 MHz and 8 MHz. The Spin Stamp uses a 10 MHz crystal w/ PLL8x, and others have reported success with higher frequencies, too. But 50 MHz, I believe, is way beyond reasonable even to attempt.

    Also, the specs mention crystal frequencies up to 60 MHz only, so a 100 MHz crystal or resonator is probably out of the question.

    -Phil
  • RaymanRayman Posts: 14,817
    edited 2008-06-11 22:56
    I've also used a 6.25 MHz crystal, no problem.

    I've heard that you can use an external 100MHz oscillator as long as you don't use the PLL.

    Note that you have to use a powered 100MHz oscillator because the series resonant, fundamental freq. type normally used don't go that fast.

    Post Edited (Rayman) : 6/11/2008 11:16:52 PM GMT
  • RaymanRayman Posts: 14,817
    edited 2008-06-11 23:12
    the uoled-96-prop is an interesting data point... Spin running on a single cog runs fine with 8MHz and 16X. But, it fails if you try to start a second cog.

    I wonder if the Hydra will run one cog at 16X...
  • hippyhippy Posts: 1,981
    edited 2008-06-12 02:26
    @ BradC : One thing I forgot to add earlier was that if USB will only work on 5MHz / 6MHz then that's good enough for me. Don't go busting yourself up to get some peculiar and not ideal system clock wedded to whatever timing the USB Cogs want ( unless you really want to of course ).
  • BradCBradC Posts: 2,601
    edited 2008-06-12 04:00
    hippy said...
    @ BradC : One thing I forgot to add earlier was that if USB will only work on 5MHz / 6MHz then that's good enough for me. Don't go busting yourself up to get some peculiar and not ideal system clock wedded to whatever timing the USB Cogs want ( unless you really want to of course ).

    I love a challenge [noparse]:)[/noparse]
    If I get the auto-timing right, then it should "Just work" for any frequency over 80Mhz.. I need a swag of different crystals to test it with though.

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    Pull my finger!
  • Phil Pilgrim (PhiPi)Phil Pilgrim (PhiPi) Posts: 23,514
    edited 2008-06-12 04:53
    Brad,

    I faced the same problem trying to compute valid FRQA and PLL divider values for any given PWM frequency. It's difficult to do accurately without native unsigned 64/32 division. I ended up writing such a routine for the code posted here, which you're wlecome to use if it helps. Today, I've been working on similar code written in Spin to get a reliable 4.5 MHz audio subcarrier for broadband video from any valid clock frequency. Spin's "all numbers are signed" motif has been a bit of a speedbump, though.

    -Phil
  • Lord SteveLord Steve Posts: 206
    edited 2008-06-12 04:55
    Ken Peterson said...
    Lord Steve: Sorry for the OT comments. BTW - where do you get 6.25MHz crystals? I didn't see anything that value at Digikey.
    I got mine at Alltronics.com, but I bought out their stock.· smilewinkgrin.gif
  • Peter HannemanPeter Hanneman Posts: 5
    edited 2008-06-13 17:05
    What kinds of input voltages are you guys needing to run to operate with 6Mhz+ crystals?
  • RaymanRayman Posts: 14,817
    edited 2008-06-14 01:56
    I've just used the +3.3 from proto/demo regulator with my 6.25MHz crystals. But, I think the latest datasheet does show an advantage at higher voltage, as I recall...
  • hippyhippy Posts: 1,981
    edited 2008-06-14 12:28
    All my work been done at 7.3728MHz has been on an otherwise unmodified ProtoBoard ( not entirely true, but only modifications which don't affect the PSU side of things ).
Sign In or Register to comment.