P2_Eval Crystal Startup and Capacitance setting tests

jmgjmg Posts: 13,609
edited 2018-12-27 - 05:13:09 in Propeller 2
Switching from RCFAST (~ 23MHz) to Xtal (20.000MHz) is tested, and the Xtal Cap settings also tested

Xtal: TSX-3225 20.0000MF20G-AC3 EPSON CRYSTAL 20.0000MHZ 9PF SMD 20MHz ±20ppm ±10ppm 9pF 60 Ohms

Docs have
HUBSET ##%0000_000E_DDDD_DDMM_MMMM_MMMM_PPPP_CCSS 'set clock mode
CC Xtal pins mode
SS Osc Select : RCFAST : RCSLOW : XI : PLL
%CC     XI status       XO status       XI/XO Z         XI/XO caps        Measures      ppm change  
%00     ignored         float           Hi-Z            OFF
%01     input           600-ohm drive   1M-ohm          OFF            -> 1000.1439Hz   +144   ppm
%10     input           600-ohm drive   1M-ohm          15pF per pin   ->  999.9933Hz   -6.700 ppm  << operating point.
%11     input           600-ohm drive   1M-ohm          30pF per pin   ->  999.9472Hz   -53 ppm

%SS     Clock Source    Notes
%11     PLL             CC != %00 and E=1, allow 10ms for crystal+PLL to stabilize before switching to PLL
%10     XI              CC != %00, allow 5ms for crystal to stabilize before switching to XI
%01     ~20kHz          ~20kHz, can be switched to at any time, low-power
%00     20MHz+          20MHz+, can be switched to at any time, used on boot-up.


Taqoz test script, to Enable PWM, then switch clock sources (after this a reset is needed to recover TAQOZ & autobaud).
Tests confirm switching without any delay is not reliable.
it may be that before the Xtal amplitude is large enough, that noise/illegal clock widths can come from Xtal Osc.
// Reset target '> '[esc] enables TAQOZ
$3E,$20,$1B
// wait appx 250ms needed, for TAQOZ to load
'16 PIN 200 10000 1 PWM',$0D
// wait ~5ms for ok echo, 
// Then Enable Xtal Osc, VAR delay for Osc Start, and switch to Xtal source.
'%11_00 HUBSET 62,000 WAITX %11_10 HUBSET',$0D

Scope image shows the boundary/marginal case, were the starting delay is too small (2.7ms here at 62,000) and the Xtal has not fully started - there is visible a longer delay, as the clock source handover occurs.
When the WAITX is sufficiently large, there is no extra delay, as the switch over is done in 1 SysCLK.

Xtal oscillates in all 3 Cap modes, even minimum C, and the %10 mode has < 10ppm error.


TAQOZ PWM has a couple of issues :
a) It generates a wide initial pulse, before the user pulses appear after ~ 6ms
b) It seems to be off-by-2, ie the output numbers, are half the sysCLK - is that a Smart Pin bug, or a TAQOZ bug ?]
... Ahh, buried far away from the PWM examples, is mention of SAW, (no examples) & some trials reveals
'16 PIN 200 10000 1 SAW',$0D
removes the unexpected 2x error & also changes the pulse hiccups.
In SAW mode, the 2.6ms HI is still present, but there is much less delay in the pulse train repeats, after the falling edge. Prr of 433us has first low of 442us

Looks like the TAQOZ docs need some corrections & clarifications on actual SysCLK cycle counts and modes.

Changed to 160MHz to look for jitter, and also measured ppm temperature effects

160MHz is SAWtooth /10k of
1-16k/15999.836 = -10.250ppm
Then change to 20MHz Xtal, and measure over next 5 mins
1-1k/999.9900 = -10.000ppm
1-1k/999.9908 = -9.200ppm
1-1k/999.9913 = -8.700ppm
1-1k/999.9922 = -7.800ppm ....cooling down,

Comments

  • cgraceycgracey Posts: 11,267
    edited 2018-12-27 - 02:45:54
    Thanks for trying these things out, Jmg.

    Now I am really wondering something. You are like a perpetual volcano of questions. Now that you've got a P2 board, you are going to start answering a lot of your own questions. This will probably elicit lots of other questions which you will get answers to. What I am wondering about is whether this whole process is going to be exothermic or endothermic. You could be on your way to supernova or neutron star.
  • potatoheadpotatohead Posts: 9,736
    edited 2018-12-27 - 03:00:26
    Big grin over here.

    Jmg armed and dangerous :D
    Do not taunt Happy Fun Ball! @opengeekorg ---> Be Excellent To One Another SKYPE = acuity_doug
    Parallax colors simplified: https://forums.parallax.com/discussion/123709/commented-graphics-demo-spin<br>
  • jmgjmg Posts: 13,609
    cgracey wrote: »
    Thanks for trying these things out, Jmg.

    Now I am really wondering something. You are like a perpetual volcano of questions. Now that you've got a P2 board, you are going to start answering a lot of your own questions. This will probably elicit lots of other questions which you will get answers to. What I am wondering about is whether this whole process is going to be exothermic or endothermic. You could be on your way to supernova or neutron star.

    I hope neither ;)
    My scope seems not quite good enough to catch the jitter, is has a bug in large +ve offsets, and I suspect large -ve offsets might not be all they claim either...
  • Any mode that sets a pin to pwm will at some point prior to this be in some other state. So i don't understand the concern that the pin is low before it pwms. You could try setting the mode and then making the pin an output (to enable the smartpin) but it'll be a few hours before i can get to try that. TAQOZ was set in ROM before we had real silicon so things are a little different.

    Tachyon Forth - compact, fast, forthwright and interactive
    useforthlogo-s.png
    --->CLICK THE LOGO for more links<---
    P2 +++++ TAQOZ INTRO & LINKS +++++ P2 SHORTFORM DATASHEET
    P1 +++++ Latest binary V5.4 includes EASYFILE +++++ Tachyon Forth News Blog
    Brisbane, Australia
  • Another big grin from me too

    Tachyon Forth - compact, fast, forthwright and interactive
    useforthlogo-s.png
    --->CLICK THE LOGO for more links<---
    P2 +++++ TAQOZ INTRO & LINKS +++++ P2 SHORTFORM DATASHEET
    P1 +++++ Latest binary V5.4 includes EASYFILE +++++ Tachyon Forth News Blog
    Brisbane, Australia
  • jmgjmg Posts: 13,609
    Any mode that sets a pin to pwm will at some point prior to this be in some other state. So i don't understand the concern that the pin is low before it pwms. You could try setting the mode and then making the pin an output (to enable the smartpin) but it'll be a few hours before i can get to try that. TAQOZ was set in ROM before we had real silicon so things are a little different.

    An initial state I am ok with, it is the 2.9ms HI then 3.5ms low I was not expecting, as well as the mysterious off by 2 effect. Hopefully, that is not in the pin ?
    ie the asked for clocks, seem to be 2x time expected from sysCLK value ?
  • Peter JakackiPeter Jakacki Posts: 8,421
    edited 2018-12-27 - 03:18:05
    Well i guess you're in a position now to find some answers!!! :lol:
    This is where the real fun is.

    Tachyon Forth - compact, fast, forthwright and interactive
    useforthlogo-s.png
    --->CLICK THE LOGO for more links<---
    P2 +++++ TAQOZ INTRO & LINKS +++++ P2 SHORTFORM DATASHEET
    P1 +++++ Latest binary V5.4 includes EASYFILE +++++ Tachyon Forth News Blog
    Brisbane, Australia
  • jmgjmg Posts: 13,609
    edited 2018-12-27 - 04:15:32
    Well i guess you're in a position now to find some answers!!! :lol:
    This is where the real fun is.

    I am in a position to make measurements, the best answers will come from the TAQOZ author, as the issue is either in TAQOZ, or in the pin itself (hopefully not).
    I can see 2 PWM modes, maybe TAQOZ chose the triangle PWM, not the sawtooth PWM ? - the docs are unclear on that detail.
    Addit : see updated #1 post - as suspected, PWM uses triangle, and there is another keyword, SAW, that must be used to give sysclk correct results.
  • Peter JakackiPeter Jakacki Posts: 8,421
    edited 2018-12-27 - 05:07:13
    I may be the author but an author only uses other peoples words to express thoughts, so other people can understand it.

    This is the current ROM code for PWM which you can see happens to be the triangle waveform,whether that's a good thing or not, it is what I initially used to play with PWM. Feel free to rearrange these words or just interact with the smartpin word by word ;
    : PWM ( duty frame div -- )	$50 
    : PWM1				L WRPIN SWAP 16<< + WXPIN WYPIN ;
    : SAW ( duty frame div -- )	$52 PWM1 ;
    

    BTW, The PWM word falls through to the next word PWM1 which just happens to be a common routine for both PWM and SAW. Even though SAW "calls" PWM1 the actual compiled code executes this as a fast jump so that there is no need for a return or an EXIT

    Tachyon Forth - compact, fast, forthwright and interactive
    useforthlogo-s.png
    --->CLICK THE LOGO for more links<---
    P2 +++++ TAQOZ INTRO & LINKS +++++ P2 SHORTFORM DATASHEET
    P1 +++++ Latest binary V5.4 includes EASYFILE +++++ Tachyon Forth News Blog
    Brisbane, Australia
  • jmg - I've just received my ES and hooked up to p32 and tried 32 PIN 100 1000 1 PWM but no matter what i do it works nice and clean, no mysterious delay. Can you test other pins? I've set p32 high as will just to check and it pulls low somewhere in the pwm cycle and on this case the first high pulse is 40us later

    Tachyon Forth - compact, fast, forthwright and interactive
    useforthlogo-s.png
    --->CLICK THE LOGO for more links<---
    P2 +++++ TAQOZ INTRO & LINKS +++++ P2 SHORTFORM DATASHEET
    P1 +++++ Latest binary V5.4 includes EASYFILE +++++ Tachyon Forth News Blog
    Brisbane, Australia
  • Cluso99Cluso99 Posts: 15,079
    edited 2018-12-27 - 09:14:11
    potatohead wrote: »
    Big grin over here.

    Jmg armed and dangerous :D

    +1 :sweat_smile:
    My Prop boards: P8XBlade2, RamBlade, CpuBlade, TriBlade
    Prop OS (also see Sphinx, PropDos, PropCmd, Spinix)
    Website: www.clusos.com
    Prop Tools (Index) , Emulators (Index) , ZiCog (Z80)
  • jmgjmg Posts: 13,609
    jmg - I've just received my ES and hooked up to p32 and tried 32 PIN 100 1000 1 PWM but no matter what i do it works nice and clean, no mysterious delay.

    This is the first thing I do, after starting TAQOZ and waiting 250ms

    '32 PIN 200 10000 1 PWM',$0D

    pulse start is the same = 2.8ms wide hi, then first pulse is at 6.56ms

    '32 PIN 200 10000 1 SAW',$0D
    has the same 2.8ms hi, and then pulses start from there (no trailing wide low)

    32 PIN 100 1000 1 PWM has faster pulses, and that indicates the trailing low is until 5.773 ms
Sign In or Register to comment.