Shop OBEX P1 Docs P2 Docs Learn Events
Propeller boot time observations — Parallax Forums

Propeller boot time observations

Don MDon M Posts: 1,653
edited 2013-03-01 05:08 in Propeller 1
I am starting to experiment with 2 Props connected together. The Slave Prop (Propeller DNA board) is powered from the Main Prop (Quickstart board).

I was having some problems in waiting from a message coming from the Slave so I connected my logic analyzer to the SDA and CLK pins on each board and observed the boot up time. I noticed that when there is a USB connection the boot up time for the Prop is considerably longer than if it is not connected to the USB and just powered by an external source.

Here is a screen shot when just using external power source:

No USB.jpg



And here are 2 screen shots- one each of them connected to USB while the other one isn't:

USB DNA.jpg


USB QS.jpg



It takes 7 seconds to boot up with the USB connected versus 1 1/4 seconds without.


Why is that?

I thought that maybe PST or the Prop Tool were trying to connect somehow but I quit both of those programs and results are still the same. Must have something to do with the PC USB port?

Edit: The bottom 2 lines on the trace are the QS board and the middle 2 are the DNA board.
No USB.jpg
1024 x 185 - 31K
USB DNA.jpg
1024 x 132 - 28K
USB QS.jpg
1024 x 137 - 23K

Comments

  • RaymanRayman Posts: 14,826
    edited 2013-02-28 16:17
    If you are USB powered, it takes some time for your computer to talk to the USB chip and authorize use of USB power...
    Could that be it?
  • Don MDon M Posts: 1,653
    edited 2013-02-28 16:21
    If you look at the trace you'll see that I triggered on the 3.3V line so they are both getting power at the same time. You'll notice that whichever board is connected to the USB port that the data coming from the EEPROM comes in bursts instead of a steady stream when there is no USB present.
  • jmgjmg Posts: 15,183
    edited 2013-02-28 16:31
    Don M wrote: »
    You'll notice that whichever board is connected to the USB port that the data coming from the EEPROM comes in bursts instead of a steady stream when there is no USB present.

    That's a little stranger, but the delayed start is easier to understand.

    I think if the boot see's a long LOW on TXD/RXD it decides there is no point waiting for PC info, and so can skip faster to i2c.
    If the lines hint at someone connected, then it needs to wait for link action.

    Once it has decided to do i2c boot, you would expect that to not care about the pin states ?
    I think the boot code was posted somewhere..
  • Martin HodgeMartin Hodge Posts: 1,246
    edited 2013-02-28 16:36
    If you have a spare input or two on your LA, watch the reset line on each board for oddities. also, how exactly are you connecting the power on the two boards together?.
  • Don MDon M Posts: 1,653
    edited 2013-02-28 16:55
    Martin Hodge wrote: »
    If you have a spare input or two on your LA, watch the reset line on each board for oddities. also, how exactly are you connecting the power on the two boards together?.

    Ground to ground and +5V to +5V. I thought about looking at the reset pin with another channel. When I get a few minutes I'll take a look.

    I don't think the issue has to do with the fact I have 2 boards connected together. Even if you use just a single board and compare the traces of DC powered connected versus having the USB powered debug port connected the result is still the same.
  • Don MDon M Posts: 1,653
    edited 2013-02-28 17:11
    Connected a Quickstart only. Monitored the SDA, CLK and Reset lines. I get the same results. The difference is that when the USB is connected the Reset line does get toggled a few times. During DC only it toggles just once at the very beginning then stays high.

    The middle 2 traces are the SDA & CLK. The very bottom trace (7) is the Reset line.

    QS with RESET Line.jpg
    QS with RESET Line.jpg
    1024 x 151 - 26K
  • Don MDon M Posts: 1,653
    edited 2013-02-28 18:09
    So I guess the question is- why is the USB port toggling the DTR line so many times to cause the resets and interruption in loading the image from the EEPROM?
  • Mike GMike G Posts: 2,702
    edited 2013-02-28 20:59
    Maybe it has to do with the documented Serial Quirk?
    http://forums.parallax.com/showthread.php/135067-Serial-Quirk?p=1043169&viewfull=1#post1043169
  • Martin HodgeMartin Hodge Posts: 1,246
    edited 2013-02-28 22:43
    Mike G wrote: »
    Maybe it has to do with the documented Serial Quirk?
    http://forums.parallax.com/showthread.php/135067-Serial-Quirk?p=1043169&viewfull=1#post1043169

    I believe that thread is discussing the problem with using pin 30 for output on a board with an FT232 on it. Something the DNA is immune to I must add.

    Don's problem is happening every time the board is plugged into USB, correct? I've always had Windows reset the FT232 multiple times when first connecting and just passed it off as a Windows "quirk". It would be interesting to see if Linux does this too.
  • Don MDon M Posts: 1,653
    edited 2013-03-01 03:31
    Mike G wrote: »
    Maybe it has to do with the documented Serial Quirk?
    http://forums.parallax.com/showthread.php/135067-Serial-Quirk?p=1043169&viewfull=1#post1043169

    The issue is not board specific. I believe it has to do with interaction of USB port and FTDI chip. And maybe Windows. I don't know. When I get a few moments I'll try it with my MAC.
  • Don MDon M Posts: 1,653
    edited 2013-03-01 05:08
    Tried it on my Mac Mini and my Google Chrome CR-48. Neither one affects the reset line line like the Windows (XP) machine did.

    So the boot up is fast using other than the Windows PC. FWIW I am using goSerial on the Mac.

    BTW this issue only shows when first powering up the board via the USB connection. If it's already powered and you press the reset button on the board it boots up fast. So it has something to do with the enumeration process with Windows.
Sign In or Register to comment.