Help, Crystal issues

steprog · 2014-03-28 17:19

Hello,
I think my crystal is not working on my board and I need a crystal tomorrow. I am in Salem, Oregon this weekend trying to use the propeller for fire suppresson control and detection in vehicles. However I used some digikey crystals and i think they are not working for my prototype. Does anyone out there know where I may pick up some tomorrow up here ?
Greg

Bob Lawrence (VE1RLL) · 2014-03-28 17:32

However I used some digikey crystals

Can you post the Digikey part number so we can tell you if they should work?

Also , it's a good idea to post a picture of your prototype if you can.

steprog · 2014-03-28 17:38

Yes, thanks good ideal to check 5 MHz 20 pf X410-ND

Bob Lawrence (VE1RLL) · 2014-03-28 17:45

From the Prop manual Page 29:

OSCMx
1 0
_CLKMODE
Setting
XOUT
Resistance
XIN/XOUT
Capacitance Frequency Range
0 0 XINPUT Infinite 6 pF (pad only) DC to 80 MHz Input
0 1 XTAL1 2000 Ω 36 pF 4 to 16 MHz Crystal/Resonator
1 0 XTAL2 1000 Ω 26 pF 8 to 32 MHz Crystal/Resonator
1 1 XTAL3 500 Ω 16 pF 20 to 60 MHz Crystal/Resonator

http://www.parallaxsemiconductor.com/sites/default/files/parallax/Web-PropellerManual-v1.2_0.pdf

Bob Lawrence (VE1RLL) · 2014-03-28 17:48

18-20pF is the correct crystal spec@ 5 MHZ

The 20 pf capacitance rating you have is ok.

I'll have a look at X410-ND on the Digikey site.

Bob Lawrence (VE1RLL) · 2014-03-28 17:58

X410-ND

ECS-50-20-4
ECS Inc
CRYSTAL 5MHZ 20PF THRU
3,652 - Immediate

0.97000
1
Bulk

HC-49US

Bob Lawrence (VE1RLL) · 2014-03-28 17:59

X410-ND

HC-49US

That should work fine.

steprog · 2014-03-28 18:02

thats it, but it is so flaky on my board. I talked to Chris at Parallax and discussed possible problems, but I think my power is rock solid lots of bypass. My crystal is not far away, one or two vias. Driving me crazy. Never had this problem in all the projects I have had before. If anyone has anything like this nearby, would love to get my hands on some extra.

Bob Lawrence (VE1RLL) · 2014-03-28 18:09

Did you test your design with some basic spin code from the library first to see if everything works first?

I guess we need to see pictures to help you.

My experience with xtals is that they usually work or not, nothing in between unless they are affected by temperature or stray capacitance.

David Carrier · 2014-03-28 18:20

Try running it without the PLL, e.g. with:

_clkmode = xtal1

If it works, then the issue may be with the PLL, and not the crystal. It is the most sensatvie part of the Propeller microcontroller to conditions outside of the absolute maximum conditions, so if the Propeller microcontroller was exposed to ESD, excessive current on a supply rail, or reverse or under-voltage conditions, a bad PLL is a likely outcome.

David Carrier
Parallax Inc.

Phil Pilgrim (PhiPi) · 2014-03-28 19:11

steprog wrote:

My crystal is not far away, one or two vias.

Are we taliking about a PCB here or a solderless breadboard? If the latter, that could be your problem:

1. Crystal wires are very skinny and may not make good contact with the solderless terminals.
2. The extra stray capacitance of the breadboard could impact reliability.

-Phil

Lawson · 2014-03-28 19:34

Another thing to check is how clean you board is. While I haven't had flux deposits shut down the crystal, I have had the crystal temporarily stop running on several boards after I washed the PCB. (they always start running once they're bone dry though) I assume that leakage currents though the last bits of moisture were messing with the crystal oscillator. Note: I've always used PCB sockets for the crystal. directly soldering the crystal is more likely to trap flux and have leakage problems.

How long are the wires to and from the crystal? Do you have a ground-plain under the crystal and the wires to/from it? On my boards so-far I've used <1cm wires to the crystal and a ground-plane under everything with rock-solid results.

Marty

steprog · 2014-03-28 19:54

Hi Guys,
It is just crazy. It is a PCB circuit. I built 4 boards and switched crystals and prop chips sometimes the boards work other times they don't. Something is just irratic I think. Coffee and another late night of troubleshooting I think. When I get something that makes sense I will report back.
Greg

Phil Pilgrim (PhiPi) · 2014-03-28 20:32

Greg,

Bob Lawrence suggested posting a picture of your board. Please do. Include both a photo and your layout artwork. David Carrier suggested running the Prop without the PLL. Have you tried that yet? We're trying to help, but if you don't follow our suggestions, there's not much we can do.

-Phil

steprog · 2014-03-28 21:18

Sorry Phil,
I am out of town sitting in hangar in Oregon. I don't have a way to send my board layout or schematic. I did try running without the pll but I can't say I understand my results yet. Part of the problem is I am running on fumes. I decided I need to go slow and think through everything. Thanks Phil I know you guys are the best. Unfortunately, I really messed up and didn't bring my latest schematics with me.

steprog · 2014-03-28 21:28

Is it possible to have a memory problem and the program still shows that the eeprom is programmed? I found I had a short on a bypass cap for the eeprom fixed it but didn't solve my problem. Thinking of replacing the chip. Any thoughts?

Bob Lawrence (VE1RLL) · 2014-03-28 21:41

The other thing was to determine if it's a Hardware or Software Problem. "Did you test your design with some basic spin code from the library first to see if everything works first?"

Just write some code to flash a LED or 2 and see if it's stable.

Do you have access to an Oscilloscope , where you are?

LoopyByteloose · 2014-03-28 21:52

It just could be added capacitance in the board construction. Maybe a different drive setting for the crystal will bring it to life.

In other words, i suspect that you should work through the other Propeller OSCMx settings regardless of it merely being a 5Mhz xtal.

Bob Lawrence (VE1RLL) · 2014-03-28 21:54

Can you post your code on here?

steprog · 2014-03-28 21:57

Yes, Bob, tried a simple on/off loop with the spin code and nothing. However I had more sophisticated code working and then it goes away. I think maybe going through a simple program and going through the crystal settings will tell me something. If your able to program your memory, that means that rcfast is working true?
Greg

steprog · 2014-03-28 22:01

_clkmode = rcfast
_xinfreq = 5_000_000
_stack = 2000

CON
i2cSCL = 28

' Card Type
Under = TRUE
' CardNum = 2 'Under
' Sensor = True
' Mister = False
' Foamer = False

'Set pins for MosFet Drivers
FOAM_Pin = 17
MIST_Pin = 16

'Set pins and Baud Rate for USB comms
pcDebugRX = 31
pcDebugTX = 30
pcDebugBaud = 115200
Mode = 0

'Set Pins and Baud Rate for XBee comms
XB_Sleep = 26
XB_Reset = 27
XB_Rx = 24 'XB DOUT
XB_Tx = 25 'XB DIN
XB_Baud = 9600

'Count for A/D
num = 8

DAT
start byte "start",0 ' start processing command
Hello byte "hello",0 'send command
stop byte "stop",0 ' same as "reset"
space byte " ",0

VAR
byte command[128] ' store command from PC
byte V[8]
long OverPressure 'OverPressure Flag
long RfOverPressure 'OverPressure Flag from external
long GoodSwitch 'Switch Flag
long RfGoodSwitch 'Switch Flag from external
long Flame 'Flame Flag
long RfFlame 'Flame Flag from external
long RfFoam 'Foam Flag from external
long HiTemp 'High Temperature Flag
long RfHiTemp 'High Temperature Flag
long FlameTimeout 'FlameTimeOut Flag
word FlameCnt 'Flame time out counter

byte SetPoint,Upper,Lower,MaxThreshold
byte Sensor1_T2, Sensor1_T1, Sensor2_T2, Sensor2_T1
byte MaxTherm
byte Therm1, Therm2
word Sensor,n,ix,Timer,CheckWord
long Cog, TestVar
byte comd,Value

OBJ

USB : "FullDuplexSerialPlus"
getadc : "ad7829"
XB : "XBee_Object_2"
OP : "OverP2"
SW : "GoodSW"

pub main
Initialize
USB.start(pcDebugRX,pcDebugTX,Mode,pcDebugBaud) ' start the PC debug object
' XB.start(XB_Rx,XB_Tx,0,XB_Baud) 'initialize and start up the XB object
' XB.AT_Init ' Initialize for fast AT command use - 5 second delay to perform
' XB.Delay(1000)
' getadc.start(@V[1],num)
' OP.start(@OverPressure)
' SW.start(@GoodSwitch)
' display(string("Online ")) ' this string goes to PC for display
' USB.tx("O") 'Transmit Overpressure signal

repeat
display(string("Online ")) ' this string goes to PC for display
outa [MIST_Pin] := 1 'Activate Mist
outa [FOAM_Pin] := 1 'Activate Foam
waitcnt(clkfreq/10 + cnt) 'wait 1/5 sec to repeat writs
outa [MIST_Pin] := 0 'Activate Mist
outa [FOAM_Pin] := 0 'Activate Foam
waitcnt(clkfreq/10 + cnt) 'wait 1/5 sec to repeat writs
{

Bob Lawrence (VE1RLL) · 2014-03-28 22:02

Yes, Bob, tried a simple on/off loop with the spin code and nothing.

As in no problems or nothing is working?

I had more sophisticated code working and then it goes away

Are you using multiple cogs? Is your stack space correct?

Bob Lawrence (VE1RLL) · 2014-03-28 22:06

[PHP]Thinking of replacing the chip. Any thoughts?[/PHP]

David Carrier suggested that it may be a bad PLL and swamping out the Propeller chip for another one will determine that.

Why aren't you able to use a cell phone to snap a few pictures of your board and load them here?

steprog · 2014-03-28 22:09

As in signals don't oscillate. Swapped memory- no difference, commented out all objects still no oscillation

steprog · 2014-03-28 22:09

Sorry Bob, no smart phone here

Bob Lawrence (VE1RLL) · 2014-03-28 22:10

When your testing with the xtal , don't forget to set it back to :

_clkmode = xtal1 + pll16x 'Standard clock mode * crystal frequency = 80 MHz
_xinfreq = 5_000_000

Phil Pilgrim (PhiPi) · 2014-03-28 22:11

Greg,

attachment.php?attachmentid=78421&d=1297987572

-Phil

steprog · 2014-03-28 22:16

'_clkmode      = xtal1 + pll16x
_clkmode = xtal1
_xinfreq      = 5_000_000
_stack        = 2000

CON    
  i2cSCL        = 28

' Card Type
  Under = TRUE
'  CardNum = 2     'Under
'  Sensor = True
'  Mister = False
'  Foamer = False

'Set pins for MosFet Drivers  
   FOAM_Pin            = 17
   MIST_Pin            = 16

'Set pins and Baud Rate for USB comms
  pcDebugRX       = 31
  pcDebugTX       = 30
  pcDebugBaud     = 115200
  Mode = 0

'Set Pins and Baud Rate for XBee comms
  XB_Sleep        = 26
  XB_Reset        = 27
  XB_Rx           = 24  'XB DOUT
  XB_Tx           = 25  'XB DIN
  XB_Baud         = 9600

'Count for A/D 
  num           =  8

DAT
  start  byte "start",0          ' start processing command
  Hello   byte "hello",0             'send command
  stop    byte "stop",0           ' same as "reset"
  space   byte " ",0


    
  
  
VAR
  byte command[128]             ' store command from PC
  byte V[8]
  long OverPressure   'OverPressure Flag
  long RfOverPressure 'OverPressure Flag from external
  long GoodSwitch     'Switch Flag
  long RfGoodSwitch   'Switch Flag from external
  long Flame          'Flame Flag
  long RfFlame        'Flame Flag  from external
  long RfFoam         'Foam Flag  from external
  long HiTemp         'High Temperature Flag
  long RfHiTemp       'High Temperature Flag
  long FlameTimeout   'FlameTimeOut Flag
  word FlameCnt       'Flame time out counter
  
  byte SetPoint,Upper,Lower,MaxThreshold
  byte Sensor1_T2, Sensor1_T1, Sensor2_T2, Sensor2_T1
  byte MaxTherm
  byte Therm1, Therm2
  word Sensor,n,ix,Timer,CheckWord                                                                         
  long Cog, TestVar
  byte comd,Value
  
OBJ

'  USB            : "FullDuplexSerialPlus"
'  getadc         : "ad7829"
'  XB             : "XBee_Object_2"
'  OP             : "OverP2"
'  SW             : "GoodSW"
   
pub main
    Initialize
'    USB.start(pcDebugRX,pcDebugTX,Mode,pcDebugBaud) ' start the PC debug object
'    XB.start(XB_Rx,XB_Tx,0,XB_Baud) 'initialize and start up the XB object
'    XB.AT_Init                     ' Initialize for fast AT command use - 5 second delay to perform
'   XB.Delay(1000)
'    getadc.start(@V[1],num)
'    OP.start(@OverPressure)
'    SW.start(@GoodSwitch)
'    display(string("Online ")) ' this string goes to PC for display
'            USB.tx("O")   'Transmit Overpressure signal

                                   
    repeat
'       display(string("Online ")) ' this string goes to PC for display
           outa [MIST_Pin] := 1  'Activate Mist
           outa [FOAM_Pin] := 1  'Activate Foam
           waitcnt(clkfreq/10 + cnt)                 'wait 1/5 sec to repeat writs
           outa [MIST_Pin] := 0  'Activate Mist
           outa [FOAM_Pin] := 0  'Activate Foam
           waitcnt(clkfreq/10 + cnt)                 'wait 1/5 sec to repeat writs

{

Bob Lawrence (VE1RLL) · 2014-03-28 22:19

_clkmode = rcfast

I don't see how you can test your code above with _clkmode = rcfast.

_clkmode = rcfast is ok for testing the PLL but you have to use simple code such as flash a LED.

steprog · 2014-03-28 22:26

I changed it to xtal1, and I use a logic probe on the input to two mosfet drivers

Bob Lawrence (VE1RLL) · 2014-03-28 22:27

but it is so flaky on my board.

Which part of your program doesn't work? Serial comms? XBee comms/ bad data etc, Please provide more details.

Help, Crystal issues

Comments