Ouch, that's hot!

Michael Caisse

So... obviously I have missed something. I have looked at the FAQ and browsed through the forum, but I'm missing something.

I have an SX28 hooked up to an SX Tech Board.

I have loaded several examples and tried them with the SXKey and all works fine.

I tried the ADC8.SXB example and it works fine with the SXKey installed. I modified the code to light an LED connected to RC.7 (through a 470 ohm resistor) when the "ADC" value is above 128. I have also modified the DEVICE and FREQ lines to read:

DEVICE SX28, OSC1MHZ, TURBO, STACKX, OPTIONX
FREQ 1_000_000

When I use the device dialog to read back the configuration it indicates that the oscillator is set to "Internal 1MHz".

I press the Ctrl-P and program the device and it works as expected: turn the pot and the light goes on or off.

If I unplug the power and then plug it back in, the device does the expected and the program runs.

If I unplug the power, remove the SXKey, then plug power back in.... the SX chip gets blazing hot. In fact... if you ignore it is getting hot you get to replace it with another chip after it melts down and lets the smoke out.

My first thought of course is that there is no clock from the SXKey and therefore no program running and the device has the pin (in/out) configured incorrectly. There are several problems with this first thought of course.... the first of which is I have set the device for an internal clock.

So.... any ideas from the experienced? I started with two chips and prefer not to toast my last one (o;

Clock Loop

Provide a schematic of your entire circuit. Chances are very good your overlooking a connection. The sx shouldn't even get warm. I leave my sx-key disconnected from my sx chip all the time, with no crystal, it doesn't heat up.

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Meh. Nothing here, move along.

Michael Caisse

So I thought I should just back up and verify something very simple.

I'm running the LED28.src file as is. The "schematic" is just as described in the source code comments: LED adode to RC.7, cathode to 470 ohm resister and other resistor leg to groud.

I do a contol-P and it loads and runs. The LED is flashing away.

I then power off the board, remove the SXKey and power the board up. No flashy LED and the SX device gets way hot.

Just to clarify, the only items I have plugged into the SX Tech Board are:

• SX28 Chip
• LED connected to the RC.7 header and the protoboard
• 470 ohm resistor connected to the protoboard and Vss
• Power to the VDC input
  

For grins I removed the two components (LED and resistor) and the SX chip gets hot when powered up without the SXKey.

Now for the fun one. This will be the comment that will jog somebody's memory (I just wish it was mine). If I put the components back like before (LED and resistor), plug the SXKey in, plug in the power and watch the LED flash.... and then remove the SXKey with the power on.... it keeps flashing and doesn't get hot.

It is acting like it is in some initial power-up state that the SXKey is able to get it past. Like... maybe when the data sheet says you don't need any external components when using the internal clock it really means don't let it float. I had a funny problem like this with a Coldfire 5282 and an "unused" pin once.

All I can say is that I must be pathetic 'cause I haven't seen any other posts will people having issues with such a simple test.


Thanks for the help!

;=======================================================================
;TITLE:         led28.src
;
;PURPOSE:       Simple Blink-an-LED example.
;
;AUTHOR:        Parallax, Inc.
;
;REVISIONS:
;  09/17/04 -   Updated to support SASM.
;
;CONNECTIONS:
;  RC.7 to LED anode.
;  LED cathode to 470 ohm resister, other leg of resistor to ground.
;
;DETAILS:
;  RC.7 is toggled high and low with a simple delay loop in-between.
;  NOTE: To simplify code, the entire RC port output latch is inverted
;  by this routine.
;=======================================================================


;-------------------------- DEVICE DIRECTIVES --------------------------

        DEVICE        SX28,OSC1MHZ,TURBO

IFDEF    __SASM  ;SASM Directives
        DEVICE        STACKX,OPTIONX
        IRC_CAL        IRC_SLOW
        
ELSE        ;Parallax Assember Directives
        DEVICE        STACKX_OPTIONX
ENDIF

        RESET        Initialize

;------------------------------ VARIABLES ------------------------------
            
        Count1    EQU    $08
        Count2    EQU    $09

;---------------------------- DEBUG SETTINGS ---------------------------

        FREQ     1_000_000
    
        WATCH    Count1,16,UDEC

;------------------------ INITIALIZATION ROUTINE -----------------------

Initialize
        ;Configure port settings
        mov    rc, #%00000000        ;Port C output zero
        mov    !rc,#%01111111        ;Port C.bit7 output dir.

;---------------------------- MAIN PROGRAM -----------------------------

Main
;Toggle RC.7 with a delay of 64k counts in-between.

        call    delay            ;delay    
        mov    W,/RC            ;toggle all of RC port
        mov    RC,W                
        jmp    Main            ;goto main                


Delay        clr    Count1          ;Initialize Count1, Count2    
        clr    Count2                                

Loop        djnz    Count1,loop        ;Decrement until all are zero        
        djnz    Count2,loop                            
        ret                ;then return        

Chris Savage

Hello,
·
·· What happens if you put a 1000uF capacitor (watch polarity) in between the V_SS and V_DD sockets on the board and try the same thing?· Does it still get hot?


▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Chris Savage
Parallax Tech Support

Michael Caisse

Thanks Chris. That did the trick.

PJAllen

So, what was going on here, a latch-up?

Chris Savage

Yes, I believe this is a latch-up situation, which is why I suggested additional filtering on the supply.

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Chris Savage
Parallax Tech Support

Clock Loop

Interesting, wouldn't a properly made power supply fix this?

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Meh. Nothing here, move along.

LoopyByteloose

When in doubt, try powering off a battery. Very smooth input. That might cool it down.
There is a condition where you can actual cycle the reset from a low power input problem, but I am not sure it would create a lot of heat at the SX.

Is there a solder bridge or a dead short on other I/O? Sometimes the exposed backplane is sitting on a wire or metal surface.

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
"If you want more fiber, eat the package.· Not enough?· Eat the manual."········

···················· Tropical regards,····· G. Herzog [noparse][[/noparse]·黃鶴 ]·in Taiwan

Michael Caisse

The chip must have latched up. Adding just a little more capacitance across the power of the SX Tech Board does the trick. Of course, the real question should be "Why doesn't the SX Tech Board already have a sufficient capacitor."

The final project (an artificial neuron that my daughter was working on) runs off of 3AA bateries. All works fine and she is a happy camper.

LoopyByteloose

My guess is low voltage.
It is not really latched up. It is looping between on and off. The capacitor allows for lower operation without the voltage swings.
You could change the Brown out configuration to a lower voltage, but don't turn it off [noparse][[/noparse]looping occurs unless you have a really good supply].

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
"If you want more fiber, eat the package.· Not enough?· Eat the manual."········

···················· Tropical regards,····· G. Herzog [noparse][[/noparse]·黃鶴 ]·in Taiwan