Check protocol for FRED 1.4

Check Protokoll for FRED 1.4 Testprogram

First English version 5/2/99.

to FRED Homepage

This test software for FRED exists since version 1.4. It is to be written into the 16F84 of FRED. The goal is

Test the hardware components step by step.
Determine the limits of the A/D conversion, which result in callibration values to be written into the data flash memory.

For the test you need

A pull-up for the Loconet (a command station will do, but there has to be no trafic on the net)
A Loconet monitor (I use the one Martin Pischky wrote).
A programmer hardware, see here: http://www.jdm.homepage.dk/newpic.htm
Programmer sofware FCALIB (from Martin Pischky)

Step by step proceeding

Compile FTEST_F.ASM (or use the HEX file shipped with FCalib)
Program the hex file in the FRED (for example with FCalib)
Plug the FRED into Loconet and observe the LEDs and the trafic on the net. The tables below show the steps that are performed by the software step by step.
After FRED has completed all tests successfully, determine the limits of the A/D conversion.
Compile FRED_F.ASM (or use the HEX file shipped with FCalib)
Program the hex file and the calculated callibration values into FRED (for example with FCalib)
Verify normal operation of FRED.

Test steps

1. Reset-condition: immediately after power up

Event	Explanation	Removal
LED lights up immediately, Loconet is idle	ok	-
LED stays dark	PIC does not run or LED not connected properly	Check oscilator and reset circuit, check LED and resistors.
Net-LED of Chief turns off and Loconet-Monitor repors framing error	Short on the Loconet or transmit transistor is open all the time	Check transmit transistor. It might be nescessary to cut conductors to encircle the fault. Normal condition: Q1 off, collector is on Loconet-High-Level.

2. LED-Test: First three seconds

Event	Explanation	Removal
LEDs flash red (short), green (long), off, red, green, off, ...	that's the way it should be	-
Flashing sequence twisted: green, red, off, green, red, off, ...	LED soldered twisted	Guess what?
Only red or green flashes.	Something is not soldered properly.	Check LED and resistor soldering.
Flashes yellow	Short cut...	Beat up the guy who soldered.

3. Loconet-idle-test: 1 second

Event	Explanation	Removal
LED is 1s yellow	perfect	-
stays yellow for eternity (>>1s)	carrier detect faulty	Check hardware carrier detect. Should be that way: R13 on Loconet-High (no Trafic!!!!!) PIC Pin 9 (RB3=loconet_rx) on high C2 empty PIC Pin 6 (RB0/INT=loconet_cd) on high
LED goes red and stayes red	Loconet-receive sticks to 0V	Check PIC-Pin 9 (RB3=loconet_rx), should be high

4. CD idle ok: 1 second green

5. Measuring CD-time: LED is yellow for a short time

Event	Explanation	Removal
LED goes red and stayes red	loconet_rx does not go low within 4us after loconet_tx gets high	Check transmit transistor Should be this way: PIC Pin 11 (RB3=loconet_tx) on high Q1 durchgeschaltet PIC Pin 9 (RB3=loconet_rx) on low

6. CD-Measuring finished: 1 second green

7. Loconet transmit: FRED sends measured CD-time

Now there should happen a few things:

The LED flashes red/green.
FRED transmits a message [A8 2A xx cc]
xx is the measured CD time in steps of 10us. This value should be close to 100, this means 1ms. xx=0 means Time-Out
cc is the corresponding xor-Check-Byte
After that there should be this messages every half a second [AF xx yy cc]
xx are the least significant 7 bits of adc_raw
yy are the most significant 3 bits of adc_raw
cc is the corresponding xor-Check-Byte
Whenever a function key or the emergency stop key is pressed or released or you toggle the direction switch, you should see the message [A8 2B xx cc]
xx is the key event as defined in button.inc
cc is the xor-Check-Byte

Event

Explanation

Removal

DC-time=1

C1 not connected

check C1

CD-Zeit st not near to 1ms=100*10us

RC-circuit not correct

check C2,R8
if the value is in the interval 100..130, it's the compounents tollerances => replace R8 with 110k or 100k (CD time of about 107 is the upper limit)

LED does not flash (sticks to red or green), no messages

Programm might hang in ADC-routine; Pressing Fc or ESTOP releases the hang for a moment.

Check ADC-circuit. Should be:

C1 full
Pin 4 of comparator on high
Pin 2 of comparator on low
Pin 10 of PIC=RB4 on low

ADC-Werte <30 oder >1000

RC circuit faulty

Check R2,R1,C1

No messages for keys/switch

Keys not soldered properly???

Key	Pin-name	Pin-#
F_A	RB1	7
F_B	RB2	8
F_C	RB6	12
ESTOP	RB7	13
Direction	RA3	2