W1209: problems connecting half duplex to sensor connectors

[TabbieSuperb]

Hi Thomas!
First of all thank you for your incredible work, you have opened for a newbie like me a whole new world.
Sorry if those are noob questions but I think having some feedback from me can be interesting.
I've successfully flashed W1209 by soldering some pins to the unpopulated swim interface and using the usb dongle, with LED display showing a wonderfull "4th". After that, I've removed the C1 capacitor. Next, I have built with a breadboard the half duplex circuit that you kindly published, and correctly wired it to a working ttl dongle (very happy console user of 4th in the STM8 discovery board), but most of my keyboard presses are translated to zeros and some cursor chars. I think console speed may be too fast for my W1209, because when I change speeds to 2400 I start to see correct chars. However, I suspect W1209 firmware is currenty only working at 9600.
Could you please give me some ideas about what to do/test/build?
Thank you very much.

[TG9541]

Hi TabbieSuperb!

Great to hear that it almost worked for you!

Let's check if your half-duplex circuit works with the SWIMCOM image:

1. flash SWIMCOM.ihx
2. connect your circuit to pin2 (PD1/SWIM) and pin4 (GND) of the ICP connector
3. try again with 9600/8-n-1

Edit: also please tell the type of your serial adapter (I'd like to learn what works best).
... and on a 2nd thought: it's trivial to change the serial port to 2400 baud (changing the TIM4 prescaler is all it takes):

         TIM4RELOAD = 0xCF       ; reload 0.104 ms (9600 baud)
         MOV     TIM4_ARR,#TIM4RELOAD
-        MOV     TIM4_PSCR,#0x03 ; prescaler 1/8
+        MOV     TIM4_PSCR,#0x05 ; prescaler 1/32

Here is a binary for testing:
W1209_2400baud.zip

By the way, which kind of STM8 discovery board do you use - is it the one with the STM8S105C6T6?

[TabbieSuperb]

Thank you Thomas!
For your reference, my ultra cheap serial adapter as identified by linux
[ 2636.719707] usb 1-1: New USB device found, idVendor=10c4, idProduct=ea60
[ 2636.719712] usb 1-1: New USB device strings: Mfr=1, Product=2, SerialNumber=3
[ 2636.719716] usb 1-1: Product: CP2102 USB to UART Bridge Controller
[ 2636.719718] usb 1-1: Manufacturer: Silicon Labs
[ 2636.719720] usb 1-1: SerialNumber: 0001
[ 2636.728066] cp210x 1-1:1.0: cp210x converter detected
[ 2636.729641] usb 1-1: cp210x converter now attached to ttyUSB0
The "discovery board" that I'm using is this one https://es.aliexpress.com/item/J34-Free-Shipping-ARM-STM8-Development-Board-Minimum-System-Board-STM8S103F3P6-Module-for-Arduino/32351992629.html?spm=2114.13010608.0.0.MPsrwT, which is a STM8S003F3P6. It works great. I love the relay of the cheap W1209 and I wanted a closer to it STM8 version for testing...

Circuit works: I've flashed my W1209 to SWIMCOM and I have a working 9600 console wired to SWIM pin! Thank you!

However, after flashing to the W1209_2400 version still no luck with the sensor's pins. May be it is just broken after a bad removal of C1, who knows. I will continue testing it.

Now for the programmable voltage regulator!!! ;)

[TG9541]

Hi TabbieSuperb,

the problem with your W1209 is interesting - maybe there is a short circuit. You can test it from "within" by reading the ADC using the SWIMCOM image:

: ain   ( -- )   6 ADC! ADC@ . ;

I've been thinking about changing the W1209 to a SWIMCOM interface for a while because this would make developing code that uses the analog port much easier, and also modifying the board would be less of a trouble. The W1209 uses PD1/SWIM to a segment of the LED display, but the display is "common anode", and might be able to work around it. I'll reopen this ticket, and give it a try.

Good luck with the DCDC - I'm looking forward to hearing about it!

[TabbieSuperb]

After running some "ain":
: ain ( -- ) 6 ADC! ADC@ . ; 1023 ok 1022 okain 1022 okain 1023 okain 1023 okain 1021 okain 1023 okain
I've tested for shorts, didn't find anything. I've cleaned up some excess solder on the board, no luck with sensors. Stranger things! :)

[TG9541]

When you connect the NTC sensor the ADC value should be 350-ish (and drop when the sensor gets warmer). That would confirm that the circuit works normally, and make this thing even more mysterious ;-)

By the way, I've been playing a bit with a serial interface solution for the W1209 ICP with (mostly) working LED display. It works nicely at 3.3V but I had no luck with 5V operation so far.

Edit: combining LED display and serial interface won't work without wither injecting a small current (which would activate the LED), or a modification of the hardware. The reason is the LED common anode resistor (hidden below the 7S-LED display) which is a bit to low for 5V operation. Activating a column (anode) GPIO triggers a com interrupt, even if multiple segments are active). A possible work-around is the replacement of the 5V voltage regulator AMS1117-5.0 with a 3.3V regulator AMS1117-3.3.

[TG9541]

One more thing about your USB serial interface converter:
I've never tested the CP2102 with the half-duplex hack. I often use the CH340, but its parameters vary a lot. I found that PL2303 clones work well. If you have access to a scope it would be great if you could have a closer look at the signals. What you can also try is powering the W1209 through the ICP header using the ST-LINK adapter (without applying 12V to the W1209 power terminals).

[TabbieSuperb]

As you stated, but with may be 10ºC temperature?
531 ok
531 okain
531 okain
489 ok
475 okain
463 okain

I've myself tested SWIMCOM using +12v in board mains and by the ICP header with CP2102 providing +5V and +3V.
I think the simplest solution is to do as instructed by you from the beginning and buy a CH340, and just try! I will try with the scope if somebody led me one while it arrives... ;)
Thanks for everything!

[TG9541]

The circuit seems to be OK, I guess that the signal levels of the hacked com port are to blame. Better get a PL2303 clone instead of a CH340.

It's also possible to do something "completely different" and communicate e.g. through the GPIO of key "-". In order to make waiting for the new USB adapter easier, I quickly made a W1209 binary with half-duplex communication through PC5 ("-" key) for you:

W1209km.zip

If you'd like to reproduce this, here is the changed part of the globalconf.inc file:

        PSIM     = PORTC        ; Port for UART simulation
        HAS_TXSIM        = 1    ; Enable TxD via GPIO/TIM4, word TXGP!
        PNTX             = 5    ; Port GPIO# for HAS_TXDSIM
        HAS_RXSIM        = 1    ; Enable RxD via GPIO/TIM4, word ?RXGP
        PNRX             = 5    ; Port GPIO# for HAS_RXDSIM

By the way, the "-" key still works, as you can see with the help of the following code:

FILE
: keytest ?KEY IF EMIT THEN ;
' keytest BG ! 
HAND

[TabbieSuperb]

Totally awesome! It works as you designed on the minus key! Thank you very much! I've made a photo with the rehacked cable... ;) Best wishes! http://es.tinypic.com/r/wvbp5j/9

[TG9541]

Cool! Great that it works. Maybe I'll uses this as the default method 👍

[TG9541]

Please refer to issue #42 for a different approach.