Sammy C21 support for 2 TMC2209 drivers with UART com

[dc42]

PS - lines 1482-1484 are currently:

#if TMC22xx_HAS_MUX
	SetUartMux();
#endif

I think you may need instead:

#if TMC22xx_HAS_MUX
	SetUartMux();
#elif TMC22xx_USE_SLAVEADDR
	delay(2);							// give the previous TMC22xx driver time to get off the bus
#endif

[Gixxerfast]

@dc42

constexpr uint8_t TMC22xxSercomNumber = 3;
Sercom * const SERCOM_TMC22xx = SERCOM3;

constexpr Pin TMC22xxSercomTxPin = PortAPin(22);
constexpr GpioPinFunction TMC22xxSercomTxPinPeriphMode = GpioPinFunction::C;
constexpr Pin TMC22xxSercomRxPin = PortAPin(20);
constexpr GpioPinFunction TMC22xxSercomRxPinPeriphMode = GpioPinFunction::D;
constexpr uint8_t TMC22xxSercomRxPad = 2;
constexpr IRQn TMC22xxSercomIRQn = SERCOM3_IRQn;

I really hope I haven't messed up here.

Also, earlier I had to change this one too:

 inline void TmcDriverState::UartTmcHandler() noexcept
 {
-#if !(TMC22xx_HAS_MUX || TMC22xx_SINGLE_DRIVER)
+#if !(TMC22xx_HAS_MUX || TMC22xx_SINGLE_DRIVER || TMC22xx_USE_SLAVEADDR)
 # if TMC22xx_USES_SERCOM
        DmacManager::DisableCompletedInterrupt(TmcRxDmaChannel);
 # else

[dc42]

@gixxerfast yes that change to UartTmcHandler is correct.

The stepsticks need to be wired such that DRIVES_UART is routed to pin 14 of the TMC2209 chip as it is on the tool board. The MS1 and MS2 pins of the chip need to be both grounded for driver 0, whereas for driver 1, MS1 should be connected to +3.3V and MS2 grounded as you have illustrated.

Did you see my message about the SetUartMux call?

[Gixxerfast]

@dc42 said in Sammy C21 support for 2 TMC2209 drivers with UART com:

Did you see my message about the SetUartMux call?

Yes, if you mean this I have added that:

#if TMC22xx_HAS_MUX
	SetUartMux();
#elif TMC22xx_USE_SLAVEADDR
	delay(2);							// give the previous TMC22xx driver time to get off the bus
#endif

OK, I'll check the pin on the TMC2209 chip, but I think it should be alright.
EDIT: Yes, DRIVES_UART (via PDN_UART) is connected to pin 14 on the TMC2209 chip.

On the first step stick both MS1 and MS2 are connected to ground.

[dc42]

@gixxerfast lines 569 and 626 should also have || TMC22xx_USE_SLAVEADDR appended.

[Gixxerfast]

@dc42 OK thanks, done. Behaviour unchanged though.

[dc42]

@gixxerfast have you disabled I2C support in your SAMMYC21.h file? That also uses SERCOM3.

[Gixxerfast]

@dc42 Yes, I have. There are pin conlicts as well IIRC.

[dc42]

@gixxerfast please share your SAMMYC21.h file. Have you changed any other files except TMC22xx.cpp ?

If you set the number of drivers to 1, set TMC22xx_SINGLE_DRIVER to 1 and TMC22xx_USE_SLAVEADDR to 0, are you able to control driver 0?

[Gixxerfast]

@dc42 Sorry for the delay, I was just testing running it as a single driver but unfortunately it didn't work any better.
So either I have a fundamental hardware fault (even though i measured that all the lines goes to where they are supposed to according to the schematics) or the configuration is somehow still wrong.

My initial tests with step/dir worked so that part should be OK, but not the sercom/uart part.

I'll attach the SAMMYC21.h file (renamed to fool the forum software)

SAMMYC21.h.stp

Thanks for your help.

Forgot about the changed files. These are the files I have changed:

$ git status
On branch 3.4-dev
Your branch is up to date with 'origin/3.4-dev'.

Changes not staged for commit:
  (use "git add <file>..." to update what will be committed)
  (use "git restore <file>..." to discard changes in working directory)
        modified:   src/Config/SAMMYC21.h
        modified:   src/Movement/StepperDrivers/TMC22xx.cpp
        modified:   src/Platform.cpp
        modified:   src/Tasks.cpp

Untracked files:
  (use "git add <file>..." to include in what will be committed)
        src/Platform.cpp.save

[dc42]

@gixxerfast two suggestions:

1. It's not going to work without the voltage monitoring if you power up 5V before VIN, because the drivers won't respond to the UART data until VIN is present. You may be able to get round this by sending M999 B124 to restart the board after powering it up.

2. I suggest you concentrate on getting it working with a single driver and the UART to start with, because that configuration is known to work on the tool board.

What changes did you need to make to Platform.cpp ?

[dc42]

PS - in your SAMMYC21.h file you should not need the line "constexpr IRQn TMC22xxSercomIRQn = SERCOM3_IRQn;". Also you should enable stall detection in that file, to match the configuration used by working TMC2209 configurations.

[Gixxerfast]

@dc42 The board has a LMP78 regulator so VIN should be there at the same time of possibly before 5V if not the 100uF capacitors delays things.

Here's the changes I made in Platform.cpp to make it compile.

I agree, should continue with one driver and see

[Gixxerfast]

@dc42 Hi again. One thing before I put this aside for a while.

I see now that you (1LC and mini5+) pull PDN_UART high whilst the step sticks pulls it low.

That suggest to me that I might have to invert the signal to make it work. I might be barking up the wrong tree here but is that possible to do?

1LC:

BTT TMC 2209 Stepstick

[Gixxerfast]

@dc42 Yes!!! Win

Well, sort of anyway. I removed the pulldown resistor on the step stick and added this ugly 20k pullup resistor and now it works anyway.

2022-01-30 22:46:11	M122 B124
Diagnostics for board 124:
Duet SAMMYC21 firmware version 3.4.0beta7+8 (2022-01-30 19:28:27)
Bootloader ID: not available
Never used RAM 4808, free system stack 2745 words
Tasks: Move(notifyWait,0.0%,127) HEAT(notifyWait,0.0%,117) CanAsync(notifyWait,0.0%,65) CanRecv(notifyWait,0.0%,74) CanClock(notifyWait,0.0%,65) TMC(delaying,1.9%,49) MAIN(running,97.8%,471) IDLE(ready,0.0%,40) AIN(delaying,0.3%,154), total 100.0%
Last reset 00:05:31 ago, cause: power up
Last software reset time unknown, reason: HardFault, available RAM 224, slot 2
Software reset code 0x0060 ICSR 0x00000003 SP 0x20007f00 Task MAIN Freestk 6126 ok
Stack: 20001048 200037b4 ff0030e8 200030e0 00000000 00017cf5 00018a28 0100000f 00000000 20003064 20001d70 a5a5a5a5 a5a5a5a5 a5a5a5a5 a5a5a5a5 00000000 20001d40 00000000 a5a5a5a5 00018307 00000000 00017471 00000000 fffffffd 20003058 00000000 00000002
Driver 0: pos 16000, 160.0 steps/mm,stalled, standstill, SG min 0, read errors 0, write errors 0, ifcnt 14, reads 6229, writes 2, timeouts 0, DMA errors 0, CC errors 0, steps req 16000 done 16000
Moves scheduled 1, completed 1, in progress 0, hiccups 0, step errors 0, maxPrep 401, maxOverdue 0, maxInc 0, mcErrs 0, gcmErrs 0
Peak sync jitter -1/12, peak Rx sync delay 200, resyncs 0/0, no step interrupt scheduled
MCU temperature: min 24.5C, current 26.8C, max 26.8C
Last sensors broadcast 0x00000000 found 0 114 ticks ago, 0 ordering errs, loop time 0
CAN messages queued 2327, send timeouts 0, received 3743, lost 0, free buffers 37, min 37, error reg 0
dup 0, oos 0/0/0/0, bm 0, wbm 0, rxMotionDelay 300, adv 37058/37058

2022-01-30 22:50:56	m569.2 P124.0 R1
Register 0x01 value 0x00000001

Now, I don't know what to do about this later, can't remove resistors from step sticks and running jumpers.

[dc42]

@gixxerfast try using a pullup resistor of 4.7K or 3.3K. That might be enough to defeat the two 20K pulldown resistors without causing a problem for the Tx signal passing through the 1K series resistor.

[Gixxerfast]

@dc42 Thank you for the suggestion.

I added a 3.3K pullup and it seems to work great..

I have now two working motors with UART control.

Thank you for taking an interest in this and for all the help.

(Now, on to the next problem, I need to get a servo motor)

2022-01-31 11:57:21	M122 B124
Diagnostics for board 124:
Duet SAMMYC21 firmware version 3.4.0beta7+8 (2022-01-31 11:32:38)
Bootloader ID: not available
Never used RAM 192, free system stack 3843 words
Tasks: Move(notifyWait,0.0%,107) HEAT(notifyWait,0.1%,117) CanAsync(notifyWait,0.0%,64) CanRecv(notifyWait,0.0%,73) CanClock(notifyWait,0.0%,64) TMC(delaying,3.0%,45) MAIN(running,96.6%,490) IDLE(ready,0.0%,41) AIN(delaying,0.3%,154), total 100.0%
Last reset 00:10:25 ago, cause: software
Last software reset time unknown, reason: HardFault, available RAM 224, slot 2
Software reset code 0x0060 ICSR 0x00000003 SP 0x20007f00 Task MAIN Freestk 6126 ok
Stack: 20001048 200037b4 ff0030e8 200030e0 00000000 00017cf5 00018a28 0100000f 00000000 20003064 20001d70 a5a5a5a5 a5a5a5a5 a5a5a5a5 a5a5a5a5 00000000 20001d40 00000000 a5a5a5a5 00018307 00000000 00017471 00000000 fffffffd 20003058 00000000 00000002
Driver 0: pos 112000, 160.0 steps/mm,stalled, standstill, SG min 0, read errors 0, write errors 1, ifcnt 28, reads 25370, writes 14, timeouts 0, DMA errors 0, CC errors 0, steps req 112000 done 112000
Driver 1: pos 96000, 160.0 steps/mm,stalled, standstill, SG min 0, read errors 0, write errors 1, ifcnt 26, reads 25370, writes 13, timeouts 0, DMA errors 0, CC errors 0, steps req 96000 done 96000
Moves scheduled 7, completed 7, in progress 0, hiccups 0, step errors 0, maxPrep 695, maxOverdue 0, maxInc 0, mcErrs 0, gcmErrs 0
Peak sync jitter -1/11, peak Rx sync delay 210, resyncs 0/0, no step interrupt scheduled
MCU temperature: min 25.8C, current 26.5C, max 26.6C
Last sensors broadcast 0x00000000 found 0 119 ticks ago, 0 ordering errs, loop time 0
CAN messages queued 5042, send timeouts 0, received 8123, lost 0, free buffers 37, min 37, error reg 0
dup 0, oos 0/0/0/0, bm 0, wbm 0, rxMotionDelay 426, adv 36934/37057

[Gixxerfast]

@dc42 said in Sammy C21 support for 2 TMC2209 drivers with UART com:

It's not going to work without the voltage monitoring if you power up 5V before VIN, because the drivers won't respond to the UART data until VIN is present. You may be able to get round this by sending M999 B124 to restart the board after powering it up.

Hi again. I'm exploring the voltage monitoring but as expected I run into trouble.

As seen above the Vin value is way to large (should be 24 V ish).

The only AD-port left is PB08 and I'm using that for voltage monitoring and as I am running the Sammy with 3.3V I'm using this in the configuration:

Edit: Tried the PA03 with the same result

I measure 1.7V at the monitoring point which matches the calculation with () the top resistor at 61Kohm and the bottom at 4.7Kohm.

(4.7*24)/61+4.7) = 1.71V

So, now I'm left with the question if PB08 is unsuitable for measuring the voltage?
I also see that PB08 is supplied from VDDANA whilst the PA08 is supplied from VDDIO.

Any clue to why I don't get the correct reading?

Also, why is there so little memory left compared to the 1LC? They should have the same amount of memory as it's the same MCU (ATSAMC21G18A).

The firmware binary for the Sammy is small 118 kB and the 1LC binary slightly larger at 129kB

[Gixxerfast]

@DC42 Might be of interest.
I have now temporarily replaced the Sammy C21 with a Seeduino XIAO extending Klipper.

It has no problem communicating via UART with the TMC2209 without any pullup resistor (and with the 20k pulldown intact).

I'm going to look at the UART signal later. The Duet UART transmission without the pullup is shifted somehow. I can see that it's transmitted but it's sort of shifted so that the startbits comes last in the start byte instead of first. I'll add some pictures from the logic analyzer tomorrow.

22:00:07
SG_RESULT: 00000000
22:00:07
PWM_AUTO: 000e0024 pwm_ofs_auto=36 pwm_grad_auto=14
22:00:07
PWM_SCALE: 0000001d pwm_scale_sum=29
22:00:07
PWMCONF: c80d0e24 pwm_ofs=36 pwm_grad=14 pwm_freq=1 pwm_autoscale=1 pwm_autograd=1 pwm_reg=8 pwm_lim=12
22:00:07
DRV_STATUS: c0190000 cs_actual=25 stealth=1 stst=1
22:00:07
CHOPCONF: 34030053 toff=3 hstrt=5 tbl=2 vsense=1 mres=4(16usteps) intpol=1 dedge=1
22:00:07
MSCURACT: 00f7000c cur_a=12 cur_b=247
22:00:07
MSCNT: 00000008 mscnt=8
22:00:07
TSTEP: 000fffff tstep=1048575
22:00:07
FACTORY_CONF: 0000000c fclktrim=12
22:00:07
IOIN: 21000041 enn=1 pdn_uart=1 version=0x21
22:00:07
OTP_READ: 0000000c otp_fclktrim=12
22:00:07
IFCNT: 00000008 ifcnt=8
22:00:07
GSTAT: 00000001 reset=1(Reset)
22:00:07
GCONF: 000001c0 pdn_disable=1 mstep_reg_select=1 multistep_filt=1
22:00:07
========== Queried registers ==========
22:00:07
SGTHRS: 00000000
22:00:07
TPOWERDOWN: 00000014 tpowerdown=20
22:00:07
TPWMTHRS: 00000003 tpwmthrs=3
22:00:07
IHOLD_IRUN: 00081919 ihold=25 irun=25 iholddelay=8
22:00:07
SLAVECONF: 00000200 senddelay=2
22:00:07
========== Write-only registers ==========
22:00:07
DUMP_TMC STEPPER=gear_stepper

With Duet and Sammy C21 and no pullup on UART-line:

With Klipper and Seeeduino XIAO and no pullup:

[dc42]

@gixxerfast I think the reason is that in the TMC22xx driver on the SAMC21 we disable the transmitter for a short while in function TmcDriverState::StartTransfer. My guess is that this sets the output to high impedance. You could probably rewrite it to just reset the receiver, if you also defer enabling the transmit DMA until the very end.