RM46L852PGE GIOB[2] input always high

Dmitri Snejko

Other Parts Discussed in Thread: HALCOGEN

Hi,

I am using GIOB[2] as output mapped to default pin 142. Whatever is written to DOUT the input DIN register is reading always 1. Looks like a silicone bug. Also the reference manual spnu514.pdf on page 230 in Table 4-1 describes PINMMR29[16] register only responsible for GIOB[2] pin multiplexing. The following note below the table doesn't make any sense in this case:

• The input to a module comes from the dedicated pad when the condition [ not(A) or (A AND B) ] = TRUE.
• The input to a module comes from the multiplexed pin when the condition [ A and not(B) ] = TRUE.

since A and B are the same: PINMMR29[16]

Regards, Dmitri.

over 13 years ago

0 Anthony F. Seely over 13 years ago

TI__Guru 69070 points

Dmitri,

Sorry this is an obvious quesiton, but did you program the pin DIR register to an ouput (1)?

-Anthony

0 Dmitri Snejko over 13 years ago in reply to Anthony F. Seely

Intellectual 745 points

Hi Anthony,

Yes I did. I have a led attached and I see the actual pin state is changing. Not DIN state. DIN is always 1 for bit 2. I have also another 2 leds attached to GIOA[6] and GIOA[7] configured as outputs as well. These two bits are matching in DIN and DOUT.

Dmitri.

0 Anthony F. Seely over 13 years ago in reply to Dmitri Snejko

TI__Guru 69070 points

Dmitri,

I think you alluded to this in your other post http://e2e.ti.com/support/microcontrollers/hercules/f/312/t/262132.aspx.

Probably, the LED current is high enough that even though you have the pin configured as an output, the voltage doesn't get below VIT- for the input buffer, which is usually somewhere in the ~1.3V neighborhood. In this case even if DOUT is set to drive '0' the pin may read as a '1'.

Would be interesting to get a voltage reading on the LED pin when you are trying to drive low to confirm this.

Best Regards,

Anthony

0 Dmitri Snejko over 13 years ago in reply to Anthony F. Seely

Intellectual 745 points

No, the led current ( < 1mA ) is fine and not different from the others: high state is 3.3V, low state is 20mV which is way below VIL (0.8V)

Dmitri

0 Sunil Oak over 13 years ago

TI__Mastermind 49120 points

Hi Dmitri,

Pin 142 is the default input for the GIOB[2] signal. You do not need to configure any multiplexing control register for being able to use this pin as the GIOB[2] input.

PINMMR29[16] is '0' by default, and this selects the GIOB[2] input from pin 142. You need to set this bit in order to take in the GIOB[2] input from pin 55.

Can you please confirm the configuration of PINMMR29[16]?

Regards,

Sunil

0 Dmitri Snejko over 13 years ago in reply to Sunil Oak

Intellectual 745 points

Hi Sunil,

I was thinking this way as well. This is how I came to this confusion statement from the ref manual I posted in the original message. My original PINMMR29[16] was set to 1 (it was the way how HalcoGen configured pin 142). After I changed it to 0. No difference found. In both configurations I am reading 1 form the pin. For my application it is not a big deal (don't need the input) but it looks suspicious.

Dmitri.

0 Sunil Oak over 13 years ago in reply to Dmitri Snejko

TI__Mastermind 49120 points

Hi Dmitri,

Did you remember to unlock the write permission to the PINMMR control registers before trying to clear the PINMMR29[16]?

To enable write access to the PINMMRs, the CPU must write 0x83e70b13 to the kick0 register (0xFFFFEA38) followed by a write of 0x95a4f1e0 to the kick1 register (0xFFFFEA3C).

Regards, Sunil

0 Dmitri Snejko over 13 years ago in reply to Sunil Oak

Intellectual 745 points

Actually if I read PINMMR29[16] in the debugger it is always 0.

Dmitri.

0 Dmitri Snejko over 13 years ago in reply to Sunil Oak

Intellectual 745 points

Yes there is the sequence.

Dmitri.

0 Sunil Oak over 13 years ago in reply to Dmitri Snejko

TI__Mastermind 49120 points

Can you please paste a snapshot of the GIOB port control registers? You can get this from the CCS "View --> Registers" command and will look like the one below.

Regards,

Sunil

0 Dmitri Snejko over 13 years ago in reply to Sunil Oak

Intellectual 745 points

Sure.

Pin written 0:

Pin written 1:

PINMMR29:

Dmitri

0 Dmitri Snejko over 13 years ago in reply to Sunil Oak

Intellectual 745 points

Hi Sunil,

I found the problem source. Actually it is opposite: PINMMR29[16] should be set to 1 to be able to read it from pin 142. PINMMR29[16] in my case was overwritten in latter code so it was always 0 no matter that it was set to.

Thanks, Dmitri.

0 Sunil Oak over 13 years ago in reply to Dmitri Snejko

TI__Mastermind 49120 points

Hi Dmitri,

Please refer to the erratum number DEVICE#70 on page 24 in the document at: http://www.ti.com/lit/er/spnz187/spnz187.pdf

The PINMMR29[16] implementation is being corrected from RM46 revision A to revision B. I have a revision B part, hence the opposite information.

Please note this change when you receive the revision B of the RM46x MCUs.

Regards, Sunil

0 Dmitri Snejko over 13 years ago in reply to Sunil Oak

Intellectual 745 points

Hi Sunil,

How can I detect which silicon rev is running?

Dmitri.

0 Sunil Oak over 13 years ago in reply to Dmitri Snejko

TI__Mastermind 49120 points

Hi Dmitri,

You can read the silicon revision (bits 7 to 3) in the device id register at address 0xFFFFFFF0. The device id for revision A reads 0x8046AD05.

Revision B would have a device id of 0x8046AD15.

Regards, Sunil

0 Dmitri Snejko over 13 years ago in reply to Sunil Oak

Intellectual 745 points

Thanks Sunil,

Dmitri

0 Andy Kay over 11 years ago in reply to Dmitri Snejko

Intellectual 765 points

Hi there,

this may also help:

In my case, I could setup HalCoGen (V3.5.2) to set PINMMR29[16] to 1:

HalCoGen -> PINMUX -> Special Pin Muxing
-> "Use GIOB_2, for disabling selected HET2 PWM outputs" = YES

Regards, Andy

Arm-based microcontrollers

Arm-based microcontrollers forum

RM46L852PGE GIOB[2] input always high