Slew Rate Limiting Option - Teensy 4.1

[PGPH]

As shown in the attached picture, I have generated with a teensy 4.1 pulses of about 125 nanoseconds which show suprisingly quite high initial and final oscillations.

To test the impact of the SRL, I have then tried without success to Disable/Enable the option with the following instructions (commented in the here below initial code):
CORE_PIN2_CONFIG &= ~PORT_PCR_SRE;
CORE_PIN30_CONFIG |= PORT_PCR_SRE;

Compilation of the modified code stopped with following error message:
'PORT_PCR_SRE' was not declared in this scope
17 | CORE_PIN2_CONFIG &= ~PORT_PCR_SRE;
| ^~~~~~~~~~~~
exit status 1
Compilation error: 'PORT_PCR_SRE' was not declared in this scope

In an attempt to fix this, a few libraries where included in the code, but did not solve the problem.

I am curently stuck!

Thank you in advance for helping me on this subject.

___________________________________________

#include <Wire.h>
#include <WireIMXRT.h>
#include <WireKinetis.h>
#include <TeensyThreads.h>
#include <buffer.h>
#include <color.h>
#include <gfxfont.h>
#include <_Teensy.h>
#include <imxrt.h>

void setup()
{
  pinMode(2, OUTPUT);

  // Disable SRL
  // PORT_PCR_SRE set to 0 (default).
  // CORE_PIN2_CONFIG &= ~PORT_PCR_SRE;

  // Enable SRL
  // CORE_PIN30_CONFIG |= PORT_PCR_SRE;

  test();
}

void test()
{
  while(true)
  {
    digitalWriteFast(2, HIGH);
    delayNanoseconds(100);
    digitalWriteFast(2, LOW);
    delayNanoseconds(400);
  }
}

void loop()
{
}

 

[joepasquariello]

Compilation error: 'PORT_PCR_SRE' was not declared in this scope

I don't see PORT_PCR_SRE defined anywhere in the T4 core. When I search for SRE, I do see IOMUXC_PAD_SRE, so perhaps that is what you're looking for.

 

[PGPH]

Thank you Joe.

Based on your suggestion, I tried the following command line to disable the SRL:
CORE_PIN2_CONFIG &= ~IOMUXC_PAD_SRE;

Supprisingly, the pulses disappear completely!

Any idea as to why?

 

[PaulStoffregen]

What happens if you just go with the pinMode() alone, no special writing to the IOMUXC registers?

 

[PGPH]

Hello Paul.
Without writing to the IOMUXC register (the command line I just mentioned) I get the pulses shown before.

 

[MarkT]

Are you not using a 10:1 low impedance probe (450 ohms at the probe tip, 50 ohm load at 'scope end of the cable?

Like figure 3 here, but with 450 or 470 ohm tip resistor (then you can use 10x setting on the scope): https://www.sigcon.com/Pubs/straight/probes.htm

 

[PGPH]

Thank you MarkT for your suggestion to make the connection cable to the scope a resistive probe.
I understand you imply that this could help filtering out the high frequency oscillations.
I will certainly do this later, but for the time being I am trying to see the effect of the Slew-Rate-Limiting option.

I already noticed in the past that inserting a 220 ohms resistor just after the Teensy improved significantly the signal, which made me believe that the oscillations were part of the generated signal from the Teensy.

Since my last post, I have tested the following change, which had surprisingly still no impact on the pulse shape:
replace
CORE_PIN2_CONFIG &= ~IOMUXC_PAD_SRE;
by
IOMUXC_SW_PAD_CTL_PAD_GPIO_B0_12 &= ~IOMUXC_PAD_SRE;

 

[MarkT]

If there's an unterminated cable it will resonate all be itself - resistance will damp any resonances. But for full bandwidth you have to terminate with a matched load.

 

[PGPH]

Some further investigation to check the impact of the Slew-Rate-Limiting option on the pulse shape.

I found that Pin 2 native name on Teensy 4.1 is not B0_12, as I inittially thought, but EMC_04 (source is KurtE's pinout-chart referred to on PJRC Web site).

With the help of ChatGpt , I also found how to manipulate the PAD register.
Here is what I did to get the pulse shape shown on the attached picture.

  // Set all bits of the PAD Register to mofify the options, including SRL/SRE. 
  IOMUXC_SW_PAD_CTL_PAD_GPIO_EMC_04 =
  (0 << 16) |   // Bit#16     HYS = 0    (Hysteresis : OFF)
  (0 << 14) |   // Bit#15-14  PUS = 0    (Pull-UP / Pull-Down : 100 kohms pull-down)
  (0 << 13) |   // Bit#13     PUE = 0    (Pull Selection : KEEPER)
  (0 << 12) |   // Bit#12     PKE = 0    (Pull Keeper : OFF)
  (0 << 11) |   // Bit#11     ODE = 0    (Open-Drain : PUSH-PULL)
  (0 << 6)  |   // Bit#7-6    SPEED = 00 (Commutation Speed : 50 Mhz)
  (2 << 3)  |   // Bit#5-3    DSE = 010  (Drive Strength : ~130 ohms)
  (0 << 0)  ;   // Bit#0      SRE = 0    (Slew Rate Limiting : LOW)

After experimenting on different bit settings, I concuded the following:
1) By default, all bits are set to 0 , except DSE bits which are 111 (~37 ohms).
2) SRE set to 0 or 1 doesn't change anything.
3) DSE is the only parameter that has an impact; decreasing its value reduces signal oscillations/ringing.

Any comment is welcome.

 

[NewLinuxFan]

If I'm following this correctly, you added that code after pinMode(2, OUTPUT); which enabled 2 internal resistors?