Teensy 4.1 vs. ESP32


Peter Boreland
 

I was wonder what the relative advantages of the Teensy 4.1 processor has over the ESP32.

For instance, does the greater processing power affect:

1. Sidereal tracking accuracy?
2. Slewing rates?
3. Step to step accuracy?
4. Ability to microstep without losing steps (16 seems to the the upper limit)?
5. Relative performance of pulse vs square  #define STEP_WAVE_FORM? (Pulse does not loose steps while square does at 16 microsteps using the MaxESP32)?
6.Other things?

Peter


Howard Dutton
 

On Thu, Nov 24, 2022 at 03:04 AM, Peter Boreland wrote:
1. Sidereal tracking accuracy?
No.

2. Slewing rates?
I'd call it Step rates, but basically, Yes.  Only matters in practice if you can't microstep mode switch.

3. Step to step accuracy?
Yes, but I'd say you'll never notice it in use except that slews might be a bit rougher (which can limit speeds) on an ESP32... especially if WiFi or Bluetooth are enabled.

4. Ability to microstep without losing steps (16 seems to the the upper limit)?
5. Relative performance of pulse vs square  #define STEP_WAVE_FORM? (Pulse does not loose steps while square does at 16 microsteps using the MaxESP32)?
These fall into the categories of misconceptions and/or bugs that may or may not have already been fixed.

I'll add that responsiveness in OnStepX should be better with a T4.1 than ESP32 as commands are processed faster.
Another example the sidereal clock ticks are 5x faster on a T4.1 (at the 0.015 arc-second scale.)
And the motion controller supervisor operates at 5x shorter timing intervals so it's more responsive as well (rate changes for guiding, acceleration, etc.)

Really don't get me wrong the ESP32 is good enough to make an excellent controller, but it is not a T4.1


George Cushing
 

Speed, more UART channels and GPIO pins.


Howard Dutton
 

On Thu, Nov 24, 2022 at 09:24 AM, George Cushing wrote:
Speed, more UART channels and GPIO pins.
Double precision FPU.  Wonderfully stable libraries.


Peter Boreland
 

Thanks Howard for the detailed response.

Would you be kind enough just to say a bit more about "And the motion controller supervisor operates at 5x shorter timing intervals so it's more responsive as well (rate changes for guiding, acceleration, etc.)" : rate changes for guiding.

Peter