Minimum Pulse-Width Resolution for nRF52840 PWM Module

RE: Minimum Pulse-Width Resolution for nRF52840 PWM Module

Kirk Wayland — Fri, 01 Aug 2025 21:04:10 GMT

According to the nRF52840 Errata list, at 400Khz operation, the clock is at 1.25 usec. But this is too short and violates the spec of 1.3 usec. Only workaround is to slow the frequency to 390Khz, but this means the standard macros in the SDK cannot be used.

Don't forget about sensing the SDA line to see if it is stuck low. If it is, you have to clock the I2C until it is no longer stuck (1 to 9 clocks).
So, GPIO to sense line, then enable and cfg TWIM, and the use the bus. This must be done for every transfer.

RE: Minimum Pulse-Width Resolution for nRF52840 PWM Module

johnwest — Wed, 15 Dec 2021 22:32:03 GMT

@hmolesworth,

Sure - assuming a 50% PDC (positive duty cycle) if the GPIO hardware allows this and indeed the minimum period is defined as 1 / (base clock) - still would be nice to see this called out in detail in the specs.

Thanks,
John

RE: Minimum Pulse-Width Resolution for nRF52840 PWM Module

hmolesworth — Wed, 15 Dec 2021 22:18:52 GMT

Welcome, but 125nSec period or 62.5nSec pulsewidth :-)

RE: Minimum Pulse-Width Resolution for nRF52840 PWM Module

johnwest — Wed, 15 Dec 2021 21:20:41 GMT

hmolesworth,

OK - so 125 nS then,

Thanks,
John

RE: Minimum Pulse-Width Resolution for nRF52840 PWM Module

hmolesworth — Wed, 15 Dec 2021 21:14:43 GMT

No idea, but with a x1 divider each step is 1/16MHz; the 16MHz is fixed in stone (well, silicon), no way of changing that. To get 8MHz PWM at least 4 steps are required 0-1-0-1 as period (COUNTERTOP) cannot be less than 3. A COUNTERTOP value of 4 with the sequence 0-1-0-1 gives 8MHz PWM signal.

RE: Minimum Pulse-Width Resolution for nRF52840 PWM Module

johnwest — Wed, 15 Dec 2021 21:03:08 GMT

So, min pulse width is 1 / (base clock) - is that in the spec someplace?

RE: Minimum Pulse-Width Resolution for nRF52840 PWM Module

hmolesworth — Wed, 15 Dec 2021 20:29:20 GMT

16MHz clock gives 62.5 nSec minimum pulse width.

RE: Minimum Pulse-Width Resolution for nRF52840 PWM Module

johnwest — Wed, 15 Dec 2021 18:59:02 GMT

haakonsh,

So, is the minimum pulse width 1uS? 1nS? 1pS?

Thanks,
johnwest

RE: Minimum Pulse-Width Resolution for nRF52840 PWM Module

haakonsh — Wed, 15 Dec 2021 09:57:59 GMT

I'm not quite sure what exactly you are asking for, so I'll just list the relevant timing specs:

16MHz base clock
1x - 128x PRESCALER (16MHz - 125kHz)
15-bit COUNTERTOP value (3 - 32767)
15-bit compare values (1 - 32767)

Full specs: PWM — Pulse width modulation