[HUB75]Leds do not light up as expected

[ehime-iyokan]

I wrote program using hub75 package and raspberry pi pico. but, it didn't light up. So I have tried to fix hub75.
As a result. It light up as expected. And Environmental information is here. -> https://github.com/ehime-iyokan/tinygo_hub75_sample

First, I added a time.sleep(1 * time.Microsecond) after latch-High.
By doing so, the light came on. However, I don't know why, so if you know, please let me know.
Moreover, the red lights are on in different places.

Second, I fixed buffering of red data.
I guess it has to do left-shift first.
d.buffer[c][offsetR] = d.buffer[c][offsetR] &^ 1 << bitSelect
-> d.buffer[c][offsetR] &^= 1 << bitSelect

[ehime-iyokan]

I added a time.sleep(1 * time.Microsecond) after latch-High. By doing so, the light came on. However, I don't know why
If I figured out why it worked correctly, I'll create a pull request.

[sago35]

@ehime-iyokan
The GPIO of the RP2040 might be too fast. Without adding time.Sleep(), the pulse width is 110ns, which might be too short. With time.Sleep(), it is around 5us.

I think the fix you are considering is necessary.

with time.Sleep(1 * time.Microsecond) : 5.046us

without time.Sleep(1 * time.Microsecond) : 110ns

[sago35]

Including other parts as well, it is better to use delay.Sleep() for microsecond sleeps. I think it would be good if you could create a Pull Request with that change.

https://github.com/tinygo-org/drivers/blob/v0.28.0/delay/sleep.go

with delay.Sleep(1 * time.Microsecond) : 1.16us

[ehime-iyokan]

I created pullrequest. => (#722)

Second, I fixed buffering of red data.
I guess it has to do left-shift first.
d.buffer[c][offsetR] = d.buffer[c][offsetR] &^ 1 << bitSelect
-> d.buffer[c][offsetR] &^= 1 << bitSelect

(#691 (comment))

[ehime-iyokan]

Thank you for providing the information. @sago35
For now, I've decided to use a microcontroller board other than the RP2040. Next time I use the RP2040, I’ll consider creating a pull request.