'level shifting' is a nuisance as there are too many cases of needing to interface 3.3v to 5v and vice versa.
while looking around for 'solutions' I found this little rather interesting logic gate
74lvc1g17
https://www.ti.com/lit/ds/symlink/sn74lvc1g17.pdf
https://assets.nexperia.com/documents/d ... VC1G17.pdf
https://www.diodes.com/assets/Datasheets/74LVC1G17.pdf

apparently, both TI and Nexperia (and Diodes) makes them (probably more), and aliexpress vendors seemed to stock both part numbers quite commonly.
this apparently can be used for translating between 3.3v - 5v upwards.
but do note that the upwards trip point is close to 3v when running off 5v vcc.

I'd think shifting down is similar in a sense that it can take 5v inputs and run off 3.3v VDD
so the output would be 3.3v.

quite a handy chip

(note: an 'old' way of interfacing 3.3 - 5v on the output is 74HC*T* series, the HC series won't do it, 74HCT can translate 3.3v output to 5v
but that I think quite a few datasheets for 74HCT states a min VDD of 4.5v, hence translating down is an issue)

these chips matter are they are practically 5v logic level mosfets 'drivers', and many '(stepper) motor drivers' chips has a 5v specs, so do WS2812B Neopixels etc.

 

It really depends upon the logic series and standards in use, for example when going from a 3V3 output to a 5V input it's ok if the 5V part is TTL as the logic '1' threshold is low enough (2.4V AIRC) but not if the 5V part is CMOS where the threshold is 90% of vcc. Translators also loose speed of course if that is important.

 

a trouble I'd guess is many designs these days are CMOS even for fairly 'simple' chips such as motor drivers
and my guess is simply because CMOS normally sports a rather 'flexible' VDD requirements.
this makes it 'easier' to make LDOs, voltage doublers etc that runs off batteries but do not assure a 'fixed' VCC.

 

I recall that the CD4009 and CD 4010 were logic level shifter ICs many years ago.. At the time they were called "Hex-buffers", with the CD4009 having six inverters while the CD4010 provided six level shifting buffers. So those devices may provide the required level shifts.

 

the other one of my favourite 'level shifter' is actually 74lvc1g17
https://www.nexperia.com/products/a...ers/schmitt-trigger-ics/series/74LVC1G17.html
https://assets.nexperia.com/documents/data-sheet/74LVC1G17.pdf
https://www.ti.com/product/SN74LVC1G17
https://www.ti.com/lit/ds/symlink/sn74lvc1g17.pdf
apparently Ti has a better part

and 74LVC1G14 inverting
https://www.nexperia.com/products/a...-transceivers/inverters/series/74LVC1G14.html
https://assets.nexperia.com/documents/data-sheet/74LVC1G14.pdf
https://www.ti.com/product/SN74LVC1G14
https://www.ti.com/lit/ds/symlink/sn74lvc1g14.pdf
TI parts seemed to have a lower trigger \[ V_{T(+)} \] threshold.
The good thing about the LVC is that it is a full cmos gate hence push-pull, possibly 'much' faster than using a mosfet and a pull up resistor.
It'd seemed it is possible to use the 74lvc to convert from 5v to 3.3v and vice versa

 

I used it to translate 3.3V for WS2815B (5V logic) and it worked fine. Mine are from Diodes Inc.

 

At sufficiently low frequencies (1-2MHz), a logic level N channel MOSFET and 2 10k resistors can be used as a bi-directional level shifter. SparkFun sells a quad channel version for $3.50 (you can get them on AliExpress for about a tenth of that, but you risk having them handled improperly).
https://www.sparkfun.com/products/12009