Hi all - I've got a 3.3v 1-Wire device (DS28E05P, multiple) that needs to communicate on a 5v 1-Wire bus. Would the following circuit work well for this?
Is it overkill and should use a voltage or capacitor divider as a better solution? I don't have any in front of me until next week so I can't just test it.
According to the DS28E05 datasheet the maximum pullup is 1.5K
If your controller can tolerate the minimum logic high of the 3V3 1-wire device and you are not using strong pull-up in the 1-wire library it might work without any other interface.
You would still need the 3V3 power source.
According to the DS28E05 datasheet the maximum pullup is 1.5K
Ah thanks I missed that.
You would still need the 3V3 power source.
Yes I was hoping to avoid the need for a 3.3v power source, hence my comment about using a voltage divider. Perhaps this is a better route to take, as I was reading the BSS138 might be unreliable at higher frequencies (though this 1-wire device isn't particularly fast). Would be nice to find a level shifter that could take in a 5vdc signal line, and output a 3.3vdc signal, I don't know if such a thing exists.
well apparently it does exist! The TXB0104 used in this Sparkfun adapter can take in a 5VDC supply and a programmable output voltage of 3.3v data signal - bidirectionally. The 1 channel version TBX101 seems like what I'd need. It's less only $0.40, making it cheaper than a AZ1117-3.3v + SN74LV1T34DBV. Anyone used it before? I don't really see a downside unless I'm missing something in the datasheet
What is the minimum Vih of the controller you are using?
I'm using a esp32, so 3.3, however I'm using a DS28E05P DS2484R+T 1-Wire master which translates the 1-Wire bus to 5v for distance purposes.
EDIT: I'm actually not sure what VIH refers to, thought you meant supply voltage.
This device can only translate push-pull CMOS logic
outputs. If for open-drain signal translation, see TI TXS010X products.
Thank you I wasn't aware of that detail, totally would have chosen the TXB and have realized my mistake a week later 
I'm not sure the TXS would work either, just know the TXB would not.
For 1 wire devices that get their power from the dataline the current is delivered through the pullup. If you have multiple devices like you said you would the pullup may need to be even smaller than 1.5k
Level converter?
got it. I suppose I should order some parts and actually test this before designing a board. Was hoping it would be simple enough to whip out a quick board, but since they are parasitically powered I'm not sure what the roadblocks were going to be. I've never used 1-wire before.
Can you not use a DS2485 1-wire master that runs off 3v3?
I was referring to the minimum voltage that is recognized as a logic 1 for your controller before I knew you were using a 5V 1-wire master.
The DS28E05 is an EEEPROM it doesn't translate anything. Whats this 5V translate for longer distance stuff?
1 wire was developed by Dallas Semi which was bought by Maxim which was bought by Analog Devices. Along the way many of the app notes seem to have disappeared
I done for today
The DS28E05 is an EEEPROM it doesn't translate anything. Whats this 5V translate for longer distance stuff?
Ack my bad, copy-paste error. The 1-wire master is a DS2484R+T. The DS28E05 is an identification device with built-in eeprom that is being used and there are many on the 1-wire bus.
The DS2484R+T has a built-in level shifter which is being powered by a 5V supply, I went this route to increase the distance and since there was already a 5V supply as well on the same cable being run.
I believe the DS2484 will run off 1.71V to 5.25V. Is 3v3 simply not an option?
Now I see
If you read that app note you will see that designing a large 1-wire network is not so simple.
5V won't necessarily gain you distance and your idea with the translator defeats the whole purpose of 1 wire since you now need to distribute 5V with the 1 wire signal.
Maybe 1 wire was not the best choice for your idea.