making sense of DC Characteristics (or: 5V USART @ 3.3V?)

[Pavius]

A project i'm working on uses a 3.3V USB to serial device (CP2101) connected directly to a PIC16F887 running on 5V. The PIC's Tx line will be 5V, which as i understand can be tolerated by the CP2101. The CP2101's Tx line, however, will be 3.3V. I tried to see if this is tolerated by the PIC and all i can gather is that the VIH (Input High Voltage) range for TTL buffered I/O ports is 2.0 - VDD (page 245 of the datasheet). This puts this well within range.

1) Am i reading the datasheet correctly?
2) Will the PIC's USART function properly if the Rx pin receives 3.3V?
3) Any other insights (the PIC can't run @ 3.3 due to other limitations and the project manager still insists on using the CP2101 due to familiarity on the PC side)

 

[Mr RB]

Hi Pavius. Check out the 16F887 datasheet page 56 (58) it shows the internal workings of the USART RX pin RC7.

It shows the input is a schidt trigger type (see the little _//~ symbol inside the input buffer device). RC7 is not a TTL input. They don't use TTL inputs much on PICs these days, and USART RX is one place they will always use schmidt trigger input anyway to clean up noisy serial data.

Now see page 251 (253) for table 17.5 the input high voltage for schmidt trigger inputs is specced at 0.8*Vdd. So that's 4v.

I can't say for the CP2101 but i've had success interfacing other devices with 3.3v outputs to PIC schmidt inputs by just attaching a 10k pullup resistor to +5v. The 3.3v devices use output FETs that have integral protect diodes so using a 10k pullup resistor pulls them about 0.7v above the 3.3v rail and got reliable triggering. Just check the CP2101 specs and make sure you are not sourcing too much current to the 3.3v rail to pull it up, a 10k resistor from 5v->4.0v sources about 100uA which won't hurt the 3.3v device output FET, but it might raise the 3.3v rail if it draws less than 100uA.

 

[blueroomelectronics]

The FT232R will work with a 5V CPU.

 

[Pavius]

Thanks a ton, RB. I did notice that schmidt trigger inputs were indeed differently spec'ed than TTL'd ones but didn't know where to search to know what type the input was. Hadn't thought about looking at the internal workings diagram (always been a software guy and that part rarely helped, except for looking at muxes along the way of a signal).

Bill, i have already suggested the FT232 (has a very wide operation voltage). He wanted to stick with the CP2101 because he made many projects with it and is familiar with the PC side aspects (drivers, etc).

Thanks again!

 

[ericgibbs]

Thanks a ton, RB. I did notice that schmidt trigger inputs were indeed differently spec'ed than TTL'd ones but didn't know where to search to know what type the input was. Hadn't thought about looking at the internal workings diagram (always been a software guy and that part rarely helped, except for looking at muxes along the way of a signal).

Bill, i have already suggested the FT232 (has a very wide operation voltage). He wanted to stick with the CP2101 because he made many projects with it and is familiar with the PC side aspects (drivers, etc).

Thanks again!

hi,
Some tips in this pdf are useful.

 

[Pavius]

Thanks, Eric. Some good reading for the weekend