I'm making a FSR drum shield for the leonardo, it's wireless, will be using Xbees.
For sake of not having to translate 5V TTL down to the xbee's 3.3V, I'm going to swap out the VReg on the board to a 3.3V.
I know my way around AVR Studio enough that this will not be a problem with fuse bits and all that but I'm looking at the schematic for the board and wondering about all this 5V Autoselect jive and wondering if there might be a trace I need to cut or something or if this will not be a problem.
Are you lowering the clock speed as well?
Figure 29-2 of the datasheet shows that ~4.5V is needed for guaranteed operation at 16 MHz.
(vs 3.8V for a '328)
You bring up an excellent point. Before their U suffix became the norm I never liked 16Mhz for stability purposes anyway. I'll pop that crystal off and change that out while I'm at it.
For sake of not having to translate 5V TTL down to the xbee's 3.3V, I'm going to swap out the VReg on the board to a 3.3V.
An XBee shield that does that seems like a much simpler solution. The pin spacing on the XBee is not compatible with the Arduino or breadboard, so you need something between the XBee and the Arduino anyway.
For sake of not having to translate 5V TTL down to the xbee's 3.3V, I'm going to swap out the VReg on the board to a 3.3V.
An XBee shield that does that seems like a much simpler solution. The pin spacing on the XBee is not compatible with the Arduino or breadboard, so you need something between the XBee and the Arduino anyway.
2mm pitch does not concern me, however, the size of an entire shield does. I have 2mm protoboard, 10 Pos 2mm female headers and xbee socket breakout boards at my disposal, but after I proto this up it's going to be a board spin shield with it's own Xbee socket.
FWIW, I have a board where I can jumper Vdd for either 3.3V or 5V, and I've
never had a problem running Duemilanove or UNO chips with a 16-Mhz xtal.
I think the specs may be a little conservative. I suppose it might not run so
well at -40C or 150C.
Or at all
Even the automotive '328 parts are only spec'ed to run at +125C.
Regular chips only to +85C.
I don't see an automotive temperature rated ATMega32U4.
The Vcc curve requires 4.5V for 16 MHz.
Its not the same as a '328 where the parts can do 20 MHz.
oric_dan(333):
FWIW, I have a board where I can jumper Vdd for either 3.3V or 5V, and I've
never had a problem running Duemilanove or UNO chips with a 16-Mhz xtal.
I think the specs may be a little conservative. I suppose it might not run so
well at -40C or 150C.
If you check the table on page 308 of the datasheet under "Speed Grades" for the Atmega328 you'll see you're right on the edge of stable for 16Mhz at VCC of 3.3V. CrossRoads is giving good advice here.
Works right now isn't good enough, from personal experience troubleshooting crystal instability is extremely frustrating.
If you check the table on page 308 of the datasheet under "Speed Grades" for the Atmega328 you'll see you're right on the edge of stable for 16Mhz at VCC of 3.3V. CrossRoads is giving good advice here.
Now I'm kind of intriqued by all of this. I interpolated Figure 28-1 of my 328P d/s
[note - Table 28-2 is 'not' applicable], and it indicates the 328P can run at 13.3Mhz
with Vcc=3.3V. Is the chip suddenly going to croak if I use 16-Mhz xtal, or maybe
just run a wee tad warmer? Or will it give up completely.
At this point, I guess my best course of action is just keep using it at 3.3V with
16-Mhz xtal, and watching for the day it stops. People have been overclocking
chips like this for years.
