I find every time that I go to manually configure the Uno timers I have to keep flicking back and forwards through various pages in the datasheet, in particular for the timer mode which is split across two registers. I've made up some charts which summarize the registers on one page.
Thumbnails below, click on them to get the full-size image:
If you spot any errors, please let me know.
These images are licensed under a Creative Commons Attribution 3.0 Australia License.
Thanks. He does much better charts than me, though. I must find out how he does it. 
Thanks! Very useful. I'll bet the folks on AVRFreaks would appreciate them, too.
Hello. I also am trying to program the Timer 1 through their records , to operate an Arduino Uno as an incremental encoder.
My problems come when TIFR1 write to the registry to try to reset to 0 once the TOV1 Bit Overflow occurs . This allowed to write to that record ? I can change the bit TOV1 individually?
Thanks in advance and sorry for my English , is a translator
Hello. I also am trying to program the Timer 1 through their records , to operate an Arduino Uno as an incremental encoder.
My problems come when TIFR1 write to the registry to try to reset to 0 once the TOV1 Bit Overflow occurs . This allowed to write to that record ? I can change the bit TOV1 individually?
Thanks in advance and sorry for my English , is a translator
Please start a new post with your question, and include your code.
Nick--
Thank you so much for this. I have been driving myself crazy navigating through the data sheet(and the registers).
Do you know why Atmel did not make bits 2 and 3 of TCCRxA R/W instead of R and place WGMx2 and WGM13 there?
No idea. It seems an odd design decision in light of the fact that those two bits are spare. I can only guess that if there was a reason, it was to do with the way internal routings on the chip are, or for some sort of timing reason.
Bits seem to end up in odd locations for backwards compatibility and to avoid conflicts. There was very likely an earlier processor that used bits 2 & 3 in the A register but not bit 3 in B. By doing that sort of thing they reduce the chance of failure if a customer is not careful when transitioning from an older processor to a newer one. Better to have a 5000 horsepower engine not run optimally than to have it reduced to slag.
cattledog:
Nick--Thank you so much for this. I have been driving myself crazy navigating through the data sheet(and the registers).
Do you know why Atmel did not make bits 2 and 3 of TCCRxA R/W instead of R and place WGMx2 and WGM13 there?
These things are always a sign of later extensions added in a new processor that is constrained to
be back-compatible with previous chips. Additions and improvements to the timer units were no
doubt added after the original register layout was defined. Do you redesign the unit and break
existing code or add the functionality in a non-disruptive way which is messier?
Nice, thanks for that Nick!
Thanks! I've amended the charts a bit (you might need to refresh this page, or the linked page) to add in the bits for the timer compare output bits.