You are getting close to the maximum. How about not sending the spaces? Even then you will be pushing it to condense your data into 11 bytes.
A more reasonable approach might be to do some calculations and send an average every second, or a delta (eg. rather than the temperature, the change in temperature).
I managed to get 460800 baud with the uno. Above that, I would just see garbage.
You could also use Serial.write rather than print so you just print 2 bytes rather than somewhere between one and six (and you'd save another because you wouldn't need a deliminator). Then you'd have to write a program on your computer to interpret it, though
I optimized my analogRead to get the maximum speed out of it.
This page (near the bottom) shows you how to get 77 kHz reading speed out of analog read. So the analog read times is not a concern to me.
@WizenedEE
How did you manage to read that speed on the serial monitor of PC?
I would separate the reading of sensors from communicating their values.
i.e. use the ISR just to communicate the most recent sensor readings.
Then do the reading of sensors somewhere else.
Store the sensor values in some volatile globals and then use the ISR to slam
out the most recent readings.
Either that or depending on how fast the data readings really change, only communicate
the changes to the data vs the actual data.
This could dramatically reduce the amount of data transfered.
But since you have not described the overall system and its real needs or what
else might be going on in the foreground it is pretty difficult to make suggestions.
The big concern I have over using a serial port interface like this particularly at
a high data rate is that there does not appear to be any sync sequence or checksum on the data.
So if there is ever dropped character because of noise on the line, the two
ends would be hopelessly out of sync with each other and the receiver would
be "seeing" totally garbage information.
@ Nick Gammon:
Thank you for the info on your website.
The temperature change of my system is like this: room temp to 100 deg C and back down to room temp in under 500 ms... as you can see I really need to capture and send this data.
@ bperrybap:
I did what you suggested in other way around: readings in ISR and sending in the loop. I will definitely experiment with your suggestion.