as a first project i try to plan an IMU. An ATmega328 should do the data processing of the sensors (adxl345, itg-3200, hmc5883). With the help of an n-channal Mosfet the XBEE can be powered on for wirless data transmission instead of using the ftdi cable.
The 2 different red planes are 3.3V planes (one for xbee & Atmega, the 2nd for the sensors).
The blue planes are connected ground planes.
I read some tutorials before planing und routing the board. Now i hope to get comments to improve the layout and my knowledge.
I have now rerouted the board. The ground planes are now on the top layer, the power planes on the buttom layer. I think it is easier to rout/decouple the power pins?!
Hami:
I have now rerouted the board. The ground planes are now on the top layer, the power planes on the buttom layer. I think it is easier to rout/decouple the power pins?!
Are there any other bugs?
Thanks for reading this post!
Hami
[attachment=0]brd2.png[/attachment]
You cannot route a trace around the center pad on your gyro. The datasheet asks for it to not be soldered at all. So you shouldn't have a pad on the PCB, but you should also not route any traces in that area due to the risk of shorting (soldermask really shouldn't be counted on in a place like that).
Your power connector seems to have some mounting pads that you’re just leaving floating. I’d recommend either grounding them (if you can) or putting some vias on them to give them some more mechanical strength.
A cap on your reset line would be good too, unless you plan on using debugwire.
I didn’t do a full review - these were just the things that stood out to me.
NleahciM: Thank you verry much. I read the datasheet of the itg3200, but i did not remember that.
I will correct this. Furthermore i will add the cap.
NleahciM:
Your power connector seems to have some mounting pads that you’re just leaving floating. I’d recommend either grounding them (if you can) or putting some vias on them to give them some more mechanical strength.
Do you mean the battery connector? shoul i put vias on the left pads? I am sorry, i did not understand. Will they not be solderd otherwise?
NleahciM:
Your power connector seems to have some mounting pads that you’re just leaving floating. I’d recommend either grounding them (if you can) or putting some vias on them to give them some more mechanical strength.
Do you mean the battery connector? shoul i put vias on the left pads? I am sorry, i did not understand. Will they not be solderd otherwise?
The connector on the lower left side of the PCB is what I am referring to. It has two surface mount pads that do not appear to be connected to anything else. Thus my assumption about them is that they're some sort of mechanical mounting pads that are designed to give the connector more strength. Problem is, the lamination between copper and FR-4 is not very strong and those pads can peel of very easily. Adding vias to connect them to the bottom layer will dramatically increase their strength.
Hami:
I am now searching for the correct cap types. Hopefully this is my last step.
Is a tantal cap ok for decoupling the ldo and the battery?
I have to use low esr-caps with the magnetometer. I will use ceramic caps. Are ceramic caps ok for dcoulping power supply, reset-cap, etc?
For the resistors (Pull ups, etc.), i want to use thick film resistors. Would this be ok?
Thanks!
Hami
Ceramics are good for decoupling. They will also work well for your reset line. I would suggest a large tantalum (ie 10uF) in parallel with a small ceramic (ie 100nF) for your battery, especially if it connects with long leads (low ESR of ceramics can cause ringing in your battery leads when you connect the battery - this can destroy components). Hard to say about the LDOs - ceramics are normally preferred but I'd check the datasheet in case they call out anything in particular. It also looks like you're missing a bypass cap on U2.
NleahciM:
Hard to say about the LDOs - ceramics are normally preferred but I’d check the datasheet in case they call out anything in particular.
I am sorry, i should have wrote. I use the micrel Mic 5205LDO. I did not found somethink obout the esr of the input cap. Therefore i will use a tantalum. (And i have learned about ringing leads ;-) ).
Furtermore they propose Caps with an ESR of 5 Ohm or less as ouput cap. I know that cermic have a much lower ESR. They write, that “Ultra-low-ESR”-Caps can cause the oscillation. So i think i will use tantalums instead of cermic caps (next time i will have done my homework better…). By reading the datasheet the 5th time, i realised, that i probably should reduce the capacitance of the Output-cap of U4, cause of the low values of the output current to earn more stability.
NleahciM:
It also looks like you’re missing a bypass cap on U2.
I choose two different designs from the datasheet (micrel MIC5205 LDO): "Low noice [...] " and "Ultra-Low noice [...]" design (p. 9). I thougth this is ok for just powering the xbee and the atmega, is it not?
Are thick film resistors OK for the use in this project?