I'm working on an arduino clone with built in protoboard area. Can someone double check my schematic and let me know if there's any glaring errors? I have it at
Also, I'd appreciate any suggestions about anything I should add to the board.
Thanks.
(at a very first glance)
hope this helps
C
Thank you for your reply.
I completely missed the reset pin pullup.
I was definately planning to add the decoupling capacitor, but I might well have forgotten, so thanks :).
As far as the regulator I'm planning to use a 7805 which doesn't really need caps (the datasheet says they're optional, I didn't realize they were mentioned at all until a week ago). I was considering allowing for LDO regulators by adding space for 2 caps on the output which I just wouldn't have to install with a 7805.
As far as the regulator I'm planning to use a 7805 which doesn't really need caps (the datasheet says they're optional, I didn't realize they were mentioned at all until a week ago). I was considering allowing for LDO regulators by adding space for 2 caps on the output which I just wouldn't have to install with a 7805.
From my experience, I never used a regulator without at least the output capacitor... if its Vout has to power "stuff" all around the board (and the proto area you are planning to build in) I would definitely place it.
For the input cap (usually I place it too) it depends on the stability of the Vin. But again depends on application, this is why "the datasheet says they're optional"
good luck
C
The Atmel recommendations for the reset pin are a 10K pullup in parallel
with a diode (cathode to Vcc). They also suggest a cap to gnd for noisy environments.
The cap to ground will not work with the FTDI reset trick.
Atmel also recommends ADC filtering using a 10uH inducttor and decoupling cap.
The FTDI reset is done with a 0,1uF ceramic cap in series between #RST and /RESET.
The schematic for my board is in the datasheet at Loading...
The only change I need to make is the /RESET diode and the optional reset cap
to GND.
(* jcl *)
For the FTDI reset, do you mean the 0.1uF cap goes from RTS# on the FTDI cable to pin 1 of the ATmega168? Right now I'm not connecting to that pin or CTS# on the cable. I've looked at the BBB and Boarduino schematics and both seem to do it as I did in my schematic, though it looks like something is missing as far as the reset.
I thought about the inductor. I was considering leaving space for a jumper or coil to be installed but was leaning towards leaving it connected permanently to make assembly simpler. I'm not sure how much this hurts the ADC accuracy though.
You are correct on the #RTS.
Check my schematic but I believe I have #CTS tied to GND.
The inductor is about the same size as a resistor. The part number I use
is in my datasheet. I would add it and use a jumper if you don't need it.
The important thing is to keep the caps close to the uC. I would use a ground
plane and surround the clock with a ground ring.
An excellent application on grounding and decoupling is
"An IC Amplifier User¡Çs Guide to Decoupling, Grounding,
and Making Things Go Right for a Change" by Paul Brokaw
http://www.analog.com/static/imported-files/application_notes/135208865AN-202.pdf
(* jcl *)
The ADC on the ATmega performs pretty poorly at best.
I second Brokaw's excellent app note. I see it has been updated since its original publication.
The ADC on the ATmega performs pretty poorly at best.
In what ways? I filter the AVcc, use a ground plane and keep the loop area
on all decoupling caps small. On some designs I add a ground ring around
all ADC channels. I have not noticed any problems.
The impedance of the ADC inputs is low so some signals could require
a buffer.
(* jcl *)
But it is only guaranteed to be good to 8 bits. As a 10 bit converter, I think it's a waste of a perfectly good integer. 
Since it uses 62.5% of the unsigned integer I am going with the glass half full
(* jcl *)
That's a good attitude.
Thanks everyone for your help.
To summarize (and make sure I haven't missed anything), I'm adding:
a pull-up resistor to the reset pin
a 0.1 uF cap from RTS# on the FTDI cable to the reset pin
a decoupling cap near the power pins on the MCU (under the socket?)
1 cap on the input, 2 on the output of the voltage reglator
space for a coil or jumper between Vcc and AVcc
As far as the diode with the cathode to Vcc, anode to the reset pin, I'm not clear what it's for. It seems like some sort of over-voltage protection so if anything higher than Vcc ends up on the reset pin it gets shunted to the power rail. But why?
jluciani: I will be using a 2-layer PCB, so I don't think I'll be able to run a ground plane around the clock. I was going to have the crystal next to the MCU and the caps on the other side of the crystal. Are you suggesting the caps go even closer?
over-voltage protection so if anything higher than Vcc ends up on the reset pin it gets shunted to the power rail. But why?
In some types of programmers the reset line can be connected to +12V for short pulses. I wonder if this is to protect the system against this happening and accidentally corrupting the flash.
As far as the diode with the cathode to Vcc, anode to the reset pin, I'm not clear what it's for. It seems like some sort of over-voltage protection so if anything higher than Vcc ends up on the reset pin it gets shunted to the power rail. But why?
It is an over-voltage clamp. Since the devices that connect to the reset pin
typically go open-circuit you can get a transient that exceeds the Vcc.
The reset cap from the FTDI chip generates quite a nice transient 
Fortunately
there is not a lot of energy stored in the cap and the 10K helps to clamp it. I would
include the diode.
jluciani: I will be using a 2-layer PCB, so I don't think I'll be able to run a ground plane around the clock. I was going to have the crystal next to the MCU and the caps on the other side of the crystal. Are you suggesting the caps go even closer?
It is not difficult to get the ground ring around the clock. I have done a number
of TH and SMD two-layer layouts and have not had a problem.
(* jcl *)
It's no trouble to add the diode to the PCB, so I may as well, it doesn't force me to install it :). It really shouldn't be too hard to add the ground circle around the clock, the IC ground is right there anyway. So just one one side of the PCB? It really makes a difference? (I know of high frequency circuit mysticism, I just didn't think it applied here).
Now for part B of my original question. Once I have a design for a basic Arduino+protoarea, what else would be good to add to the board. I'm trying to build a foundation to make quick and easy semi-permanent installations. I'm going to add an LCD header, and maybe a space for a SOIC chip like the adafruit protoboard. But anything else I can think of would be almost as easy to add to the protoboard area as it would to add to the hardwired area.
It really shouldn't be too hard to add the ground circle around the clock, the IC ground is right there anyway. So just one one side of the PCB? It really makes a difference? (I know of high frequency circuit mysticism, I just didn't think it applied here).
Whether it makes a difference depends upon how the board is used in a system.
The ground ring reduces the radiated noise. If there is no circuitry close to the
noise source it won't matter. It is good design practice to put it in.
I usually do both sides but one side should be fine.
Now for part B of my original question. Once I have a design for a basic Arduino+protoarea, what else would be good to add to the board. I'm trying to build a foundation to make quick and easy semi-permanent installations. I'm going to add an LCD header, and maybe a space for a SOIC chip like the adafruit protoboard. But anything else I can think of would be almost as easy to add to the protoboard area as it would to add to the hardwired area.
What type of installations do you do? Sensors? Motor drivers? High current LEDs?
Hard to say it's your installations
(* jcl *)
I usually do both sides but one side should be fine.
Does it actually help to be a circle or any closed loop?
What type of installations do you do? Sensors? Motor drivers? High current LEDs?
As far as what I've already got in mind, high current LEDs, IR receivers, keypads, and rotary encoders will be common. I want to play with stepper motors at some point, but I have nothing in mind so I'll just get a shield eventually.
But I'm hoping to produce a platform that will be useful to others. The main advantages to my board will be the perfboard area and it will be routed to fit a standard, easy to get enclosure which has a battery compartment and optional IR panel.
Also, what's your favourite perfboard pattern? I'm planning on rows of 3 pads connected with power down the sides of the board, but 2 pads also has some appeal. Both ways give me 1 hole on each IC pin.
My original plan was power running between every grouping of 3 like the cheap radioshack boards, but that spreads everything out so much and it's a fairly compact board because it's a small handheld enclosure.
Is the diode connected from VCC to reset so that the capacitor on the reset pin is discharged quickly when power is removed?
That way the micro still gets a proper reset if power is applied again faster than the capacitor can discharge through the pull up resistor...?
Is the diode connected from VCC to reset so that the capacitor on the reset pin is discharged quickly when power is removed?
The diode is connected with its cathode to Vcc. As long as the voltage on the capacitor is less then Vcc+0.6 volts, the diode will not conduct, so the diode will not have any effect to discharge the capacitor.
It is only a clamp as jluciani said to protect the pin from overvoltage. Nasty things can happen when electric fields collapse.
