On their schematic pins 2 & 3 cross over, Vbat goes to pin 3 but on your schematic Vbat is pin 2. Just thought I’d point that out.
run it through FreeDFM…its checks for all issues and sends you an email with a report of where they all are
I can’t see any major drawbacks in your design. Moreover, you will learn over the road 
1 Like
Your reply…
Yea, you have seen 'em too. 
1 Like
I personally have not experience a wonky electron, but I am 100% sure that I will within a year or two.
I was going to use 10K … but I am new to this. The math indicated that 10K should meet the requirements at the lowest specified bus speed.
The original recommended value was 4k7, giving 1mA sink current in a 5V system
I ran into a subtle quirk with 4pcb/FreeDFM.
They truncate drill diameters to mils. So a 0.3mm metric via drill, giving 0.01181 inch (which is perfectly fine for practically any cheap board manufacturer) ends up being classified as 0.011 inch drill. This exceeds their standard prototyping board specifications and you end up with a quote of some 79 bucks or so for 5 boards. So be aware of that.
Same goes for Silkscreen width. KiCad standard is 0.00476. They require 0.005. This is a ridiculous difference, but their automatic check complains. It will be automatically fixed by them, however, infering no extra cost.
1 Like
Being pull-up resistors they would provide a source current.
I’m not sure what you mean by “original recommended” but there is no such recommendation in the I2C Bus Specification, and hasn’t been since at least version 1.0 (1992) when fast-mode was added. There are a series of charts provided for determining the min and max values based on speed, supply voltage and bus capacitance. The minimum specified sink current for devices supporting standard-mode (100 KHz) and fast-mode (400 KHz) is 3 ma. Therefore the minimum pull-up resistance for a 5V system would be ~1K5. The maximum resistance depends on bus capacitance. Assuming 100pF bus capacitance, the maximum resistance for fast-mode would be ~3K5. 4K7 would be okay for a bus capacitance of 250pF in standard-mode but only 70pF or less in fast-mode. Even then you wouldn’t choose the max resistance, but rather the mid point of ~3K1.
Therefore in a 5V system supporting fast-mode with a bus capacitance of ~100pF the optimum value for pull-up resistors would be ~2K5. (All of the ops devices support fast-mode).
My recommendation of 1K was for devices supporting high-speed mode (up to 3.5 MHz) where the min sink current is 20ma. But I see now that the ATMega only supports fast-mode.
In summary, the op should be using ~2K5 pull-up resistors.
So the boards arrived. I have not had time to assemble them, but it turns out that I made the holes for the adafruit products too small, so I either need to get smaller headers, or drill the holes out. I will look into smaller headers first. No biggie.
Once assembled, or found to have problems, I am going to write a full out post of things to look out for, I think that will help beginners.
Are we talking about common “pin headers” (a.k.a., “Berg strips”, “post headers”, et al) with 0.025" square pins? Because the square post mounts into a round hole, they easily push into a hole of nominal 0.035" diameter - and 0.032" or less with a little persuasion.
Again, if I know the part you’re working with, enlarging the drilled hole is probably NOT a good idea in general. The drill will remove the plated-through “barrel”, disrupting connectivity between front and back layers.
-
If traces connect to the header pads ONLY on the solder-side, this isn’t a significant problem. The mechanical integrity will be slightly reduced, since the header post isn’t anchored by the solder joint within the hole.
-
If you can solder to BOTH sides of the board (and there are NO internal layers connected to the header), then soldering both sides of the header pins ensures top-to-bottom connectivity. This is seldom possible with header strips where a plastic “body” spaces the individual pins and prohibits solder access to one side of the joint.
Dale
@dchisholm I removed the pin headers, and replaced them with the leads I trimmed off of the resistors, it seems to be working ok. There is a 7000hz whine coming from the power supply, and I have found a 50mV/10kHz sawtooth wave in the power line, do those sound ok, or should I be worried?
50mV of ripple on the output of a switching supply isn’t anything to brag about, but it’s not unusual. Does your supply have a published specification for ripple?
However, the audible whine sounds like the supply may be overloaded, or possibly underloaded. Have you calculated your circuit’s current draw, and verified that it’s within the supply’s capability?
Dale
p.s. - Or maybe you simply have sensitive ears. Are you a 13-year old girl? I think they are supposed to have the best hearing of all people.
2 Likes
Its my own project, so it passes my QC this one time.
So it is OK? I will look up the ripple.
I think that it is linked to a coil expanding, I saw that somewhere online, I know I have not overloaded it, but since I have not fully tested every element, under loading is possible.
It is likely that I do, based on prior experience, but I know I am not a 13 year old girl.