OP may like the following simplified design of UNOR3.
The PB DOES NOT have a full swing crystal option and will not be able to drive xtals that require full swing.
You still have the capacitors wrong.
I done for today.
I cannot argue about "reinventing the wheel". At the time I started the thought process I wasn't aware that the UNO (and Nano) schematics were available. However, digging into them once I was underway, I found there were some differences. At the worst, I've learned quite a bit along the way. The original thought process was to try to reduce the footprint of the whole solution and tightening up the Nano footprint seemed a good place to start.
C8 is included per my reading of the Microchip AN2519. In section 3 of that document there is this quote:
"To protect the Reset line from further noise, connect a capacitor from the RESET pin to ground. This is not directly required since AVR devices internally have a low-pass filter to eliminate spikes and noise that could cause reset. Using an extra capacitor is an additional protection."
I've got the UNO and Nano schematics open. The UNO shows a 10k pullup and the diode as I have shown. The Nano only has a 1k pullup. I'm still improving: if I understand you correctly, as well as the UNO R3 schematic, C5 on the UNO is the hi-pass filter. Then, when that 200ms reset pulse hits, as you say, the reset pin only sees a brief pulse as, in my schematic, D4 filters the reverse pulse...if I understood it all.
I understand about the physical reset button. There will be one in the implemented circuit.
Thank you.
--HC
AN2519 - OK, but they also mentioned that AVRs have internal low-pass filter for noise elimination, so it is not needed.
The pull-up resistor value is not critical. It can be anything between 1k-12k, with 328P for sure.
Looking at that schematic, I see that C11 and C1 are parallel to R4: do they act as filter capacitors? That seems very similar to the Nano schematic which just shows a pullup on the reset.
If that board is impressive. That was not easy.
--HC
That is the same one I recommended you read way back in post #5
It only took you 85 post and 5 days to finally getting around to reading it.
Yes, I saw that mentioned, too. I didn't have clear design in adding it, I did it simply based on the portion I quoted. Looking at the UNO schematic which you mentioned: I believe I see what C5 does coming from the (shown) ATMega16U2 Pin 13. I believe I see how that would give a shorter pulse to the reset than the 200ms you mentioned AVRDude does. Therefore, I see that, what you've brought up, to be a better way.
I chose the 10k value in part because that's what the UNO schematic shows. That was underscored by this quote from the same AN:
"For DebugWIRE to function properly, the pull-up must not be less than 10 kΩ."
I don't have/intend, at this time, to use DebugWIRE, but it didn't seem a bad idea to not preclude its use.
Thank you.
--HC
No, they don't. Both are connected between 5V and GND filtering VCC, of course. R4 is pull-up for RESET. (Golam's schematics)
If you want to understand, the datasheet contains an information about valid reset cycle, how many clock cycles etc. => min. reset pulse length.
As I see in the recent UNO schematics https://content.arduino.cc/assets/UNO-TH_Rev3e_sch.pdf, I didn't see it for years, L2 10uH was added probably for better VCC filtration.
I was checking other thing. There is R2 in parallel with crystal/resonator. It is not needed and if you check the photo of UNO it is not placed on the board, never was.
Why are you still using global lables for power and ground? It's poor design practice.
What's this "connect only at PS" mean?
What't way too long?
200ms pulse?
During ICSP programming the RESET line is held LOW the whole time.
1 Like
What? ICSP connector is behind the capacitor I wrote about, see the UNO schematics. It is about entering the bootloader or resetting the chip. Did I make some wrong assumption?
But yes, as I looked at OP's schematics, he has only ICSP connector so definitely without capacitor. The C8 confused me. Sorry.
I can see why. Comming Coming in on post #81 and not reading all the previous posts would confuse anyone.
1 Like
???
Congratulation. You found a typo.
You are entitled to keep it or sell it.
Google also has an alternate opinion. Let @jim-p to clarify.
Really? The fact that Arduino is "Open Source Hardware" is one of its main claims to fame. Back in ~2008, "Modern Devices Inc" sold a "bare bones board" that looked a lot like what you're doing. (They don't sell it any more, but there is still my "spin" on it: https://www.instructables.com/Single-Sided-Really-Bare-Bones-Board-Arduino-in-EA/ )
Mostly pedantic nonsense.
As if most vendors ever paid that much attention, compared to "oh, here's a design with 18pf caps that works; I guess I'll do that too!" (not to mention many crystals coming without full specs.
After much math and verbal haranguing, you have a value that it a whopping 2pf off of the OP's schematic. Sources say a crystal can be "pulled" "as much as 100ppm" (0.01%) by varying the load caps by a quite substantial amount (assuming that you don't prevent the circuit from oscillating, which appears to require being WAY off), so I have a lot of doubts that the result is worth the effort.
(Finding semi-quantitative info is difficult: "if my caps should optimally be 17.2pF but are actually 22pF, how far will the frequency be off?", but for maximal pedantry I did find: https://www.allaboutcircuits.com/technical-articles/assessing-effect-of-load-capacitance-on-frequency-of-quartz-crystal/ )
(anyone want to do the actual measurements? Hope you have a GOOD frequency counter!)
(Or you could just use a ceramic resonator, like on the Uno. Absolute accuracy may be less than a crystal with "incorrect" load caps, but it'll cut down parts count and remove a "worry.")
You are so pathetic that I feel sorry for you. Your sense of interpretation is ridiculous.
I've never said that OP should strictly use that value.
Ugh, yes, of course, you're right. Sorry...a little stupid falls out of me from time to time. 
--HC
1 Like
Thank you, yes, I looked that up. 2.5uS.
Re L2: I've seen on a number of schematics the use of an inductor as part of a power filter for AVcc. As with many things, there seems to be some debate over the necessity of it. One bit I read somewhere suggested that it wasn't necessary below some frequency, that an inductor of the size necessary to choke low-frequency signals would be "impossibly large", and to just use a resistor. That's outside my wheelhouse. I would suppose it would be helpful if one knew which frequency one was trying to remove.
While playing Where's Waldo with that schematic (looking for L2), that resistor caught my eye. It hadn't really jumped out at me before. Where I've seen high-value resistors like that before, somewhere, was for bleeding down capacitors. Maybe someone overly eager put it in to bleed down those man-killer crystal caps. It also may be a belt-n-suspenders approach to trying to ensure that the crystal starts oscillating without fighting residual capacitance.
--HC
1 Like