| Author |
Message |
|
|
 Posted: Apr 18, 2012 - 10:33 AM |
|
Joined: Dec 13, 2010
Posts: 13
|
|
I bought two different of these Yamaichi socket adapter boards from ebay. None of which is suiting the ATmega32u4 =P The pinouts are for other chips. So I took the opportunity to try some hot air de-soldering of the components. That was a breeze.
I thought, why not make my own adapter board. But I am a little bit unsure of exactly which pins the different programmer interfaces are supposed to use. I think I got it correct but would love to get it confirmed, please =)
I think JTAG should be pretty obvious from that data sheet. I had some quandaries about the clock signals on the ISP interfaces but I think I got it sorted out from the AVR-ISP500 programmer data sheet.
I added the schematic for the Teensy2.0 as well, since that is pretty much what I have been copying so far...
Other things to add or figure out:
E2 needs to be pulled low to enter the boot loader on reset. Will get a switch/resistor/jumper pr whatever.
Reset will get a switch to ground of course. The internal pull-up is enough?
UCap should not be connected to VCC, even when running the chip at 5V?
AREF is getting an input pin. The capacitor to ground is enough filtering? No extra inductors or anything?
It should be possible to add a low-pass filter between VCC and the AVCC inputs? or is that just overkill?
Any thing else I should be considering? |
|
|
| |
|
|
|
|
|
|
Posted: Apr 18, 2012 - 11:42 AM |
|
Joined: Feb 03, 2010
Posts: 33
Location: Minnesota USA
|
|
|
Quote:
Other things to add or figure out:
E2 needs to be pulled low to enter the boot loader on reset. Will get a switch/resistor/jumper pr whatever.
Reset will get a switch to ground of course. The internal pull-up is enough?
UCap should not be connected to VCC, even when running the chip at 5V?
AREF is getting an input pin. The capacitor to ground is enough filtering? No extra inductors or anything?
It should be possible to add a low-pass filter between VCC and the AVCC inputs? or is that just overkill?
Any thing else I should be considering?
I'd recommend a PU resistor on the /RESET line - but some designs out there don't have one. (see app note AVR042 - Hardware Design Considerations)
UCAP should be connected to VCC only when device is powered from an external 3.0-3.6V voltage regulator. All other scenarios will use the chip's internal regulator for the USB buffers power supply (See datasheet, section 21.3)
AREF filtering with capacitor is enough. Again, see AVR042 for Atmel's analog supply recommendations.
One more thing - don't assume that your crystal capacitors must be 10pF; check the requirements for your specific crystal for the capacitor size.
Good luck. |
_________________
Beau
www.goEWS.com (Wireless Modules/Breakout Boards -and- AVR Breakout Boards)
|
| |
|
|
|
|
|
|
Posted: Apr 21, 2012 - 04:46 PM |
|
Joined: Dec 13, 2010
Posts: 13
|
|
What about this? (schematics below)
I am pretty well educated in math, physics, and computer science, but never really learned anything about electronics...
My mental image of how the oscillator works is that it is like a piezoelectric swing set hanging in a sea of electric charges. The charges are pushed back and forth between the capacitors. Too much capacitance - the oscillations are too large, and the crystal breaks. Too little capacitance - the oscillations are too small for the magic to happen.
Identical capacitors connected in series doesn't usually do anything to the capacitance. There is the same amount of charge able to run into the system before it becomes saturated. The voltage is divided across them, but that should be fairly uninteresting in this case.
So what do I look for in the crystal data sheet? The series load capacitance sounds like a good value.. Should my capacitors both be equal to the load capacitance. Sounds correct to me, but I am only guessing =D
I.e. with the crystal in the attached pdf, I should have two 10pF capacitors if I have this right? |
|
|
| |
|
|
|
|
|
|
Posted: Apr 21, 2012 - 06:17 PM |
|
Joined: Feb 12, 2005
Posts: 16310
Location: Wormshill, England
|
|
As a silly question, why use a clam-shell?
You might just as well solder your TQFP-44 directly onto the board. The cost of a clam-shell is equivalent to about 10 complete pcb + headers.
If you want to change the chip family, solder on a new pcb.
David. |
|
|
| |
|
|
|
|
|
|
Posted: Apr 21, 2012 - 07:34 PM |
|
Joined: Dec 13, 2010
Posts: 13
|
|
I found no simple solder DIP adapter boards for QFN44, so I bought one with the yamaichi adapter. That one didn't map out all the pins I need. So I bought another one that was meant for some other chip. It has all GND pins connected together, and all VCC, and capacitors, and LEDs, and stuff where I don't want them.
I don't overly like etching my own boards, but a simple DIP adapter I probably should have done myself =) Now on the other hand I have two of the yamaichi adapters lying around doing nothing... |
|
|
| |
|
|
|
|
|
|
Posted: Apr 21, 2012 - 11:28 PM |
|
Joined: Nov 01, 2005
Posts: 6324
Location: Hilversum - the Netherlands
|
|
|
|
|
|
|
|
Posted: Apr 22, 2012 - 11:25 AM |
|
Joined: Dec 13, 2010
Posts: 13
|
|
If I had no higher intentions I would just go MacGyver in my kitchen, and make an extremely simple adapter like below. I found adapters like that for the QFN32 package. That was my first drag soldering attempt. I haven't had much luck trying to program it though.. Avrdude doesn't seem to support the ATmega32u2, and FLIP seem to be near impossible to get to run on 64 bit Ubuntu. Today might be the day to set up a new Windoze machine =P
That is an interesting idea however to solder the double row pin headers to either side of the PCB. |
|
|
| |
|
|
|
|
|
|
Posted: Apr 22, 2012 - 05:40 PM |
|
Joined: Dec 13, 2010
Posts: 13
|
|
| And, that would look something like this. It is a semi-autorouted draft so far, but I don't think it will take too much more tweaking. |
|
|
| |
|
|
|
|
|
FAQ
Search
Register
Log in to check your private messages
Log in
Maximize
RSS
