Global Moderator
Boston area, metrowest
Offline
Brattain Member
Karma: 270
Posts: 17040
Available for Design & Build services
|
 |
« on: June 23, 2011, 01:10:26 am » |
I was looking at Digikey pricing to get a parts list/price list together and looking at price kit options or something, thinking offer a lesser priced kit with a lower memory part for someone that wanted a data logger with dual serial but didn't need huge amounts of memory for a sketch. Now, I'm a pretty smart guy, and I know how to read data sheets, but I can't for the life of me figure out what these variations mean on the part numbers. I am guessing they have to do with screening as the price goes up. I don't see any explanations, and the graphs in Section 28 all look numbingly the same.
These are all 40 pin DIP parts, 20 MHz, dual serial, 3 interrupts, -40C to +85C operation
164A-PU, $4.75 16K Flash, 0.5K EEPROM, 1K RAM 164PA-PU, $5.21 164P-20PU, $5.60 << not on datasheet 164P-20PQ, $5.77 << not on datasheet
324A-PU, $5.36 32K Flash, 1K EEPROM, 2K RAM 324PA-PU, $5.88 324P-20PU, $7.16 << not on datasheet
644A-PU, $6.45 64K Flash, 2K EEPROM, 4K RAM 644PA-PU, $7.50 644-20PU, $7.98 << not on datasheet 644P-20PU, $8.19 << not on datasheet 644P-20PQ, $8.36 << not on datasheet
1284-PU, $7.33 128K Flash, 4K EEPROM, 16K RAM 1284P-PU, $8.13
Anyone have an explanation?
|
|
|
|
|
Logged
|
|
|
|
|
Global Moderator
Dallas
Offline
Shannon Member
Karma: 129
Posts: 10379
|
 |
« Reply #1 on: June 23, 2011, 02:20:16 am » |
164A-PU, $4.75 16K Flash, 0.5K EEPROM, 1K RAM 164PA-PU, $5.21
...I think a "P" in that position indicates "picoPower".
|
|
|
|
|
Logged
|
|
|
|
|
Global Moderator
Dallas
Offline
Shannon Member
Karma: 129
Posts: 10379
|
 |
« Reply #2 on: June 23, 2011, 02:26:29 am » |
164PA-PU, $5.21 164P-20PU, $5.60 << not on datasheet
...In the ATtiny family, a number in that position indicates maximum clock speed (dash 10 is 10MHz / low voltage range; dash 20 is 20MHz / high voltage range).
Looks like we're both confused. DigiKey has "picoPower™ Technology" attached to the "-20" processor and the non-20 processor has an extended voltage range "Voltage - Supply (Vcc/Vdd) 1.8 V ~ 5.5 V" versus "Voltage - Supply (Vcc/Vdd) 2.7 V ~ 5.5 V".
|
|
|
|
|
Logged
|
|
|
|
|
Global Moderator
Boston area, metrowest
Offline
Brattain Member
Karma: 270
Posts: 17040
Available for Design & Build services
|
 |
« Reply #3 on: June 23, 2011, 02:46:52 am » |
I looked at all of them for typical current draw in Active condition at 5V - there was no difference between A and PA, or 1284 and 1284P. I looked at atmel.com for picopower, no joy there either.
|
|
|
|
|
Logged
|
|
|
|
|
Global Moderator
Dallas
Offline
Shannon Member
Karma: 129
Posts: 10379
|
 |
« Reply #4 on: June 23, 2011, 03:21:12 am » |
At Mouser, this... ATMEGA164P-20PU ...has this... Operating Supply Voltage: 4.5 V to 5.5 V But this... ATMEGA164P-20PQ ...has no voltage range specified.
There is NO explanation of "Q" in that datasheet.
|
|
|
|
|
Logged
|
|
|
|
|
Leighton Buzzard, UK
Offline
Edison Member
Karma: 11
Posts: 1049
|
 |
« Reply #5 on: June 23, 2011, 03:55:43 am » |
as it happens I just stumbled across this list in an earlier thread.... all the variants and then some
|
|
|
|
« Last Edit: June 23, 2011, 03:57:51 am by mmcp42 »
|
Logged
|
there are only 10 types of people them that understands binary and them that doesn't
|
|
|
|
Humboldt, CA
Offline
Full Member
Karma: 1
Posts: 220
Arduino BBB
|
 |
« Reply #6 on: June 23, 2011, 09:54:38 am » |
I believe that the PicoPower chips are ones that will run at exceptionally low voltage levels for a given clock speed, hence consuming less power.
For our purposes it doesn't matter much, as we're using a full 5v even if we don't need to do so to get the clock speed up. I suspect the PicoPower chips would have an easier time running at 16/20mhz at 3.3v though.
|
|
|
|
|
Logged
|
|
|
|
|
Global Moderator
Boston area, metrowest
Offline
Brattain Member
Karma: 270
Posts: 17040
Available for Design & Build services
|
 |
« Reply #7 on: June 23, 2011, 06:08:14 pm » |
From mmcp42's link, I was able to find this app note on picoPower: http://www.atmel.com/dyn/resources/prod_documents/doc8349.pdfThe things they did to conserve power while running off a charged capacitor (simulate a battery, but with a controlled source) were able to extend the running times by a lot! Some things have been discussed in the forum, others not, resulting in a greater than 36x increase in runtime: "In this example, the runtime was increased from 6 to 217 seconds running off the same power source." Gonna see what else I can find ...
|
|
|
|
|
Logged
|
|
|
|
|
Global Moderator
Boston area, metrowest
Offline
Brattain Member
Karma: 270
Posts: 17040
Available for Design & Build services
|
 |
« Reply #8 on: June 23, 2011, 06:30:23 pm » |
I looked at the 164s, 324, 644s, and 1284s. The only thing I see different is that parts without P don't have picoPower. And the straight 644 only has 1 UART. No explanation of what the Q means.
|
|
|
|
|
Logged
|
|
|
|
|
SF Bay Area (USA)
Offline
Faraday Member
Karma: 80
Posts: 5513
Strongly opinionated, but not official!
|
 |
« Reply #9 on: June 23, 2011, 08:24:37 pm » |
The letters BEFORE the dash are (usually) part of chip type. "A" implies a die shrink since the original release, "P" means picopower. There's pretty much a separate datasheet for each combination of letters before the dash (separate datasheet for ATmega644P and ATmega644PA.) The letters AFTER the dash designate the packaging, which will normally be described in the "Ordering information" section of the datasheet.
So the atmega644P datasheet says that 644P-20AU is a TQFP-44, -20PU is a DIP-40, and -MU is a QFN-44 The 644PA datasheet add a couple packages but is basically the same. I couldn't find a -20PQ anywhere :-(
|
|
|
|
|
Logged
|
|
|
|
|
Western New York, USA
Offline
Faraday Member
Karma: 17
Posts: 3507
|
 |
« Reply #10 on: June 23, 2011, 10:03:47 pm » |
..."A" implies a die shrink since the original release,... I think I read on avrfreaks that the 'A' signifies a change in die supplier. The die shrink was a result of this change in supplier but may or may not have been the stimulus for the change.Don
|
|
|
|
|
Logged
|
|
|
|
|
Global Moderator
Boston area, metrowest
Offline
Brattain Member
Karma: 270
Posts: 17040
Available for Design & Build services
|
 |
« Reply #11 on: June 23, 2011, 10:37:35 pm » |
Ok, So consensus is A is a different die somehow. Q still the mystery letter! Only seems to affect price, so must be some extra screening thing.
|
|
|
|
|
Logged
|
|
|
|
|
|
|
Left Coast, CA (USA)
Offline
Brattain Member
Karma: 282
Posts: 15443
Measurement changes behavior
|
 |
« Reply #13 on: June 23, 2011, 11:07:53 pm » |
Ok, So consensus is A is a different die somehow. But sometimes it also can mean additional features and maybe not a die shrink ? Like second serial port for 644p Vs 644. I would be interested in if there is a way to determine if the internal device codes (used by AVRDUDE) are different between the different 'flavors' of chips within a family. Does Atmel have a single document where device numbers Vs internal device codes is shown? AVRDUDE has those codes listed in it's avrdude.conf file, but where did they find this information? Lefty
|
|
|
|
|
Logged
|
|
|
|
|
|
|
|