|
When Atmel obsoletes a part there is a pin compatible replacement that
has more functionality. Examples of this would be replacing the
90S1200/2313 with the ATTiny26 and 90S8535 with the ATMega8535.
Comparable tools are also less expensive in general. Compare the cost
of an STK500 with that of a DV164006. You're correct that they're getting closer with the 18F series, but there are still the issues of 4:1 clock:instruction cycles and the accumulator (W) register. I'll take 32 working registers over an accumulator any day. To be sure, Microchip's sample policy is probably a bit easier but 5 or 10 dollars for a part isn't a big deal. REB Dennis Clark wrote: The AVR has a better instruction set for high level programming I think. Although the new 18F PIC series has greatly improved their instruction set giving you access to the stack (also check out the dsPIC stuff for even better support) and allowing you to get access directly to FLASH memory for table moves and such. The AVRs don't segment their data and code spaces requiring bank switching (again, the 18F does not bank switch the code, but does bank switch some of the data memory). The AVR is faster due to a better pipeline I think, but again, the PIC seems to be catching up there too with the 18F parts. The one BIG difference that gives the AVRs a negative to me is that Atmel doesn't hold a candle to Microchip for engineering support and while Microchip never obsoletes a part, Atmel seems to change parts and obsolete old lines almost at the drop of a hat. That and I STILL seem to have problems getting some AVR parts... IMO, DLC --- In p...@yahoogroups.com, David Duley <dduley@d...> wrote:On Thu, 19 Aug 2004 09:39:46 +0100, you wrote:----- Original Message ----- From: "Trent" <scoobytw@y...> To: <p...@yahoogroups.com> Sent: Thursday, August 19, 2004 9:11 AM Subject: [piclist] Re: PIC10F competitionI don't want to get flamed or start a cpu war but I checked out AVR's (lightly, looking at info on their website) and it seemed that the AVR's didn't have any peripherals except timers and AD, did I miss something? I have heard a lot of people say they liked AVR's better for architecture. What is better about them? do you not have to keep track of things like which bank you are accessing or running code from, stuff like that? Being your standard enginerd I'd get a few to play with but I didn't see the appeal at first glance. What'd I miss?Some advantages are: (mostly) 1 clock per instruction 32 registers no banks easier ISP LeonDon't forget a way better instruction set. Dave Duley------------------------ Yahoo! Groups Sponsor --------------------~--> Make a clean sweep of pop-up ads. Yahoo! Companion Toolbar. Now with Pop-Up Blocker. Get it for free! http://us.click.yahoo.com/L5YrjA/eSIIAA/yQLSAA/dN_tlB/TM --------------------------------~-> to unsubscribe, go to http://www.yahoogroups.com and follow the instructions |
|
|
PIC10F competition
Started by ●August 19, 2004
Reply by ●August 29, 20042004-08-29
Reply by ●August 29, 20042004-08-29
|
On Sun, 29 Aug 2004 06:36:31 -0400, Roy E. Burrage wrote: >Comparable tools are also less expensive in general. Compare the >cost of an STK500 with that of a DV164006. To be fair, you're not comparing apples to apples. With the STK500 you are getting a "development board" that can program *some* AVR parts, some require an additional cost STK501 or STK502 module. The STK500 is not exactly compact and portable either. On the Microchip side, the DV164006 includes an ICD2 *and* a PICDEM2 development board. The ICD2 will pretty much program any PIC through ICSP and debug any flash based part with on-chip debugging, no adapters required except for some of the very low pin count devices. You also get the ICD2 which is in a case and very portable. They are both very similar in cost -- $100-$250 in cost to get started, both low cost by my account. You get a bit more flexibility with Microchip, Atmel gives you a slightly lower cost entry (but I can't see the difference being significant for anyone that's serious about doing development work on an ongoing basis). There's a real question as to the availability of parts though. Atmel doesn't have the fab capacity of Microchip. Atmel will put parts on allocation at the drop of a hat once their fab facilities are loaded. Over the 15 years or so I've used Microchip parts I've never had a problem getting production quantities of parts, I can't say the same about Atmel. In fact, I had a couple good years in the late 90's re- designing products for clients that had Atmel parts made of "Unobatanium" and replacing them with equivalent PICs that were readily available. Microchip parts tend to be slightly cheaper in production quantities than Atmel chips with the equivalent functionality too. Also, Microchip has a much better peripheral spread in their chips -- you can almost always get a PIC with *exactly* the peripherals you need for a given design. Atmel is more limited in your options since they have much fewer chip models to choose from. I know there are a lot of hobbyists out there in both camps, but quite honestly neither chip company is there to serve them. It's all about *selling* chips and they'll both go where they sell the most chips. Yes, it's an advantage to get hobbyists and students on-board with their chips, but it's a minor consideration next to where the volume sales are. Both companies are relatively "hobbyist friendly", as compared to Motorola (or whatever they are calling themselves these days) for instance. That all said, they are both good microcontroller families that each have their strengths and weaknesses. Matt Pobursky Maximum Performance Systems |
Reply by ●August 29, 20042004-08-29
|
Number wars can be interesting. Especially in the software arena. Figures don't lie, but liars can figure. :-) Software benchmarks, I don't put much trust in. It is what you are comfortable with, and what gets the job done. One instruction set may be better at one type of application, the other better at a different application. They are probably are both reasonably good having been developed specifically for microcontroller, unlike the Intel line which has maintained compatibility(to their credit), but at an extreme cost. The little quirks of the original 4004 which were there to fully utilize hardware(forget software for now) still haunt your P4. I am more concerned about hardware. There you can see some posturing in data sheets. I glanced at one(I hope represtentative) data sheet from Atmel and Microchip, and just a few parameters. I see Atmel has max IO current of 40 ma. I thought great, I could use that. However, IO port current limit is 100 ma vs. 200 ma. and VSS/VDD current limit is 200 ma. vs. 250/300 ma., so be careful out there. The really interesting number I saw on the Atmel sheet was that maximum leakage current on a typical pin is 980 na. and Microchip is way up there at 1 ua. Chad --- "Roy E. Burrage" <> wrote: > When Atmel obsoletes a part there is a pin compatible replacement > that > has more functionality. Examples of this would be replacing the > 90S1200/2313 with the ATTiny26 and 90S8535 with the ATMega8535. > Comparable tools are also less expensive in general. Compare the > cost > of an STK500 with that of a DV164006. > > You're correct that they're getting closer with the 18F series, but > there are still the issues of 4:1 clock:instruction cycles and the > accumulator (W) register. I'll take 32 working registers over an > accumulator any day. > > To be sure, Microchip's sample policy is probably a bit easier but 5 > or > 10 dollars for a part isn't a big deal. > > REB > > Dennis Clark wrote: > > >The AVR has a better instruction set for high level programming > >I think. Although the new 18F PIC series has greatly improved > >their instruction set giving you access to the stack (also check > >out the dsPIC stuff for even better support) and allowing you to > >get access directly to FLASH memory for table moves and such. > >The AVRs don't segment their data and code spaces requiring > >bank switching (again, the 18F does not bank switch the code, but > >does bank switch some of the data memory). > > > >The AVR is faster due to a better pipeline I think, but again, > >the PIC seems to be catching up there too with the 18F parts. > > > >The one BIG difference that gives the AVRs a negative to me is > >that Atmel doesn't hold a candle to Microchip for engineering > >support and while Microchip never obsoletes a part, Atmel seems > >to change parts and obsolete old lines almost at the drop of a > >hat. That and I STILL seem to have problems getting some AVR > >parts... > > > >IMO, > >DLC > > > > > >--- In , David Duley <dduley@d...> wrote: > > > > > >>On Thu, 19 Aug 2004 09:39:46 +0100, you wrote: > >> > >> > >> > >>>----- Original Message ----- > >>>From: "Trent" <scoobytw@y...> > >>>To: <> > >>>Sent: Thursday, August 19, 2004 9:11 AM > >>>Subject: [piclist] Re: PIC10F competition > >>> > >>> > >>> > >>> > >>>>I don't want to get flamed or start a cpu war but I checked out > >>>>AVR's (lightly, looking at info on their website) and it seemed > that > >>>>the AVR's didn't have any peripherals except timers and AD, did I > > >>>>miss something? I have heard a lot of people say they liked > AVR's > >>>>better for architecture. What is better about them? do you not > >>>>have to keep track of things like which bank you are accessing or > > >>>>running code from, stuff like that? Being your standard enginerd > > >>>>I'd get a few to play with but I didn't see the appeal at first > >>>>glance. What'd I miss? > >>>> > >>>> > >>>Some advantages are: > >>> (mostly) 1 clock per instruction > >>> 32 registers > >>> no banks > >>> easier ISP > >>> > >>>Leon > >>> > >>> > >>Don't forget a way better instruction set. > >> > >>Dave Duley > >> > >> > > > > > > > > > >to unsubscribe, go to http://www.yahoogroups.com and follow the > instructions > >Yahoo! Groups Links > > > > > > > > > > > > > > > > > ===== My software has no bugs, only undocumented features. __________________________________ |
|
|
Reply by ●August 29, 20042004-08-29
|
> The really > interesting number I saw on the Atmel sheet was that maximum leakage > current on a typical pin is 980 na. and Microchip is way up there at 1 > ua. Am I too rusty in my powers of 10 or is this just a difference 2% ? Wouter van Ooijen -- ------- Van Ooijen Technische Informatica: www.voti.nl consultancy, development, PICmicro products |
|
|
Reply by ●August 29, 20042004-08-29
|
It looks like 2% to me? 18 pins = something like < 15uamps on a PIC if ground and Vcc do not have leakage. Same size Atmel would be < 14.7 uamps? Why is that interesting? Does either have a 'sleep mode' with zero leakage? That would be interesting? --- Wouter van Ooijen <> wrote: > > The really > > interesting number I saw on the Atmel sheet was > that maximum leakage > > current on a typical pin is 980 na. and Microchip > is way up there at 1 > > ua. > > Am I too rusty in my powers of 10 or is this just a > difference 2% ? > > Wouter van Ooijen > > -- ------- > Van Ooijen Technische Informatica: www.voti.nl > consultancy, development, PICmicro products > > > ------------------------ Yahoo! Groups Sponsor > --------------------~--> > Make a clean sweep of pop-up ads. Yahoo! Companion > Toolbar. > Now with Pop-Up Blocker. Get it for free! > http://us.click.yahoo.com/L5YrjA/eSIIAA/yQLSAA/dN_tlB/TM > --------------------------------~-> > > > to unsubscribe, go to http://www.yahoogroups.com and > follow the instructions > Yahoo! Groups Links > > __________________________________________________ |
|
|
Reply by ●August 29, 20042004-08-29
|
> > > The really > > interesting number I saw on the Atmel sheet was that maximum leakage > > current on a typical pin is 980 na. and Microchip is way up there at 1 > > ua. > > Am I too rusty in my powers of 10 or is this just a difference 2% ? > > Wouter van Ooijen > I had the same thought and I am not too old to be rusty about anything! *Carlos Marcano* -Guri, Venezuela- --- Outgoing mail is certified Virus Free. Checked by AVG anti-virus system (http://www.grisoft.com). Version: 6.0.745 / Virus Database: 497 - Release Date: 27/08/04 |
|
|
Reply by ●August 29, 20042004-08-29
|
I thought it interesting that they would bother to take the 2% and 2 extra digits, so as to be able to use nano rather than rounding it to 1 micro. :p Chad --- Carlos Marcano <> wrote: > > > > > > > The really > > > interesting number I saw on the Atmel sheet was that maximum > leakage > > > current on a typical pin is 980 na. and Microchip is way up there > at 1 > > > ua. > > > > Am I too rusty in my powers of 10 or is this just a difference 2% ? > > > > Wouter van Ooijen > > > > I had the same thought and I am not too old to be rusty about > anything! > > *Carlos Marcano* > -Guri, Venezuela- > --- ===== My software has no bugs, only undocumented features. __________________________________ |
Reply by ●August 29, 20042004-08-29
|
Hi Matt, I agree completely with what you say. In my case, Atmel has chosen local reps that completely ignore potential costumers, requests for samples made via Atmel's site are automatically forwarded to them, which means they are promply ignored. So, any request for sample parts is an ordeal which makes a costumer fear what he'll have to go through to get production quantities. Imagine what a student will have to go through. Microchip, however, has an EXCELLENT sample parts request system, it's convenient, fast (most of the time if parts are in stock, longest it has taken for me was 2 1/2 weeks), and you can get quite a few different parts for testing what's best for your design, no questions asked. One thing I learned when I was an engineering student was that companies that give students facilities to use their products or services in their research or projects are likely to get bussiness from them when (and if) they end up in some company's development team. "Yeah, when I was a student, Microchip was excellent, they sent me sample parts with no problems, whereas Atmel completely ignored me". I've seen this happen, not with Atmel/Microchip, but with other manufacturers. I think manufacturers should be very aware that students will be tomorrow's engineers, managers, purchase department heads, etc. etc. Best regards, Mike ----- Original Message ----- From: "Matt Pobursky" <> To: <> Sent: Sunday, August 29, 2004 5:13 PM Subject: Re: [piclist] Re: PIC10F competition > To be fair, you're not comparing apples to apples. With the STK500 you > are getting a "development board" that can program *some* AVR parts, > some require an additional cost STK501 or STK502 module. The STK500 is > not exactly compact and portable either. <SNIP |
|
|
Reply by ●August 29, 20042004-08-29
|
I figure the companies know all this and don't give it a lot of thought. If it is easy to be accomodating, fine. If not, oh well. The devices chosen for a design will ALWAYS come down to the least expensive device capable of doing the job at hand. Engineers don't choose devices, Purchasing does. Who cares how nice they treated you in college - if the company needs 100,000+ parts Purchasing will be making the choices, with or without input from Engineering - engineers probably won't even be invited to the meetings. Ever notice how the Sales people come to talk to engineers but always want to know who is making the decisions? They know full well it isn't Engineering. At the end of the day, its about profit margin - for everyone. A product has a price target and competition to keep it there; every bit of manufacturing cost that can be driven out, will be. So what if the engineering time and programming expense is a little higher if a company can get lower production costs and higher margins? Engineering is an expense; a side issue of getting the product out the door. Time to market is another metric - so get rid of the Brand X chip programmers and get some Brand Y chip programmers, they'll understand the devices. We're using Brand Y on this job... Cause Purchasing said so... What the heck, outsource the design or get some 'rent-a-grunts' over here to write the code. Don't put them on our payroll - we'd have to provide compensation AND benefits. Don't be overly impressed with your engineering credentials - no one else is... Mine either, for that matter. --- In , "Michael Puchol" <mpuchol@s...> wrote: > Hi Matt, > > I agree completely with what you say. In my case, Atmel has chosen local > reps that completely ignore potential costumers, requests for samples made > via Atmel's site are automatically forwarded to them, which means they are > promply ignored. So, any request for sample parts is an ordeal which makes a > costumer fear what he'll have to go through to get production quantities. > Imagine what a student will have to go through. > > Microchip, however, has an EXCELLENT sample parts request system, it's > convenient, fast (most of the time if parts are in stock, longest it has > taken for me was 2 1/2 weeks), and you can get quite a few different parts > for testing what's best for your design, no questions asked. > > One thing I learned when I was an engineering student was that companies > that give students facilities to use their products or services in their > research or projects are likely to get bussiness from them when (and if) > they end up in some company's development team. "Yeah, when I was a student, > Microchip was excellent, they sent me sample parts with no problems, whereas > Atmel completely ignored me". I've seen this happen, not with > Atmel/Microchip, but with other manufacturers. I think manufacturers should > be very aware that students will be tomorrow's engineers, managers, purchase > department heads, etc. etc. > > Best regards, > > Mike > ----- Original Message ----- > From: "Matt Pobursky" <ccsuser@m...> > To: <> > Sent: Sunday, August 29, 2004 5:13 PM > Subject: Re: [piclist] Re: PIC10F competition > > To be fair, you're not comparing apples to apples. With the STK500 you > > are getting a "development board" that can program *some* AVR parts, > > some require an additional cost STK501 or STK502 module. The STK500 is > > not exactly compact and portable either. > > <SNIP> |
|
|
Reply by ●August 29, 20042004-08-29
|
> The devices chosen for a design will ALWAYS come down to the least > expensive device capable of doing the job at hand. Engineers don't > choose devices, Purchasing does. Yeah, in all jobs I have had my boss made the decisions. But the various bosses very rarely decided outside my recommendations :) Wouter van Ooijen -- ------- Van Ooijen Technische Informatica: www.voti.nl consultancy, development, PICmicro products |
