Microcontroller comparison chart?

xanatos

I am going to be selling a LOT of people on the value of the Basic Stamps and the Propellers, and I keep getting questions about the ATMEGA, ATTINY, and PICAXE controllers, for which I have no answer.· Is there a side-by-side feature comparison chart anywhere that would allow me to show the advantages of the Parallax products over the systems these folks are more familiar with?

Thanks,

Dave

·

HollyMinkowski

Who are these people? and what qualities are they most interested in?

If it's for educational use then basic stamps and arduinos are nice and easy starting
points for beginners.

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
- Some mornings I wake up cranky.....but usually I just let him sleep -

SRLM

Probably a computer class at school...

Anyway, you probably won't find comparisons built to order. Comparing the basic stamp to anything is like comparing a rock to a Ferrari. The BS2 value is in the documentation and support, not the hardware. For the Propeller, the architecture is exotic enough that people probably just can't find enough points to make a valid comparison.

There is a good instructable on how to choose a microcontroller. I'd read through that just so that you can understand (and counter) the arguments for the other controllers.

xanatos

Industrial clients. No school... that's a distant memory!

Dave

Mike Green

For industrial clients it's really difficult. How do you make a case for high levels of support? How do you make a case for shortening development time when certainly the big names in the business (Atmega and Microchip) all claim to do that? As Holly noted, it all depends on what types of applications you want to deal with. ATmegas and PICs really are better than Propellers and Stamps for some applications. Stamps are quite good for most applications where the limited memory and speed are not a factor. They're reliable, low power, and simple to use. If you use the modules rather than the OEM form, they are pricey although that's not a big issue for prototypes and other limited quantities. The Propeller is very good at signal processing including audio I/O and video output. It's very good at a variety of types of communications. It's marginal at best for USB unless you use an external low level controller like the MAX3421E. It can do simple internet interfacing limited by available memory. The YBox2, for example, can act as a web server with a TV status display, retrieve short e-mails from a POP server, do Telnet, and can even download compiled Spin applications and the whole thing fits in an Altoids tin using DIP packages. With the 8 separate processors, coding multiple functions is much simpler and therefore more reliable with shorter development time than doing the same kind of thing using a single processor and interrupts. The ATmegas and PICs do have their separate I/O processing units, but what's available doesn't always fit your needs and they can be buggy in newer devices and not really fixable where the Propeller does it all in software where it is fixable.

Post Edited (Mike Green) : 9/1/2009 4:37:04 AM GMT

HollyMinkowski

Good advice as always from Mike

Remember that many projects can profit by having 2 controllers on board.
The sm prop in quantity is about $6 and can add a lot of bang for that $6!

If mostly the shop uses C on AVR, ARM and other uc's and adds the Prop chip when its special
strengths are needed then that is the best setup I can think of. Only a few people
would need to train up on the prop to make use of it in projects. We are doing that
and it is saving $ and time

Industrial clients just love to save time and money!

Is this a new startup or something like that?

I'm available, they can make me an offer
I'm deep into AVR and ARM with C and asm
and I'm making steady progress as a propeller noob.
I could use a change of scenery....
And it's time to stop working for My Uncle's and his friends businesses.
It reeks of nepotism....and the projects while cool are just not
all that interesting to me anymore.

I find myself working on arts and crafts projects...not a good sign...

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
- Some mornings I wake up cranky.....but usually I just let him sleep -

xanatos

Mike - "Prop does it all in software where it is fixable" may be the single most relevant point in my case. That was the proverbial nailhead I believe, thank you.

Holly - Yes, new startup. Added it as a second business to run. My clients so far have all been companies who need very unusual and highly individualized solutions to problems and have been unable to find off-the-shelf or even custom designed solutions. Additionally, client installs tend to be highly install-specific, requiring numerous special parameters unique to the setup of the location (in this case the installs are nationwide). Hence the tremendous value in having a fixed hardware solution with a custom software configuration for each, as Mike described the prop above.

I am indeed looking into the ATMEGAs and other uCs, but Parallax offers such a robust user community that it is a real joy working with their products and other users. It also makes development of some truly unique applications possible.

Holly, what does your uncle's company do? What kinds of applications do you develop on your platform?

Dave

Agent420

I'm curious how the hardware platform creates such a big portion of your selling point?·It sounds as if you are providing an engineered solution for your clients.· Though there may be economic differences bewteen Basic Stamp and Propellers vs some other alternatives such as AVR's, I would think those differences become quite small in terms of your finished product, which must be much more influenced by your development, installation and support·rather than your oem module cost.· Why should your clients should care much whether a box providing a needed function contains a Basic Stamp or a Mega128?

So why the question of comparison may be valid, I would think it actually applies more to you than your clients.

As for "Prop does it all in software where it is fixable", that argument doesn't really seem important to me...· For one, how many avr's, pic's and the like suffer from irrevocable firmware bugs?· Secondly, similar internal design flaws may be present within the prop that we have yet to discover; it is a relatively new chip on the block.

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔


Post Edited (Agent420) : 9/1/2009 12:13:58 PM GMT

Leon

When something goes into production there is a vast difference between a $7 MCU and one costing $1 - $2. For hobbyists and niche market products it doesn't matter.

Leon

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Amateur radio callsign: G1HSM
Suzuki SV1000S motorcycle

Post Edited (Leon) : 9/1/2009 12:19:26 PM GMT

Agent420

Still, that seems a decision more for the designer than the client.· If you are designing something that addresses their needs and are able to offer it at a price that is agreeable to them, so be it.· Again, the question is more can you justify the profit/cost for your materials.

Unless the original question relates to something like, "I can accomplish this with a selection of controllers, but the Basic Stamp version will be $7 more each.· And the reason you want to pay $7 more for something that does the same thing is..."

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔

NetHog

Ageed with Agent420.

If you're selling hardware for its own sake, then the argument is valid.
If you're selling solutions, then the metrics are:

1) Does it do what I want to do?
2) Is it reliable?
3) Is it ready quick enough?
4) What is the initial and unit cost?

1) Is a must
2) should be a must (sadly people play 2 against 3 & 4)
3) and 4) tend to play against each other

When picking hardware:
Picking a Stamp over, say a PIC, will possibly make something ready quicker but increase unit cost. However it may not always do what the customer wants.
Picking a PIC/SX may address the "does what I want to do", reduce "unit cost" but increase "initial cost" and not ready as quickly

Personally I would not limit yourself to Stamp/Prop if you're making a business out of it. However that doesn't stop you considering these two platforms "up front" because it reduces your development time (which I believe would be the main benefit of sticking to these two platforms). Most successful businesses I've seen don't bind their hands for no good reason.

For example, a project I was doing recently, I initially started using an SX. Then realized that I both needed to reduce the board size, and needed some functionality that none of the Parallax line currently provides (Fast ADC / Analog comparison on chip). I ended up going with an 8 pin PIC chip. Another project I'm working on is using a Prop.

Agent420

^ Well phrased.

In researching these alternative platforms, you may find you sell yourself on some of them ;-)· I guess much of it depends on what kind of quantity you are dealing with and your development time needed to create the product.

Personally, I have pretty much always viewed the Basic Stamps as an educational or hobby item...· Cheaper, more capable alternatives exist for industrial applications.

Now the Propeller, it is kind of unique and may well find some niche areas, but I wouldn't rule it as the perfect general purpose chip for any occasion.

In all fairness, there are some good user communities for avrs and pic as well.

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔

Leon

Choosing a PIC over a Propeller or Stamp doesn't necessarily increase development time. Microchip has a vast number of software libraries and application notes with software that do everything from reading an ADC to speech synthesis and recognition. I can often put together most of an application in a couple of hours, using C or assembler. Then there are the excellent development tools including assemblers, compilers, simulators, debuggers and emulators. Support is excellent - via email and a forum. There are hundreds of different PIC devices, and it is nearly always possible to find one that has exactly the right features for a particular application, minimising cost. Microchip has a completely different marketing strategy from that Parallax's. It's very successful, BTW; Microchip is the market leader in 8-bit MCUs, and has sold over seven billion chips since 1990, when Steve Sanghi became their CEO. They were a basket case before that.

Leon

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Amateur radio callsign: G1HSM
Suzuki SV1000S motorcycle

Post Edited (Leon) : 9/1/2009 1:28:53 PM GMT

photomankc

For a truly one-off custom solution then I would think the cost of the controller is going to be a non-issue when compared against the cost of your development time and thus if the Propeller can do the job and you can code it faster for the Prop then it makes sense to go that way even if you have to shoe-horn it a bit. If my solution costs $5,000 in time and materials or $4994 is really not going to make or break it. I think if you are going to produce it in any quantity then the equation changes and the per-unit cost becomes much more important as does your ability to be flexible in selection and code for any platform.

I definitely agree with the others about the client not caring about the guts. All I want as the client is a device that will perform the control function I need, will work properly, and cost as little as nessisary. It's not really a big selling point to explain that the X micro is used and that lets you fix bugs in software... however bugs are fixed isn't of great importance to me, I'd expect that no matter who was selling it or how they were developing it.

dMajo

IMO Propeller can be compared to anything else in its price range. It's a strenge animal, at beginning for someone used to interrupts and hardware pheripherals not so intuitive but when you understand its architecture you find it more powerful than the mips it have.

SX as a low cost is well supported by parallax and also here I think his force is the single clock instruction time with the up to 75MHz clock capability.

For the BasicStamp I have a total different opinion. When I have neded something similar I have always chosen the picaxe. For les than half the price of the weakest BS2 module you have the most powerful picaxe (40X2) in 5V and 3.3V options running up to 64MHz. I appreciated very much the I2C slave feature. In this mode the internal scratchpad ram is available externally as a i2c eprom: all done in the background, transparently to the user. I have found this a magnificent way to change/adjust the picaxe program parameters externally through the i2c bus without any programming effort on the picaxe side.

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
· Propeller Object Exchange (last Publications / Updates)

CounterRotatingProps

> SX .... up to 75MHz clock capability.

dMajo,

with heatsinking, they can go up to 100MHz safely. I ran one at 120MHz for several weeks before the poor little guy started exhibiting problems. (And I think if it's heatsink had been bigger, it would have been OK for much longer.)

- H

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔

WBA Consulting

A colleague of mine that uses Cypress PSoC devices in most of his designs still can't understand the usage of the propeller due to it's cost. He recently did a 4 pad cap-sense board that controlled 5 relay breakout boards using a QFN 32 PSoC that ran $2.65. The catch is the amount of programming time that went into it was substantial. I could do the same project with a Stamp and QT240 chip in about a day. He's been playing around with some ARM processors and what they can do is pretty amazing, but I know how much time he spends learning them also. He saw this setup in Circuit Cellar and will be looking into it. Pretty amazing, but imagine the learning curve and "non-Parallax level" of support/forums.

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Andrew Williams
WBA Consulting
WBA-TH1M Sensirion SHT11 Module

Agent420

^ I got the ARM bug a while ago, had to know what all the fuss was about, and it was a bit appealing to have that much power, speed and memory in one chip...· I've got the MikroElektronika board which is great and very reasonably priced.

Arm's do have their place, but they are not for the faint of heart to start out on and a good ide/compiler can be expensive.· I know there are a few free GCC compilers out there, but if you're doing work then it's worth the investment to get a known working platform going.· I'm using the Keil uVision setup and it is very·nice.


▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔

CounterRotatingProps

> PSoC that ran $2.65. The catch is the amount of programming time that went into it was substantial

Hi Andrew,

Did your friend use PSoC Express? That particular app should have taken <30 minutes with just a few mouse clicks. (Unless there was something more complex in the breakout boards.) PSoC's are very cool.

- Howard

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔

HollyMinkowski

@photomankc

You are right, cost is the most important thing.

@xanatos
I used to write code for medical equipment.

Now I am doing custom projects for an outfit that builds turnkey solutions for
gov and security groups mostly... some of the projects are pretty strange.

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
- Some mornings I wake up cranky.....but usually I just let him sleep -