Migrating from PIC18 to ARM: which MCU to choose?

I've pretty much settled on the STM32 at this point, while keeping the LPC as a backup plan.

As for development tools, I've looked at Raisonance but as far as I can tell (I could be wrong here) their development tools will only work with their own RLink JTAG debugger and not with anything else like a Wiggler or a J-Link or a USB-OCD. While a standard RLink is reasonably priced, the non-crippled Pro version is about seven times more expensive which totally blows the budget when you need to buy three of them. This also makes it the most expensive item in the laptop bag, double the price of the laptop itself. Maybe this is just my own personal bias coming from the world of PIC18 where a debugger with a two-digit price tag is the norm and you don't feel too bad if it gets zapped during a field service call.

I have been doing some testing with an STM32F103 and a parallel port Wiggler and so far the loading & debugging is MUCH faster than any PIC18 tools that I've ever used. Seeing as anybody can build their own Wiggler clone with a few cents worth of spare parts, I'm not exactly sure what more you get for an extra 700 Euro.

Reply to
Stimpy
Loading thread data ...

How much does a parallel port cost these days? I don't think I can even have one on my laptop that will work reliably with a wiggler. On the other hand, there are decent USB debuggers for $150 the last time I checked.

Rick

Reply to
rickman

Standard PCI parallel cards are still readily available for around $20 but new laptops don't seem to have parallel ports anymore. It looks like a good alternative might be a USB JTAG adapter made from a FTDI FT2232 chip. These are also cheap, readily available with published schematics, supported by OpenOCD and Rowley, and more likely to be around ten years from now than some proprietary device.

Reply to
Stimpy

I would not discount the 120MHz LPC1769 M3. STM32 peaks out at 70MHz.

Reply to
linnix

There seems to be 120MHz ones in the pipeline, anyone heard any more information?

--

John Devereux
Reply to
John Devereux

The SAM3S has no CAN, like the SAM7S does not have CAN. That does not mean that it is true to claim that the SAM7 does not have CAN. There is a SAM7X which has CAN and Ethernet.

Obviously Atmel has the knowledge to design a SAM3 as well as AVR with both Ethernet and CAN.

While they are nice to have onboard, the device without them may have more applicability and someone has made a judgement that Atmel will be best benefited by introducing the SAM3S before any other product range.

I have been working closely with the ARM team on many things, but in this case, the peripheral was developed an implemented first on a 32 bit AVR.

Atmel uses the same peripheral library on both ARM and 32 bit AVR so it is no surprise that the ARM team (which did the actual implementation) decided to use it on the SAM3S.

You can get samples from the production lot now (not engineering samples) of the AT91SAM3S4AA,so it is very close.

The AT45DB642D dataflash has been in production for years, but that does not mean that there is Digikey stock. It also does not mean that it is impossible to get right now, just that production cannot fulfill all demand right now. We have a component crisis and most colleagues in the industry has long lead times.

I think that is the whole idea.

Obviously the professional version is better. You can easily get source code control elsewhere free of charge.

It all depends on the requirements and boundary conditions. The AVR has a free widely available compiler.

See above.

close

Those AVR chips are getting into some decent volume in the mobile industry at the moment so Samsung, Motorola, HTC etc. all probably disagree with you.

--
Best Regards
Ulf Samuelsson
 Click to see the full signature
Reply to
Ulf Samuelsson

That is their 90 nm parts.

--
Best Regards
Ulf Samuelsson
 Click to see the full signature
Reply to
Ulf Samuelsson

That is their 90 nm parts.

--
Best Regards
Ulf Samuelsson
 Click to see the full signature
Reply to
Ulf Samuelsson

rickman expounded in news:7b2ead92-131f-4893-84f1- snipped-for-privacy@l17g2000vbf.googlegroups.com:

That reminds me of something that has been nagging at me. How long is USB going to be around?

I've read some noise about the future of Light Peak or some such, and I keep wondering how much of an investment in USB tools/peripherals I should be making? I suspect it will be a while yet, but how long is that?

I suppose obsolescence is a good excuse to the better half for upgrades later, OTOH.

Warren

Reply to
Warren

In message , Warren writes

As virtually EVERY electronics device has USB and especially non- consumer items with a life of many years USB will be around for another decade.

I know many people putting USB into new systems that have a life of at least 20 years.

If it was just consumer items that have a life of 1-5 years the it could go "overnight" but non-consumer systems have far to much tied up in it for USB to go in less than a decade.

--
\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\
\/\/\/\/\ Chris Hills  Staffs  England     /\/\/\/\/
 Click to see the full signature
Reply to
Chris H

[snipped]

I used a Luminary Cortex-M3 device with 3 CAN interfaced with a number of AT90CAN devices via CAN. The AT90CAN devices had no trouble handling the CAN messages at 1Mbit with the CAN Bus loaded to 80%.

In the end I had to throw out the Luminary device because of the exceptionally poor CAN interface implementation. The CAN interface is an old 8-bit design which was clobbered together with the Coretx-M3 core. The Cortex-M3 core is very nice, but poor peripherals kills any and all advantages of the faster core.

Far too many of the ARM7, ARM9 and Cortex-M3 devices suffer because of the exceptionally poor periphals which are added.

I find the best peripherals on MCUs are from Atmel and Freescale.

[snipped]

Regards Anton Erasmus

Reply to
Anton Erasmus

In theory usually the peripherals from Atmel are good, but as in all chips, there are bugs, see e.g. for the TWI interface:

formatting link

I think I have found another bug: Sometimes a TWI interrupt was generated, even if I didn't enable it with the interrupt mask bit (in combination with PDC). I had to implement a workaround with my own simulated interrupt mask and testing it in the interrupt function against the actual source bit and ignoring unwanted interrupts. Now it works stable, but it needed some time, GPIO test points and a 4 channel digital oscilloscope to prove the bug and to implement the workaround.

--
Frank Buss, fb@frank-buss.de
http://www.frank-buss.de, http://www.it4-systems.de
Reply to
Frank Buss

I agree that a buggy implementation is a absolute pain to deal with, but at least if the design is sound, then it is just a matter of fixing the bugs in later revisions. If the actual design is bad, then even if it completely bug free, then it is still a pain to work with. I have found the NXP peripherals by-and-large to be badly designed AND very buggy.

On which Atmel device did you experience this TWI bug ?

Regards Anton Erasmus

Reply to
Anton Erasmus

Which NXP chips do you mean? At least the LPC series looks good in the datasheets.

It was the AT91SAM9G20.

--
Frank Buss, http://www.frank-buss.de
piano and more: http://www.youtube.com/user/frankbuss
Reply to
Frank Buss

Chris H expounded in news:n+Cv51Iur+ snipped-for-privacy@phaedsys.demon.co.uk:

..

Interesting. I also see USB 3.0 is coming. These things concern me from a test equipment pov, as parallel and serial ports are disappearing, and USB is everywhere (for now). I suppose adapter cards will always be available, but so much hinges on popular O/S support in your desktop of choice.

Warren

Reply to
Warren

In message , Warren writes

"Disappearing"? Gone on laptops a while ago. PC's I am not sure about but as you say...

This is true but I am sure that RS232 will be around in windows a long while yet.... remember embedded XP etc will still be used on embedded gear that will want serial.

--
\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\
\/\/\/\/\ Chris Hills  Staffs  England     /\/\/\/\/
 Click to see the full signature
Reply to
Chris H

USB 3 (and USB 2 before it) is backwards compatible - it's one of the rare occasions where the industry has learned from its mistakes.

It's still possible to get parallel port on laptop docking stations, but I haven't seen one on a new laptop for a long time. I've seen serial ports on the occasional fairly recent laptop.

Desktop PC's seldom come with parallel ports, but a few still have a serial port. However, it's very cheap and easy to put a parallel and/or serial port PCI card into a PC - they are fully supported by any decent OS and even by the BIOS.

The market for embedded XP is tiny, and windows is not much used by embedded developers looking for long term availability of software or hardware.

But I agree that serial port support will probably be around for a long time in windows. There is little reason to remove it, and serial ports are used by a huge number of devices. USB to serial port converters work fine in the majority of cases, so I don't see it being a problem for a good while.

Parallel port is less likely - most current use of parallel ports on windows involve "cheating" in some way with drivers that give programs direct access to the hardware. As windows is becoming more and more of a control freak that limits what you can do with your own machine, these will get harder to use. I don't know if it is at all possible to get such programs working on Win 7 64-bit with it's paranoia about driver signing, but it would certainly be a challenge.

Of course, it's always possible to use Linux, which is far better at working with older hardware, and is still available in every version ever released.

Reply to
David Brown

My HP has one (2 years old) but most "consumer" laptops don't

I agree but it is still cost in a VERY cost orientated market to put tracks and connector on a mother board.

But MS is pushing very hard into this market and it is a no cost thing to have... it would probably cost more to take it out than keep it.

Older devices.. Newer things are tending to have USB... this is a pity as in many cases RS232 is all that is needed and it is a doddle for any embedded engineer to do a serial link.

I thought it had gone!!!!

That is not true... over half the version of Linux are now obsolete and unsupported. Even if you can find them the drivers are still a problem and they may or may not work on other HW..... also of course it is no more stable than Windows and has absolutely appalling documentation if you can find it. As for maintaining a components to make a Linux..It may be necessary to acquire these from multiple independent source locations. For example, one distribution has to aggregate code from over

200 different open source projects for its distributions. Once downloaded, these components have to be ported to the target hardware, and then integrated with each other and maintained throughout the product life-cycle.

Development of significant Linux expertise may be required to create a new distribution in-house. This often includes adding staff to focus on maintaining the open source software, not doing new development. This issue is frequently overlooked, as existing staff and management may underestimate the effort required to do this in a production environment, believing that open source software is an off-the-shelf solution.

The time required to develop such a project is also often miscalculated. Training of staff may be slower than expected due to poor, out-of-date, or non-existent documentation. Open source or semiconductor vendor code may be less robust than required for a production environment, or may not account for the limitations of an embedded environment, requiring additional development time and/or staff. Linux drivers may be unavailable for new or proprietary hardware, requiring additional development time, and perhaps highly skilled developers with specialized expertise.

Linux is not the cheap or easy solution you suggest.

--
\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\
\/\/\/\/\ Chris Hills  Staffs  England     /\/\/\/\/
 Click to see the full signature
Reply to
Chris H

USB 3.0 Looks totally different from an interface perspective. It looks more like PCI Express than USB 2. The USB 3 connector seems to be basically 2 totally seperate and different interfaces on one connector. When plugging ina USB 2 Cable it uses the USB 2 interface. When plugging in a USB 3 cable it uses a totaly new and seperate interface. So to support both one will need totally separate interfaces. Of course if USB 3 becomes popular enough, it might be combined into one peripheral. This is my understanding after reading a little bit about the subject on the internet.

Regards Anton Erasmus

Reply to
Anton Erasmus

Anton Erasmus expounded in news: snipped-for-privacy@4ax.com:

..

Apparently it will do full duplex I/O. The only thing that I don't like is the increasing complexity of the interface. But I suppose it is already complicated enough to just warrant a "USB interface" pre-built for a hobby AVR project. I currently use the USB Boarduino, which is convenient in that respect.

Warren

Reply to
Warren

ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.