Why HCS08 not popular with hobbyists.

[Circlotron]

Just last night I bought the software and ISP programmer for Motorola / Freescale / NXP HCS08 micros and was left with 58 cents change from AUD$1K... It’s probably a rare hobbyist that is going to buy one of those. The point is, if the cost to get started with them was a whole lot lower, like with PIC for example, then a whole lot more people would start with them and probably continue to specify them if they made a career out of it. You would think the semi manufacturer would understand such a simple concept!

[nctnico]

They are not popular with companies either. None of the companies I worked for ever paid for development tools. If there where no free tools then a microcontroller is not used. Plain and simple.

I'm not sure if I ever worked with the HCS08 but I have used an 8 bit (automotive) Freescale microcontroller for an existing project of a customer. Getting the programming hardware to work was a major crap shoot. Mostly because Freescale (now NXP) uses their own proprietary protocol and the (expensive!) USB devices used for programming are flaky at best. Out of desperation I ordered a second USB programming device because I thought the one I had was broken but the second one behaved just as bad and the supplier (PE Micro) couldn't fix it.

A long time ago a Freescale rep. tried to get my employer at that time to try Freescale's 16 bit microcontrollers (coldfire IIRC) but that also stopped at needing the right programming hardware. So the development board went back for a refund.

[langwadt]

They are not popular with companies either. None of the companies I worked for every paid for development tools. If there where no free tools then a microcontroller is not used.
I'm not sure if I ever worked with the HCS08 but I have used a Freescale microcontroller for an existing project of a customer. Getting the programming hardware to work was a major crap shoot. Mostly because Freescale (now NXP) uses their own proprietary protocol and the (expensive!) USB devices used for programming are flaky at best.

yeh, I'm sure the money STM spent on making a debugger and including it on every dirt cheap dev board has paid for
itself many times 

[nctnico]

What has worked really well for me are simple serial (UART / RS232) port bootloaders. UARTs have internal filtering so a spike doesn't screw up programming like regulary happens when using JTAG or SPI like protocols.

[nctnico]

Clarification: I'm not saying I never used commercial tools but these had to be provided for free. For mainstream stuff there is enough competition to make this happen.

[westfw]

left with 58 cents change from AUD$1K

$1k like $1000?Why?  There OSHW/OSSW implementations of at least the basic programmer/debugger (http://usbdm.sourceforge.net/USBDM_V4.12/html/index.html)Implementations sell on aliexpress/etc in the $15-20 range.And I've heard that the (cheap) Freescale (now NXP) "Freedom" eval boards can have their programmer section used for the 8bit CPUs as well as the ARM chips.(Not that I've actually DONE either of these things. :-(  )

[PCB.Wiz]

yeh, I'm sure the money STM spent on making a debugger and including it on every dirt cheap dev board has paid for
itself many times

Yes, the big players like Atmel, SiLabs, ST, Microchip, Infineon, Renseas, Nuvoton offer low cost eval boards, with built-in Debug-Bridge MCUs, and many also allow the same board to use that  Debug-Bridge MCU as a external device programmer.

The fastest growing 8-bit MCU right now, looks to be the 8051 series, with many new Asian variants being added.
 

[james_s]

Some companies just never seemed to figure out that the way to sell chips is give away the software for free to absolutely anyone and make sure cheap development boards are available. The FPGA vendors didn't seem to get the memo either although at least they offer crippled versions of the software for free if you jump through a few hoops. A lot of engineers use the parts that they are familiar with so if they tinkered with AVRs or PICs or STM32 or whatever in their hobbies they'll use that family for a paid project.

[techman-001]

Just last night I bought the software and ISP programmer for Motorola / Freescale / NXP HCS08 micros and was left with 58 cents change from AUD$1K... It’s probably a rare hobbyist that is going to buy one of those. The point is, if the cost to get started with them was a whole lot lower, like with PIC for example, then a whole lot more people would start with them and probably continue to specify them if they made a career out of it. You would think the semi manufacturer would understand such a simple concept!

Wow, that's a lot of money to feed a Motorola ISA obsession

Whilst my first ISA love affair was with the MC6800 it eventually gave way to the STM32F0 ARM Cortex M0 for which almost all the tools are free, and STM even give away the complete unrestricted, un-crippled Atollic Studio, https://atollic.com/truestudio/ which is very impressive if you're into C or assembly. Just make sure you have a 4 core I7 with 32GB ram to run it all!

Even a 'low end' 32 bit STM32F051 with 64KB Flash, 8KB Ram, 48MHz internal clock and 37 onboard peripherals can be had for $0.56 each, which makes them very hard to resist.

While some of the worlds largest MCU manufacturers (such as STM) do give away the necessary tools, most don't and still cling to the old ways of charging a arm and a leg for tools to use their hardware, which has always struck me as crazy since I started doing embedded around 1976.

[ehughes]

You can get USB to BDM sub $20

There is even an open source project to convert a FRDM board into BDM.

http://usbdm.sourceforge.net/USBDM_V4.12/html/index.html

A Commerical BDM is $199.

I used an HCS08 in a commercial project about 3 years ago.  There are still good for a really cheap MCU application.

I never really did understand why companies will not pay for decent tools.   If you are sinking 10's of k$ a month into your engineers, even a $500 tool (i.e. JLINK) is a nothing burger.

[nctnico]

The problem is that if you use a few types of microcontrollers the total cost can easely add up to tens of $k just for tooling. At that point a cheap MCU no longer is cheap.

BTW: I think I have bought that open source USB BDM interface from someone and it also didn't work (and it wasn't $20; more like $120).

[james_s]

But why would a company spend money on tools if suitable tools for competing products are available for free? It would be different if the paid tool saved considerable engineering effort or allowed the use of a substantially superior part, but in most cases it doesn't. Then there is the fact that many of the engineers will probably choose parts they have used before in personal projects and the free tools greatly increase the chances of being chosen for that.

[nctnico]

But why would a company spend money on tools if suitable tools for competing products are available for free? It would be different if the paid tool saved considerable engineering effort or allowed the use of a substantially superior part, but in most cases it doesn't.

True. Another downside is that each professional tool comes with it's own crappy IDE made by the 'intern du jour'. Nowadays Eclipse is pretty much standard so there is some form of consistency.

[jaromir]

As a person who actually uses HCS08 - I'm using USBDM for a few hobby projects, it works with no hiccup. Apart from 20USD for hardware (because I was lazy to build one) I haven't paid a single cent for anything else. I tried their free of charge Eclipse based IDE, but settled on SDCC compiler and text editor later.

[Yansi]

Clarification: I'm not saying I never used commercial tools but these had to be provided for free. For mainstream stuff there is enough competition to make this happen.

Definitely. But it seems most corpos think otherwise

[blacksheeplogic]

But why would a company spend money on tools if suitable tools for competing products are available for free? It would be different if the paid tool saved considerable engineering effort or allowed the use of a substantially superior part, but in most cases it doesn't. Then there is the fact that many of the engineers will probably choose parts they have used before in personal projects and the free tools greatly increase the chances of being chosen for that.

Training costs, team size,  functionality, support, standards, project complexity, compatibility, perceived advantage, resource requirements, performance  to name a few.

The problem with hobbies is that everyone got a different hobby and make different choices, in a dev team everyone need to be on the same tool and taking a vote on who's preferred tools set you should use this month does not get the job done.

Also, cost has to include down time when you need support or run into a bug - typical forum support "reinstall and see if that fixes you problem" is only OK for a hobbyist on a budget.

[james_s]

Sure but that's going to be true for any tool, and at least the companies I've worked for have had the attitude that if we're gonna train everyone on a tool, we'll do so on the one that's free unless it is clear that a paid tool offers substantial advantages. If it's a case where an expensive paid tool is worthwhile for saving time or enables doing something that can't be done reasonably by the free tools then by all means use the paid tools. In quite a lot of cases though the expensive stuff is just as crappy or sometimes worse than the free options.

I don't work with hardware anymore in my day job but as much as ever the bulk of the tech industry is moving toward open source or bespoke internal solutions. The entire back end infrastructure of my current employer is on Linux using open source tools. The stuff we pay for is on AWS which is itself on Linux. There's not an expensive commercial development package in the whole joint.

[legacy]

BlackFin+, which is not supported by GCC yet.

BF537 here, but Gcc is not good for it.
I am using VisualDSP++ v5.1.1, with a B1K cable.

[legacy]

In quite a lot of cases though the expensive stuff is just as crappy or sometimes worse than the free options

 

[splin]

In quite a lot of cases though the expensive stuff is just as crappy or sometimes worse than the free options.

Quite. At least with the free stuff you do have a chance of connecting with the original developer(s) for support, or at least have access to the schematics and firmware. With many 'professional' solutions you have almost no chance of getting worthwhile support once the supplier has moved onto their next and greatest version. And because the developer has been made redundent and/or the software was lost in one of the many company reorganisations and takeovers that occurred shortly after you bought it.

Best you might get is the recommendation to upgrade to the latest and greatest Mk 7 bugfest - which might, or might not, fix the problem you have with the current version but will likely bring you a bundle of new improved and decidedly more 'interesting' problems. All for only $12K, less a generous 5% discount for being a naive schmuck loyal customer and to compensate for the $8K written off on the previous POS version.

[blacksheeplogic]

Professional tools from manufactures which are be partly funded by the cost of the tools and generally have much better support than a bunch of hobbits on a forum. Not all open source/free tools are bad but may really are quite fractionanalized and with very arrogant forum support (if any).

If it's my free time and a hobby/interest, sure digging though open source tools and figuring out some issue's might work. But if I have a delivery deadline I don't have the this time.

I spent my entire weekend working on a open source driver, it's diagnostics suggest rebooting the system. What do you think the forums/online support community think is the solution? I'm left to dig though thousands of lines of code trying to figure out what's wrong. It's my time, my project, my timeline so I'll deal with it but there's no way I would use it in a commercial project.

I'm not saying all open source projects are bad nor that all professional tools are the best. But on a project I want support and if the issue delays the project I want trace-ability of the actions taken and a next step owner, not a vague hope someone knowledgeable will read my post and provided something useful. I don't want next step on a delayed project being me digging though 10K lines of shitty code.

In quite a lot of cases though the expensive stuff is just as crappy or sometimes worse than the free options.

Quite. At least with the free stuff you do have a chance of connecting with the original developer(s) for support, or at least have access to the schematics and firmware. With many 'professional' solutions you have almost no chance of getting worthwhile support once the supplier has moved onto their next and greatest version. And because the developer has been made redundent and/or the software was lost in one of the many company reorganisations and takeovers that occurred shortly after you bought it.

[legacy]

I did mean BlackFin+, not the old BlackFin.

I know, and I said that even for the old BlackFin line gcc sucks.

[langwadt]

In quite a lot of cases though the expensive stuff is just as crappy or sometimes worse than the free options.

Quite. At least with the free stuff you do have a chance of connecting with the original developer(s) for support, or at least have access to the schematics and firmware. With many 'professional' solutions you have almost no chance of getting worthwhile support once the supplier has moved onto their next and greatest version. And because the developer has been made redundent and/or the software was lost in one of the many company reorganisations and takeovers that occurred shortly after you bought it.

Best you might get is the recommendation to upgrade to the latest and greatest Mk 7 bugfest - which might, or might not, fix the problem you have with the current version but will likely bring you a bundle of new improved and decidedly more 'interesting' problems. All for only $12K, less a generous 5% discount for being a naive schmuck loyal customer and to compensate for the $8K written off on the previous POS version.

with the added bonus of having to mess with license keys etc.

[Tomorokoshi]

In quite a lot of cases though the expensive stuff is just as crappy or sometimes worse than the free options.

Quite. At least with the free stuff you do have a chance of connecting with the original developer(s) for support, or at least have access to the schematics and firmware. With many 'professional' solutions you have almost no chance of getting worthwhile support once the supplier has moved onto their next and greatest version. And because the developer has been made redundent and/or the software was lost in one of the many company reorganisations and takeovers that occurred shortly after you bought it.

Best you might get is the recommendation to upgrade to the latest and greatest Mk 7 bugfest - which might, or might not, fix the problem you have with the current version but will likely bring you a bundle of new improved and decidedly more 'interesting' problems. All for only $12K, less a generous 5% discount for being a naive schmuck loyal customer and to compensate for the $8K written off on the previous POS version.

with the added bonus of having to mess with license keys etc.

Yes, getting licenses set up in corporate servers is almost as miserable as getting software tools. License and purchase requests require layers of cost justification and interaction with management that is deterrent enough to shift to less optimal tools and platforms.

Motorola/Freescale/NXP/PDQ/etc. in particular was difficult this way. For industrial embedded projects, dealing with a network license is very troublesome due to the difficulty of even getting a network connection in some equipment location. FlexLM in particular is detestable.

I brought this up several times to local Freescale representatives at the time years ago. They never really figured it out. It's no surprise that ARM and PIC took much their actual and potential business away as new products got developed and old ones got updated. This was unfortunate, because my personal preference was for M68K derivatives when possible.

Microchip is also moving towards a network license model, which combined with the Java-burdened MPLAB X IDE results in painful experiences.

The question I asked them at the time was whether they were in the hardware business or the software business. Processor vendors need to make up their mind.

[nctnico]

That is a good point about the licenses. Earlier this year I had to install a piece of software which uses a dongle. Getting it going took the better part of a day. Primarily due to the wrong drivers being installed  .