CS540 accelerator

[Daedalus]

Originally Posted by project23

This is also a concern.

My Furia came with a counterfeit freescale CPU and three dodgy 18 year old RAM chips of unknown origin (new old stock or salvaged). Whilst this is unlikely to be a health concern, it has been noted several times that risk of overheating and resultant burning would violate the CE marking.

In what way? Electronics like this are exempt from all directives except for EMC, which does not cover electrical safety. There are separate rules about that, but they're not CE directives and therefore not associated with being CE marked.

These things don't conform to CE marking at all, really, they're just stamped with the mark.

Are you sure? Have you asked the manufacturer for their declaration of conformance (which they're legally obliged to make available to customers)? Ask nicely and they should also happily provide all the supporting documentation from their DMR.

This is another reason why open hardware is the way to go for hobbyist systems. In amateur radio you're allowed to design and build your own equipment because it isn't intended for mass production. You can even sell it from one ham to another, because as Stephen suggested its a prototype of sorts.

If you're selling it to an end user for any purpose other than testing or evaluation (i.e., if they use it as a finished item), then it's not classed as a prototype.

Try and sell it to a mass audience however, and in this country at least you'll have ofcom, trading standards and the whole shebang right up you're a$$. You'd be in for a lot of trouble never mind the loss of your license.

Define mass audience? One user? 10? 100? What if you sell 10 hand-made prototypes? Or 100 hand-made prototypes? The thing is, there's no lower limit on what's considered a production run. If you make a bespoke, one-off device for a manufacturing line in a factory for example, that needs to be CE marked. As for regulatory bodies, they'll only get involved if there's a reported issue, so in most cases there's no checking. Radio is one area where you may get reported however due to the high possibility of interfering with a neighbour's equipment.

There's a lot of these low cost accelerators coming out of Europe lately. They're not bad, they're fantastic. But they're not sold with the legal protection and quality assurance of products built on a commercial scale like the ACA boards. Take a look at lothareks videos. It's a guy in a room at a desk putting these things together and then shipping them to either ebay buyers, or online Amiga stores. That isn't company, with liability, and assurance if something goes wrong. It is legally questionable.

How big does your desk have to be before it means you make better quality products? You also put Vampires in the same category as ACA boards, when it's well known that these are essentially assembled by hand, by a guy on a desk. (For entertainment value, you should look on Youtube for some of their soldering videos...)

I know i'm ranting here and i'm really not having a go at the designers or builders of these boards. A lot of work has gone into them and they work well, very well in some cases. All i'm saying is that designing a proprietary board, operating as a 'company' and selling these things en-masse it not a good idea.

In the UK and every other European country, if you even sell a single, completed electronic product, it must at the very least conform to the WEEE regulations, and to do that you must declare yourself as a company. You might not report any profits or loss and therefore not be a trading company, but you need to at least be registered.

Designing open hardware projects, and yes - selling the odd prototype privately whether for a profit or at cost - is both legal and fair to all.

Citation needed. As I already pointed out, it's pretty dubious whether selling a completed accelerator to an end user who is going to use it as an end product can be classed as a prototype, and therefore legal to sell.

You know what you're getting, and if you're up for the learning curve you can do it yourself.

I don't have a solution for you, but you're taking risk when you're buying these 'out of nowhere' proprietary cards with little to no technical support.

So what's the difference? Unless you're already well versed on electronics, the fact that a card is open-source doesn't make it any more likely that you can solve any problems that crop up. Also, by being open-source means that literally anyone can attempt to build it and sell it, meaning that buying one from a hobbyist on eBay is probably even less likely to be reliable than buying from the original developer, or a commercial product sold through Amiga stores.

Just search these forums for how many users have had trouble with their Furias, for example, and i'm not talking 'poor socket seating' - and that's just on the English forums! At least if it was open-source, you could have a friend or a professional fix the problem for you.

Given how very popular those boards are, and how they operate at the limits of the Amiga's capability, it's not surprising that there are many users having problems. For comparison, search these forums for people having problems with TerribleFire accelerators. There are a lot there too, and that's not at all a reflection on plasmab's design or the board itself, that's just how it goes in these situations when a board is very popular and people use them to push 30-year-old machines beyond their original designs. ACA accelerators also crop up commonly with issues of their own too, and Apollo accelerators were infamous for causing problems back in the '90s.

You would have experienced forum users able to look at the schematics and help you troubleshoot. The solution offered for my problem? I was asked to send my Furia back to Eastern Europe (forget the country, i'd have to check) for 'testing', covering postage myself, no legal assurance that i'd even get it back, and with a charge for labour despite the fact that the board had a counterfeit processor and faulty RAM chips. The response from the designer was one line, requesting i send the board back and that go to PM for more details. I'm still awaiting a response for the possibility of a blank PCB so I can transfer my working parts over from a broken board. (Long story, i fixed all the problems myself, then had a bit of an accident and bloody destroyed it. See my other posts for details).

You really do have a bee in your bonnet about the Furia, don't you? Well, who did you buy it from? Presumably a commercial outfit, who have a legal obligation to their customers. If it was in warranty, get it replaced by the distributor under warranty. If it was out of warranty - well, how did it take you that long to find it doesn't work?

[plasmab]

Originally Posted by Daedalus

Are you sure? Have you asked the manufacturer for their declaration of conformance (which they're legally obliged to make available to customers)? Ask nicely and they should also happily provide all the supporting documentation from their DMR.

I've never understood how CE marking enforcement works in the EU. It seems to me either people have a really cheap source of EMC testing that they're not telling anyone about or they're not bothering.

Given the penalties for violation I opted not to sell my boards myself.

[Daedalus]

Originally Posted by plasmab

I've never understood how CE marking enforcement works in the EU. It seems to me either people have a really cheap source of EMC testing that they're not telling anyone about or they're not bothering.

Well, for most cases of CE marking (including the EMC directive), self-certification is the normal approach. Normally companies will get a lab to do the testing, which is where the big money comes into it as you know. But there's no requirement to register the product with a reporting body in self-certification situations, which means the enforcement doesn't enter the picture unless a product is reported to the manufacturer's country's enforcement body, and that's only likely to happen when someone reports a legitimate complaint about a product interfering with their TV or radio.

The other side of things is that the directive doesn't prescribe any limits, targets or test methods. It's up to the manufacturer to select an appropriate approach that documents their compliance with the directive. Generally this will be testing for the limits of a standard (typically ISO or local harmonised equivalent), but any approach is allowed so long as it's justified. If you can write a good justification as to why your device doesn't need to be tested in an EMC lab, that will be enough. Of course, it needs to stand up to scrutiny from the reporting body should you ever be audited, so "I didn't feel like it" doesn't wash here. A detailed, comprehensive write-up on the frequencies used in the device, including all the calculations for the worst-case traces on the board and the expected radiated and conducted interference, how you addressed these (calculations for bypass capacitor size), comparative test results performed with a spectrum analyser and everything else like this should be a reasonable justification. Of course, compiling such a justification can take days or weeks, but if it's sound and does actually justify the lack of testing, it could save you thousands in unnecessary lab fees.

Given the penalties for violation I opted not to sell my boards myself.

Yeah, some of the penalties can be pretty scary indeed. But if you can show that you've taken reasonable steps to meet the requirements of the directive and weren't deliberately ignoring the standard, normally you'll just be told to stop selling the product and maybe recall those in the field for corrective work. You won't get a fine or prison time unless you really take the p1ss on a grand scale.

Note: obviously don't take this as legal advice - it's clearly not. But wrangling ISO standards and CE & FDA approval for medical devices (including electronics) is a significant portion of my day job.

[plasmab]

Indeed. Thanks for that. The issue i had was that unless if fixed the crystal in place you couldn't be sure what frequencies might get generated. Which is one reason i fought against people who wanted a replaceable crystal. The other reason was space obviously.

I guess I could argue that it would sit inside the A500 shielding but i'm not sure if that would fly. It doesnt physically fit.

[plasmab]

Originally Posted by FeRu

Seriously??? This is one of the most advanced and promising project since Vampire,.
and I see so much negative opinions or some trash speaking like discussion about CE logo....

Its not trash talk. Its a genuine question. By the book its illegal to mark something like as CE in the EU without EMC testing. If someone has a way round the EMC regulations then I want to know about it because thats the reason I cannot do an assembly run of my cards.

I dont want to get banned from being a company director or do prison time.

It seems that everything that comes out of Poland has CE on it. But the vampire doesnt have CE marking on it for example. They respect the rules. Is there some dude in Poland that does EMC testing for $100???

Originally Posted by project23

For the benefit of non-polish speakers - this is the google translate of FeRu's last couple messages from the Polish form:





Oh and that's just him - there are worse, it's just that he's the guy who posted here.

I actually thought his registering here to make a really rather balanced and resolute post was admirable, but clearly now we see what's going on at the other end.

I'll say this again - i have no problem with this new accelerator, i'm sure its great.

The problems I have had are genuine problems with the way business is done around these new accelerator boards and how they are manufactured, sold, and most importantly the aftercare and attitude towards those with issues. These problems are real, and they are not personal. On EAB I see people conducting themselves rationally about rational issues. That's more than can be said for the thread I just quoted from.

Thanks for your understanding

John

Yeah google translator is sh*t. In those posts I mean something else and those opinions are simply jokes or sarcasm. If you misunderstood those words because of translator then sorry but it's not our fault.
You or anyone can simply sent email or post in ppa.pl and ask about project details in english, and everything will be clear instead of trash talking and joking. Maybe we overreact a little but but put yourself in our position.....
Polish Amiga community imho is one of the strongest in world ( right behind UK and GER),so you shouldn't be in state of shock that someone with skill and good idea made new expansion for our beloved computer line.
Cheers

[project23]

Originally Posted by Daedalus

You really do have a bee in your bonnet about the Furia, don't you?

There is a reason, and I will concisely explain why shortly - i'm a bit busy for the next hour or so, so this post is a place holder.

A lot of your response to me seems quite heated, and that's fine - I note that you're in the very trade that I am making assumptions around. If i am wrong, then obviously i'll concede to better knowledge, but there really is no need for us to get heated over any of this, agreed?

You've actually raised some very interesting points that have made me re-think my outlook a little. Especially regarding 'home project' vs 'commercial' etc.

I'll get back to this thread when I can, and respond to you properly - but please, this 'bee in your bonnet' stuff - it just isn't necessary or even justified. I have tried the warranty system you refer to, and it has proven to be, shall we say - not up to standard? Or at the very least i've been left with zero confidence. As we speak I am talking via email to Lotharek also seems to take issue with the same 'commerical supplier' i bought the Furia from. So it isn't as simple as you might think.

Speak later - kind regards

John

[project23]

Originally Posted by FeRu

Yeah google translator is sh*t. In those posts I mean something else and those opinions are simply jokes or sarcasm. If you misunderstood those words because of translator then sorry but it's not our fault.
You or anyone can simply sent email or post in ppa.pl and ask about project details in english, and everything will be clear instead of trash talking and joking. Maybe we overreact a little but but put yourself in our position.....
Polish Amiga community imho is one of the strongest in world ( right behind UK and GER),so you shouldn't be in state of shock that someone with skill and good idea made new expansion for our beloved computer line.
Cheers

Again let me make this clear...

I respect the Polish amiga community. I own a Furia and it is an incredible card. I also respect the Polish. This is nothing personal at all.

I am also not in a state of shock over your accelerator card. I'm sure its brilliant. I don't understand why you think this is some kind of jealousy issue?

I haven't trash talked once. Once. I have been respectful whenever I have shared my opinions and even my frustrations.

I think thats as clear as I can be, and i think google translate can do a good enough job of translating it to Polish if necessary.

I really should get back to work, lol.

John

[Daedalus]

Originally Posted by project23

A lot of your response to me seems quite heated, and that's fine - I note that you're in the very trade that I am making assumptions around. If i am wrong, then obviously i'll concede to better knowledge, but there really is no need for us to get heated over any of this, agreed?

I'm really not, I'm just clarifying things since there are a lot of assumptions being made (and to be fair, are made by many people everywhere, every day, not just in Amiga or retro computing context). From my point of view it's your post that was getting heated - indeed, you even call it a rant.

You've actually raised some very interesting points that have made me re-think my outlook a little. Especially regarding 'home project' vs 'commercial' etc.

that was the hope.

I'll get back to this thread when I can, and respond to you properly - but please, this 'bee in your bonnet' stuff - it just isn't necessary or even justified.

Relax, it's a pretty tame phrase compared to a lot that get bandied about. You're frustrated, perhaps justifiably so, but I think you're being clouded in your judgement because of that.

I have tried the warranty system you refer to, and it has proven to be, shall we say - not up to standard? Or at the very least i've been left with zero confidence.

Unfortunately, that's more common than it should be, perhaps due to the small nature of the suppliers in the market. But that's where the problem is, not with the manufacturer, and not with the regulation (or lack thereof). A commercial seller is obliged to comply with all relevant laws regarding warranty, suitability for purpose and so on.

As we speak I am talking via email to Lotharek also seems to take issue with the same 'commerical supplier' i bought the Furia from. So it isn't as simple as you might think.

Initially, the suppler must honour the warranty by law, simple as that. The minutiae of that (length of warranty etc.) varies from country to country, but at least initially, the supplier must honour the warranty. The repair or replacement of that board is a matter between the supplier and the manufacturer once it has been returned, but that's nothing to do with the customer. After a time (typically 6 months to a year), the warranty changes over and becomes the manufacturers responsibility, but the vast majority of faults with electronics either happen in the first few weeks or after many years. So, if the fault was reported to the supplier during the first 6 months (taking the UK as an example, and you have to comply with the laws of the territory into which you sell, not where you are based), then it really should be as simple as that.

[hooverphonique]

Originally Posted by Daedalus

Well, for most cases of CE marking (including the EMC directive), self-certification is the normal approach. Normally companies will get a lab to do the testing, which is where the big money comes into it as you know. But there's no requirement to register the product with a reporting body in self-certification situations, which means the enforcement doesn't enter the picture unless a product is reported to the manufacturer's country's enforcement body, and that's only likely to happen when someone reports a legitimate complaint about a product interfering with their TV or radio.

The other side of things is that the directive doesn't prescribe any limits, targets or test methods. It's up to the manufacturer to select an appropriate approach that documents their compliance with the directive. Generally this will be testing for the limits of a standard (typically ISO or local harmonised equivalent), but any approach is allowed so long as it's justified. If you can write a good justification as to why your device doesn't need to be tested in an EMC lab, that will be enough. Of course, it needs to stand up to scrutiny from the reporting body should you ever be audited, so "I didn't feel like it" doesn't wash here. A detailed, comprehensive write-up on the frequencies used in the device, including all the calculations for the worst-case traces on the board and the expected radiated and conducted interference, how you addressed these (calculations for bypass capacitor size), comparative test results performed with a spectrum analyser and everything else like this should be a reasonable justification. Of course, compiling such a justification can take days or weeks, but if it's sound and does actually justify the lack of testing, it could save you thousands in unnecessary lab fees.

If you can show that you've taken reasonable steps to meet the requirements of the directive and weren't deliberately ignoring the standard, normally you'll just be told to stop selling the product and maybe recall those in the field for corrective work. You won't get a fine or prison time unless you really take the p1ss on a grand scale.

This is exactly how the workings of this has been explained to me also (by somebody I trust).

[Daedalus]

Originally Posted by plasmab

Indeed. Thanks for that. The issue i had was that unless if fixed the crystal in place you couldn't be sure what frequencies might get generated. Which is one reason i fought against people who wanted a replaceable crystal.

Yep, and further complicated by the harmonics that might be generated by oscillators at frequencies other than the design frequency. If people modify their own boards then that's their own lookout though, and you can easily wash your hands of any resulting issues.

I guess I could argue that it would sit inside the A500 shielding but i'm not sure if that would fly. It doesnt physically fit.

You might be better off arguing that it makes no difference to the level of interference produced since the noise from an unshielded 500 would swamp it anyway, resulting in no measurable difference between a 500 with a TF board fitted and one without.

By the book its illegal to mark something like as CE in the EU without EMC testing.

Well, that's not strictly true, though it's generally accepted as the standard process in industry. The situation is like this for these sorts of products:

• It's illegal to mark something with CE if it does not fall under the scope of any CE directives, so very simple, passive devices should not be CE marked (for EMC compliance at least - they may still be subject to other directives)
• It's illegal to mark something with CE without a formal (and legal) declaration that guarantees that:
  • It does not cause adverse interference with other electronic products
  • It does not react adversely to interference from other electromagnetic sources

That's it. If you personally declare that these two conditions are true for your product, then you can stick a CE mark on it, and it's legal. The key thing is your declaration, which is a legal document, and is why manufacturers either test or justify their guarantee instead of simply saying "it's all good". A manufacturer needs to be able to stand by their declaration, documenting why they believe it's true, but what gives them that belief is entirely up to them.

Is there some dude in Poland that does EMC testing for $100???

I doubt it, but there might be a guy there writing justifications for $100

[plasmab]

Originally Posted by Daedalus

That's it. If you personally declare that these two conditions are true for your product, then you can stick a CE mark on it, and it's legal. The key thing is your declaration, which is a legal document, and is why manufacturers either test or justify their guarantee instead of simply saying "it's all good". A manufacturer needs to be able to stand by their declaration, documenting why they believe it's true, but what gives them that belief is entirely up to them.

Interesting. However the original MC68030/040s are not RoHS compliant and therefore clearly violate the CE marking? Translation: Can you mark something as CE if its shipped with products that might have lead in them?

[Daedalus]

Originally Posted by plasmab

Interesting. However the original MC68030/040s are not RoHS compliant and therefore clearly violate the CE marking? Translation: Can you mark something as CE if its shipped with products that might have lead in them?

Yeah, things get a little murky there. Parts manufactured before 2006 won't be RoHS certified, but that doesn't automatically mean they contain restricted substances. If you can prove that they don't contain lead, cadmium etc., then you can use them - effectively having them RoHS certified yourself. This of course is gonna cost you money in lab fees, but might be worth it if you have boxes of chips lying around the place.

Perhaps more relevant is the fact that there are many valid exemptions from the RoHS directive. Certain classes of medical devices are exempt, as are certain aspects of automotive, aviation, space and military equipment. One exemption in particular states that parts and products manufactured with the sole purpose of upgrading or repairing electronic equipment placed on the market prior to 2006 do not need to meet RoHS regulations. This is essentially for keeping old electronics out of landfill by extending their useful lives, but I'd say there's a pretty good argument for using this when it comes to Amigas. In this case, the RoHS directive requirements aren't applicable and a CE mark still applied based on the EMC directive, though it would be wise to have that decision and justification well documented in the design files. The WEEE regulations still apply in this situation however, so the wheelie bin symbol should be applied in any case.

[project23]

Originally Posted by Daedalus

I doubt it, but there might be a guy there writing justifications for $100

Quick question - where do counterfeit components and NOS components used in a new design fit into this, or other legislation.

Now please note the problem with my Furia was not CPU based, it was a few bad ram chips as i'm sure you'll know if you've read my rants.

But say you have EMC testing or a legal declaration for a particular prototype, and then you sell that product with varying types of NOS (or even just salvaged) components as well as counterfeit components different from those used in the original prototype. Does that in any way invalidate the declaration/EMC test?

Also, with regard to counterfeit chips coming out of China - must you be able to show that the components used are genuine and thus themselves conform to their own degree of whatever testing is required? Can you EMC test or write a declaration for one board and then sell a bunch of other boards with different chips on - all under the same declaration?

I'm genuinely just curious here.

See the Furia I and Solderbro have (and i'm sure a few others) has a CPU mask that idrougge claims is fake, and a date code that as far as i can research is patently impossible. Freescale stopped production of the 020 in 2010, these are stamped 2015. They carry the Freescale logo. They work!.. ish. They just shouldn't exist!

I know this is splitting hairs - again its not related to my problems just a question of curiosity - but presumably if the chips used on your board are provably fake, then any EMC/safety/whatever requirements Freescale/Motorola were required to comply with (if any?) are no longer valid?

What about second hand parts? If its genuine NOS then i'm going to guess its fine? What if its pulled from an old device? Does that change anything?

I'm in a particularly hyper mood this week, as i'm sure you may have noticed , so if i'm bombarding you with irrelevant questions just tell me to shut the f&(k up, haha - but as i say i'm just curious. Particularly with reference to the fake chips - which unfortunately i believe we are increasingly unable to avoid. (I don't for a second believe Lorathek et al knew or in any way chose to use counterfeit parts).

Cheers,

John

[Daedalus]

Originally Posted by project23

But say you have EMC testing or a legal declaration for a particular prototype, and then you sell that product with varying types of NOS (or even just salvaged) components as well as counterfeit components different from those used in the original prototype. Does that in any way invalidate the declaration/EMC test?

In principle, if they're different, then they do invalidate the declaration, yes. What should really be done is that the design allows for a variety of configurations, say 2 or 3 different types of RAM chip. This would let you chop and change at any point in the production. Changing to a different chip without that preparation would indeed require at least a rationale as to why the declaration still stands. Most of the time, with genuine parts, a comparison of the datasheets of two similar parts will give you enough information to justify the change without further issues. For example, changing a chip from a National Semiconductor to a Texas Instruments equivalent will likely be no problem so long as their specs match well enough (current consumption is a key thing to look at, as are transient times and so on, as these will have a bearing on the generated noise). Such "like-for-like" changes are expected and accepted without too much trouble.

If there is a difference that might result in a different noise profile, the calculations made in the original justification would have to be re-done and the board adjusted to suit if necessary, or if it was tested in a lab, retesting would be required. Again, some good justification documentation here could save a lot of money, e.g. testing just the affected area of the board for changes from the original measurements instead of retesting the entire system.

Also, with regard to counterfeit chips coming out of China - must you be able to show that the components used are genuine and thus themselves conform to their own degree of whatever testing is required? Can you EMC test or write a declaration for one board and then sell a bunch of other boards with different chips on - all under the same declaration?

Yeah, with counterfeit chips, all bets are off. Genuine chips will all work closely enough to each other that testing of a sample is applicable for the entire production run, but there are no guarantees of that for counterfeit chips. Knowingly using dodgy chips like that is asking for trouble alright, though if the chip supplier provided authentic-looking documentation it could be a genuine oversight.

See the Furia I and Solderbro have (and i'm sure a few others) has a CPU mask that idrougge claims is fake, and a date code that as far as i can research is patently impossible. Freescale stopped production of the 020 in 2010, these are stamped 2015. They carry the Freescale logo. They work!.. ish. They just shouldn't exist!

Ugh... It would be interesting to see what paperwork (if any) was available from the chip suppliers here.

I know this is splitting hairs - again its not related to my problems just a question of curiosity - but presumably if the chips used on your board are provably fake, then any EMC/safety/whatever requirements Freescale/Motorola were required to comply with (if any?) are no longer valid?

Yep, that would be true.

What about second hand parts? If its genuine NOS then i'm going to guess its fine? What if its pulled from an old device? Does that change anything?

NOS parts are indeed fine, though things like RoHS could be an issue depending on whether it's deemed applicable. Pulling from old devices is fine too, and is explicitly covered by the RoHS directive in the same way. That is, if it prevents those parts ending up in landfill, it's a good thing. But given that they're probably not RoHS compliant, the same applies as the NOS comments.

so if i'm bombarding you with irrelevant questions just tell me to shut the f&(k up, haha - but as i say i'm just curious.

Haha, I wish I had the energy to he hyper

Particularly with reference to the fake chips - which unfortunately i believe we are increasingly unable to avoid. (I don't for a second believe Lorathek et al knew or in any way chose to use counterfeit parts).

Yeah, and it's not a new problem either. With most semiconductors being so cheap nowadays it's less prevalent at the lower end of the market, but a couple of decades ago there were instances of standard logic chips being faked and even making it into the supply chain through industrial suppliers. Some of these still circulate unfortunately, for example the popular MAX232 regularly shows up on eBay from a faulty batch that were likely taken from a scrap pile in some factory and sold online. These have a tendency to latch to the rails and eventually burn themselves out, and resulted in lots of complaints and a damaged reputation.

FTDI (a Scottish company incidentally) have also had their USB-serial chips cloned and used worldwide in thousands of different devices, often unknowingly by the device manufacturer, and only being discovered years later when an updated driver stopped the counterfeits (and thus the devices built around them) from working.

Unfortunately for the manufacturers in these cases, buying chips directly from China means there's very little discourse when they're suddenly swamped in returned faulty devices. Buying them from EU or US sources would offer some protection, but that's not always possible.

[supaduper]

Originally Posted by plasmab

Is there some dude in Poland that does EMC testing for $100???

Sorry Steve that cracked me up lol

[project23]

Originally Posted by Daedalus

Yeah, and it's not a new problem either. With most semiconductors being so cheap nowadays it's less prevalent at the lower end of the market, but a couple of decades ago there were instances of standard logic chips being faked and even making it into the supply chain through industrial suppliers. Some of these still circulate unfortunately, for example the popular MAX232 regularly shows up on eBay from a faulty batch that were likely taken from a scrap pile in some factory and sold online. These have a tendency to latch to the rails and eventually burn themselves out, and resulted in lots of complaints and a damaged reputation.

FTDI (a Scottish company incidentally) have also had their USB-serial chips cloned and used worldwide in thousands of different devices, often unknowingly by the device manufacturer, and only being discovered years later when an updated driver stopped the counterfeits (and thus the devices built around them) from working.

Unfortunately for the manufacturers in these cases, buying chips directly from China means there's very little discourse when they're suddenly swamped in returned faulty devices. Buying them from EU or US sources would offer some protection, but that's not always possible.

I don't know much about chip counterfeiting other than the old 'sand off the label and print a better one' method. I know actual fake fabrication happens and comes out of China, but just not to what degree.

In your opinion, would it be difficult/worthwhile for a fab in China to clone the likes of an 020?

See - if its just a re-branded 16mhz model with a dodgy date, then it shouldn't work at 33mhz. Lotharek is right about that. It must be a 25mhz model. But at the same time it couldn't possibly have been manufactured in 2015.

What about rejects? Could companies in China get hold of reject or returned 25mhz chips and re-label them with a recent date?

It's just baffling to me. Seems like a lot of effort just to sell to a niche motorola 020 market - unless like I say its rebranded rejects/returns, but 25mhz models... I really don't get it.

John

Haha, I wish I had the energy to he hyper

Oh don't worry, i'll get the downside too! haha

[project23]

Originally Posted by Daedalus

In the UK and every other European country, if you even sell a single, completed electronic product, it must at the very least conform to the WEEE regulations, and to do that you must declare yourself as a company. You might not report any profits or loss and therefore not be a trading company, but you need to at least be registered.

A thought just came to mind...

If a friend sends me his TF530 PCB, and his parts, and i build it for him, and return it to him. For free. Thats just hobbyist electronics, right? I mean there's no harm in that?

Okay so what if I charge him for assembly? I'm not selling him a product - i'm giving him his product back having tinkered with it. I'm assembling it for him in exchange for money - a service.

I'm talking about threads the likes of 'send me your parts and i'll build your 530 - it'll cost £30 a board for me to build and send it back' for example.

Would that require the same sort of CE certification as you've already mentioned?

John

[Daedalus]

CPU speeds are (or perhaps were) often determined based on testing and market demand. So (assuming the same package) all 68020s start off the same, and during testing it's found that some work fine at 33MHz, others not. Those that work are labelled 33MHz, others are labelled 16MHz, even though they're basically the same part. But if there's a batch where all the chips work at 33MHz, and an order for 16MHz parts is waiting, a certain number of the parts will be marked 16MHz anyway, just to fill the order. These parts will happily "overclock" to 33MHz, and so depending on your luck you could end up with something like that. Relabelling chips is probably fairly common (I once saw an 030 accelerator where they didn't even bother, just ground away the frequency designator so you couldn't see it was overclocked), and the selling of reject or "beta" test parts through side markets is definitely a big thing with semiconductors valued at more than a few quid each.


Yeah, that's a work-around I guess. It all gets a bit hazy, but kits are exempt, and charging just for labour is entirely out of the scope of CE marking. I've even heard of fully built boards that just need one connector soldered to it being sold as "kits", which is probably a bit of a stretch, but seems to be a used work-around.

[project23]

Originally Posted by Daedalus

CPU speeds are (or perhaps were) often determined based on testing and market demand. So (assuming the same package) all 68020s start off the same, and during testing it's found that some work fine at 33MHz, others not. Those that work are labelled 33MHz, others are labelled 16MHz, even though they're basically the same part. But if there's a batch where all the chips work at 33MHz, and an order for 16MHz parts is waiting, a certain number of the parts will be marked 16MHz anyway, just to fill the order. These parts will happily "overclock" to 33MHz, and so depending on your luck you could end up with something like that. Relabelling chips is probably fairly common (I once saw an 030 accelerator where they didn't even bother, just ground away the frequency designator so you couldn't see it was overclocked), and the selling of reject or "beta" test parts through side markets is definitely a big thing with semiconductors valued at more than a few quid each.

The thing is, the 020 didn't have a 33mhz model. Just 16 and 25 (i'm talking the PQFP model here, the ec020). So right off the bat the Furia is overclocked from 25 to 33.

What you're saying could still be very true though. 16's re-badged as 25's could well still happily run at 33. At the end of the day if they're the essentially identical, as you say, to the 25's - then they'd overclock just as much as genuine 25's.

It's the unlucky few for whom that doesn't work so well. As I say Solderbro had to take it down to 25 again to get it stable, but mine seems okay at 33 despite them having the exact same print on the chip.

Yeah, that's a work-around I guess. It all gets a bit hazy, but kits are exempt, and charging just for labour is entirely out of the scope of CE marking. I've even heard of fully built boards that just need one connector soldered to it being sold as "kits", which is probably a bit of a stretch, but seems to be a used work-around.

Yeah exactly... This isn't the kind of thing i'd do as a business, but as a favour to some forum members who didn't have the tools or skill to build their accelerator, ya know?