English Amiga Board


Go Back   English Amiga Board > Support > support.Hardware

 
 
Thread Tools
Old 26 September 2018, 00:29   #21
project23
Used Register

 
Join Date: Jul 2018
Location: Liverpool
Age: 35
Posts: 260
Please see follow up post here:

http://eab.abime.net/showpost.php?p=...&postcount=251

The problem was with three RAM chips. Replacing them fixed the board - and it ran *amazingly*. Unfortunately i then accidentally destroyed some of the traces to the 020, and it is useless.

I am hoping I can buy a blank furia PCB so that I can transfer the components.

Cheers all,

John
project23 is offline  
Old 10 October 2018, 00:28   #22
project23
Used Register

 
Join Date: Jul 2018
Location: Liverpool
Age: 35
Posts: 260
Okay, bit of an update on the counterfeiting thing:

I've actually enjoyed investigating this little mystery, so again there is no hard feelings towards either Lotharek or Boris who I think have created and sell a great board. I have hard feelings for their supplier, who I think is knowingly selling counterfeit chips. Whether this is a problem for accelerators, or as replacements in already existing older machinery, I really don't know.

Anyway... After finding the news articles already mentioning a cease of manufacture in 2010, I went to NXP themselves to ask direct about the chip. (I did not mention what the chip was included on, and would not.) At first they were not so helpful, you can read the full thread here, but the relevant excerpt is quoted below:

Quote:
Freescale stopped taking orders for MC68EC020FG25 in 2006. I cannot tell you when Freescale made the last MC68EC020FG25.
Now obviously that's orders, not chips rolling off the production line. So it stands to reason that the 2010 date mentioned in the news article is at least possible.

Certainly a date of 2015 on a chip last ordered by suppliers or manufacturers in 2006 is ridiculously unlikely.

There's more:

I found this official document from Freescale noting product discontinuance in 2010. There is no 020FG25 or 020FG16, but there are 020AA25 and 020AA16.

Now every datasheet for the AA variant at every distributor gave the same FG variant (the general EC020 datasheet). The chip was functionally identical. It took me a while (because i'm stupid) to realise that the AA variant is the RoHS compliant version of the chip. RoHS 1 took effect in 2006, and guess when the AA variant was released? (I've lost the source for this but i'll dig it out). 2006.

So you can bet that after 2006 there was no possibility of an 020FG chip being legitimately manufactured for distribution in the EU. (but possibly outside of EU).

The official Freescale discontinuance document however, lists fabrication plants from Hong Kong to Belgium, and all of them list the EC020AA variant of any kind as discontinued. Again, this is in 2010, so you can also bet that after 2010 there wasn't a single FG25 or FG16 built by Freescale

As for the chip itself, physically, it is a reasonable fake at first glance (it looks legit). However even a cursory glance at this document about how to spot counterfeit components, gives some obvious clues.

The most obvious is that the usual 'mottled' (bumpy, uneven) texture of the chip surface is completely missing on the top only. In fact it is so smooth that it's somewhat satisfying to run ones finger over it. The whole chip should be 'mottled', as noted in the article, and even if it legitimately isn't - there should be no change of texture around the chip. In my chip the bottom has the usual mottled texture as my genuine Motorola 020.

This is evidence that the chip has been 'blacktopped' (as referred to in the article). The font used on my chip vs a genuine Motorola is also different, and so far as i can tell from images of genuine Freescale chips, the logo is too big.

I intend to do a proper thread about this at some point, when I have more information/sources at hand. I intend to use my microscope to take some close pictures of the surface of both chips.

I have also ordered an FG16 from ebay. Why? Because a very good point made by Lotharek, was that if these were blacktopped 16's - sure they might run at 25, but the Furia runs stable at ~32. How is that possible??

Well... it turns out that at least some 16's were made with the same photomask as the 25's. If the demand at some point during the use of this photomask was for the 16's, then a lot of 25's would have been binned as 16's as was often the case.

My intention (if I get my Furia working again), is to try and ramp up the speed from 25 to 32 on a 16mhz EC020 to see if it will handle it. Maybe it will, maybe it wont. Maybe what I receive in the post will be a lump of lead, because you wont believe how difficult it is to get a genuine looking FG16, even from China, whilst for some reason China seem to be swimming in FG25s! I guess we'll find out when it arrives.

Anyway that's all I have for now - like i say this 'investigation' has been something of a short term hobby, not a complaint. It also serves as a warning - to buyers of chips and in particular designers of accelerators.

PGA chips are easy to remove, clean, and re-sell. SMD chips are not (to the layman). In my eyes the only way genuine discontinued SMD chips can be bought en-masse is through - for a time - people in the industry who have found a few trays of NOS chips in the back.

Like I said above - i know i'm repeating myself - when I have everything sorted out and written more concisely, with more sources and microscope images etc, i'll write a thread warning of counterfeit QFP 020 chips on ebay.

Until then, to the few who actually made it to this point, thanks for ya time and take it easy!

Cheers all,

John

Last edited by project23; 10 October 2018 at 11:28.
project23 is offline  
Old 10 October 2018, 18:05   #23
8bitbubsy
Registered User

8bitbubsy's Avatar
 
Join Date: Sep 2009
Location: Norway
Posts: 1,278
My Furia EC020 v2 (white PCB) is as stable as it gets, I have run the Amiga for hours at the time. I guess it has a real 68020 then...
If you want, I could remove the heatsink and check what the text on the CPU says, for science.
8bitbubsy is offline  
Old 10 October 2018, 22:28   #24
project23
Used Register

 
Join Date: Jul 2018
Location: Liverpool
Age: 35
Posts: 260
Quote:
Originally Posted by 8bitbubsy View Post
My Furia EC020 v2 (white PCB) is as stable as it gets, I have run the Amiga for hours at the time. I guess it has a real 68020 then...
If you want, I could remove the heatsink and check what the text on the CPU says, for science.
That's entirely up to you, though the data would be appreciated if it isn't too much trouble!

Again, i'm not suggesting these CPUs are causing problems for the Furias. I spent a lot of today (took some pics first) scraping/sanding the blacktopping in the hopes of finding another engraving underneath - i didn't, so they've done a good job,

One theory is this - some 16's are 25's binned as 16's. Probably a lot. All the counterfeiters need to do is run 16's at 25mhz and check temperature and stability. Then again, I don't know how difficult it would be to set up a rig to do this systematically. The ones that qualify is 25's are sold as 25's, and for all intents and purposes are 25's. This would make the effect on the Furia negligible.

Another theory is that because the 0E13G photomask is shared by both the 16 and the 25, if the yield is high enough, then most 16's with the 0E13G mask would run at 32 as well as a 25 with the same 0E13G mask. They're essentially the exact same chip - just binned according to performance. Don't forget these chips were sold for years in embedded systems - well over 10 years, maybe 15? I imagine they had the process pretty efficient.

Thats why i've ordered a 16 in the hope that its a 0E13G so I might test that theory.

Oh, and again, its near impossible to find an EC020FG16 on ebay that doesn't cost the world - even from China. Yet seriously, there are multiple sellers with what appear to be full on trays of FG25's. Suspicious to say the least.

OH! And a note on CE compliance. Again not that i'm having a go specifically at the Furia here. But it is my understanding that products sold after 2013 cannot be CE certified if they do not pass RoHS 2. The EC020FG line is non RoHS product, and was replaced by the EC020AA line (which is). This AA line was discontinued as per the link in the previous message.

So at the very least , any FG16/25 and indeed any AA16/25 with a date code after ( I think April?) 2010 is unquestionably fake.

And any recent product with a CE marking and a non RoHS compliant component (like an FG25), is not CE complaint at all.

Fun stuff, this investigation rubbish!

Regards!

John

Last edited by project23; 11 October 2018 at 13:24.
project23 is offline  
Old 10 October 2018, 23:13   #25
project23
Used Register

 
Join Date: Jul 2018
Location: Liverpool
Age: 35
Posts: 260
Don't forget it was an underclocked FG16 in the A1200, which best guesses (the data just isn't there following C='s collapse) give around (or at least) 200,000 units sold. Probably a lot more built.

Approx 100,000 units of the CD32 was sold in Europe alone, and again used the FG16.

So - and I have no idea here - but lets say globally the 16 was more popular in the mid 90's. That'd make the majority of the 0E13G mask clockable up to 25, let alone 32.

If you know otherwise, please, i'd appreciate the information. (Like maybe 40 million 25's were used in cars or something).

John
project23 is offline  
Old 11 October 2018, 10:30   #26
8bitbubsy
Registered User

8bitbubsy's Avatar
 
Join Date: Sep 2009
Location: Norway
Posts: 1,278
Ok, I'll leave my accelerator be then, I only wanted to take off the heatsink if I could contribute with important information.
8bitbubsy is offline  
Old 12 October 2018, 13:53   #27
project23
Used Register

 
Join Date: Jul 2018
Location: Liverpool
Age: 35
Posts: 260
Quote:
Originally Posted by 8bitbubsy View Post
Ok, I'll leave my accelerator be then, I only wanted to take off the heatsink if I could contribute with important information.
Thanks anyway - if you do ever have your A600 open again for any reason - and can easily remove the Furia (I don't want to ruin your setup), then taking the heatsink off takes just a few seconds and the details would be interesting.

Interesting for sure - especially if you have a genuine chip and your Furia is stable - but not something i'd ever ask any owner to actually do. So please do not unless you are interested too!

Cheers again!

John
project23 is offline  
Old 12 October 2018, 22:46   #28
Stedy
Registered User

Stedy's Avatar
 
Join Date: Jan 2008
Location: United Kingdom
Age: 41
Posts: 619
Hi,

I might have 3 or 4 genuine Motorola EC020s in my store of parts, will check tomorrow. Would you be interested?

The last 68EC020s were shipped 31st May 2012, http://cache.freescale.com/files/sha.../PCN13644D.htm, so anything with a later date code is fake.

Some other 68K series processor shipment dates, for completeness:
MC68030 last shipments, 31st May 2012.

MC68030RP discontinued Dec 30, 2000.
MC68030RC16C-40C, MC68030FE16C-33C, Jan 15 2011.
MC68030RC33C/RC50C, Jan 15 2011. Last shipment May 2012.

MC68040FE33V/RC40V Obsolete Oct 11, 2013.
MC68040FE25A/FE33A Obsolete Nov 21, 2014, last ship date 21 Nov 2015
See http://cache.freescale.com/files/sha...n/PCN15939.htm
Stedy is offline  
Old 13 October 2018, 00:01   #29
project23
Used Register

 
Join Date: Jul 2018
Location: Liverpool
Age: 35
Posts: 260
Quote:
Originally Posted by Stedy View Post
Hi,

I might have 3 or 4 genuine Motorola EC020s in my store of parts, will check tomorrow. Would you be interested?

The last 68EC020s were shipped 31st May 2012, http://cache.freescale.com/files/sha.../PCN13644D.htm, so anything with a later date code is fake.

Some other 68K series processor shipment dates, for completeness:
MC68030 last shipments, 31st May 2012.

MC68030RP discontinued Dec 30, 2000.
MC68030RC16C-40C, MC68030FE16C-33C, Jan 15 2011.
MC68030RC33C/RC50C, Jan 15 2011. Last shipment May 2012.

MC68040FE33V/RC40V Obsolete Oct 11, 2013.
MC68040FE25A/FE33A Obsolete Nov 21, 2014, last ship date 21 Nov 2015
See http://cache.freescale.com/files/sha...n/PCN15939.htm
Precisely!

Anyway this is a quick reply from my mobile as I am in work, but how much would you sell per chip? Take to PM if necessary.

If I can get (currently trying) blank furia board or complete my trace repair (which I may now prioritise) then I'd be very interested!

Cheers,

John
project23 is offline  
Old 13 October 2018, 04:03   #30
project23
Used Register

 
Join Date: Jul 2018
Location: Liverpool
Age: 35
Posts: 260
Just a proper follow up now that i'm home.

First of all thanks for the response, the help, and the offer. It's appreciated

Quote:
Originally Posted by Stedy View Post
Hi,

I might have 3 or 4 genuine Motorola EC020s in my store of parts, will check tomorrow. Would you be interested?
In short - definitely maybe, haha. Depends on price and whether Boris (designer of the furia) gets back to me about a blank board. Failing that it depends on me actually fixing this damaged PCB which is just about 6 thousand feet further up my hardware learning curve.

That being said, if the price is good I might just pick one or two up anyway, provided theres at least one FG25. We're talking the PQFP here not the PGA variant, as i'm sure you know. I'm not sure the forum rules on selling things so get in touch by PM and let me know more details.

I currently have one EC020 - a cheap FG16 ($8) ordered from China of all places - the thinking being well, why fake the lowest available model. The interesting thing is that it looks like its in the original packaging, has the 1E13G photomask that the FG25 uses and a proper date code (0050). It has most of the 'holes' in the right place and the right texture but there are parts that make it questionable.

First, accoring to cpu-world it was the 0E13G mask that was shared by both chips, but that information might just be incomplete. I mean why wouldn't they use the same mask its the same bloody chip. Further, on the bottom it looks like two extra 'holes' (there's usually two, not four, right?) have been tampered with (i.e it looks like its been filled in then sanded over). Which I just don't understand. I'll post a pic tomorrow. It'll wake the GF now if i go grab the microscope. My genuine (but sadly busted FG25 has two holes on the bottom. One I expect some kinda line number and the other contains the country of origin. The fake I have has 4, and this has 2 with another 2 clearly sorta - 'dealt with'.

I mean if its fake, why is it fake? Are they taking all 0E13G's apart completely (ie die removal - it does apparently happen in the counterfeiting world), then repackaging them themselves as 16's or 25's - essentially binning them as what would have originally happened although this time its purely performance based whereas last time customer demand played an important role? I mean Commodore wanted what looks like at least half a million, all underclocked 16's. What if the majority ran happily at 25?

Interestingly the 'pin 1' hole is right in the corner, but his ebay picture has it up and to the right a little. Maybe they changed it between 92 and 2000 (the 92 from Dave Haynie's high res A1200 pic. which matches the ebay pic - thanks Dave), so that the 'dot' was more in the corner and the writing changed from paint to etch or something. Maybe different FABS had the dots in different locations. Perhaps i can find a motorola changes statement if or something.

I know i seem crazy about this stuff, haha - it's just surprisingly fun to be investigating something like this and following the clues, learning on the way. A few weeks ago i'd have simply trusted whatever chip came through the door!

If it's a fake, its a good one, but why. That's what i can't understand. Personally, as i say, i think they're binning them all again, resulting in the lower end not mislabelled chips still having to be relabelled due to the process.

Just a thought.

Quote:
The last 68EC020s were shipped 31st May 2012, http://cache.freescale.com/files/sha.../PCN13644D.htm, so anything with a later date code is fake.
Yeah, that's the document I have too. Interestingly it lists the MC68EC020AA16 and AA25. But these are just the RoHS compliant versions. There is another document noting the change but I don't have a link to hand.

Quote:
Some other 68K series processor shipment dates, for completeness:
MC68030 last shipments, 31st May 2012.

MC68030RP discontinued Dec 30, 2000.
MC68030RC16C-40C, MC68030FE16C-33C, Jan 15 2011.
MC68030RC33C/RC50C, Jan 15 2011. Last shipment May 2012.

MC68040FE33V/RC40V Obsolete Oct 11, 2013.
MC68040FE25A/FE33A Obsolete Nov 21, 2014, last ship date 21 Nov 2015
See http://cache.freescale.com/files/sha...n/PCN15939.htm
Thanks dude i'll take a look at that. I'm thinking about putting together a large thread with all of this information, and what to look out for when buying chips for accelerators and the like. But hey, if they're binning them correctly and it works then is it really so bad? The problem is there is obviously no regulation, so some people are just selling hot air. Another problem is expecting something binned at 25mhz to overclock as well as the genuine 25's. You'd expect them to if its the same mask, but what if they didn't ramp the clock up so high on those batches because they didn't have to. (Motorola i mean, in making sure they took 16, rather than 25, so why push as high as the mid 30's as you might do if testing for 25 minimum).

Anyway, thanks again,

John

Last edited by project23; 13 October 2018 at 04:11. Reason: clarity
project23 is offline  
Old 13 October 2018, 05:16   #31
project23
Used Register

 
Join Date: Jul 2018
Location: Liverpool
Age: 35
Posts: 260
Okay I realise its 4:09 in the morning, and i'm mental. But when you work night shifts you wind up doing things like this.

So right now, although i cant get access to the microscope, I have two apparently genuine Motorola EC020FG's side by side.

One I bought in England for around 20 pounds. It worked fine until I think I let it overheat like an idiot and blew it (it was overclocked, no heatsink).

The other I just received from China for the cost of a biscuit and so far has been untouched.

What's important is that apparently they were made by Motorola merely 46 weeks apart, both in the year 2000, and both with the exact same photomask - 1E13G.

The differences are staggering. They both claim from the marks at the bottom to be from Indonesia, so they are presumably from the same FAB.

They both have two bottom markings, in the same places, but the FG25 has a different texture (more rough) to the FG16, and the FG16 also has those afforementioned weird circular 'marks' in the other two corners.

The pins are a different length, not the part that touches the PCB but the part top part of the pin. It's difficult to see because i've knackered my FG25's pins, but its there. The FG16's pins widen toward the top, whereas the FG25's pins are a consistent width all the way into the chip.

The font for both the 'line/batch number' and the 'indonesia' stamp is different from chip to chip.

Moving to the top of the chips themselves, the FG16 has a smoother surface than the FG25. Apparently a smooth surface is clear evidence of blacktopping. The holes are in the right places but couldn't be more different.

The 'pin 1' hole is curved like a hemisphere on the FG25. On the FG16 it's a simple flat drill of the same width. There is also a chip (no pun intended) on this corner of the FG16 worth mentioning.

As far as lettering goes it goes from bad to worse. They're the same colour but the FG16 looks like paint while the FG25 looks like an etch. On the FG16
the Motorola logo is a little too big. The font, compared with the FG25 is way off. It isn't the right size or shape. The 25 has crossed zeroes (like the symbol for the empty set in mathematics), the 16 does not. The serial number of the 16 in 9 digits, for the 25 its 10 digits.

Remember - these chips were apparently made in the same year, by the same company, in the same factory, just months apart!

What's scary is that if i didn't have what I think is a genuine 25, i'd have no problem proclaiming this 16 as real. I'm just so gutted I ruined my 25! Haha...

Tomorrow i'll have pics with more details. It's just crazy though - why on Earth would anyone fake the slowest available model of a CPU from 20 years ago, unless as a biproduct of a much bigger process?

If these are all perfectly functional EC020 chips, I believe the fake ones have been completely repackaged. I mean the die itself here. It's just too good a job! The bloody pins are different!

Anyway, for those who have read, all three of you, thanks for ya time

Night all, i really should sleep now, and stop connecting string to pictures of chips on a map of the globe surrounded by newspaper clippings

John
project23 is offline  
Old 13 October 2018, 16:56   #32
Stedy
Registered User

Stedy's Avatar
 
Join Date: Jan 2008
Location: United Kingdom
Age: 41
Posts: 619
Hi,

The MC68EC0202 parts I have are 16 MHz parts, photo attached. I have 5 of them, it was for an accelerator project that did not happen. I bought them from an Amiga repairer a few years ago now. If you're interested in a CPU or two, contact me, no rush, I won't be selling these on other forums.

You're questioning why 68000 series processors are counterfited?
They are still used by the bucketload. I work in the Aerospace industry, there are thousands of aircraft using the 68000, 68020 and 68040 processors still in service. There may be between 3-50 68K processors per aircraft, across the various avionics boxes. Changing the CPU to a newer one and re-certifying it would cost >£1 million per box. The parts used in aircraft are from reputable sources, don't panic!

I've seen printers, test equipment, network equipment and embedded systems with a 68000 series CPU, still in use. If people want to maintain this equipment, they need spares, especially as embedded equipment has 20-30 year lifespans.
Attached Thumbnails
Click image for larger version

Name:	MC68EC020_Ian.jpg
Views:	33
Size:	92.4 KB
ID:	60248  
Stedy is offline  
Old 13 October 2018, 17:01   #33
plasmab
Registered User

plasmab's Avatar
 
Join Date: Sep 2016
Location: Glasgow
Posts: 2,716
Yup. My logic analyser is 68K based. Very little need to replace it.
plasmab is offline  
Old 13 October 2018, 18:29   #34
project23
Used Register

 
Join Date: Jul 2018
Location: Liverpool
Age: 35
Posts: 260
Right, so the market is there then! And the money! Stedy i'm not sure about the 16's... If they're confirmed genuine then some date in the future I might be interested, just to see how they clock. If you know that one pictured to be genuine, then we can add 1E13G to the list of photomasks shared by the 25. Again, thanks for the offer!

I've taken some photos with my cheap microscope. (I have a series of LED lamps and a bracket of sorts that enables me to solder with it). I'm sorting them out and putting them side by side - chips that are supposed to be from the same place just months apart, and their differences. I'll add them to this thread when i'm done.

Last edited by project23; 13 October 2018 at 18:36.
project23 is offline  
Old 14 October 2018, 20:14   #35
project23
Used Register

 
Join Date: Jul 2018
Location: Liverpool
Age: 35
Posts: 260
Okay!

So i've got a couple of images here that are pretty interesting.

They are side-by-side images of the FG25 and FG16 i have apparently manufactured in the same year and in the same place.

First, lets get the boring imaging conditions out of the way!..

The chip on the top left is the FG25 and the two other chips are the same FG16. (There are only two individual chips in the pic). The images of both chips were taken with similar lighting (for that zero interested persons, this includes the LED's from the cheap scope, and two further LED torches on adjustable beams meant for microphones - seriously, these are great for electronic work). So the lighting was the same intensity but I did move the torches to make the etching legible. I did not move the torches in such a way to make the surface texture appear different. What I can say in terms of consistency is that all pics were taken from the same height. Obviously that height was reduced for the image comparing the bottom stamps. I have a 'science lab style' thingy that you would use to mount a burette. The scope is tightly mounted in this thing, and cannot move. Obviously this is important for an accurate comparison of alignments etc.


Okay with that crap out of the way here is image one:
The first thing that jumps out is the texture - the 16 is far smoother and contains less 'glass' like substance than the 25. Also whilst the pin 1 indent is similarly aligned in both chips, the 25 is bowl-like whilst the 16 is merely a flat indent. It is difficult to see from the pic but the texture of the indents on the 25 are smooth. The texture of the indents on the 16 is also smoother than its surface (this makes it a pretty good fake, if it is fake, as this awesome guide explains). Remember, these things are supposed to come out of the same (or at least identical) mould in the same year.

The second thing that leaps out is the font. It's different from chip to chip - clearly so. Of note is that the 25 has lines through the 0's whilst the 16 does not. I have no idea at this stage which would be accurate for 2000. At least in the 90's (i'll show genuine 90's FG16's in a following post) the standard was not to have a line through the zero. So now i'm questioning the legitimacy of the 25 as well!

Then we get on to the alignment. I have used the same bottom right hand indent to align the chips horizontally and vertically, and these lines are in blue. This should align everything else. I have forgotten to add the alignment lines for the top left corner indents - but if i recall they are remarkably close. Again, one is a good fake or both are fake?

The yellow lines show vertical alignment and the red lines the horizontal. The results are pretty self explanatory. The FG25 etch is much longer than that of the 16. It is also much wider. We also see that at least in comparison to the 25 (which has a line running through the centre of the Motorola symbol), the 16 etch is off centre.

EDIT: I see how the alignment of the logo may be seem to be accounted for by slight rotation of one of the chips. I don't think this is the case, but this wouldn't effect the stark difference in height, width, and font of the overall mark anyway. In my defence though the original GIMP file I used had a good dozen or so more guide lines to match up the other indents etc. I realise the bottom one looks slightly rotated but its just the lighting on the top and the way i've cropped it. I've taken this a little obsessively thorough

Okay! So... on to the bottom. This image is much less confusing - each part is labelled. When I say 'CHINESE' and 'UK' I am referring to the place I ordered the parts from, recently on eBay. The four parts at the top compare the two bottom indents of both chips, and the bottom two parts compare the same chip (FG16) with different contrast settings.

I didn't bother with alignment issues here. The font difference speaks for itself. These did not come out of the same mould, and as I said above there are marks on the FG16 in opposing corners (the one shown is the worst) which are just damn weird. I don't know if the circular nature of the mark is just coincidence, but a genuine chip should have a homogeneous surface texture, top and bottom.

So which is the fake?

At first I thought the FG16 - but honestly I think they both are. (I'll explain why in a follow up post). I don't know why you'd fake a 16, but remember these chips were apparently from the same place in the same year. It's just all utter nonsense.

Again, this link is a great guide to identifying counterfeit chips and what to expect from a genuine one.

So! There we have one great example of two very real looking chips with same year date codes and same FAB location and two very different appearances. Apparently (again see above link to counterfeit identification guide) this is simply just not possible.

I know i'm crazy, taking this stuff so seriously, haha

I've been off work all week and that's my excuse - but honestly it's been really interesting learning about just how big a problem this counterfeiting business is.

And again, for anyone who has missed me saying it - if it works, and you're happy with it, then i've nothing in particular against a counterfeit. These chips are so old, and the applications so embedded that any genuine ones would surely have been bought up years ago by those in the aircraft/medical industry as noted by someone above.

If you've gotten this far down the post, well done, haha.

Cheers,

John
Attached Thumbnails
Click image for larger version

Name:	ChipAlignment.png
Views:	82
Size:	202.1 KB
ID:	60265   Click image for larger version

Name:	HoleComparison2.png
Views:	82
Size:	485.1 KB
ID:	60266  

Last edited by project23; 14 October 2018 at 20:46.
project23 is offline  
Old 19 October 2018, 16:02   #36
project23
Used Register

 
Join Date: Jul 2018
Location: Liverpool
Age: 35
Posts: 260
So look what arrived in the post today:





It's an old PCI SCSI adapter from around 1997. Got it for £12 with free postage. Not bad when you consider what processor it has onboard!

A genuine 9719 datecode FG25. I think Lorathek has posted me a blank Furia board so when it arrives i'll be using this FG25 on it (I destroyed the original during the accident).

As far as price goes, I think that's pretty good considering its nature. The current £18 listing for 3pcs wasn't up at the time and there's no pics provided so i'd be unsure of authenticity.

This thing by the way is like a time capsule... the box, manual, warranty card etc etc - it's a shame to take parts off it but i'll keep it as clean and re-usable as I can.

By the way the SRAM on the bottom right are 32k x 8, so may be of some use in another project - probably not.

Anyway! Thought i'd post this update as a great way to get good authentic chips these days is to research what products they were used in and then find them on ebay.

Can't wait to get started on it all.

John
Attached Thumbnails
Click image for larger version

Name:	scsicard.png
Views:	59
Size:	803.0 KB
ID:	60337  
project23 is offline  
Old Yesterday, 13:40   #37
project23
Used Register

 
Join Date: Jul 2018
Location: Liverpool
Age: 35
Posts: 260
For the aid of Mick and anyone else interested, here is a closeup of the FG25 from the SCSI board I posted about above.



Do remember though that this is from 97 and the one from your ebay link is from 99 - mine claims to be from 2000.

So it is possible that the entire etching/printing/moulding method changed some time after 97 (It had been changed before, circa the A1200). But I also have a genuine looking FG16 from 2000 that has a different font (see the above posts). So at least one of them is fake.

One thing I can tell you is that the one pictured in this post is, in my view, 99% likely to be genuine.

As for the current market - well, i've kind of given up on trying to identify the genuine ones - at this point in time its anyones bloody guess. They're probably just as likely to be fake as they are genuine.

But in my searches that font only pops up twice - in your ebay batch and in the one I recently bought from a different seller.

For what its worth the CPU mask is correct, thats about all I can say for certain without having a genuine FG25 from 99 or 2000 to compare against.

Hope this helps...

Remember though - it'll probably work - it might even work well - but mine may not have burned out if it had been genuine. At the same time it may have lasted forever if I had used adequate heat sinking etc.

John

EDIT: Also, this confirms that Stedys FG16 chip from 95 is also likely to be genuine. On both, the 'pin 1' locator is flat, and not rounded. However, on the A1200 the FG16 has a more rounded locator and the font is smaller, so either Commodore used fakes or Motorola did change their processes over the years. If you're interested in taking this investigation further, what you want is a pic of genuine product containing an FG16 or FG25 from around 2000.
Attached Thumbnails
Click image for larger version

Name:	SCSIControllerFG25Closeup.png
Views:	26
Size:	411.0 KB
ID:	60400  

Last edited by project23; Yesterday at 14:03.
project23 is offline  
Old Yesterday, 14:00   #38
Mick
Registered User

 
Join Date: Jan 2004
Location: England
Posts: 271
Okay thanks for the detailed response and linking me up with this thread, I think I'll avoid them unless some can confirm.
Mick is offline  
Old Yesterday, 14:10   #39
project23
Used Register

 
Join Date: Jul 2018
Location: Liverpool
Age: 35
Posts: 260
Quote:
Originally Posted by Mick View Post
Okay thanks for the detailed response and linking me up with this thread, I think I'll avoid them unless some can confirm.
Yeah, what I did in the end was research what products the FG25 was used in, then hit ebay for those products.

The SCSI adapter i found was only £12 delivered, but its pretty much guaranteed to be genuine, and similarly priced to other chips (not the super cheap lot of 3 you posted - that wasn't around at the time and the price is suspiciously low if you ask me.

Try and look for pics of the AA25 - thats the RoHS compliant version released later.

By all means take a risk, depending on application - a lot of ebay listings will be remarked FG16's I reckon. But if you're dead set on the genuine article, just find a cheap old product with one inside. They sold a lot for use in embedded systems etc.

John

EDIT: Check out the following link!

https://www.eevblog.com/forum/repair...1d-mso-repair/

It's concerning the repair of an oscilloscope, i think?

But look at the pic - an FG25 from 2004! Missing the slashed zeros!

This is much more consistent with my FG16 and very different (in terms of font) to the FG25's from ebay.

So there we go, if you plan on overclocking this chip, perhaps steer clear from the ebay ones unless they look similar to the one in the link above. For what its worth, i ordered my similarly looking FG16 from China. My thinking was 'why bother 'faking' the lowest speed chip' - and I think i may have been right. Also, there are far fewer 16's than 25's coming out of China right now. Then again, my 16 is missing the 'bowl' pin 1 locator - mine is flat, so who knows.

Last edited by project23; Yesterday at 14:29.
project23 is offline  
 


Currently Active Users Viewing This Thread: 1 (0 members and 1 guests)
 
Thread Tools

Similar Threads
Thread Thread Starter Forum Replies Last Post
Strange A4000 problem( s) macce2 support.Hardware 17 09 May 2017 23:54
Strange problem (or I'm stupid) arti Coders. General 5 14 October 2016 19:10
Very strange problem Arnie support.WinUAE 5 26 February 2015 20:55
Strange problem ami_junkie support.Hardware 0 30 April 2010 13:06

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off

Forum Jump


All times are GMT +2. The time now is 02:53.


Powered by vBulletin® Version 3.8.11
Copyright ©2000 - 2018, vBulletin Solutions Inc.
Page generated in 0.08395 seconds with 15 queries