The MAX7219 LED display controller – real or fake?

Introduction

[Here's the Arduino tutorial]

If you’re experimenting with various Arduino or other projects and working with LED matrices or lots of LEDs – you may have come across the Maxim MAX7219 “Serially Interfaced, 8-Digit, LED Display Driver” IC. It’s a great part that can drive an 8 x 8 LED matrix or eight digits of seven-segment LED displays very easily. However over the last few years the price has shot up considerably. Supply and demand doing their thing – and for a while there was also the Austria Microsystems AS1107 drop-in replacement, which could be had for a few dollars less. But no more.

So where does the budget-minded person go from here? Charlieplexing? Lots of shift registers? Or dig a little deeper to find some cheaper units. With a MAX7219 heading north of US$10 in single units, they may turn to ebay or other grey-market suppliers in the Far East. Everyone likes to save money – and who can blame them? However with the proliferation of counterfeiting, “third shift” operations and other shifty practices – is buying those cheaper examples worth it?

A few people have been asking me of late, and there’s only one way to find out … so over the last month I ordered eight random “MAX7219s” from different suppliers on ebay and will compare them to the real thing using somewhat unscientific methods, then see how they work. The funny thing was that after five weeks only six of the eight arrived – so there’s risk number one: if it doesn’t come from a reputable supplier, it might not come at all. Funny stuff. Anyhow, let’s get started by looking at the differences between the real MAX7219 and the others. (Or if you want to learn how to use the MAX7219 with Arduino – click here).

Pricing differences

The easiest hint is the price. The non-originals are always cheaper. And if you wonder how much the real ones are in bulk, the quickest indicator is to check the Maxim website and that of a few larger distributors  For example the Maxim “sticker price” for 1000 units is US$4.18 each:

maxpricing

How much at Digikey? Lots of 500 for US$4.67 each:

digikey

And you wouldn’t buy just one from element14 at this price:

aue14pricing

However in fairness to element14 they will price match if you’re buying in volume. So if you can get a “MAX7219″ delivered for US$1.50 – there’s something wrong. Moving on, let’s examine some of those cheap ones in more detail.

Visual differences

If you’ve never seen a real MAX7219 – here it is, top and bottom:

realtopss

realbottomss

And here’s our rogue’s gallery of test subjects:

testsubjectsss

In a few seconds the differences should be blindingly obvious – look at the positioning of the printed bar across the part, the printing of the logo, and the general quality and positioning of the printing. Next, those circles embedded in the top of the body at both ends of the part, and the semi-circle at the top end. And if you turn them over, there’s nothing on the bottom. Furthermore, there isn’t a divot indicating pin 1 on the fakes, as shown on the real part:

divot

Oh – did you notice the legs on the real one? Look closely again at the image above, then consider the legs on the others below:

fakelegsss

Finally, the non-originals are shorter. The Maxim width can fall between 28.96 and 32.13 mm – with our original test MAX7219 being 32 mm:

realwidthss

and all the test subjects are narrower, around 29.7 mm:

fakewidthss

Fascinating. Finally, I found the quality of the metal used for the legs to be worse than the original, they were easier to bend and had trouble going into an IC socket. You can find all the physical dimensions and other notes in the data sheet available from the Maxim website. Finally, this packaging made me laugh – knock-offs in knock-off tubes? (Maxim purchased Dallas Semiconductor a while ago)

faketubingss

Weight difference

Considering that they’re shorter, they must weigh less. In the following video I put the original on the scales, tare it to zero then place each test subject – you can see the difference in weigh. The scales are out a bit however the differences are still obvious:

However over time the manufacturers may go to the effort of making copies that match the weight, size and printing – so future copies may be much better. However you can still fall back to the price to determine a copy.

Do they actually work? 

After all that researching and measuring – did they work? One of the subjects came with a small LED matrix breakout board kit:

matrixassembledss

… so I used that with a simple Arduino sketch that turned on each matrix LED one at a time, then went through the PWM levels – then left them all on at maximum brightness.

Here’s the real MAX7219 running through the test:

And test subjects one through to six running it as well:

And from a reader request, some current measurements. First the current used by the entire matrix module at full PWM brightness, then with LEDs off, then the MAX7219 in shutdown mode:

current

Well that was disheartening. I was hoping and preparing for some blue smoke, dodgy displays or other faults. However the little buggers all worked, didn’t overheat or play up at all.

Conclusion

Six random samples from ebay – and they all worked. However your experience may vary wildly. Does this tell us that copies are OK to use? From my own personal opinion – you do what you have to do with respect to your own work and that for others. In other words – if you’re making something for someone, whether it be a gift or a commercial product, or something you will rely on – use the real thing. You can’t risk a fault in those situations.  If you’re just experimenting, not in a hurry, or just don’t have the money – try the cheap option. But be prepared for the worst – and know you’re supporting an industry that ethically shouldn’t exist. And at the end – to be sure you’re getting a real one – choose from a Maxim authorised source.

I’m sure everyone will have an opinion on this, so let us know about it in the moderated comments section below.  And if you made it this far – check out my new book “Arduino Workshop” from No Starch Press.

LEDborder

In the meanwhile have fun and keep checking into tronixstuff.com. Why not follow things on twitterGoogle+, subscribe  for email updates or RSS using the links on the right-hand column? And join our friendly Google Group – dedicated to the projects and related items on this website. Sign up – it’s free, helpful to each other –  and we can all learn something.

The following two tabs change content below.

John Boxall

Founder, owner and managing editor of tronixstuff.com.

33 Responses to “The MAX7219 LED display controller – real or fake?”

  1. Jose Lopez says:

    Good article John;
    Always a pleasure to learn…

  2. perhof says:

    Excellent article. I totally agree with your conclusion.

  3. Jensa says:

    I’ve used both copies and originals. I can’t really tell the difference from just playing around with then. However again – the originals are really, really pricey so I wouldn’t blame those that rather opt to buy 5 of the kits you have here (chip, led matrix, pcb, connectors) for less than the price of a single, original chip.

    You can also get 30 clone-chips for the price of a single original, so Maxim must be trying to milk the market here?

  4. denzel says:

    Awesome read! Can’t quite agree that the industry should not exist, because I’d think that in itself would not be moral.
    But–I do think it is a disgrace to run what is essentially a scam.. or not, what about decapping to see what is really inside? :)
    And is this patented tech? Idk, they could at least slam their own logo on it but I guess they don’t care much about the marketing.

  5. stu says:

    Interesting. I’ve been wondering about how these knock-offs of electronics bits come about. Do you have any more info that throws light on the ‘industry that ethically shouldn’t exist’ – how much is actual fraudulent 3rd party manufacture, how much is 3rd shift production of the same part. Is there much actual theft and resale of parts. I’d be really interested to read more about how all this happens. Great articles you are doing by the way

  6. Dilum says:

    I have also used a fake one. The IC was used in a AC control system, when the high current contractors were turned on the MAX7219 reset it self. They worked perfectly when the controllers were off…I have ordered a original IC from futurelec but still i haven’t received it.

  7. Mats Engstrom (@matseng) says:

    So how did you come to the conclusion that the cheap ones are fake? They seem to work just as well as the more expensive ones. The only difference is small discrepancies in the printing and the weight/size of the encapsulation.

    I’m sure Dallas/Maxim have more than one chip fab and they might even outsource the manufacturing of simple bread&butter products to a chip foundry in order to free up their own factories for the more funky and exotic stuff. Or they might outsource just the encapsulation part of the manufacturing.

    Even if they do it all themselves I don’t think that the printing would be absolutely identical between different factories and production lines – especially over a period of a few years.

    It doesn’t seem logical that the Chinese would set up a chip fab plant (with a price tag starting from 1 billion USD and upwards) to produce clones of chips and selling them with a retail price of $1. And it they can reverse engineer and then design a new set of production lithographic masks for doing a chip I think they would be able to reproduce the printing on the encapsulation 100% as well.

    • John Boxall says:

      Hi Mats
      Maxim said the first thing they look at is the printing, size and weight – and that bad functional copies exist. So *someone’s making them* If there’s a third-shift happening at the plant that’s responsible for the originals Maxim should have picked up on that by now. And the Chinese don’t often care about matching things 100%. Example – https://www.sparkfun.com/news/350http://www.retronix.co.uk/counterfeitdetection etc.

      • Mats Engstrom (@matseng) says:

        Mmmm… Well, there must be two kinds of Chinese then, One bunch that can clone a chip to (almost?) 100% and one that does like 85% correct printing :-)

        My guess is that the chips actually come from Maxim, and they are pillaged from the factory either as complete wafers or just the separated chips that tested out-of-spec and should have been destroyed. And then there might be some shady back-alley shops that have bonding and encapsulation tech to make them sellable.

        If Maxim makes a few millions of these every year and 0.1% ends up stolen that might account for the ones available at ebay and other marketplaces. It’s hard to tell how many cheap you actually can but from those sellers.

        And of course some of the cheap ones might be fully legit, but been bought very cheap from as leftovers from large runs of pcbs where a few thousands of parts end up unused.

      • John Boxall says:

        Exactly. I’ve asked Maxim for their 5c as well, should be interesting.

  8. Polytech says:

    I’d be honest, it would not have even occured to me (in 2013) that any of the chips from China would preform any worse or simply different than what you might call “legit” Maxim. Is there anything left in the electronics world that is not made in China? These are often called “factory overruns” and it would be extremely difficult (to avoid saying impossible) to find out the legitimacy of the way the eBay seller acquired the ICs. But as far as the physical properties of the thing, I would not doubt for a minute that they are the exact same IC. I’ve been getting 7219s, motor control ICs, MOSFET drivers, ULN Darlingtons and anything else you can imagine from ahem… Far East … and never once had any issue.

  9. n0idea says:

    Thanks for the info, I actually bought 10 Max7219 for 5$(including shipping) from aliexpress.com and they all worked.

  10. Mats Engstrom (@matseng) says:

    John, can you make some more “scientific” tests on these chips to try to determine if they actually are different or not.

    *Measure the quiescent current (active with all leds off).
    *Measure the quiescent current (sleep).
    *Turn on all leds 100%, set the current to slightly higher than the max spec (50mA?) mthen easure the current through one of the segment outputs and the peak voltage there.

    This might show any differences in the chip itself. But you should probably have two genuine chips datecoded with different years to do a proper test.

    Anyways, you first test was interesting enough. There are so many people saying that all chips bought at taobao or Chinese markets are totally fake and non-functioning. Apparently medium priced chips are not necessarily just fake crap.

    • John Boxall says:

      Interesting. I took some quick current measurements today and put the results at the end of the review. However with only one original and six others, it’s more of a comparison between the others. Otherwise that will do for now.

  11. Salut,
    I just checked the price here in Germany at an official retailer and the DIL24 chip is 3.89 Euro for a single chip. So for me, going for a fake one atz ebay with a 25% chance of loss is not very attractive. Furthermore, according to EU law, importing and owning forged products is illegal and the customs office is checking this where I live.

    Ciao, Mathias
    PS: Amazon just told me your book is on the way (the original of course, no ebay copy)

  12. jayfdee says:

    Thanks,very interesting article,I have bought numerous chips,boards etc from China/HK and never had a problem,always received the product,and they have all worked OK.
    The only problem is the delay in delivery. Keep up the excellent articles.

  13. pavan says:

    Great article regarding LED Display Controllers.. Thanks for the post..!

  14. Dirk says:

    Great article ive subbed on youtube

  15. ubiyubix says:

    Very interesting read. Why is the AS1107 no more? I have successfully used the similar chip AS1100 which I got as a sample from austria microsystems. So far I did not find a supplier where you can actually buy any of the AS11xx chips in small quantities.

  16. daitomodachi says:

    This might be a silly question, but can you include a schematic of the breakoutboard?

  17. ninjanear says:

    But, what if Maxim’s factories are placed in China? Just an example.. many manufacturers have their factories placed in Asia, so if the copies are working like the original and electrical characteristics are fair similar (as far as we can test them), i think they could be the same. The difference is Maxim requieres some level of Q&A, at least in the serigraphy and package finish, that the copiers don’t do.
    I know this happen with other fields in the techonological industry so.. why not here?
    An example, prices of appliances in Bangkok is 5 times lower than in many countries in Europe (verified personally). The key is the taxes are increiblely low there and the low salaries (aka, laboral abuse), so many factories are placed in Thailand (Philips, LG, Toshiba, etc.).

    Yes, sending oversas due to customs, shipping, etc. will increase prices but.. with eBay it almost doesn’t work due to the quantities or just they aren’t inspected by customs, so it can explains the lower price (one of many of them).

    Anyway be careful , because from Maxim you have a warranty of what you are buying but from unverified sellers you don’t. Maybe they are ok, maybe they are not.

    I have many ICs from eBay for testing purpopses (buying here is impossible or shipping rates are insane), and never had a problem. Even with power elemens.

  18. Miguel Dávila says:

    In my country (Ecuador) to buy a clone is not an option but a requirement, I mean, if I want to to buy a original chip from digikey, that will cost me $ 14 plus $ 100 shipping, if I buy the clone, it costs me $ 4 including shipping. Therefore, not only is a bit more expensive to buy the original, for someone who does this as a hobby, it’s too expensive. With the clone I have fun and learn, with the original’m broke.

  19. Jeff says:

    For the price, it’s worth it to just get the PC board and the LED display. Saves time on assembling it yourself, even if I have to swap in a genuine chip!

  20. I just took two days to come to the conclusion that the chips I have work only for the display test and not for anything else, and yeah, they look like the fakes that you show.

    Thanks for the article; I wish I had found it two days ago.

    Have you heard anything from Maxim?

    Cheers,
    Steve

  21. Paul B. says:

    As previously stated – just who is using the cheap versions? How many?

    Suppose the sellers on eBay sell a few thousand a year in all. Or ten or twenty thousand even.

    Now just who really is going to set up a compete fabrication facility for that relatively trivial quantity, and do it in competition with a major manufacturer?

    The suggestion is frankly totally absurd. The only plausible conclusion is that there is one manufacturer (or possibly a couple of facilities operated by that manufacturer) in production and that part of that production – whether overruns, out-of-specs or simply a deliberate allocation on a marketing basis – diverts through some process to the cheaper market.

    And let’s not misunderstand things here – the manufacturing cost is well under a dollar a chip. The mark-up is – as it always is – in distribution. So there is no reason to imagine that Maxim is not in control of the whole business.

    The interesting question that has to be asked is – what is the actual primary application of these chips? Presumably monochrome scrolling message boards (“Times Square” displays). It would be interesting to know what proportion of these actually use this particular chip.

  22. Ralph says:

    My guess is the manufacturing cost is well under 1c/chip, and the clone chips could be sold by the tens of thousands per month. I could see them being used for lots of high-volume electronics. A VCR that sold for $49 20 years ago most likely didn’t use a $5 led driver chip to flash 12:00 all day.

    There are dozens of IC fabs in the world where you can get cheap wafers made. 10 years ago a 1.5 micron 125mm wafer cost about $170.
    http://smithsonianchips.si.edu/ice/cd/CEICM/SECTION2.pdf
    Today I’d guess you can get .5 micron 200mm wafers for ~$150.
    The 6502 processor took 3500 transistors, so I’d expect a device like the 7219 to use a couple hundred.
    http://research.swtch.com/6502
    For argument sake, I’ll be generous and say the 7219 takes 1000, and I’ll be generous on wafer costs and say a 1 micron 200mm wafer costs $150. That 200mm wafer has a bit more than 30,000 square mm of area usable for dies. 1000 transistors in a 1 micron process technology equates to ~.15 mm^2 per die, or about 200,000 chips per wafer, and a cost of 0.075c per chip.

    Here’s an article I found interesting on the scope of the clone chip market:
    http://www.bunniestudios.com/blog/?p=918

    I intentionally use the word clone, since using the same part numbers has been an accepted practice in the chip industry. I don’t think the clones should be copying the Maxim logo. I don’t think it even helps their sales; I think people would be more likely to by an Acme Semiconductor Max7219 than one that is labeled and sold as a real Maxim but is clearly a clone.

Trackbacks/Pingbacks


Leave a Reply

Receive notifications of new posts by email.

The Arduino Book

Arduino Workshop

Für unsere deutschen Freunde

Dla naszych polskich przyjaciół ...

Australian Electronics!

Buy and support Silicon Chip - Australia's only Electronics Magazine.

Use of our content…

%d bloggers like this: