Opinions re: MCU vendors

[crap! I meant to crosspost this here and s.e.d :< ]

Hi,

I'm looking at a couple of different parts from Atmel and TI for an upcoming design. I'd appreciate comments (not *rants*) folks might have about both firms in terms of:

- quality of their products

- accuracy of their documentation

- ability to meet commitments

- responsiveness during the design process etc.

I often have to approach vendors as a "one man shop" during product development with production buys done by clients

*after* the design is finished. So, I usually can't carry the clout that their buying power will (NDA's).

And, I tend to use devices in unusual ways :> So, I tend to end up digging around in dark corners of implementations uncovering things that they didn't expect (or *know*) would ever occur. This is particularly important nowadays when everything is spec'ed with "typical" numbers and damn few "worst case" numbers ("Yikes! The power supply went into current limit! I guess the *real* power consumption is considerably more than the 'typ' figures you guys published...")

I.e., I don't want to waste time with firms that are either slow to respond to *highly* technical inquiries (in the past, I've had face-to-faces with the actual designers of many components that I've "stetched" in unusual ways) or simply brush them off thinking it's not worth their time/effort (Motogorilla has lost *many* designs over the years on account of this).

Thanks!

--don

Reply to
D Yuniskis
Loading thread data ...

One simple litmus test, is to look at the Errata sheets, and also confirm the target device is actually in real production.

- and feel free to ask a 'test question' ;)

DY: [And, I tend to use devices in unusual ways :>]

Hehe, doesn't everyone do that to some extent!!.

It's true that average data sheets have gone backwards in information, and so getting a low cost Eval board, is often more important than the data.

There is no real substitute for actual working silicon, as a reference point.

Example: I wanted to know the Transistion current of a device (Icc vs Vin) on a digital pin. You'd be dreaming to ask that via the average support channels [The what?] - so faster to just measure a device.

-jg

Reply to
-jg

Like Jim said, we all do that :-). That's what we do for a living, I guess.

I can only speak of Motorola (Freescale) and TI, never used Atmel processors.

My experience with TI is almost 10 years old now - but back then I was considering their 5420 DSP. They sent me the data on paper for free (one of the books is still valuable as it carries a few notes I penciled over some opcodes when I was implementing the assembler...), but they were just horrible when it came to getting prototype quantities. They had nothing at distributors. The support person I was in contact with told me "we cannot send you any free samples as the part is too expensive, and we cannot sell you that small a quantity". Terrific. This was at the time their silicon was at revision D or E, mind you. A friend of mine in the US had friends who used the part and sent me some (the part is not easy to replace in this design to this day, their tech support engineer said I would have no chance to make it work... OK, he did not know me so he can't be blamed). (

formatting link
)

With Freescale, it has been easier for me to get in touch with some guy either involved with the design of the part I use or someone who knows someone etc.; getting data under NDA has also worked. But going through the direct channels will be no better than TI, I am pretty sure.

Generally I have learned to make my choices based on technical and price reasoning (the latter has rarely been important, I don't design many mass-consumer products) and part survival expectations/gut feeling; one way or the other one can get the needed data - even from a place like Bulgaria.

Perhaps the worst adviser when making a choice can be some unpleasant experience with sales or support personnel, one should simply keep being above all the usually related nonsense.

Dimiter

------------------------------------------------------ Dimiter Popoff Transgalactic Instruments

formatting link

------------------------------------------------------

formatting link

Original message:

formatting link

Reply to
Didi

I've always had good support on tech stuff direct from Atmel. Ther is also a huge support forum- avrfreaks.net

Reply to
TTman

You say look at the errata sheets, but don't say why.

Is a small errata the sign of a quality product, or a dishonest vendor. Or simply that the part is too new for the bugs to have surfaced?

Is a long errata the sign of a bad product, or an honest helpful vendor that is trying to prevent you wasting time on known bugs?

I think you have to know the company well *before* you can come to conclusions simply by looking at the errata.

--
Regards,
Richard.
 Click to see the full signature
Reply to
FreeRTOS info

That is a pretty poorly defined metric. It reminds me of the Volvo commercial many years ago when the US makers were still building crap. It was a huge presentation room with hundreds in the audience. The speaker talked about quality and then started chanting, "Quality, Quality,..." and the audience started chanting with him. The point was that the US makers were just chanting quality without doing anything measurable. But then quality is not a measurable thing. Can you tell me what you mean by quality? Can you define it in some terms that someone else could respond in a way that would have meaning to you?

My experience is that both Atmel and TI have quality documentation. TI seems to be less willing to modify their documentation if you have questions it does not answer. Once I got Atmel to add some info to their SAM7 data sheets to define crystal requirements. But then they provided a table with data points at specific frequencies that didn't say what to do in between. I pointed out the issue and they wouldn't even comment verbally much less clean up the issue. So I still think they are about equal.

I like this one. What sort of commitments are you talking about? Production? Updates? Support? If you are a small player, don't expect too many commitments in the first place, much less holding to them. All IC makers have a handful of major customers for whom the sun rises and sets. Everyone else in secondary or even tertiary.

Again, I have not found a significant difference between the two companies. I have found the occasional FAE who is exemplary. I've had a TI FAE who would walk across burning coals (he's no longer with TI unfortunately). My current Lattice FAE is pretty durn good although I think he would want to good pair of boots before the walk on coals. I want to say all the support I've gotten from Atmel was through distributor FAEs and was other than great. Disti FAEs are typically trained in many product lines from many makers and just don't have the time to dig into them all in detail. But they can be helpful in getting the ear of the factory support people. So you need to gauge your local FAEs yourself.

I use contract manufacturing and get the same result. If you aren't buying your own parts, it is much harder for the support people to get credit for your orders, assuming you have any. In your situation, you will be invisible or nearly so. It helps if you customer's procurement people are willing to cooperate. But you need to get them in the loop with your local salesperson. Of course this is a PITA, but it helps to get credit to the right people and I think they really appreciate it.

Yep, if they can't see the dollars, they don't respond very well. Isn't that true of us all?

All in all, I don't think you will see a big difference between Atmel and TI. I think Atmel will be slightly better in terms of support. But it also depends on the products. If you are looking at the CM3 CPUs, TI doesn't make them. They've bought Luminary Micro and I think much of LM still works the same as it did. So you would need to check out LM as a distinct entity from TI.

Rick

Reply to
rickman

It is poorly defined only in the sense of how it is *measured*. But, you *know* a quality product when you see it. You also know

*crap* when you see that!

Sure it is! Perhaps not in some "internationally acceptable interchangeable unit of measurement".

E.g., I spent a few summers working for a (huge) hand tool manufacturer. Part of my job was quality related. Judging the quality of their current products and how they compare to other manufacturers of hand tools.

Tell me, how do you *measure* the quality of a hammer? A screwdriver? A tape rule? A saw??

First, you verify it meets all of your mechanical specifications. E.g., what is the draft angle of the two faces of the (slotted) screwdriver's tip? If it is a "cabinet tip" screwdriver, are the

*sides* of the tip parallel? Are the flutes in a Phillips screwdriver free of buildup from the plating process? Are the markings on a tape rule legible and accurate?

Second, you verify that it meets the other design specifications. E.g., what is the (Rockwell) hardness of the shank? How thick is the plating?

Third, you verify that it meets the appropriate "bogo-units". E.g., if the shank is held fast, how much torque is required to "strip" the handle off of the shaft? If the screwdriver tip is pushed into a precision die with a force of X, how much torque is required to *sheer* the tip off? How many times can you "bang" a hammer before the handle fails?

Of course, the third group are, by far, the most fun to devise and implement. They are somewhat arbitrary. Bogus. Yet, they each pertain to QUALITY FACTORS that the USER WILL PERCEIVE.

Ever been pissed off because you *tore* the tip off a Phillips screwdriver? (the average man can do this *easily* for a #0 screwdriver, "with some effort" for a #1, and "only if really determined to do so oand with the assistance of leverage enhanceers" for a #2. #3 requires hydraulic assistance. :> )

Notice how *long* the finish on a tape rule lasts? Despite the fact that it is *continuously* experiencing friction? Ever wonder why the plastic case on that 25' tape rule doesn't shatter when the tape retracts from its full extension?

How do we measure quality in *software*? Count the number of bugs? Are all bugs of equal "value" (within an application -- lets ignore the complication of dealing *between* application domains)? I was drawing a schematic last night. The "part editor" has a bug in it that causes placement of pins not to track the cursor's position. Annoying. OTOH, abandoning a newly created schematic sheet causes the application to *crash*. Potentially very expensive!

And, for the most part, we don't even have a deterministic way of *testing* software -- at least I can set up a machine to bang a hammer continuously until it breaks! (whatever a "bang" unit happens to be)

Sure, obvious things:

- parts failing to meet their published specification (you would agree that this would be "crap"?)

- parts that *technically* meet their published specification but not "in good faith". Especially nowadays with datasheets full of "typ" numbers. ("Yes, typically Icc is 10mA. The fact that *all* of the devices you have purchased from us draw 100mA is still within the (unspecified) maximum that we publish for that part.") But, also, remember that their "product" isn't limited to bits of plastic made in the far east:

- documentation that is grossly and obviously incorrect

- errata that are not kept current and/or are "hidden" for fear that someone would "think ill" of their product

I was reviewing an SiLabs part last week. Application data in the datasheet (sample schematic) had to be one of the worst drawn documents I'd ever seen! As if someone had pasted little *squares* (don't folks know that there are symbols for things like optoisolators? transformers??) on an Etch-a-Sketch and asked a two year old to "connect the dots".

This document speaks for your (their) component. It is reproduced in perhaps 6 other app notes verbatum. So, each time you see it, you remember, "Oh, crap. Not this again..."

Documentation for a reference design using that same component had obvious flaws in the schematic. "What did you guys use to layout the *board*? Why publish anything other than the

*actual* schematic that drove your PCB layout tool??"

Sure, these are nits. But, everyone who reads these documents runs an increased risk of making a mistake *or* spending extra time to sort out what is *really* intended in the documents.

Presumably, they would want customers to have *success* with their products. This suggests they would want to do everything reasonably possible to facilitate that success (proofreading a document doesn't seem to be a *huge* undertaking!). So, if something like this that is in *their" "best interest" is treated so casually, what does that suggest about the "quality" of their silicon? their support? etc.

It seems that manufacturers are moving more into just becoming foundries -- or IP houses (ARM). Trying to get a mix between them seems to be difficult.

Wanna buy a Z380? :>

Of course! I've bought components "by the pound". And, size shouldn't affect their *ability* to meet their commitments. When I'm buying 1M of a particular device / year, I have a different level of expectation (tell me I have to wait 3 years and I'll smile and turn a new crank on thedesign so your name isn't on any of the components). When I'm buying 1K of that same device / year, I *don't* expect the same *timeliness* of a commitment -- *but*, if you tell me 18 weeks, then I expect it to *be* 18 weeks and not 24 weeks or 36 weeks or "gee, your request must have got lost in the cracks"!

I.e., I expect a vendor to cater to a (known) bigger client more promptly (allocation, etc.) than a small fish. But, I expect him to meet those commitments (that *he* is defining!) equally.

E.g., I expect a banker to kiss the ass of someone with $1M on deposit. Walk in the door and perhaps he even gets up and holds your chair for you, etc. OTOH, someone with $100 in a savings account maybe can HELP THEMSELVES to a complimentary lollipop from the big candybowl at the teller's window.

BUT I EXPECT EACH OF THEIR TRANSACTIONS TO BE PROCESSED WHEN PROMISED (maybe an instant wire transfer "on the house" for the big shot and "10 days" for the small fry)

Agreed.

I think these folks (i.e., the ones that are "good") tend to either have a ood work ethic or are engineers that crossed over into "support". I'm sure it must be a double-edged sword for them: on the one hand, you are free of the pressures of coming up with something "never done before" under a deadline; on the other, you never experience the depth of learning that requires; on the *other* other hand, you get exposed to a wider range of product ideas (which you can always mull over at your leisure without *having* to do so). For the right sort of person, it could be a dream job. For others, a nightmare.

Agreed. I've had friends who have outright told me that their goal is to "shmooze". Informal get-togethers (golf, ball games, etc.) with clients to build personal friendships -- the product isn't important. :<

Makes you wonder what happens to those accounts when FAE Joe moves to another disti! :>

Of course! I'll buy 100 pieces and they'll never hear from me again. They'll never (by design!) connect my efforts with larger purchases later on by Company X.

In my case, that is what they *don't* want. Often, I design a product because a client doesn't want anyone to know they are looking into a particular market or product offering. As much as people like to

*think* secrets are kept, you'd be amazed at how much "proprietary" information leaks out of sales reps, FAE's, etc. I have a client who loves to describe how he frequents sandwich shops (etc.) located near his competitors during the lunch hour and casually "reads his newspaper" -- all the while listening in to the banter going on around him (i.e., engineers and sales people from those firms chatting about work).

Look how much you can discern on USENET about the types of products people are working on. Even if you *don't* know which firm employs them, etc.

I like to think it isn't true of me! I try to do my best regardless of the size of the contract or the vendor. But, that's because I'm doing this "for me" and not just "for pay". YMMV.

Thanks for your comments!

--don

Reply to
D Yuniskis

Yes. I am especially wary of "new parts". I've bought plenty of vaporware over the years and have no desire to go down that road...

I have started to use Digikey as a barometer of whether or not I want to embrace a particular product. Yeah, its an arbitrary criteria and can hurt me just as much as help... but, you need

*some* way of "sorting things"

"We can't discharge you unless you are insane. You must file the paperwork declaring your insanity. If you file the paperwork, you must be sane."

(apologies to Mr. Heller)

I think a problem, nowadays, is that there are *so* many different parts that picking the right horse for longterm availability is easiest done by examining wet tea leaves. I try to design so that I can reuse key aspects of the *design* (not the "particulars") and hope that those things can be easily carried over to a new "platform".

Understood. Though, if they are the gatekeepers, it effectively muddies *all* of your transactions!

Reply to
D Yuniskis

Agreed. In the late 70's, I can recall having problems with some Intel parts. Each call to the manufacturer resulted in a fuzzy denial of any problems with the silicon. You were left with the distinct impression that they were chuckling and thinking you obviously didn't know what you were doing... "be kind to him", etc.

Eventually, we found a smoking gun. Documented exactly how to reproduce the problem. When confronted, they replied, "Oh, sure. That's ________". I.e., they had known all along of the problem and were probably hoping they would have a fix or workaround before it jeopardized *this* sale -- as well as other sales to us and other folks who were just adopting MPU's at the time.

This left a distinct sour taste in my mouth.

OTOH, when I was evalating the 32032 for a design, the NatSemi rep handed me the errata lists for the silicon, development tools, etc. in the first batch of documentation provided. Not to say that NatSemi (or the '032) was "better" than Intel, but their *frankness* won them a disprportionate amount of good will!

Too bad NatSemi has such an abysmal record with processors. :<

I'm a big boy. I *expect* their to be problems with your parts. I can live with that. *If* I know about them and can plan for them (double fault on 68000?). But, if you try to hide them -- or, worse, aren't even documenting them -- then Ii have no way to estimate my risk exposure.

Reply to
D Yuniskis

You can "know" what you will, but if you can't tell me how you measure the thing, it means nothing to me. Likewise, if I tell you a part or a company is "crap" that should mean nothing to you unless I tell you why/how I decided it was "crap". So unless you tell me what "quality" means to you, I can't tell you which companies have it.

Ok, I'll bite. How do I measure the quality of a CPU chip or a CPU maker?

And did you rate them on a scale of 1 to 10 without any sort of a guide from your company? Or did you write about the things that you decided made up "quality"? If neither, then what was the point of your subjective evaluation?

Didn't I say that "quality" was not measurable... yes, I think that is what I said.

Ok, now you are evaluating it to your specs. Is that quality? What if someone else has different specs? Does the quality change?

Ok, more specs...

Again, more specs...

That is interesting. So there is a tradeoff between strength of the tip and how well the screwdriver fits your screw. So which has higher quality?

... snip ...

If this is what you wanted to know, why didn't you ask this? By saying "quality" you conveyed none of this to me.

... snip much ranting ...

How does this impact you? If it has not changed, Xilinx and Altera are both still pure IP houses using the big fab houses as their only/ primary source of silicon. Other companies, like TI and Atmel run their own foundries while also designing their own chips. I have never had a problem with either. What is it that you are saying about it?

This also means nothing to me. I am asking you to tell me what you mean by "commitments". There are any number of types of commitments makers make, both implied and explicit. None hold to them all.

Well you are living in a fantasy world. It has been more than once that I was quoted (not promised) parts and they failed to arrive because a bigger fish ate them. The disti said "sorry" and did I still want to keep the PO open for delivery in another month. It happens because if there is a hiccup, they will piss you off rather than the big guy. What would you do, piss of your new, very small customer, or your old, established customer that is 30% of your annual business?

... snip ...

Rick

Reply to
rickman

Most of the processors made by widely used instrumentation makes are quality products. I have a LOT of other problems with manufacturers, but truly poor quality microcontrollers hasn't been my experience -- except perhaps in one (or two) rare cases.

Atmel makes some good products. I won't use them except for hobbyist use, though. There are reasons that may only be local to me. But the reasoning has nothing to do with poor quality of the Atmel parts I've received.

Ah. On this last point, I'll make a side comment. Both TI's MSP430 line and Microchip's PIC line(s) have silicon bugs. The big difference I see between them is that Microchip _fixes_ their silicon bugs. You can see that in the subsequent errata sheets on steppings of their parts. While TI _never_ seems to fix theirs. Simply documents them. Now this isn't hard and fast, as I haven't done an exhaustive evaluation across all the various instances at both companies. But on the parts I've used from each, it's the case.

For example, TI has a CPU4 bug which means that an entire subset of their instructions don't work. It's been around _forever_. Is it fixed? No. Will it be? No. Live with it. Is it a game changer? No.

Some may prefer TI's approach -- document and move on. That way everyone knows what to expect and once they learn it they can depend on it never changing. Some may prefer Microchip's approach and simply buy at or beyond the stepping that works for them.

Personally, I prefer Microchip's general approach to repairing silicon bugs.

I might add, those providing published specifications that say "TBD" nearly everywhere.

The lack of maximums on some devices in some areas is pretty bad, I agree. Seems to be a trend, too.

I have had the experience and tools to really help find silicon bugs with only one manufacturer, since I have their ICE2000 system and two and a half decades of experience with them -- Microchip. In that time, those I've found that were not yet documented (I've found some that, on reflection, were indeed in the docs had I taken the moment to read them) were documented in the very next rev of the errata, which came out within about 30 days of my notifying them. I can't speak to other companies about this.

Okay. I've been very recently using a SiLabs part. Very nice, runs well for my application and I consider it a good quality part. However, the documentation on the DMA controller was very much lacking, to my mind. It wasn't possible, for example, to "see" how it interacted with the ADC peripheral. By this, I mean "exactly how." Which control lines were passed over for observation by the DMA? Were any flags controlled by the DMA unit? How was the counter applied, N or N+1, when fetching data? Etc. Not complex stuff. Just missing. I sent along a request for clarification.

Mid-September I posted a request for information on their support web site. At the end of the month, not having heard a word on it, I contacted my local representative and asked for help. The correct answers to the question were finally given on November 19th. This, after involving VP level staff in the support discussion. We are talking roughly two months to get relatively simple answers. There were some reasons on their end. But regardless, it took a while. And the questions were basic ones, not arcane.

You want better. I agree. And they are in a far better position to spread out the costs of producing more complete documentation, too. Of course, they are also in the business of comfounding the marketplace with so many choices, too, that it makes rational choices more difficult and this leads to more emotional decision making, which they may enjoy (or not.)

In my experience, Atmel cares about what you represent to them in numbers. Microchip never asks me. They just support the products, no questions asked. Microchip's profits speak how well their approach works, too.

I've had delays from nearly everyone. From Analog Devices, we got ADSP-2111's that failed basic memory tests on their internal RAM and we had a 50% failure rate on basic tests. They asked us for our testing code, which I provided, and after shifting production to a different FAB we started getting good parts. It took time. From Atmel, I put in a request in February for two samples of a part when Atmel said they would begin to become available and I didn't get the two parts until December, that year. They had been shipping samples to other customers and, as I got from my local FAE who grilled me several times on what kind of numbers I could promise them, allowed me to understand that since I couldn't promise more than 5000 a year at first we were at the back of the line despite the fact that I put in my request in the very first week when they were taking requests. It wasn't designed in, obviously. I was long into the project with another part, by then. From Microchip, we were looking forward to the PIC18F252 as a change from the C part, the PIC18C252. They were about two months late getting us those parts, as I recall. But they worked well when we got them and dropped them into the new rev without any trouble.

I like having contact back at the factory, as well. My local Atmel FAE has been in the job nearly forever. So I know that Atmel provides a career path (or, at least, one that my FAE considers adequate.) In other cases, turnover seems almost like 150-300%! Yet despite having a long term local Atmel FAE, the behavior of his company has meant I don't use their parts. So his length of service, which I sincerely respect, does me little real good.

The only company I know of which has not asked me, as a consultant doing this work, of numbers of parts is Microchip. Local distributers may ask. But I get my support from corporate. And they have _never_ troubled me with such a question. They have simply provided excellent support, without question. To rickman, it is NOT true of all.

Jon

Reply to
Jon Kirwan

It's been a while since I read "Zen and the Art of Motorcycle Maintenance". This takes me back.

Steve

Reply to
steve_schefter

You've missed the point. *You* define how *you* define quality. Then, *you* tell me how vendors have met *your* standards. I can then decide if that is important to me or not.

If *you* define quality as "their sales rep was well dressed, had an excellent vocabulary and took me out for a fantastic lunch at a four star restaurant", then *I* can decide if any of those things are important to me or not.

"How many times you can bang a hammer" is a bogus metric.

*But*, it is something that can be used to compare hammers (sales rep A dressed like a slob and took me to McDonald's and went "dutch treat"). And, other people can relate to it (to varying degrees) to see how it fits their own needs.

You can apply any number of metrics to "quality". I'm asking folks to share *their* criteria and their observations of various manufacturers in their experiences with them. If

*you* don't know which firms *you* consider to be have quality products, then how do you approach designs? By tossing darts at databooks? By picking the lowest price? By picking the best published feature list? C'mon, you can't be that naive...

Of course! How else do you judge quality? You have to judge it

*against* something. Some (set of) criteria. "Specs".

How a homeowner evaluates a hammer is very different from how a carpenter evaluates a hammer. How you evaluate an "Engineer's hammer" is different from how you evaluate a hammer used for bodywork. Each has a set of specifications -- criteria by which it is manufactured and appraised.

And the smoothness of the plating, and the color of the handle, etc.

That depends on the criteria on which the buyer/evaluator places most emphasis. If you like a good meal at a fancy restaurant (perhaps because product availability is not your concern: that's manufacturing's problem!), then you (personally) would give higher marks to the vendor who best filled your belly.

OTOH, the shop foreman -- driven by production quotas -- would probably be more interested in how quickly he can ramp up delivery of parts to meet the changing whims of his Sales Department, etc.

Because I want to know what aspects of their relationships with (these) vendors have risen to the point of "being notable". (deliberate choice of words, there!). This carries more meaning than asking "who has the least number of errors in their published data" -- could you actually *answer* a question that was that specific? Have you made a point of actually

*measuring* this? Have you some standardized procedure for codifying your results (fasten hammer into test fixture; using #13 spring, set fixture to a displacement of X inches corresponding to Y pounds of force; set repetition rate to 15 per minute; begin test)?

On the other hand, if customer X has found problem Y to be "notable enough" that it "sticks in his craw", then it may be something (for me) to think about. Likewise, if that same customer had a particularly pleasant experience with a vendor in solving some problem (availability, specification errors, application support, etc.) then that, *too*, is worth noting.

Ah, but you see, that is exactly my point! *I* have had a NOTABLY unpleasant experience with SiLabs in this one instance. It rose to a level of annoyance that I will remember. It has been detrimental to my design process. I will make a special point of noting how they respond to my comments when I send them.

As I said, "I'm a big boy". I'm used to things going wrong. OTOH, I *avoid* Zilog because of their past performance -- despite giving them a sh*tload of Z80/Z180 designs over the years. I'd want to know if folks have had problems with products never materializing, or being discontinued quickly, or being unsupported, AS A RULE, for a particular vendor. That doesn't mean that I will *avoid* the vendor or a particular product. But, I may tell a client to do a lifetime buy *now* if I was fearful of a device going prematurely obsolete.

In this case, the disti won't be in the picture. These are large accounts that the factory will ultimately handle. Again, to put it in real context, it wouldn't matter if I wanted *1* Z380 or 1,000,000 of them. Or, i432. Or...

What happens when you can't meet *either* of those?? (Z380)

Reply to
D Yuniskis

I'm curious as to what you found about Atmel that you didn't like. There are any number of things with any company that would make you want to use other parts, but this sounds like it is a show stopper for you and that sounds serious. Care to share?

I would like to see all makers fix their parts. But fixing bugs also creates problems in that now there are "revs" of a part and I have yet to find a way to order a part rev other than verbally. They don't seem to include this in the part number typically, although I saw that recently on a part I was looking at, but can't remember which one. They had both revs available in case you really wanted to stick with the earlier one for the short term.

Atmel is the only one I have used where I needed to spec a rev and couldn't. I had to get the disti inside/outside sales people on the phone and they got the factory in the loop. All in all I think it took nearly two weeks to get verification of the part rev they had available.

That's very interesting. I never saw that from Microchip before. I was hot to use some of their parts once, but I couldn't get the right combination of temp range, power consumption and I/O count (or was it voltage range?) in one part. I got limited support on trying to match my needs to their product line at that time. I understand their products have come a long way in the last five or so years. Maybe I should look at them again.

Rick

Reply to
rickman

Well, I guess that depends on how you are using them. I.e., with everything being "typ" numbers nowadays, how do you decide how well a part complies with its published specification? :-/

(I understand your point, though)

Local as to particular country? Or, city/geographical area?

(Note my comments re: software were intended as "how do we measure our *own* "product" -- I have tried not to bring up the quality of software *tools* herer -- yet!)

But, this is a double-edged sword. Will your original workarounds be "forward compatible"? Or, do the fixes force you to turn the crank again on the software to "un-workaround it"?

OK, that's valid. What is not excusable is failing to disclose these problems. You're just forcing *every* customer to rediscover the same problems -- possibly with financial liabilities as well!

Refering, once again, to the hand-tool manufacturer... one of their tools had a subassembly consisting of several parts outsourced to third parties. One of these parts was consistently made "wrong" (well, let's just say "not as originally intended" -- for whatever reason). As a result, they changed the design for the other parts to fit with this "wrong" part.

The manufacturer of the "wrong" part eventually started making it "to print" (perhaps under new management, etc.). As a result, all of the *other* parts no longer worked! :-/

First lesson learned in business: try REAL HARD to do things right! :>

Ah! ^^^^^^^^^^^^^^^^^^^^^^^^^^^^ that is acceptable. More acceptable than having it change unexpectedly in the middle of a production run!

I like to get things working the way they were intended. E.g., I will pester a compiler vendor to fix a particular bug before I will resort to "/* WORKAROUND */". Otherwise, you end up doing double work (i.e., coming up with the workaround and then documenting what it was *supposed* to be and why it

*can't* be as it should)

Yes. The Cynic thinks: "they're just covering their *ss so they can't be held accountable to anything -- not suitable for ANY purpose or application" :<

I don't know if this is because they are shifting resources and don't sort out these details anymore (i.e., the "increased productivity" in our workforce -- not only do we use less money to do things, but we actually DO LESS! :> ) or because the fab processes are evolving so quickly that the numbers are just too slippery.

From your comments here -- and others -- folks seem to hold Microchip (hard not to think of them as GI :> ) in pretty high regard. They seem to have the Digikey model when it comes to dealing with their customers: try to treat every customer as a valued customer. (i.e., sure, its great to kiss big accounts' *sses -- but, you lose one of those big accounts and you're stuck with all those *little* accounts that you've not been cultivating)

Yes, this should be something that an FAE should be able to handle without escalation. "Is this a 5V part?" "Gee, let me get the CEO on the line..."

In my case, I was looking at a PoE PD controller. This sort of thing has to "play nicely" with other devices. And, since PoE has the potential to toast things that don't follow the rules, you *really* want to NOT make any stupid mistakes and end up causing collateral damage ("Your d*mn device toasted my gigabit switch!").

From the published documentation, even some "simple" questions were hard to answer definitively. I was using the controller with an isolated supply -- so you have at least two sets of power/gnd to keep track of, not counting other supplies on the isolated side. And, the device implements hot plugging so it can further isolate the load (current, not just the "actual load") from the network. So, you've got three sets of "ground candidates" intimately relateed to the chip. Figuring out which one should be used for each purpose required a fair bit of thinking as the documentation just shows crude block diagrams (using the wrong "ground" will either bridge the isolation you are trying to achieve or subvert the hot switch capabilities). I shouldn't have even had to do that thinking as what I was trying to do was

*intended* in the chip's design -- just never fleshed out in their descriptions or suggested application!

If I couldn't get that sort of information from the datasheet, how was I realistically going to get more subtle information (i.e., using the device in less conventional -- though perfectly acceptable -- ways)? How long should I budget for an answer? Or, should I just move on to find an alternative implementation that I have confidence in? :<

Exactly. AND THIS IS THEIR PRODUCT (i.e., the documentation) JUST LIKE THE SILICON ITSELF!

Much as The Cynic would like to agree, I think it's just they are all cutting corners. The days of massive application handbooks with truly imaginative applications are long gone. Now its all cookie cutter: use it like this.

I recall working with a Japanese firm in the 80's in a new product development. They were a potential vendor for one of the largest ($$) components in the product (i.e., this component drove the rest of the design). When we asked them for specs on their component's capabilities, they replied (unabashedly) "What would you like them to be?"

This set us back on our heels as our traditional way of doing things is (was?) to figure out what the capabilities of the components were and then trying to mix and match to get the results we wanted -- or, revise those expectations. I think many of us thought (at the time) "they" were just going to write up a datasheet with our numbers on it. (cynics, eh?) Of course, they *were*! But, they would also have testeed to those specifications and guaranteed all of the units that they delivered would meet those limits. However, we would never know if our specifications were too loose or too tight -- prefering, instead, for them to tell us what they could *readily* produce (sweet spot) and then we could decide how much more it was worth to move that spec point to where we wanted it (or, move our end goal back to fit within their spec).

Different cultures.

Understood. I did a project many years ago where we priced for quantities of 30M. It was a real PITA! Quantities that big distort reality -- I mean they literally bend the fabric of space-time! :>

First, you end up with way too much "attention" from vendors. "Go away, stop bothering me. If I need something, I'll ask. And please wipe your saliva off the desk before you leave..." But, you also have a hard time making good decisions. Everything is "free" (i.e., you are literally buying by the pound). So, keeping track of what you are actually trying to achieve becomes a problem. Feeping creaturism sets in big time!

It was actually easier to scale back my thinking to "Qty 100". The differences between circuit topologies become much clearer (you can actually measure a penny -- microcents are just "a bit of noise" :>).

Of course! As the bumper sticker proclaims: "Sh*t Happens!"

Yikes! I.e., my "Digikey Standard" may not be a bad rule to follow! :<

And, that experience sticks in your mind for *future* products, right? So, that 5,000 piece order may have cumulatively been a 25,000 piece order when other designs get factored in (out)

[FAE's]

Understood. I.e., his service hasn't given him any "clout". (to use on your behalf)

This is worth noting. Unfortunately, I don't think any of the Microchip parts are up to the task. :< But, I have another product in the pipeline that may be an excellent fit...

Thanks!

--don

Reply to
D Yuniskis

There are several issues.

(a) Their (from my experience, now, and to be honest I don't want you to imagine that my experience speaks to the experience of everyone --- it is indeed an isolated data point of sorts) _rabid_ interest in quantity predictions _before_ they were willing to do anything for me. It was clear that before they'd consider getting me samples earlier that I needed to "use the right words" to convince them.

I think my Atmel FAE is a great guy who was placed in a difficult position. He was asked to press me for numbers, so he did. And I could see his own anguish over the already existing delays and the insistence of his management that I somehow tell him BIG numbers in order to allow him to place my request earlier in the queues.

(b) Their support process, which is funneled via the local FAE, who at the time didn't really know anything about the parts and had no experience with them. I'm sure he is more informed today and it is NOT his fault, then. But the actual support resources were in France (AT91) and there was the usual two day turn around that you'd expect when I have to send a request to my FAE, he has to pass it along (with whatever he adds to it) to France, it gets read the next day, someone there writes my FAE a response, then he passes it on to me. This really made a long train out of something that could have been shorter. It was later shorted up a little, where I wrote more directly. But there still wasn't much support.

(c) Dropping parts on a very short leash when it appears they aren't making profit on them. No willingness to "bridge" over the hard times for very long.

Finally, it was those and many others issues related to tools and parts and support and samples and having to be constantly grilled over "numbers" to see if I mattered to them, or not. Especially, when compared to Microchip.

I've talked about Microchip before. It's almost a miracle to me. This is a company that somehow gets everything right on the business side. They support their parts into infinity. They support their tools into infinity. They don't pester you, they are partners who want everyone to succeed. I like to think that I place my clients first, ahead of my own interests where I can, and do what is _right_ for them, not for me. And I see that mentality at Microchip. They work for me, they care about my end. I see it at every single level, every interaction, every moment of each year. I have NO idea how they do it. It's just something I stand back and marvel at.

This doesn't come from some glancing experience. This goes back to my first use in the late 1980's. They have never placed me in a difficult situation. Not once. That doesn't mean they haven't made mistakes with others. But my experience has been one pleasant surprise after another.

I've mentioned before that when I had so much as a flaky switch on a ProMate II, a few years back, when they had dropped selling them years before, they sent me a replacement overnight with shipping for the old unit and a box included. I had it the next morning. Did they fret over proof of original purchase? No. Never asked. Did they say it was no longer supported (as Analog Devices did for a similar unit I had for their ADSP-21xx series)? No. Never even hinted at it. They just jumped. By comparison, I've never seen such good support from other companies that seems to continue nearly forever, so far as I'm able to tell. They never say "sorry about that." They have _solutions_ and have obviously considered their customers well in making business decisions.

Why would I choose someone else? Unless there was a need for a unique combination of peripherals that I required? (Like the SiLabs part I'm now using, for example.)

Sheesh.

Take a look at an errata sheet for some Texas Instruments MSP430. Then take a look at the A3 silicon errata for the Microchip PIC18F2525 part and compare it with the B5 stepping's errata, for example. Quite a difference.

Microchip _fixes_ the silicon bugs. As I said, that may either be good or bad, depending on your point of view. I usually prefer it.

Mainly, I like them for their business model and the quality of their technical support personnel (at least, those I've experienced in the past.) Sometimes, very dry personalities. But very well informed, regardless, and supportive where I've needed that support.

People complain bitterly about the instruction sets. And for the older processors, you bet. A naked cpu, if ever there was one. But I never really count that high on my list of worries. Some of my products are still being manufactured and supported after 25 years. And I _like_ having a company around making microcontrollers that sees things on those kinds of time scales and supports products that are as old as my oldest, still manufactured, instruments are. We share similar visions and they make a great partner that way.

Jon

Reply to
Jon Kirwan

but lack of an errata sheet is also suspicious - look at other parts - if there are no errata sheets for any product chances are they are the sort of company that likes to hide its mistakes and should be avoided.

Atmel and Microchip probably have by far the largest number of users, hence different applications than the others put together, so high probability that any issues have been discovered by _someone_. Number of different users is also important for availability. 10000 users buying

100 parts each means wide availability. 10 users buying 100000 parts each means the manufacturer probably only cares about those 10 users, their applicaitons and the bugs that matter to them. findchips.com is a very quick 'finger in the air' test for availability.
Reply to
Mike Harrison

I don't want to suggest that my experience is broad-based with Atmel. It isn't. It may be that the model that applies to my area in Oregon is different on the east coast or in a different country than the US. Or, to be honest, even with other customers in my area. It's merely _my_ experience. I can't pretend to know more than that much.

Well, as I've consistently written, this may be good or bad depending on your point of view. In my case, I think it is better to tether my cart to a horse that fixes them. It gives me more options, that way. Some of the problems may be serious ones. Others may have satisfactory work-arounds. On the satisfactory side for one Microchip part (PIC18F252's eeprom), the solution was to insert a NOP. Quite tolerable and wasn't a problem when the new parts fixed the defect. But sometimes there is no workaround, the issue isn't acceptable as it is, I've invested a lot of time already into the part, and the idea of tying up to Microchip means I can not infrequently look to seeing a remedy in time for the product release. Not always can I rely on that, of course. But at least there is some risk _coverage_ there. Something that just lets me know that if one of these unusual events does take place, there is at least some chance that I can keep going knowing that Microchip has a practice of fixing them and that when I talk with them I can believe them when they tell me they are working on the issue.

With TI, I'd cut my losses right away and find something else, since I'd know in advance they won't fix it in my lifetime. That can be good, too, I suppose. But I don't find much problem ordering the stepping I need, either, from Microchip. So maybe I'm just more comfortable there.

Don't mean to make this a defining issue. It's not. It's just one of those things. If you are interested, take a look at the A3 silicon errata for their PIC18F2525 part and then compare it with the B5 stepping's errata.

Well, that's a long discussion in its own right. I've been pretty impressed with Intel's Specification Updates, listing every stepping of every processor and the various issues with each. I know it can be done well. Some companies never seem to document errata, or almost never. In the few cases where I have experience, I know for a fact that Microchip disclosed what I found, though. And sooner than later.

Cripes!

Well, there is that.

The thing this misses is when the workaround is NOT acceptable for the application, yet you've invested heavily up to that point in tools and time. If it is NEVER going to be fixed, you are stuck finding another part and going back nearly to step 1. With Microchip, I've been able (in the one case where it did matter) to have a discussion about when it would be fixed, ignore the issue while completing the rest of the product and to get an order in for the new stepping in time for production runs. Software was finished except for the known issue, but I got the new stepping in time to test adequately before release. So all was good. With TI? Eh.

I don't want to speak for those who know all this better than I do. But from the outside, there is a trend there. And I'm sure there is more than one thing influencing it. But perhaps someone with first hand knowledge can comment more about the details.

I can only speak from 25 years experience, starting when they decided to go beyond rice cooker volumes and look to supporting "the rest of us." Whoever those people are who struggled to design their approach at that time to a new market for them, they are brilliant in their ability to both forsee and to deliver on their promises. I could only hope to learn from them, on bended knee. They've done right by me, anyway.

And don't ask me how they got it right, time after time. I keep waiting for the other shoe to drop.

Interesting story. Thanks.

Yes. And I talk about it here, from time to time, when the mood strikes. Whether that affects anyone is another story

-- I mean, who uses newsgroups anymore? But there it is.

Right. Or he chooses not to place himself at risk on my account for whatever reasons he may have about that.

Well, they have quite a range. But as you point out, they are no panacea, either. I _had_ to choose the SiLabs part or else go to a separate (and equally expensive) external part that would affect board size, power consumption, parts cost and assembly cost, etc. It's a novel experience with SiLabs and I have mixed feelings about them and their products, now. Time will tell how that all plays out into the future. I may design them out in a couple of years. Or not.

Jon

Reply to
Jon Kirwan

There is evidence vendors well understand this very yardstick IS widely used.

The NXP M0 rollout was a good example. That part 'hit the ground running', in terms of Eval Boards, and Disti Stocks.

The Cypress PSoC3 is scoring not quite as well: Yes they have Eval Boards, but full production indicators are not as good.

-jg

Reply to
-jg

Well they have to be doing _something_ right :-). But the characteristics you list are at the bottom of my list, if they make it there at all, that is. IOW, I care about the substance, and I would live with a less pretty envelope.

It is not just the instruction set, I remember almost 20 years ago I was asked to consider them. It was so twisted and awkward (at the time I had 68HC11E9 for MCU purposes and an MC68340 for the "larger" things) that I have not bothered to look at them to this day (just as I don't look much at x86). BTW, both Motorola parts I mentioned (had chosen back then) are also still in production. Of course I am and will be praying they do similarly with their newer products I have used... :-) . What an anti-utopic world it would be, just x86 and PIC - brr... The bad news is, things seem to be going antiutopic overall faster than I could afford to not care about it .... :-).

Dimiter

------------------------------------------------------ Dimiter Popoff Transgalactic Instruments

formatting link

------------------------------------------------------

formatting link

Original message:

formatting link

=A0Some of my products are still being manufactured

Reply to
Didi

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