Jim Thompson wrote:> On Wed, 16 Oct 2013 00:11:34 -0500, John Fields > <jfields@austininstruments.com> wrote: > > > > --- > > Have you seen the movie: "Changeling"? > > --- > > Which one? 1980 or 2008 version?I was sure he meant the new one, but the old one with George C. Scott is a pretty good ghost story, thanks mostly to him. -- Reply in group, but if emailing remove the last word.
Never Buy Maxim (again)
Started by ●October 13, 2013
Reply by ●October 17, 20132013-10-17
Reply by ●October 17, 20132013-10-17
On 10/17/2013 10:27 AM, John Larkin wrote:> On Wed, 16 Oct 2013 23:38:24 -0700, miso <miso@sushi.com> wrote: > >> On 10/16/2013 10:21 PM, John Larkin wrote: >>> On Wed, 16 Oct 2013 19:58:40 -0700, miso <miso@sushi.com> wrote: >>> >>>> On 10/16/2013 12:37 PM, John Larkin wrote: >>>>> On Wed, 16 Oct 2013 14:29:29 -0500, John S <Sophi.2@invalid.org> >>>>> wrote: >>>>> >>>>>> On 10/16/2013 12:40 PM, Tom Del Rosso wrote: >>>>>>> John Larkin wrote: >>>>>>>> the technical guy said that this is a single-supply opamp and might not be >>>>>>>> qualified for dual-supply >>>>>>>> operation. >>>>>>> >>>>>>> I don't have a degree and I'm hardly the most competent hobbyist, but maybe >>>>>>> I could work there. >>>>>>> >>>>>>> >>>>>> >>>>>> Perhaps you should written "I don't have a degree and I'm hardly the >>>>>> most competent hobbyist, so I'm sure I could work there." >>>>>> >>>>>> We should go apply. >>>>> >>>>> "Not getting back to potential customers" is another required skill. >>>>> >>>>> >>>> >>>> Then again, maybe Maxim apps reads SED and knows who they are dealing with. >>> >>> I especially enjoyed the Maxim engineer's "not characterized for dual supply >>> operation" comment. I now know who *I* am dealing with. >>> >>> Last time I used Maxim parts, we had to recall hundreds of timing modules and >>> replace about 3000 pieces of MAX9690 comparators. They started failing after >>> roughly a year of use, and eventually every one would have failed. The parts >>> were actually fabbed (and maybe designed?) by a subcontractor, in Minnesota I >>> heard. >>> >>> We had to make 3000 of these to replace them: >>> >>> https://dl.dropboxusercontent.com/u/53724080/Circuits/Break2.jpg >>> >>> https://dl.dropboxusercontent.com/u/53724080/Circuits/OnBoard.jpg >>> >>> It was rather an expensive nuisance. >>> >>> >>> >> >> And what did FA say? Well assuming you contacted them. It is quite easy >> to find damage to chips given tools like emission microscopes. Generally >> failed parts are due to the customer, and you can get some good >> diagnostics by asking the factory how the part failed. > > The outputs got erratic at high temperature, looked like a problem in the latch > circuit. As the parts aged, the fail temperature descended, ultimately reaching > room temp after about a year. All the parts did this. A good hi-temp bake would > push the fail temperature up, and it would begin slowly descending again from > there. What would cause that?MOSFET threshold shift due to crappy or too-thin oxide, maybe. Or possibly die stress due to the package swelling from moisture. If it were a passivation problem, it would be unlikely to be (temporarily) fixed so easily, and anyway it isn't very difficult to fix a passivation step.> > Our first indication that something was wrong was when parts wouldn't ship, with > no explanation. Then the part was discontinued, delivery infinite days ARO, > again without explanation. That was about the time they started failing. > > I eventually found a guy at Maxim who explained the Minnesota problem. He sent > me 3000 samples of the MAX9691. The 9691 is a *comparator* with back-to-back > diodes across the inputs (!!!!???) which is why we had to make the adapter > boards.Sounds like somebody realized they had a threshold shift problem. Cheers Phil Hobbs -- Dr Philip C D Hobbs Principal Consultant ElectroOptical Innovations LLC Optics, Electro-optics, Photonics, Analog Electronics 160 North State Road #203 Briarcliff Manor NY 10510 USA +1 845 480 2058 hobbs at electrooptical dot net http://electrooptical.net
Reply by ●October 17, 20132013-10-17
On 10/17/2013 10:51 AM, Jim Thompson wrote:> On Wed, 16 Oct 2013 23:38:24 -0700, miso <miso@sushi.com> wrote: > > [snip] >>> >> >> And what did FA say? Well assuming you contacted them. It is quite easy >> to find damage to chips given tools like emission microscopes. Generally >> failed parts are due to the customer, and you can get some good >> diagnostics by asking the factory how the part failed. >> >> I had Moto bitching about one of my parts. They used 4 per board. Only >> one position on the board would have a failed part. Moto felt a bit >> stupid not noticing that and simply went away, presumably finding the >> problem since they kept buying the parts long after the incident. Pilot >> error. It is nearly always pilot error. >> >> Designing ESD devices for interface chips, I spent some time in the FA >> lab destroying parts to understand what exactly fails. Emission >> microscopy detects stressed junctions, but some failures cause the >> juction to do ohmic. You can find those with a simple CCD, but liquid >> crystals is a better technique. The crystals will "boil" over the hot >> spot. You can pulse the chip to make the hot spot toggle, or my >> technique was to just leave it powered and bring a soldering iron close >> to the chip. The iron would get the crystals just ready to change state, >> making it a very sensitive technique. >> >> Minnesota parts were from VTC. To my knowledge, they didn't have any >> designers, just a fab. > > That would be correct. Quite a nice fab. Excellent BiCMOS processes. > > ...Jim Thompson >They used to have own-brand parts, including a really nice 70 MHz dual OTA. Cheers Phil Hobbs -- Dr Philip C D Hobbs Principal Consultant ElectroOptical Innovations LLC Optics, Electro-optics, Photonics, Analog Electronics 160 North State Road #203 Briarcliff Manor NY 10510 USA +1 845 480 2058 hobbs at electrooptical dot net http://electrooptical.net
Reply by ●October 17, 20132013-10-17
On Thursday, October 17, 2013 2:53:43 PM UTC-5, Spehro Pefhany wrote:> On Thu, 17 Oct 2013 12:31:56 -0700 (PDT), ccon67@netscape.net wrote: > > > > >On Sunday, October 13, 2013 2:34:32 PM UTC-5, John Larkin wrote: > > >> Maxim makes an opamp, MAX44280, that has a specified input capacitance of 0.4 > > >> > > >> pF. That would be great in some apps where Cin causes noise-gain peaking. > > >> > > >> > > >> > > >> I figured I'd give Maxim another chance, after being s*****d by them in the > > >> > > >> past. > > >> > > >> > > >> > > >> So I went online and filled out their sample request. The process makes no > > >> > > >> sense, website runaround, but eventually I got it done. > > >> > > >> > > >> > > >> Two days later I got an email to the effect that the sample request was waiting > > >> > > >> for a product-line manager approval. I responded and asked why, and was told > > >> > > >> that it was standard internal procedure and would take only another 24 hours. > > >> > > >> > > >> > > >> OK, a week later I got two samples, packaged in a short tube and wrapped in a > > >> > > >> ball of sticky masking tape. > > >> > > >> > > >> > > >> These opamps, like most Maxim parts, have nonstandard pinout. I built a test > > >> > > >> fixture to measure Cin. With power off, I measured 1.4 pF on the non-inverting > > >> > > >> input pin. I powered it up and the opamp didn't act anything like an opamp. I > > >> > > >> checked everything three times and figured the amp was blown, so I tried to > > >> > > >> replace it. It's an SC70, really tiny, and my tweezers slipped and I pogoed it > > >> > > >> into the carpet, never to be seen again. > > >> > > >> > > >> > > >> The sales engineer emailed me to see how the samples were coming along, and I > > >> > > >> told him the situation. He promised to get me 10 more ASAP. Then he called back > > >> > > >> with a technical guy. I explained my setup. I'm using +-2.5 volt supplies, and > > >> > > >> the technical guy said that this is a single-supply opamp and might not be > > >> > > >> qualified for dual-supply operation. He said he'd contact the designers or > > >> > > >> something. I asked him how they defined and measured the 0.4 pF Cin, and he said > > >> > > >> he'd find out. They were due to call me last Monday. > > >> > > >> > > >> > > >> So far, no more parts, no call. > > >> > > >> > > >> > > >> > > >> > > >> -- > > >> > > >> > > >> > > >> John Larkin Highland Technology Inc > > >> > > >> www.highlandtechnology.com jlarkin at highlandtechnology dot com > > >> > > >> > > >> > > >> Precision electronic instrumentation > > >> > > >> Picosecond-resolution Digital Delay and Pulse generators > > >> > > >> Custom timing and laser controllers > > >> > > >> Photonics and fiberoptic TTL data links > > >> > > >> VME analog, thermocouple, LVDT, synchro, tachometer > > >> > > >> Multichannel arbitrary waveform generators > > > > > > > > >Ahhh, this subject remind me one of the funny Max device... let say deviceX > > > > > >The spec saying that device X has an output pin to indicate its input signal presents or not (kind of LOST signal)... and somewhere the spec say it's an open collector > > > > > >We didn't care it's open drain or not, since we do not use that damn pin for any purpose.. just simple as we don't care like an X in logic, no micro controller on board, or anyone who cares about that LOST... So we let it float > > > > > >Guess what? we get the board built - and power it up, device X kept sleeping as it has been slept for age since its birth > > > > > >Anyone knows why it didn't wake up? > > > > Because you let that "output" float? What was the p/n?that's right Input has valid signal, power pins has good clean supply all are normal, but the chip refuse to work because the LOST pin doesn't have pull up Their logic is this, if LOST signal is low, output must shutdown... they don't care if the input present or not.. even they don't care to say something about the auto-shutdown at the outputs in their spec :-))) Anyway, we gave up our "don't care' and we must use their 'don't care' after that I dont care anything from Max
Reply by ●October 17, 20132013-10-17
On Thu, 17 Oct 2013 15:46:37 -0400, Spehro Pefhany <speffSNIP@interlogDOTyou.knowwhat> wrote:>On Thu, 17 Oct 2013 07:27:54 -0700, John Larkin ><jjlarkin@highNOTlandTHIStechnologyPART.com> wrote: > >> >>The outputs got erratic at high temperature, looked like a problem in the latch >>circuit. As the parts aged, the fail temperature descended, ultimately reaching >>room temp after about a year. All the parts did this. A good hi-temp bake would >>push the fail temperature up, and it would begin slowly descending again from >>there. What would cause that? > >AFAIUI, MOSFETs slowly shift threshold a bit when you apply continous >DC voltage to the gate. Something to do with trapped charges. >Annealing at high temperature can reverse some of that. > >It was a big problem with CMOS-input op-amps used as precision >comparators, but less of a problem with ideal op-amp circuits because >the inputs would usually be balanced, so the shifts would cancel out.The 9690 was a bipolar, ECL comparator, so I doubt that it had any CMOS inside. It could have been some sort of pinch resistor thing or something, some charge migration in an oxide later sort of like the CMOS gate thing. But I'm not an IC guy, and I don't know the physics. -- John Larkin Highland Technology, Inc jlarkin at highlandtechnology dot com http://www.highlandtechnology.com Precision electronic instrumentation Picosecond-resolution Digital Delay and Pulse generators Custom laser drivers and controllers Photonics and fiberoptic TTL data links VME thermocouple, LVDT, synchro acquisition and simulation
Reply by ●October 17, 20132013-10-17
On Friday, 18 October 2013 04:05:17 UTC+11, Greegor wrote: <snip> =20> G > Has your wife been catching on to you, slow man?=20 > =20 > BS > I wonder what that was intended to mean?=20 > =20 > G > It was open ended. > =20 > BS > Which is to say more of Greegor's contentless drivel. > =20 >Flatterer.<snip>> G > Slow Man: When avowed socialists from=20 > G > other countries start manipulating =20 > G > information to try to sell Socialism =20 > G > to people in the USA, we don't like it.=20 > =20 > BS > It's strictly for your own good. > =20 > Socialists and Stateists say that a lot!But I went on to say exactly why Greegor - amongst others - needed further = education.> BS > The shameless information manipulatio=20 > BS > that lead you to believe that socialism > BS > and communism are the same thing needs=20 > BS > to be shown up. =20 >=20 > I compared you to another avowed=20 > SOCIALIST, not a communist and > what was your reaction? > =20 > BS > <snipped irrelevant stuff about someone who posts elsewhere>It was - at best - anecdotal evidence. Greegor knows an avowed socialist th= at he doesn't like. So do I. So what? It doesn't say anything about the uti= lity of the political philosophy. =20> I'm starting to think you are more=20 > dishonest than I thought.It would be nice if you could start to think, rather than parade transparen= tly emotion-based reactions, but this ins't exaactly evidence that the proc= ess has begun. <snip>> BS > This thread started off on the defects of > BS > Maxim. A little bit of off-topic deviation > BS > is tolerable, but you are away with the > BS > fairies - of the old-fashioned asexual kind. >=20 > You ad-hommed 3 regular posters all at once.Not for the first time - and they have all posted unpleasant misinformation= about me from time to time. =20> Typical of Socialists and AspiesIf there is - in fact - a correlation between suffering from Aspergers Synd= rome and believing in socialism, I've yet to hear about it. Calling people Austic or Aspergers is a moderately popular form of mindless= abuse used by people who don't really know what the words mean.> you=20 > frequently put on airs of superiority,=20 > disrespect others and present YOU=20 > prejudices as enlightenment.Naturally. Unlike you - for instance - I know what the words "Aspergers Syd= nrome" and "socialism" actually mean, which does make me you superior. It's= not a broad spectrum superiority, but it's significant in this exchange. Any show of respect for somebody as ignorant and intemperate was you would = be quite inappropriate. My opinions do reflect my prejudices, but they also reflect a program of re= ading and thinkign which has clearly gone further than anything you have ev= er managed. =20> When I pointed out that you are an avowed=20 > Socialist, you tried danced around that=20 > fact, comparing different political parties > in Australia, none of which had any > bearing on the simple fact that you are > an outright Socialist.What you actually claimed was that I was an Australian Socialist which happ= ens to be an extinct species. It was one more bit of evidence that you don'= t really know what you are talking about, and I was happy to exploit the op= portunity to make that even more obvious. =20> Why do you try so hard to weasel out of that?Why do you think I'm trying to weasel out of anything? You posted the line "Has your wife been catching on to you, slow man?" but = couldn't explain what you thought you meant by it, which fall somewhat shor= t of constructive debate. --=20 Bill Sloman, Sydney
Reply by ●October 17, 20132013-10-17
On Friday, 18 October 2013 05:09:17 UTC+11, Greegor wrote:> > Slow man: In 5,000 words or less could > you please explain why you think your > fecal matter doesn't smell bad?Greegor's idea of a useful contribution to the usegroup ... -- Bill Sloman, Sydney
Reply by ●October 17, 20132013-10-17
On Friday, 18 October 2013 06:18:27 UTC+11, Greegor wrote:> > JL > The outputs got erratic at high temperature,=20 > JL > looked like a problem in the latch circuit.=20 > JL > As the parts aged, the fail temperature=20 > JL > descended, ultimately reaching room temp=20 > JL > after about a year. All the parts did this. > JL > A good hi-temp bake would push the fail=20 > JL > temperature up, and it would begin slowly=20 > JL > descending again from there. What would cause that?=20 > =20 > Hygroscopy - tending to absorb moisture from the air =20 > Did somebody get the wrong polymer for the chip packages?It seems unlikely - that's a very old problem. An even older problem is that reverse biasing base-emitter junctions (this = was a John Larkin circuit) into breakdown produces cumulative damage, which= can be largely reversed by baking - the legendary HP technician used to cu= re damaged transistors by stubbing out his cigarette on them. Greegor doesn't know much, Neither do I, but I do know more than Greegor. --=20 Bill Sloman, Sydney
Reply by ●October 17, 20132013-10-17
On Thu, 17 Oct 2013 14:14:28 -0700 (PDT), the renowned ccon67@netscape.net wrote:> >> Because you let that "output" float? What was the p/n? > >that's right > >Input has valid signal, power pins has good clean supply >all are normal, but the chip refuse to work because the LOST pin doesn't have pull up > >Their logic is this, if LOST signal is low, output must shutdown... > >they don't care if the input present or not.. even they don't care to say something about the auto-shutdown at the outputs in their spec > >:-))) > >Anyway, we gave up our "don't care' and we must use their 'don't care' > >after that I dont care anything from MaxOkay, but if you don't tell us the part number we can't be warned, nor can we snicker at you for missing the fine print at the bottom of page 37 where they explain not to do that (added in a revision 6 weeks after you gave up on it). Best regards, Spehro Pefhany -- "it's the network..." "The Journey is the reward" speff@interlog.com Info for manufacturers: http://www.trexon.com Embedded software/hardware/analog Info for designers: http://www.speff.com
Reply by ●October 17, 20132013-10-17
On Thu, 17 Oct 2013 15:52:04 -0700 (PDT), the renowned Bill Sloman <bill.sloman@gmail.com> wrote:>On Friday, 18 October 2013 06:18:27 UTC+11, Greegor wrote: >> >> JL > The outputs got erratic at high temperature, >> JL > looked like a problem in the latch circuit. >> JL > As the parts aged, the fail temperature >> JL > descended, ultimately reaching room temp >> JL > after about a year. All the parts did this. >> JL > A good hi-temp bake would push the fail >> JL > temperature up, and it would begin slowly >> JL > descending again from there. What would cause that? >> >> Hygroscopy - tending to absorb moisture from the air >> Did somebody get the wrong polymer for the chip packages? > >It seems unlikely - that's a very old problem. > >An even older problem is that reverse biasing base-emitter junctions (this was a John Larkin circuit) into breakdown produces cumulative damage, which can be largely reversed by baking - the legendary HP technician used to cure damaged transistors by stubbing out his cigarette on them.Okay, looked up the datasheet.. abs max differential voltage is only 3.5V, for a 5V supply device. And ft 6GHz devices.. which often have pretty low Vbe breakdown. http://pdf.datasheetcatalog.com/datasheets/37/441441_DS.pdf Eg, this discrete one that has a 1.5V abs max (!) rating:- http://www.cel.com/pdf/datasheets/upa800t.pdf Best regards, Spehro Pefhany -- "it's the network..." "The Journey is the reward" speff@interlog.com Info for manufacturers: http://www.trexon.com Embedded software/hardware/analog Info for designers: http://www.speff.com
