Chips/parts you hate working with...

[GK]

This part was a nightmare:

http://www.analog.com/static/imported-files/data_sheets_obsolete/311507236AD9501.pdf


Analog devices made a series of these programmable delay generators, with either TTL or ECL outputs. They were garbage. They were extraordinarily sensitive to noise on any pin and next to unusable in almost any logic environment. Several 10's of uV on a logic pin and the internal comparator would trigger erratically on the internally generated analogue ramp voltage. The whole series is obsolete now - I think I know why.

[T3sl4co1l]

Connectors with unclear mechanical and footprint dimensions on the datasheet.

Tremendous afternoons passed with connector, gauge, pencil & paper.

LCD displays (hurray for cheap Chinese drafting labor?) and connectors (FPC and others) seem to be the worst offenders.

Regarding redundancy: you're not required to note dimensions that are fully defined elsewhere, and often, you don't want to.  However, there is still a recognized method to do so: the measurement is written in parenthesis, indicating a derived value, as a note, not as a design reference.  (Thus, the tolerance on that number may be substantially worse than those around it; this isn't necessarily the case, like if you have a row of pins with pin-to-pin spacing "<number> +/- <tol> NON CUM.", then call out the number of pins.  The dimension between outermost pins will be both a derived value, and having tolerance exactly <tol> because it's non-cumulative.)

I hate working with Infineon parts. I worked for them and it was a terrible experience. I was glad to get another Job. Using Infineon products brings back bad memorys.

My condolences -- that sucks, from anywhere.

That's not what I've heard about them though, at least recently.  I've even had a few questions answered reasonably and promptly, at least from those respective departments.  Their parts are generally good quality and sufficiently specified.

Tim

[Kjelt]

LCD displays (hurray for cheap Chinese drafting labor?) and connectors (FPC and others) seem to be the worst offenders.

Ouch you woke up an very old frustration:

Those LCD zebra stripes , if you do not have the perfect 100% balanced mechanical pressure over the entire display some digits will not work 
And those once clipped on connectors are also bad when they are corroded, they can break off when you are only looking at them 

[wraper]

Nothing in the last 5 years since I changed professions (I'd like to think AMD have sorted out their overheating/self-destructive features properties). I personally swore off them back when the AMD K6 was around. Since then, I've seen several crispy AMD's first-hand in other machines I've worked on over the course of 10+ years in IT industry.

(Although the YouTube videos are amusing too!)

They have internal thermal protection/throttling since Athlon 64 and that was 10 years ago. Before that they did have a temp sensor but purely relied on motherboard that it will handle overheating correctly (often they did not).

[poorchava]

PIC16, PIC18, dsPIC30 - bugs on bugs chasing bugs which are being fed with bugs. Bastards owe me several weeks of my life which I spent chasing silicon bugs that sometime weren't even in errata sheet.

Chinese 'bead' 1W LEDs - i have yet to score a success on reflowing those in an oven without lens melting.

[Whales]

Flat flex and its connectors.  I've broken a lot of them when trying to work out how they operate -- every laptop seems to use at least 2 different latch styles.  The tiny plastic arms/fingers that hook them into place or guide them commonly can't take a 'normal' opening force if you operate them in the wrong direction.

A bright light and a lot of patience pay off

Corroded flat-flex is another story.  Keep your old electronics dry!

[T3sl4co1l]

Chinese 'bead' 1W LEDs - i have yet to score a success on reflowing those in an oven without lens melting.

I've done "3W" Chinese LEDs by 'hot stove reflow' on copper clad.  Your oven isn't heated from the top by any chance, is it?

I've desoldered 7-seg displays before, that kind of gooed to bits trying to pull them out...

Tim

[Alex Eisenhut]

Anythng that looks like a glass balloon with bits of scrap metal in it...

A Hewitt Mercury Arc Rectifier?

[Gribo]

I hate Murata's DLW family of common mode chokes. Is it so hard to add an orientation mark to a symmetrical component?

[void_error]

LCD displays with M2.5 mounting holes when they have enough room on the PCB to make M3 holes, and parts with unclear pin function descriptions (does that pin have any practical use or not  ).

I also really hate inductors with crap datasheets which doesn't specify the core material and/or operating frequency range. It should be a technical document not marketing crap.

QFN packages are also on the list although TQFP are perfectly fine. The only advantage I see in QFN is the small footprint but that is usually outweighed by the difficulty to solder/desolder them. That bloody annoying thermal pad on the bottom ruins everything.

[MagicSmoker]

I've had good luck with QFN packages, yet seem to always have soldering problems with TQFP's... Guess this is very much a YMMV kind of thing.

That said, I also have lots of problems with SMT LEDs... they seem unusually susceptible to damage from reflow soldering, and this is with leaded-solder; I can only imagine how well they do with lead-free soldering.

[alanambrose]

I like the QFN/DFNs - the ones I hate are the ones only available in SOIC.

I recently designed in some ADG5412 switches in LFCSP packages from the datasheet only to find when I got the boards back to make up ... that no one actually stocks the LFCSP. My stupidity I guess, but why do they have that package on the datasheet if they don't actually make it available?

Alan

[Corporate666]

I like the QFN/DFNs - the ones I hate are the ones only available in SOIC.

I recently designed in some ADG5412 switches in LFCSP packages from the datasheet only to find when I got the boards back to make up ... that no one actually stocks the LFCSP. My stupidity I guess, but why do they have that package on the datasheet if they don't actually make it available?

Alan

YES!  That is another huge issue I hate dealing with.  A lot of companies will have tons of variations and then none of them are ever available.  When you inquire, they will tell you that there will be a MOQ of 10,000 or 20,000pcs.  Even for large companies - who would commit to 20k of something you can't even get a few of for testing?

And actually, Cree is notorious for this with LED's... they will list tons of CRI and flux bins, and they will hype their latest LED to the max about how it produces 200lm/watt at 95CRI or some nonsense.  Then you see there are a couple of reels available at the distributors' websites in 80CRI and 90lm/watt flux bins... and the sales droids will tell you they can happily supply the higher flux bins (at triple the price) if you're able to commit to 50k pcs orders.  Ridiculous.

[con-f-use]

Amen on die the Cree thing!

[SirNick]

Minor gripe:  TI's TPS54226 switching regulator.  There's an Enable pin, and a Vreg pin that supplies 5v for control circuitry.  The Enable pin is diagonally opposite from the Vcc pin (with a ground pad between) but in my layout, the Vreg pin was fairly easily accessible, so I thought "OK, I'll tie Enable to Vreg".  Predictably, it turns out Vreg is ALSO disabled when Enable is LOW, so my prototype board had a bit of a chicken and egg thing going on.

To be fair, I realized this was a possibility, and combed through the datasheet looking for confirmation one way or another -- and therein lies my chief complaint.  There isn't one word about this either way.  In the block diagram, Enable goes to a single isolated block labelled "Enable Logic".  In the application circuits, Enable goes to an external header.  In the timing diagrams, the sequencing is shown between Enable, Vout, and PowerGood, but not Vreg.  In the text, not even a single paragraph is devoted to Enable's function.

For an essential (albeit trivial) signal, there's not much attention given to how it works.

[mikeselectricstuff]

I like the QFN/DFNs - the ones I hate are the ones only available in SOIC.

As Long As you design the pads to stick out at least 0.5mm from the side of the package, DF/QFns aren't too bad, and are in some respects easier to rework than QFPs - no bent /lifted pins, and no solder wicking up between pins.
I don't like the ones that need a die slug soldering down though as this severely limits space for routing tracks

[coppice]

I like the QFN/DFNs - the ones I hate are the ones only available in SOIC.

As Long As you design the pads to stick out at least 0.5mm from the side of the package, DF/QFns aren't too bad, and are in some respects easier to rework than QFPs - no bent /lifted pins, and no solder wicking up between pins.
I don't like the ones that need a die slug soldering down though as this severely limits space for routing tracks

If you ever worked with LCCs in the 80s you will realise that the die slug is critical for sustained mechanical strength over many temperature cycles with a QFN. For RF parts it is also critical for performance, but I believe its present on most parts for mechanical reasons.

[Rufus]

Parts with shit datasheets - http://www.st.com/web/catalog/sense_power/FM89/SC444/PF250725 for example.

Has a nasty 3mm square LGA-16 package to boot.

[qwaarjet]

LT3652 in 3X3 mm DFN pacakage

[T3sl4co1l]

I hate Murata's DLW family of common mode chokes. Is it so hard to add an orientation mark to a symmetrical component?

Specifically, the -5BS/BT?  That looks terrible...

Tim

[NiHaoMike]

Anythng that looks like a glass balloon with bits of scrap metal in it...

How are you going to reheat your pizza without one of those?

One of those ceramic or nichrome heating elements.

[alanambrose]

My pet hates right now are diodes that don't have cathode markings, chips where the pin 1 orientation is marked on the bottom of the chip, datasheets which show the pinout looking at the chip from below, molex put every component in the same parameterised drawing BS - only to be beaten by datasheets where it isn't possible to tell whether the pinout is shown from the top or the bottom. I'm getting a bit worked up now (!) - you must have touched a nerve...

Alan

[BravoV]

... chips where the pin 1 orientation is marked on the bottom of the chip

Curious which chip is that ? 

[c4757p]

datasheets which show the pinout looking at the chip from below

This is not uncommon with relays - and ought to be punishable by public flogging

[Sigmoid]

Anythng that looks like a glass balloon with bits of scrap metal in it...

Aw that was just mean.