Dumb question regarding SMPS

[[email protected]]

Only for some CAD packages, if at all. My CAD has a lot of the Atmels,
just not this big one.

which one?

I insist on the same, I really hate netlist-style schematics where the
front axle is on page 17 which the left front wheel it on page 32.
Exceptions are logic gate and opamp multi-packs, of course. And I never
use large BGAs, those can spell doom in a hi-rel environment.

isn't reliabily and BGAs something that was perfected many years ago?

for something like a big FPGA I think it it makes sense to put core
power, jtag
configuration and such on one symbol and each bank on a separate
symbol

-Lasse

 

[Joerg]

which one?

It's the ATMega2560.

isn't reliabily and BGAs something that was perfected many years ago?

I do not trust them. Seen way too much grief with BGA. Most of all, I do
not like something where solder connections cannot be inspected with
reasonable effort. Large rigid structures with next to nothing in
compliance on a flexible material (FR4) are IMHO a bad idea.

for something like a big FPGA I think it it makes sense to put core
power, jtag
configuration and such on one symbol and each bank on a separate
symbol

That may be ok. But I really do not like it when, for example, the
various UARTs of a uC are sprinkled across several schematics pages.

 

[[email protected]]

Only for some CAD packages, if at all. My CAD has a lot of the Atmels,
just not this big one.

Not the CAD companies. The chip manufacturer often has CAD symbols and
footprints for the popular CAD packages. They usually support the common
packages. Others can often import the data from that data.

I insist on the same, I really hate netlist-style schematics where the
front axle is on page 17 which the left front wheel it on page 32.

Of course not. All the wheels go on page-1 (with the tires/wheels/and nuts in
the hierarchy on page 2-5) and the Engine goes on page 12, with the exhaust on
page 88.

Exceptions are logic gate and opamp multi-packs, of course.

Gates and opamps are drawn symbolically but everything else is a single
physical square box. UGH! It makes *really* ugly schematics. Impossible to
follow.

And I never
use large BGAs, those can spell doom in a hi-rel environment.

Nonsense. I did a bunch of MIL stuff (Naval weapons system) with 1k pin BGAs.
Fine pitch isn't so good but there's nothing wrong with >.8mm BGA packages.
Doom comes before the design starts, without them. No project = no paycheck.

 

[[email protected]]

which one?


isn't reliabily and BGAs something that was perfected many years ago?

Sure, at least in the larger packages. Joerg is living in the '80s. They
even worked well then but many didn't have the process down.

for something like a big FPGA I think it it makes sense to put core
power, jtag
configuration and such on one symbol and each bank on a separate
symbol

Agree 100%. Even "small" 144-pin QFPs are a mess when they're shown as one
big box on a page with wires everywhere. Joerg probably doesn't like busses,
either (someone here objected to them a while back).

I often break FPGAs up by I/O bank, if I haven't got the design very far
along. If know how it's going to flow, I'll break it up that way. I can
still do that with BGAs, but not QFPs or QFNs. The CPoE has some really
strange ideas on how a schematic is to look. The CAD system sucks, too, but
that's a different topic. ;-)

 

[[email protected]]

I get the feeling that complex uP chips contain big blocks of
purchased IP, sort of shoveled into the cement mixer. ADCs, Ethernet,
USB, PCIe, whatever. Maybe the factory can't help much.

--

I've done a few big SoCs in a previous job, and that is kinda how it
works with
both inhouse and purchased IP, and for every new design the blocks get
a few
tweaks to meet some new requirement and in the end no one really know
how they
should work

-Lasse

 

[[email protected]]

Yup :-(

Very classic omission: Do the ports have input hysteresis or not? And if
yes, how much? Once the answer from the tech support engineer, after
long head-scratching, was: "Good question! I'll have to inquire about
that at the factory". Oh man ...

I just got one that had *every* power/current value listed as TBD. Nice. It's
fun designing the power system with such a complete spec.

 

[Joerg]

Not the CAD companies. The chip manufacturer often has CAD symbols and
footprints for the popular CAD packages. They usually support the common
packages. Others can often import the data from that data.

I know that IC mfgs provide lib parts at times. But I use a CAD that is
not so popular out here. It has its downsides such as this, but it has
an upside that wipes out any and all downsides: No crashes.

Of course not. All the wheels go on page-1 (with the tires/wheels/and nuts in
the hierarchy on page 2-5) and the Engine goes on page 12, with the exhaust on
page 88.


Gates and opamps are drawn symbolically but everything else is a single
physical square box. UGH! It makes *really* ugly schematics. Impossible to
follow.

That's in fact how most people want schematics. Including myself.

Nonsense. I did a bunch of MIL stuff (Naval weapons system) with 1k pin BGAs.

On alumina and other higher tech material, yes. On FR4 I simply will not
do it. If the client insists, ok, but then the result is their
responsibility.

Fine pitch isn't so good but there's nothing wrong with >.8mm BGA packages.
Doom comes before the design starts, without them. No project = no paycheck.

Well, I am quite flooded with work despite the fact that I do not use
BGA. Simulating right now to see whether I can use an MSOP sans thermal
pad instead of a leadless QFN with pad. Nothing beats the gentle
compliance of leads on a package.

If it goes on like this through winter I won't need to heat the office,
the computer does that.

 

[Joerg]

Sure, at least in the larger packages. Joerg is living in the '80s. They
even worked well then but many didn't have the process down.


Agree 100%. Even "small" 144-pin QFPs are a mess when they're shown as one
big box on a page with wires everywhere. Joerg probably doesn't like busses,
either (someone here objected to them a while back).

Actually I do like buses a lot. Especially the ones with comfy seats and
WiFi on board ... Just kidding, I do like buses on schematics.

I often break FPGAs up by I/O bank, if I haven't got the design very far
along. If know how it's going to flow, I'll break it up that way. I can
still do that with BGAs, but not QFPs or QFNs. The CPoE has some really
strange ideas on how a schematic is to look. The CAD system sucks, too, but
that's a different topic. ;-)

Ever dealt with Asian-style schematics? Oh, there is still some white on
the page, let's cram the preamp in there. Then you start following a
line clear across the page and it goes to ... ground!

 

[[email protected]]

I know that IC mfgs provide lib parts at times. But I use a CAD that is
not so popular out here. It has its downsides such as this, but it has
an upside that wipes out any and all downsides: No crashes.

It can't import data from any other format? Ours can't but I don't expect
much from it (and I'm always disappointed).

That's in fact how most people want schematics. Including myself.

Most? GOt a cite for that? None of the companies I've worked for wanted
anything that ugly. It makes the schematic unreadable.

On alumina and other higher tech material, yes. On FR4 I simply will not
do it. If the client insists, ok, but then the result is their
responsibility.

Nope. FR4. The is *no* problem. In fact, they're more reliable than other
packages. They do have to be done right, of course. Actually, at the PPoE
(small company) they had far fewer problems with BGAs than they did with any
other package. They didn't have an XRAY machine, either.

Well, I am quite flooded with work despite the fact that I do not use
BGA. Simulating right now to see whether I can use an MSOP sans thermal
pad instead of a leadless QFN with pad. Nothing beats the gentle
compliance of leads on a package.

You do tiny stuff. Without big DSPs, I don't have a job.

If it goes on like this through winter I won't need to heat the office,
the computer does that.

Global warming is finally coming to CA?

 

[Joerg]

It can't import data from any other format? Ours can't but I don't expect
much from it (and I'm always disappointed).

It can import but let's put it that way: My experience with that is very
mixed. I wouldn't trust it.

Most? GOt a cite for that? None of the companies I've worked for wanted
anything that ugly. It makes the schematic unreadable.

Most as in pretty much all my clients. I don't think there is any
opinion poll data for that available.

I don't know what's ugly about it. Look at the schematics on the web,
they usually have opamps split out, same for 74HC14 inverters and such,
but the uC or DSP is one big block.

Nope. FR4. The is *no* problem. In fact, they're more reliable than other
packages.

As you tend to say, I do not buy it.

They do have to be done right, of course. ...

Does this mean lots of laptop manufacturers including your former
employer do it wrong? This is how it looks when BGA solder joints on
their laptops give up:

Just for giggles, someone came up with a real low-tech repair method:

... Actually, at the PPoE
(small company) they had far fewer problems with BGAs than they did with any
other package. They didn't have an XRAY machine, either.


You do tiny stuff. Without big DSPs, I don't have a job.

Then I do not want your job

Global warming is finally coming to CA?

I sure hope so. However, since I do not trust that we bought the record
qty of five cords of almond again for this year. Blew through the
complete stack the last two seasons already, literally done to the last
sticks. In 2000 I thought that was impossible, back then two cords was fine.

 

[Joerg]

Our in-house BGA placement success rate has been 100%. We have had a
couple of problems after multiple rework/replacements. BGAs seem to be
a lot more reliable than fine-pitch leaded chips.

But your stuff probably doesn't occasionally get nailed to the runway
with a loud ker-crunch sound and several tires blowing out.

Most of all, I do

Another virtue of BGAs!

Hmm, I don't see that one. I want to be able to inspect. It doesn't
protect against reverse engineering since those guys have every tool
imaginable at their disposal.

 

[[email protected]]

But your stuff probably doesn't occasionally get nailed to the runway
with a loud ker-crunch sound and several tires blowing out.

Very few Gs.

Hmm, I don't see that one. I want to be able to inspect. It doesn't
protect against reverse engineering since those guys have every tool
imaginable at their disposal.

XRAY, if you must. We have one in our lab but it's never used.

 

[[email protected]]

Actually I do like buses a lot. Especially the ones with comfy seats and
WiFi on board ... Just kidding, I do like buses on schematics.

Busses where the contents are not similarly named (not Bus[0:11], rather
BusWr, BusRd, BusEn,...)?

Ever dealt with Asian-style schematics? Oh, there is still some white on
the page, let's cram the preamp in there. Then you start following a
line clear across the page and it goes to ... ground!

I work for an Asian company. I think their schematics *suck*. No hierarchy,
even.

 

[[email protected]]

I agree that on the whole they are reliable, but they have a reputation
in some circles through some manufacturers cutting corners. You only
have to look at the reliability of Xboxes!

It can't be cutting corners, rather incompetence. My experience with BGAs is
the same as JL's - far more reliable than other fine-pitched packages. Even
the .5mm BGA went flawlessly, even though the process guy didn't think they
could do it.

 

[[email protected]]

It can import but let's put it that way: My experience with that is very
mixed. I wouldn't trust it.

Trust, but verify. ;-)

Most as in pretty much all my clients. I don't think there is any
opinion poll data for that available.

Sorta my point, though if I'd shown up with a schematic like that, no one
would have hired me. They want their documentation to describe how the widget
works.

I don't know what's ugly about it. Look at the schematics on the web,
they usually have opamps split out, same for 74HC14 inverters and such,
but the uC or DSP is one big block.

Nothing else fits on the page with the "big block". Might just as well have a
netlist.

As you tend to say, I do not buy it.

Of course you don't. It's a fact, though. BGAs are a far superior package.
There are some problems with thick boards and fine-pitched BGAs but I try to
stay at .8mm or above. .5mm can be done but it gets dicey with more than six
layers or so. The boards get pretty thin, too (1mm and less).

Does this mean lots of laptop manufacturers including your former
employer do it wrong? This is how it looks when BGA solder joints on
their laptops give up:

Probably. Nothing would surprise me from the PC Co.

Sorry can't watch videos. My laptop will blow up. ;-)

Just for giggles, someone came up with a real low-tech repair method:

Then I do not want your job

Different strokes.

I sure hope so. However, since I do not trust that we bought the record
qty of five cords of almond again for this year. Blew through the
complete stack the last two seasons already, literally done to the last
sticks. In 2000 I thought that was impossible, back then two cords was fine.

I thought you bought five cords last year? I've got two houses with heat
pumps now. No more wood stoves for me (we do have a gas fireplace in the old
house and I'll probably put gas logs in this one).

 

[Joerg]

[...]

It can import but let's put it that way: My experience with that is very
mixed. I wouldn't trust it.

Trust, but verify. ;-)

Then it doesn't save time. If I have to verify all names, pin numbers
and ERC pin functions I might as well roll my own from scratch.

Sorta my point, though if I'd shown up with a schematic like that, no one
would have hired me. They want their documentation to describe how the widget
works.

My schematics do describe how it works. Analog guys live that way.

Nothing else fits on the page with the "big block". Might just as well have a
netlist.

That's perfectly fine. There will be a net name saying, for example,
PULSECNT and the two letters uC underneath. Then everyone with enough
smarts will know that this goes to sheet 7 where the big fat uC lives.
It is not necessary and would be rather confusing to whack out timer #2
and then place it on sheet 2 where this net is. You'd be looking at
sheet 7 and ask your self "There's pins missing here. Where the heck did
all the timers go?".

Of course you don't. It's a fact, though. BGAs are a far superior package.
There are some problems with thick boards and fine-pitched BGAs but I try to
stay at .8mm or above. .5mm can be done but it gets dicey with more than six
layers or so. The boards get pretty thin, too (1mm and less).

The real fact is that whole small businesses have sprung up because of
all the BGA failures. Usually one-man shops. In our area their are
mobile, they often buy retired ambulances, put in a work bench, then
drive to the customers with the BGA problems.

Probably. Nothing would surprise me from the PC Co.


Sorry can't watch videos. My laptop will blow up. ;-)

It's an IBM T40 with BGA failure. Lots of other brands have the same issues.

Different strokes.


I thought you bought five cords last year? I've got two houses with heat
pumps now. No more wood stoves for me (we do have a gas fireplace in the old
house and I'll probably put gas logs in this one).

That's why I wrote "again". We bought five cords also last year, and
blew through all of that. Lots of people in CA also have heat pumps but
that was a major mistake. Because the state government became hardcore
leftist so now they tax the dickens out of these poor folks, via reverse
tiers on the power bill. A meager baseline quantity costs around
15c/kWh. With a heatpump or A/C you exceed that within days and then
prices quickly shoot up to something like 35c/kWh. I knew an old lady
who was very skinny, so needed it warm in the house in her 90's. She
paid north of $1000/month in winter. That's insane, I won't do that.

 

[Joerg]

Actually I do like buses a lot. Especially the ones with comfy seats and
WiFi on board ... Just kidding, I do like buses on schematics.

Busses where the contents are not similarly named (not Bus[0:11], rather
BusWr, BusRd, BusEn,...)?

If it's clear what the bus does I am ok with that. But I prefer
BUSWR[0:11], BUSRD[0:11], BUSEN[0:3], and so on. Consistency is
important in schematics.

I work for an Asian company. I think their schematics *suck*. No hierarchy,
even.

You have my sympathies

 

[[email protected]]

[email protected] wrote:
[email protected] wrote:
John Larkin wrote:
John Larkin wrote:
Jim Thompson wrote:
03:12 -0700, John Larkin
I'm on my third laptop right now. Every time I plugged my Dell's "fat
snake" into the wall, I drew quite an arc. The Lenovo's arc was not
noticeable, but now I get a noticeable arc with my new HP -- not as
big as the Dell's, however.
I know FA about switch mode power supplies, obviously, so I wonder
1. What produces the arc?
2. Why would different power supplies produce different arcs (does it
just depend on output power capability)?
3. Why is there no arc when I pull the plug from the outlet?
Dumb switching power supplies have a bridge rectifier and a big
electrolytic filter capacitor. If you plug them in near the peak of
the AC line waveform, the charging current will spark.
Better supplies, with inrush limiters, or PFC (power factor corrected)
front-ends, have much less inrush charge.
Ask Jim for details. He is *so good* at designing switching power
supplies.
Indeed I am >
I like the ucc28019a. works like a champ. Undervoltage lockout too
Cheers
So does the L6561. But poorly documented and no model; and ST ignores
my pounding on their door >
If I could make major changes I'd design ST out _forever_!
Might be the new normal. I want to design in a National video driver,
LMH6722. Has a thermal pad under its belly. In the datasheet they forgot
to mention where its s'posed to be connected to. Probably V- but I'd
rather make sure. Filed a support ticket with the new owner TI on 10/2.
Got a service request number.
Today is 10/8 and (finally! ... or so I thought) there was a message in
the inbox this morning. A form letter, merely saying that, tada, a
service request number has been issued. New number: Same as the old number.
So I responded politely as to when I might be expecting an answer. No
response all day.
Hurumph!
Get one and ohm it out.
No kidding, that may be the only way :-(
Just made a CAD model for a 100-TQFP processor. Now I know why I chose
to become an analog guy and not a digital one.
Do you mean a PCB decal, or an actual Autocad sort of thing?
PADS makes IC decals really fast, for sort of standard things with
rows of numbered pins.
Are you doing 3D Solidworks sort of physical modeling? It is fun to
finally spin that stuff around in space, or take a virtual walk under
the IC pins.
No, just the schematic library part and footprint. A hundred pins, most
of which have names like this:
(OC0A/OC1C/PCINT7)PB7
One typo and all hell can break loose because the routing resources in
those uCs are sparse and can be unforgiving. Just had a major
re-shuffling in one of them on another project, not because of an error
but for a feature change. When those get maxed out in port pins the
design can slow down as much as Van Ness at rush hour, mainly because of
routing compromises.
You *should* be able to either grab the names from a spreadsheet or
cut-n-paste from a datasheet. The vendors often have models already built
that can be used for a starting place, too.
Only for some CAD packages, if at all. My CAD has a lot of the Atmels,
just not this big one.
which one?

... OTOH, our CAD people demand that
chips look on the schematic like they do on the board - no functional
partitioning (except BGAs, for some reason).
I insist on the same, I really hate netlist-style schematics where the
front axle is on page 17 which the left front wheel it on page 32.
Exceptions are logic gate and opamp multi-packs, of course. And I never
use large BGAs, those can spell doom in a hi-rel environment.

isn't reliabily and BGAs something that was perfected many years ago?
Sure, at least in the larger packages. Joerg is living in the '80s. They
even worked well then but many didn't have the process down.

for something like a big FPGA I think it it makes sense to put core
power, jtag
configuration and such on one symbol and each bank on a separate
symbol
Agree 100%. Even "small" 144-pin QFPs are a mess when they're shown as one
big box on a page with wires everywhere. Joerg probably doesn't like busses,
either (someone here objected to them a while back).

Actually I do like buses a lot. Especially the ones with comfy seats and
WiFi on board ... Just kidding, I do like buses on schematics.

Busses where the contents are not similarly named (not Bus[0:11], rather
BusWr, BusRd, BusEn,...)?

If it's clear what the bus does I am ok with that. But I prefer
BUSWR[0:11], BUSRD[0:11], BUSEN[0:3], and so on. Consistency is
important in schematics.

Huh? Do you have twelve bus writes?

You have my sympathies

Their processed really are terrible (surprised me!). One particularly bad
point is grounds. If a connector has a shield connection, for instance,
that's not shown in the datasheet as a pin, it's not shown on the schematic.
The layout the layout person has to manually connect it to GND (or wherever).
It's up to the next layout guy to remember to do it, too. Worse, if I do
place a ground pin for it in the symbol library and someone else comes along
and uses that part, they'll remove the ground pin. *Poof*, mine is gone too.
No warning, nothing.

Oh, and there is no way to obsolete a component from the database. We have
thousands of capacitors to go through that the manufacturer no longer makes.
These have to be scrubbed every time we put a BOM together.

Yes, it sucks.

 

[[email protected]]

[email protected] wrote:

[email protected] wrote:

[email protected] wrote:

[...]

I know that IC mfgs provide lib parts at times. But I use a CAD that is
not so popular out here. It has its downsides such as this, but it has
an upside that wipes out any and all downsides: No crashes.
It can't import data from any other format? Ours can't but I don't expect
much from it (and I'm always disappointed).

It can import but let's put it that way: My experience with that is very
mixed. I wouldn't trust it.

Trust, but verify. ;-)

Then it doesn't save time. If I have to verify all names, pin numbers
and ERC pin functions I might as well roll my own from scratch.

It saves a *lot* of time. You have to type every one in and *still* check
(probably a few times by different people) to make sure you didn't make a
mistake.

My schematics do describe how it works. Analog guys live that way.

If there is a huge box on the page and nothing else, you are *not* describing
how it works. It's no better than a netlist. Worse, actually, because you
have to trace wires to find out which signals they're tied to.

That's perfectly fine. There will be a net name saying, for example,
PULSECNT and the two letters uC underneath. Then everyone with enough
smarts will know that this goes to sheet 7 where the big fat uC lives.

What it's hooked to is anyone's guess, though.

It is not necessary and would be rather confusing to whack out timer #2
and then place it on sheet 2 where this net is. You'd be looking at
sheet 7 and ask your self "There's pins missing here. Where the heck did
all the timers go?".

A hundred blocks, perhaps. A few, not so much. Separating all the power,
grounds, and JTAG, for instance, cleans up the schematic a *lot*. Breaking out
the memory controller (or two - separate pages) goes a long way to
readability. Perhaps all the A/Ds on a block and timers on another. There are
reasonable ways to divide things up functionally. It's much like the argument
for hierarchy. One big box simply sucks.

The real fact is that whole small businesses have sprung up because of
all the BGA failures. Usually one-man shops. In our area their are
mobile, they often buy retired ambulances, put in a work bench, then
drive to the customers with the BGA problems.

Nonsense. BGAs are often worth salvaging so sure there's a business in doing
so. They are more difficult to replace so maybe there's a business there,
too. They are *not* less reliable than QFPs. Just the opposite, in fact.

It's an IBM T40 with BGA failure. Lots of other brands have the same issues.

QFPs have never failed? I can tell you otherwise!

That's why I wrote "again". We bought five cords also last year, and
blew through all of that.

You said "record qty of five cords", implying that it was less last year.

Lots of people in CA also have heat pumps but
that was a major mistake. Because the state government became hardcore
leftist so now they tax the dickens out of these poor folks, via reverse
tiers on the power bill. A meager baseline quantity costs around
15c/kWh. With a heatpump or A/C you exceed that within days and then
prices quickly shoot up to something like 35c/kWh. I knew an old lady
who was very skinny, so needed it warm in the house in her 90's. She
paid north of $1000/month in winter. That's insane, I won't do that.

Ouch. My heat pumps cost $.07/kWh(GA) and $.09(AL) during the heating season.
My power bill approaches $200 (each place, if they were both running) during
January and July. It's about half that in the off months so the heat portion
of the bill is about $100 (maybe $120).

 

[Joerg]

[email protected] wrote:
[email protected] wrote:
John Larkin wrote:
John Larkin wrote:
Jim Thompson wrote:
03:12 -0700, John Larkin
I'm on my third laptop right now. Every time I plugged my Dell's "fat
snake" into the wall, I drew quite an arc. The Lenovo's arc was not
noticeable, but now I get a noticeable arc with my new HP -- not as
big as the Dell's, however.
I know FA about switch mode power supplies, obviously, so I wonder
1. What produces the arc?
2. Why would different power supplies produce different arcs (does it
just depend on output power capability)?
3. Why is there no arc when I pull the plug from the outlet?
Dumb switching power supplies have a bridge rectifier and a big
electrolytic filter capacitor. If you plug them in near the peak of
the AC line waveform, the charging current will spark.
Better supplies, with inrush limiters, or PFC (power factor corrected)
front-ends, have much less inrush charge.
Ask Jim for details. He is *so good* at designing switching power
supplies.
Indeed I am >
I like the ucc28019a. works like a champ. Undervoltage lockout too
Cheers
So does the L6561. But poorly documented and no model; and ST ignores
my pounding on their door >
If I could make major changes I'd design ST out _forever_!
Might be the new normal. I want to design in a National video driver,
LMH6722. Has a thermal pad under its belly. In the datasheet they forgot
to mention where its s'posed to be connected to. Probably V- but I'd
rather make sure. Filed a support ticket with the new owner TI on 10/2.
Got a service request number.
Today is 10/8 and (finally! ... or so I thought) there was a message in
the inbox this morning. A form letter, merely saying that, tada, a
service request number has been issued. New number: Same as the old number.
So I responded politely as to when I might be expecting an answer. No
response all day.
Hurumph!
Get one and ohm it out.
No kidding, that may be the only way :-(
Just made a CAD model for a 100-TQFP processor. Now I know why I chose
to become an analog guy and not a digital one.
Do you mean a PCB decal, or an actual Autocad sort of thing?
PADS makes IC decals really fast, for sort of standard things with
rows of numbered pins.
Are you doing 3D Solidworks sort of physical modeling? It is fun to
finally spin that stuff around in space, or take a virtual walk under
the IC pins.
No, just the schematic library part and footprint. A hundred pins, most
of which have names like this:
(OC0A/OC1C/PCINT7)PB7
One typo and all hell can break loose because the routing resources in
those uCs are sparse and can be unforgiving. Just had a major
re-shuffling in one of them on another project, not because of an error
but for a feature change. When those get maxed out in port pins the
design can slow down as much as Van Ness at rush hour, mainly because of
routing compromises.
You *should* be able to either grab the names from a spreadsheet or
cut-n-paste from a datasheet. The vendors often have models already built
that can be used for a starting place, too.
Only for some CAD packages, if at all. My CAD has a lot of the Atmels,
just not this big one.
which one?

... OTOH, our CAD people demand that
chips look on the schematic like they do on the board - no functional
partitioning (except BGAs, for some reason).
I insist on the same, I really hate netlist-style schematics where the
front axle is on page 17 which the left front wheel it on page 32.
Exceptions are logic gate and opamp multi-packs, of course. And I never
use large BGAs, those can spell doom in a hi-rel environment.

isn't reliabily and BGAs something that was perfected many years ago?
Sure, at least in the larger packages. Joerg is living in the '80s. They
even worked well then but many didn't have the process down.

for something like a big FPGA I think it it makes sense to put core
power, jtag
configuration and such on one symbol and each bank on a separate
symbol
Agree 100%. Even "small" 144-pin QFPs are a mess when they're shown as one
big box on a page with wires everywhere. Joerg probably doesn't like busses,
either (someone here objected to them a while back).

Actually I do like buses a lot. Especially the ones with comfy seats and
WiFi on board ... Just kidding, I do like buses on schematics.
Busses where the contents are not similarly named (not Bus[0:11], rather
BusWr, BusRd, BusEn,...)?

If it's clear what the bus does I am ok with that. But I prefer
BUSWR[0:11], BUSRD[0:11], BUSEN[0:3], and so on. Consistency is
important in schematics.

Huh? Do you have twelve bus writes?

That usually means how many lines a bus has. Yes, some of mine have 12
lines.

Their processed really are terrible (surprised me!). One particularly bad
point is grounds. If a connector has a shield connection, for instance,
that's not shown in the datasheet as a pin, it's not shown on the schematic.
The layout the layout person has to manually connect it to GND (or wherever).
It's up to the next layout guy to remember to do it, too. Worse, if I do
place a ground pin for it in the symbol library and someone else comes along
and uses that part, they'll remove the ground pin. *Poof*, mine is gone too.
No warning, nothing.

Yikes! That can quickly result in egg in the face, at the EMC lab.

Oh, and there is no way to obsolete a component from the database. We have
thousands of capacitors to go through that the manufacturer no longer makes.
These have to be scrubbed every time we put a BOM together.

If you look long enough you probably still find rectifier tubes,
nuvistors and Leyden jars in there.

Yes, it sucks.

I bet it does. Now I am glad to be self-employed