Simulation
Hi
I brought up the idea of a full spec at the time. Since we made the parts and used them it seemed like a reasonable idea. The price on the parts went from $ 0.0326 each to $ 18.56 when the major parameters were banded 2 to 1. Since we were using 4 in a $ 35 item (bom was about $6) the suggestion was not well received.
Bob
On Aug 14, 2010, at 11:50 AM, Magnus Danielson magnus@rubidium.dyndns.org wrote:
On 08/14/2010 05:08 PM, J. Forster wrote:
FWIW, IMO any engineer who uses undocumented or uncontrolled parameters or
instructions in a production design is a fool.
If you are that silly, you must fully specify the selection criteria.
Using non-spec aspects needs documentation, motivation and risc analysis.
I've designed a selection routine for a component which still ticks on with good statistics. The main problem is that eventually that component will be on last buy level. Obviously it seems the selection was chosen on the conservative side, so it works in shipped products regardless of batch. Trimming of the manual routine has lowered a certain failure mode of testing.
Any spec should be verified. Published specs needs verification with real components. Unpublished specs needs consistency testing or even selection testing on all components. Cost of testing needs to be understood and risc of low yield in future needs to be understood and alternative approaches could be put in place before running on flat tires.
At times it may be cheaper and safer to run with more expensive components which is within spec.
Cheers,
Magnus
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.
Hi
The PDP-8 had so much code that depended on un-documented instructions that they had to include them in later versions of the machine....
Bob
On Aug 14, 2010, at 12:01 PM, "J. Forster" jfor@quik.com wrote:
I think you missinterpret what I meant. Two examples:
I've seen programmers who use "instructions" that are not part of a uP
instruction set and are undocumented, just to be "clever". If a different
brand of chip, or even a different rev., the chip does something
completely different. These guys should be strung up by their tender
parts.
I've also seen transistors used as avalanche switches (basically a failure
mode). If a different production run has improved normal mode performance,
the avalanche function may vanish.
FWIW,
-John
===============
J. Forster wrote:
FWIW, IMO any engineer who uses undocumented or uncontrolled parameters
or
instructions in a production design is a fool.
If you are that silly, you must fully specify the selection criteria.
-John
Or, has their back against the wall and can't do it any other way.
How is this any different than using trimpots or hand select?
For years, folks have hand selected matched pairs of devices, since the
circuit requires tighter tolerances than the mfr guarantees.
Many, many RF designs have "select at test" pads to set levels or tuning
stubs depending on what the actual gain or impedance properties of the
active devices are, or for trimming temperature dependencies.
Would you say that the engineer is a fool for not just specifying
tighter tolerances.. the tighter tolerances may not be available from
the mfr (who has to respond to many customers, most of which will be
happy with the standard performance). It's sort of a tradeoff.. do you
go to the mfr and say, I need a better grade of part, or do you buy the
run-of-the-mill part, and sort them.
You might decide to do the latter for competitive reasons, e.g. rather
than the mfr producing a better grade of part, and potentially selling
it to your competitors too, you keep the "secret sauce" in house.
(Granted you could have the mfr make/select a proprietary part for you..
that's basically changing who does the work, but doesn't change the
underlying design)
Even manufacturers do this, for instance with speed grades on things
like microprocessors. They don't have enough process control to
guarantee a particular speed, so they make em all, and then sort them.
The other thing is that the selection criteria might not be knowable in
a standalone sense. That is, you have to put the part into the circuit
and see if it works, rather than measuring some device parameter. I
would agree that to a certain extent, this implies that you don't really
know how the circuit works, but it might also be that the most cost
effective approach is to use empiricism, rather than analysis.
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.
DEC code was a nightmare. Any DG Nova line code would run on any machine.
-John
=============
Hi
The PDP-8 had so much code that depended on un-documented instructions
that they had to include them in later versions of the machine....
Bob
On Aug 14, 2010, at 12:01 PM, "J. Forster" jfor@quik.com wrote:
I think you missinterpret what I meant. Two examples:
I've seen programmers who use "instructions" that are not part of a uP
instruction set and are undocumented, just to be "clever". If a
different
brand of chip, or even a different rev., the chip does something
completely different. These guys should be strung up by their tender
parts.
I've also seen transistors used as avalanche switches (basically a
failure
mode). If a different production run has improved normal mode
performance,
the avalanche function may vanish.
FWIW,
-John
===============
J. Forster wrote:
FWIW, IMO any engineer who uses undocumented or uncontrolled
parameters
or
instructions in a production design is a fool.
If you are that silly, you must fully specify the selection criteria.
-John
Or, has their back against the wall and can't do it any other way.
How is this any different than using trimpots or hand select?
For years, folks have hand selected matched pairs of devices, since the
circuit requires tighter tolerances than the mfr guarantees.
Many, many RF designs have "select at test" pads to set levels or
tuning
stubs depending on what the actual gain or impedance properties of the
active devices are, or for trimming temperature dependencies.
Would you say that the engineer is a fool for not just specifying
tighter tolerances.. the tighter tolerances may not be available from
the mfr (who has to respond to many customers, most of which will be
happy with the standard performance). It's sort of a tradeoff.. do you
go to the mfr and say, I need a better grade of part, or do you buy the
run-of-the-mill part, and sort them.
You might decide to do the latter for competitive reasons, e.g. rather
than the mfr producing a better grade of part, and potentially selling
it to your competitors too, you keep the "secret sauce" in house.
(Granted you could have the mfr make/select a proprietary part for
you..
that's basically changing who does the work, but doesn't change the
underlying design)
Even manufacturers do this, for instance with speed grades on things
like microprocessors. They don't have enough process control to
guarantee a particular speed, so they make em all, and then sort them.
The other thing is that the selection criteria might not be knowable in
a standalone sense. That is, you have to put the part into the circuit
and see if it works, rather than measuring some device parameter. I
would agree that to a certain extent, this implies that you don't
really
know how the circuit works, but it might also be that the most cost
effective approach is to use empiricism, rather than analysis.
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to
https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.
On 08/14/2010 05:48 PM, J. Forster wrote:
Simulation has some value in determining things like allowable component
tolerances and Worst Case analysis, but those are really production
engineering rather than design.
As to working at brassboard but not in production, it is prudent to check
that your parts are within the production part specs.
Simulation as such is a very useful tool. Propper use may save time,
misuse will cost you big-time. It does not replace breadboarding and
design verification, but may speed up design and "what if?" on failure
mode testing.
I've often found that proposing a solution and have a quick simulation
has helped in convincing on certain design ideas. In particular some
designer make things more complex than they need to be... but a quick
simulation gets them on the right track. :)
Cheers,
Magnus
On 08/14/2010 06:15 PM, J. Forster wrote:
DEC code was a nightmare. Any DG Nova line code would run on any machine.
This is how we have learned what is a bad idea... and it is now
documented in the guidelines.
Use of the top 8 bits in pointers caused headaches for the 68k machines
when they needed more than 16 MB of addressing. Amongst other things.
The undocumented features of the 6502 is now being maintained in
modernized versions... maybe a better choice in design could be made.
Cheers,
Magnus
Hello,
I've read at least two similar stories in "Troubleshooting Analog
Circuits" by Bob Pease.
One is that it seems that some time ago, National Semiconductor started
shipping LF411s marked as LF351s as an "improvement"... and as Bob says,
most of the customers probably were happy with that (I think that most
of them probably never noticed the change), but the gain on the trim
circuit was reversed (in the LF351, if you turn the trim pot in one way,
Vos increases - in the LF411, turning it in the same way makes Vos to
decrease), and this probably would have made some people not so happy
with the improvement :)
Another story is about 2N3771 transistor, initially a single-diffussed
part, but later an epitaxial base part - with a lot more gain bandwidth
product. But since the 'new' 2N3771 meets and exceeds original 2N3771
specs, the same part number was used - but the part is quite different
(and published specs continues being the JEDEC ones). So you can imagine
that in some applications it would be quite a lot of difference if you
breadboard with the older, and during the manufacturing phase you
(probably without knowing it) switch to the new part.
I suffered some time ago a change in a small DSP from Freescale. It is a
3.3V part with 5V tolerant I/O, and I assumed (not sure if this was in
the datasheet or not) that reset pin was also 5V tolerant. Prototypes
worked, and production worked, but after 2-3 years of production, and
from some DSP production date code, we experienced a problem with the
part reset - the part did no longer liked the 5V level at the reset pin.
I asked Freescale about any change, but never got any response.
I periodically receive PCNs (product change notifications) from EBV
Elektronik, which is a quite big european semiconductor distributor,
whenever some product that I'm purchasing or have purchased in the past
suffers some manufacturing change (manufacturing moved to other plant,
process change, case materials change, etc...), like for example this
one:
http://www.ebv.com/fileadmin/templates/scripts/pcn/data/200907002f__1248209503.pdf
But I suppose that not all manufacturers are so kind to let the
customers know in advance these kind of things :)
Best regards,
Javier
El 14/08/2010 14:30, Bob Camp escribió:
Hi
Simply a few stories I thought I would share.
Simulate design. Use manufacturer's published models. Build design. Note differences. Call manufacturer. Answer - switched die three years ago, Ft is now " much better " ( now 3x old parts ).
Odd they never mentioned that to people who work for the same company.
Simulate design, Build design, verify design, ship it for a few years. Odd things start to happen. Look at some parts. Package looks different. Ask around..... Line got moved to other side of big ocean. Process got " tweaked" beta is now 4x what it was.
Again all inside the same company. Both cases were excused by industry standard specs that had no upper limit.
We had whole departments devoted to tracking this sort of stuff. It still happened on a regular basis. 30 years later the specs on the devices and their published models are still the " old version " ones.
Bob
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.
--
Javier Herrero EMAIL: jherrero@hvsistemas.com
HV Sistemas S.L. PHONE: +34 949 336 806
Los Charcones, 17 FAX: +34 949 336 792
19170 El Casar - Guadalajara - Spain WEB: http://www.hvsistemas.com
On 08/14/2010 06:39 PM, Javier Herrero wrote:
Hello,
I've read at least two similar stories in "Troubleshooting Analog
Circuits" by Bob Pease.
One is that it seems that some time ago, National Semiconductor started
shipping LF411s marked as LF351s as an "improvement"... and as Bob says,
most of the customers probably were happy with that (I think that most
of them probably never noticed the change), but the gain on the trim
circuit was reversed (in the LF351, if you turn the trim pot in one way,
Vos increases - in the LF411, turning it in the same way makes Vos to
decrease), and this probably would have made some people not so happy
with the improvement :)
Another story is about 2N3771 transistor, initially a single-diffussed
part, but later an epitaxial base part - with a lot more gain bandwidth
product. But since the 'new' 2N3771 meets and exceeds original 2N3771
specs, the same part number was used - but the part is quite different
(and published specs continues being the JEDEC ones). So you can imagine
that in some applications it would be quite a lot of difference if you
breadboard with the older, and during the manufacturing phase you
(probably without knowing it) switch to the new part.
One side-effect is that you run into possibilities of oscillation. This
have happend and was the cause of a GPS outage in a US Harbour a few
years back. What was a wise design became an enemy due to a subtle
change in part. Don't recall if the part was replaced by an "equivalent"
or same part-number, but the new version didn't work as expected during
all conditions...
Cheers,
Magnus
One side-effect is that you run into possibilities of oscillation.
This have happend and was the cause of a GPS outage in a US Harbour a
few years back. What was a wise design became an enemy due to a subtle
change in part. Don't recall if the part was replaced by an
"equivalent" or same part-number, but the new version didn't work as
expected during all conditions...
Yes, one of the most probable problems. I've read the story about the
GPS outage in that harbour, as I remember it was caused by a failure in
one TV antenna amplifier on a boat - but I don't remember if it was an
isolated case due to a failure (or to a bad repair replacing the part
with a similar one), or was a more 'endemic' problem :)
Regards,
Javier
--
Javier Herrero EMAIL: jherrero@hvsistemas.com
HV Sistemas S.L. PHONE: +34 949 336 806
Los Charcones, 17 FAX: +34 949 336 792
19170 El Casar - Guadalajara - Spain WEB: http://www.hvsistemas.com
I think this is pretty common with transistors.
A company is making a particular part, then a new, "better" part comes
along that exceeds the spec of an existing part. So they start putting the
new die into the old package as well as making the new part. Fewer dice to
make likely means cheaper and less hassels. And they don't tell anyone.
Win-win, except the new part has a much higher Ft for example and formerly
stable circuits now oscillate badly. Tilt!
-John
=============
Hello,
[snip]
Another story is about 2N3771 transistor, initially a single-diffussed
part, but later an epitaxial base part - with a lot more gain bandwidth
product. But since the 'new' 2N3771 meets and exceeds original 2N3771
specs, the same part number was used - but the part is quite different
(and published specs continues being the JEDEC ones). So you can imagine
that in some applications it would be quite a lot of difference if you
breadboard with the older, and during the manufacturing phase you
(probably without knowing it) switch to the new part.
[snip]
Best regards,
Javier
I think it was a one-off failure:
http://www.gpsworld.com/gnss-system/signal-processing/the-hunt-rfi-776
-John
===========
One side-effect is that you run into possibilities of oscillation.
This have happend and was the cause of a GPS outage in a US Harbour a
few years back. What was a wise design became an enemy due to a subtle
change in part. Don't recall if the part was replaced by an
"equivalent" or same part-number, but the new version didn't work as
expected during all conditions...
Yes, one of the most probable problems. I've read the story about the
GPS outage in that harbour, as I remember it was caused by a failure in
one TV antenna amplifier on a boat - but I don't remember if it was an
isolated case due to a failure (or to a bad repair replacing the part
with a similar one), or was a more 'endemic' problem :)
Regards,
Javier
--
Javier Herrero EMAIL: jherrero@hvsistemas.com
HV Sistemas S.L. PHONE: +34 949 336 806
Los Charcones, 17 FAX: +34 949 336 792
19170 El Casar - Guadalajara - Spain WEB: http://www.hvsistemas.com
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to
https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.