OT: Muntzing
jb-electronics
day! His name liveth evermore..."
Reading Mike's post, though, I got the impression that "Muntzing"
circuits is something that is to be avoided. Reading the respective
Wikipedia article (which is very compact, though), I did not get the
impression of "Muntzing" to be something bad.
Any insights? ;-)
My own opinion is that Muntzing should only be done by people who really
know what they are talking about... What do you think?
Regards,
Jens
Nick
definitely a pejorative term. and strongly so.
Basically, what he did was to remove bits of a good system until it
stopped working, and then just put the last bit back. It was never
good engineering, it was about being cheap and nasty. He went bust.
No professional engineer worth his salt would ever dream of
"Muntzing". Not ever.
Nick
Quixotic Nixotic
> My own opinion is that Muntzing should only be done by people who
> really know what they are talking about... What do you think?
Well I don't often know what I am talking about in the electronics
field, but I feel that if make something for myself at home for my
own amusement and have Muntzed something, such as use a PIC output to
drive a FET and it works just fine for me, then if it fries itself,
or stops working, then I can think about fixing it and have nobody
else to blame. I would not like to Muntz something that I was
selling, or was going to give to a friend even, although I have on
occasion unwittingly done so, through pure ignorance.
John S
Eric1180
jb-electronics
> LoL Muntzing. Is this about my post about me figuring out that i can
> use a pic to drive a FET?
don't worry, it was mere curiosity on my part.
> I hope i didn't offend anybody with my self discovery, i am just
> learning and i was ecstatic when i made a boost power supply with out
> even looking at a single schematic.
I sometimes nearly make engineers cry when they look at my first design
ideas ;-)
Regards,
Jens
Terry S
COS stands for "Cost of Sales". The COS teams look for ways to take
cost out of products. It's a legitimate and common practice. The more
cost you take out of a product, the higher your margin. On a daily
basis the COS team folks come to me asking how they can reduce the
cost of an assembly, find a lower cost component, reduce the harness
cost, substitute a cheaper connector, reduce assembly time, evaluate
another vendor of a particular part, reduce test time, etc.
Sometimes these efforts save the company millions. Other times they
don't save money at all or even end up costing the company money when
things go wrong. Overall, the 2nd set of eyes and hands generally
makes the product better and more reliable. Not always at first, but
inevitably the effort pays off and the company is better for it. When
I contribute to such an effort, I indirectly share in the savings,
thru increased margins, profit sharing, stock price, etc.
Muntzing isn't haphazard today like it was back then. It's a
concerted, purposeful effort with goals and payoffs that work. It
makes business sense. Just don't call it Muntzing. No one will know
what you're talking about.
Terry
Quixotic Nixotic
> There is no question. "Muntzing", "to Muntz" or whatever is most
> definitely a pejorative term. and strongly so.
>
> Basically, what he did was to remove bits of a good system until it
> stopped working, and then just put the last bit back. It was never
> good engineering, it was about being cheap and nasty. He went bust.
I remember when I was in my mid-teens being seduced by the tacky kits
that Clive Sinclair used to put in the ads in electronics mags. I
fell for his amplifier kits that had cheap carbon preset pots with
cheap plastic spindles and knobs, presumably to save money on real
pots and knobs. I never did get those kits to work, a simple preamp
and two output channels.
Sir Clive went on to make huge amounts of money out of electronics
and he remains today a very rich man, living as he now does on the
south side of London's Trafalgar Square in a reasonable degree of
comfort and splendour. If ever I get around to taking up his invite
to lunch, I may broach the subject of Muntzing with him. He's cranky
these days, so you never know if he is going to fall asleep, throw
you out on your ear or get taken to one of his favourite Soho strip
joints.
John S
Jeff Thomas
and "Muntzing".
I've not used the term "muntzing" in many years. The last time I did
was when a marketing fleeb sat in on an engineering meeting with the
belief our EE/ME staff were taking too long on a complex project that
was running over budget. I stated that we needed his muntzing skills
to whittle down the cost and still have it work...;) He understood it
as a warm compliment.
As Nick described, it was a hack method to eliminate components to the
minimum and still maintain some level of function.
It may have cut fabrication costs by a small degree, but was also
synonymous with being cheap and poorly engineered.
Nothing a self respecting engineer would want to hang his hat on.
And like Terry says;
A Professional Engineer may spend as much time designing the product
to minimize cost, as they do in designing it's function.
Regards, Jeff
jb-electronics
Maybe this should be added to the Wikipedia article?
This Muntzing reminds me of a friend of mine who - when he had to
program something, which he was really bad at - copied a fully
functional routine from some resource and then started removing code
passage after code passage until the program either crashed or worked as
he wanted ;-)
Regards,
Jens
Frank Bemelman
of a rather elaborate opamp design done by someone who didn't understand
gain-bandwidth to begin with. I got rid of ~70% of the components, got rid
of the tedious calibration procedure, and replaced very expensive opamps
with
low cost jelly beans LM358. And it produced a much cleaner signal.
That was not "Muntzing" of cource, but I can imagine that (lousy enough)
designs
can benefit from it. That is something I could dream about - like winning a
Muntzing
Award, a 24K gold plated cutter on a nice piece of polished wood or
something ;-)
Frank
----- Original Message -----
From: "Nick" <ni...@desmith.net>
To: "neonixie-l" <neoni...@googlegroups.com>
Sent: Monday, August 08, 2011 9:56 PM
Subject: [neonixie-l] Re: OT: Muntzing
[snip]
Basically, what he did was to remove bits of a good system until it
stopped working, and then just put the last bit back. It was never
good engineering, it was about being cheap and nasty. He went bust.
No professional engineer worth his salt would ever dream of
"Muntzing". Not ever.
Nick
--
You received this message because you are subscribed to the Google Groups
"neonixie-l" group.
To post to this group, send an email to neoni...@googlegroups.com.
To unsubscribe from this group, send email to
neonixie-l+...@googlegroups.com.
For more options, visit this group at
http://groups.google.com/group/neonixie-l?hl=en-GB.
Frank Bemelman
A (cheap) stool with 4 legs will always wobble.
A stool with 3 legs is stable.
Frank
Frank Bemelman
and try with two legs, you'll have a full grasp of the Muntzing principle
;-)
Terry Kennedy
> Sometimes these efforts save the company millions. Other times they
> don't save money at all or even end up costing the company money when
> things go wrong. Overall, the 2nd set of eyes and hands generally
> makes the product better and more reliable. Not always at first, but
> inevitably the effort pays off and the company is better for it. When
> I contribute to such an effort, I indirectly share in the savings,
> thru increased margins, profit sharing, stock price, etc.
problem"
There is often a tendency for good engineers who worked for a large
company to go "into the weeds" when they form their own company to
make similar products. In one example I'm very familiar with,
engineers who worked for a company named I left and formed company E
to make disk drives. They contracted with manufacturer T (a very well-
respected name) to produce stepper motor driver hybrid modules. There
was a near-90% incoming inspection failure rate on these modules, and
the rest failed after they were installed on pre-production drives and
shipped. The company I was a partner in, P, was E's second-largest
customer, behind company D.
I decided I needed to sort things out by flying to E's California
location to see what was going on. Their engineers explained that the
drives were failing due to the very-complex hybrid modules. I said
something like "why don't you just use a quad Darlington IC instead?"
After an extended period of "ummm... errrr...", they agreed that this
might be a good idea, and their chief designer and I sat down to draw
out the re-design that night. He had to fight to get approval to build
a test batch of the redesigned boards, since others in the company
were deeply invested in the original hybrid version. I still have one
of the first 10 drives with the re-designed board (it probably still
works, but there is no use for 40MB ST506-interface drives these
days), as well as a bunch of the non-working hybrid-based logic
boards. By the time E ramped up production of the drive, D had
cancelled their purchase contract and E turned into a disk platter
supplier and was subsequently purchased by another drive manufacturer.
[Company names shortened to the initial letter to protect the innocent
(and the guilty)]
threeneurons
same word. For me, Muntzing is a bad thing. Basically engineering by
hit-n-miss method, and not knowing how the damn thing works. Its okay
to simplify a design, if you know what you're doing. "That", I don't
consider Muntzing.
As an example, take the anode driver circuit again. It usually has two
transistors: (1) the PNP on the high side (MPSA92), and a (2) NPN on
the low side (MPSA42). The low side NPN, is driven from the uC. The
resistor in "series" from the uC IO bit thru the resistor, to the
NPN's base (on the other end) is needed. But sometimes I see a
resistor from base to emitter on the NPN, just like the on the high-
side PNP. This one is not needed, and can be removed. Why ? The output
port of a uC (at least AVRs and PICs, but not 8051s) are push-pull.
That is they have two transistors on the output, that either "source"
the output to a logic-1 (hi, Vdd, ... +5V), or "sink" the output to a
logic-0 (low, Vss, ... 0V). Any charge on the BE junction has a low
impedance (close to zero ohms plus that series resistor) either high
or low (depending which way output state is). You don't have this push-
pull path, on the high side, hence the need for the resistor there. In
short, high side=yes on BE resistor, low side=no on similar resistor.
Another example: pull-up resistors on CMOS. Not needed. Also, CMOS
parts draw the least power, if the high and low signals, at their
inputs, are either all the way up at Vdd, or all the way down at Vss.
If you know how the circuit works, then you know what's needed or not.
Sometimes we take a calculated risk. Using a simpler circuit, that
does the job, but might be either, less durable (failure=smoke), or
less reliable (failure=behaves funny, sometimes). That is, you use one
circuit, if you're designing a weapons system, that might stay idle
for 20-years, then needs to function perfectly, on a moments notice.
Or, a consumer item, that will be trashed in 5-years, so there's no
need to design it to last forever. In that case, use lower rated parts
(less margin), and/or leave out protection circuits. This, too, is not
Muntzing, if we design for a certain market.
Oh, on FET driving directly from you're uC. Its a definite maybe.
There are FETs out there, that are called "logic-level" power FETs.
Most FETs won't realize their lowest ON resistance, until they see
close to 10V at the gate (relative to source). A logic level FET is
designed to be at (or very close) to that low Ron, with only 5V at the
gate. So shop around for such a FET.
Nick mentioned that capacitance from drain to gate, which may (or may
not) transfer a HV pulse to the gate. The gate (to source) voltage is
rated at 20V max for almost all FETs. Usually when the HV pulse on the
drain is generated, the gate is yanked down to zero V by a pull-down
(or "sinking") transistor on your FET driver circuit, so that coupled
pulse (to the gate) current is also sunk down to GND (0 Volts). But,
what if you turned switched that driver back to high, while that pulse
is climbing. Or even worse, what if that output goes Hi-z (tristate,
because of an unexpected reset). Then that small capacitance WILL
apply a HV pulse to the uC output bit. From experience, this usually
means a very dead uC. A reset on most uCs, causes all the IO bits to
reset as inputs (Hi-Z).
This is why, its usually recommended to stick a FET driver in between
the uC and the FET. There are many dedicated chips out there that do
this function, or it can be made from discrete components. But, if its
a quick-n-dirty toy, you only intend to use on Halloween, then its
okay to drive it directly, with no intermediate driver. Beer will
probably short it out anyways. Damn waste of beer !
Get me a Dos Equis Dark, or a Guinness Stout. After I return from the
mens room, I'll give you a Bud Light. Its yellow and has a head
anyways.
jb-electronics
> This one is not needed, and can be removed. Why ? The output
> port of a uC (at least AVRs and PICs, but not 8051s) are push-pull.
You need to be careful here: There is a quite prominent exception to the
push/pull outputs in PIC controllers: output RA4 (bit 5 of PORTA), for
example, is an open drain output with many models, so you will always
need a pullup there to create a logic 1.
With AVR controllers there isn't a problem because you can
(software-)activate internal pullups.
Regards,
Jens
Jonathan Peakall
was to avoid patent infringment. I have no real idea if that is true or not,
just something I heard. I do remember "Don't be a dunce, buy a Muntz!".