The Cool parts thread

[duffy]

Don't bother with that damn postage-stamp regulator, the spec sheet is a lie. I got some in and they are just common buck mode switchers and don't boost.

 

[Speakerguy]

Duffy, I think all non-isolated point of load regulators are understood to be bucks unless otherwise specified.

 

[duffy]

Yes, unless they are boost regulators. Or have a transformer. Or use switched-caps. Or something else. Pffft.

 

[Speakerguy]

You know what POL means, right?

 

[mneary]

I guess the letters "LDO" in the part number should be a clue. But then why is there no minimum Vin-Vout?

 

[duffy]

Yeah, they should have a Vin-Vout min spec, and they don't. And if the "LDO" prefix is to indicate low-dropout, they should state that. I got excited, should have waited until I checked the damn thing - sorry.

You know what POL means, right?

Yes - almost nothing, like the term "working voltage". Any regulator can be considered point of load, except maybe the one at the power station. Here's a "point of load" regulator that does boost:
https://www.electro-tech-online.com/custompdfs/2009/08/DesignSolution51.pdf
Now why don't you post an interesting part in this potentially-awesome "Cool Parts" thread?


Here's a seemingly mundane one that is more awesome than first appears:
AMPHENOL SPECTRA-STRIP|111-2803-020|RIBBON CABLE, BONDED, 20CORE, | Newark.com


Why is this a Cool Part?

1. It is .035" (.89mm) pitch, so there are more wires per inch than standard .05" ribbon.
2. You can peel individual wires apart with your fingers. Not all ribbon cable will do that, a lot of them will pull the bare wire through the insulation.
3. It is PVC, so it doesn't have the awful "melt back" that a lot of other ribbon cable does. Others have this disgusting "plastic backing" material that make it a pain to use.
4. It's color-coded, so you can tell which wire is going where.

What does all this add up to? The perfect hook-up wire to use in projects. They stay together in neat, compact wire runs. You peel off a set of two or five or whatever to route to LED's or switches or something. Solder them with no trouble, and keep track of them with the colors.

 

[Speakerguy]

Yes - almost nothing

Actually, quite the opposite. It means (or is coming to mean with an industry standards movement) a very particular thing about it's job in a distributed power architecture (DPA) context, usually with respect to telecom and data infrastructure systems.

Screaming Circuits: Point Of Load Power Supply

Now though, you've got 5 volt, 3 volt, 1.8 volt, 1.3 volt, 1.2 volt, 12 volt, 48 volt... depending on your application. You just can't have that many different power busses. Well, you could but I won't let you. In comes point of load. You run a higher voltage bus around and at each major section of the circuit with a different voltage requirement, you drop in a little switching regulator or DC to DC converter.

(Emphasis mine).

There's even an industry standards movement for the form factors and pinouts (POLA, the Point of Load Alliance) led by Artesyn, Texas Instruments and some other company I can't remember. The point is in a DPA, AC is turned into an intermediate voltage and downregulated at point of use. Yes you will find the odd part that is buck/boost or boost or flyback or forward because there are rare applications where you need to do something other than generate 1.8 or 2.5V for your FPGA or DSP or whatever. But >90% of the market is for buck converters (probably even higher), so if you ever see something advertised as a POL (and especially as a non-isolated POL) you should assume it is strictly a buck device.

 

[mneary]

Even if I could talk you into adding a 1.1V 50A bus for my CPU card, no way it's gonna be 2% tolerance by the time it gets here. Hence POL is the only way. I'll take 12V no sweat.

 

[duffy]

Real engineers have been putting regulators on boards for a long time now, and we don't need an industry standard to tell us when to do it. For every engineer actually designing circuits, there seem to be a dozen other people calling themselves engineers who create nothing but jargon - DPA, POLA, "Web 2.0", "Fuzzy Logic", and a dump-truck-load of other useless terms.

Instead of repeatedly trying to derail this thread, why not post a link to a "Cool Part"? Here, let me demonstrate (again):

Take this seemingly ordinary $17 non-contact IR sensor...
https://www.electro-tech-online.com/custompdfs/2009/08/0004763_3901090247P006.pdf
...and arrange them in a cubic array...
**broken link removed**
...and now you have a horizon sensor. Cool.

 

[3v0]

Duffy,

Fuzzy logic is useful stuff but the term "Fuzzy logic" is jargon.


3v0

 

[tcmtech]

Cool parts must be relative to each persons applications. I sort of fell asleep a few posts back.

For me one of my top 10 most useful parts is the Teco PLR units. I love them for their versatility and robustness.

Yea its industrial uC tech but hey its what I learned to use and its basically a plug and play device with all the necessary input and output signal processing and protection circuitry built right in.

For me its cheap, very rugged, very versatile, and compatible with industry standard control signal systems.

 

[Speakerguy]

Well since people seem to be in need of boost converters, here is one of my favorites:

8MHz, 3 Amp, Quad-Phase Monolithic Synchronous Boost Converter:

https://www.electro-tech-online.com/custompdfs/2009/08/3425f.pdf

 

[duffy]

Duffy,
Fuzzy logic is useful stuff but the term "Fuzzy logic" is jargon.
3v0

It's just a marketing term used to sell vacuum cleaners. When you study the definition, the whole thing falls apart. Ever since the first programmer wrote code for the first computer, they implemented something to decide that if A > B, then C unless D< E, etc. This is what's at the heart of so-called "fuzzy logic", and it's about as meaningful as inventing a term to describe enabling interrupts when loading a driver, or a thousand other things programmers routinely do.

Here's a cool magnetic shielding material -
https://www.goldmine-elec-products.com/prodinfo.asp?number=G16600A
The permeability is twice as high as Mu-metal, a hundred times higher than electrical steel. This can save you pulling your hair out over those rare but troublesome magnetic interference problems.

 

[3v0]

Your reaction to Fuzzy Logic is typical of how most non AI people feel about concepts that come out of Artificial Intelligence. Once AI concepts are adopted by the main stream CS the concepts are trivialized and seen as easy.

Until the invention (or discovery) of fuzzy logic there were a few processes that could not be automated. If I recall correctly the heating required to make cement was one of them.

Fuzzy logic is not simply A>B thinking.

Wikipedia

. In contrast with binary sets having binary logic, also known as crisp logic, ... Fuzzy logic is a form of multi-valued logic derived from fuzzy set theory to deal with reasoning that is approximate rather than precise.

3v0

 

[duffy]

I've studied it. I've written AI routines for robots. I didn't say it was A>B, I said it was "A>B, then C unless D< E" (kindly glance up at my post) and that's all that's at the heart of it.

Until the invention (or discovery) of fuzzy logic there were a few processes that could not be automated. If I recall correctly the heating required to make cement was one of them.

Absolute nonsense. This is hype from marketing guys, sorry to hear you fell for it.


Here's an amazing material that's more thermally conductive than copper, and a definite "cool part" -
Digi-Key - P11440-ND (Panasonic - ECG - EYG-S060910)

 

[3v0]

It is clear that marketing types are use "Fuzzy Logic" and "AI" as buzz words to sell most anything.

It is true that at the base level it still comes down to standard logic and crunching bits. Everything has to.

But the advent of fuzzy logic provided solutions to problems that had none.

This is hype from marketing guys, sorry to hear you fell for it.

Do you have any other derogatory comment you would like to make ?

3v0

 

[duffy]

But the advent of fuzzy logic provided solutions to problems that had none.

Show me an actual example. Just one. Show me where fuzzy provides solutions that aren't the same as nested conditional statements.

Or are you trying to tell me you seriously believe nobody put one "if... then" clause inside another one before "fuzzy logic" showed them the way?

 

[3v0]

Look under applications here.

https://www.austinlinks.com/Fuzzy/tutorial.htm

Concrete rotary kiln's seem to be the poster boy for Fuzzy Logic. They were unable to develop a control system for the kiln's prior to fuzzy logic. A concrete example. Often used pun intended.

Fuzzy logic was not my area and most knowledge I have of it is nearly 20 years old. I do not claim to use it.

Once you understand the idea behind fuzzy logic you can certainly write code that does what it does. If you do not (never heard of fuzzy sets or fuzzy logic) then the chances are slim.

The hard problems once solved are seen as easy and ho hum in the rear view mirror.

3v0

 

[duffy]

For a "concrete" example, there's nothing in that paragraph at the bottom of the page that gives any specific reasons why this problem was intractable until fuzzy came along and saved the day.

 

[Speakerguy]

It's a tutorial. It lists in the footnotes the paper that was published on the topic. I'd probably start there.