The Aleph-X

Official Court Jester
Joined 2003
Paid Member
GRollins said:



I'm not sure what the hippe/exclusive part means.
As far as the rest, the fault is mine. I didn't express myself clearly.
I'll get the hang of this writing thing one of these years.
Let me outline my train of thought and see if I can identify where it derailed. (Yes, that's a crude pun.)
Starting with carpenter's comment, my thinking went thusly: He expressed appreciation, presumed in context to be for the Aleph-X circuit. The way I go at circuits is that I pretty much have a fully functioning prototype before I post. Others do all their development out in public with input from a hundred people. I do mine in the Dungeon, pretty much solo. Part of that process is the making of a circuit board. At that moment, I happened to be working on a PCB layout, my eyes got tired, so I looked in on the DIY site to see what was going on and take a break.
(Whether the particular circuit I'm working on ever gets posted here remains to be seen.)
Since I regard the making of PCBs as part of the R&D for a circuit, my post was a long-winded way of saying "you're welcome" to carpenter.
And, of course, you guys have never known me to be long-winded...right?
Then my thoughts went on to other PCBs I've seen here.
Let me address that in a slightly different way and see if I can clarify what I was saying.
Pick out a new member here at DIY. Make sure it's someone who's got the fire, the passion, and the commitment to do what it takes. They start with something that looks kinda silly, like a post saying, "What amp should I build?" or "Can I substitute MOSFET xxx into this circuit?" Posts like that are a dime a dozen. But fast forward a year, and if you're watching the right person, they'll be helping others choose projects, and suggesting what MOSFETs might or might not work. Two years on, they're able to make non-trivial alterations to posted circuits and fix their own problems. In short, they learn, they improve, they grow. Their skills develop. This is good. It's how things should be.
But...
For some reason, I don't see peoples' PCB skills developing. Why, I'm not sure. Am I saying that everyone should be able to do their own boards? Not necessarily. But those who attempt it should attempt to improve their PBC abilities the same way that they put effort into learning the electronics angle. And yet, time after time after time, I see boards that are no better than the last board I saw from the same member. They're not improving. I'd like to see raised standards for circuit boards. Don't just go "Ooooh! Aaaah!" every time someone posts a picture of a board. Look at the confounded thing. No...really...I mean look at it. Does it look like it was made by a drunken monkey or a master at his craft? I see clever circuits here that make me think. I see pictures of metalwork that are models of competence. Then they pop the top and the bubble bursts. Ohmigod! You mean that marvelous chassis was built to contain that? Yuck! It's criminal...or should be.
Nelson has said that you should build a chassis as though it's a temple to contain your circuit.
I say, apply that same ethic to your circuit boards. Do not settle for "Well, it works, so I guess it's okay." Goad yourself. Make each circuit board better than its predecessor. Don't pat yourself on the back just because some other member went "Ooooh! Aaaah!" Their standards might not be high enough. Make yours higher. Do not accept mediocrity. We've got some pretty bright people here and there's no reason that they can't master the art of making printed circuit boards.
Excel, dammit!

Grey


purposely quoting entire post;

now you are clear(er) .

just one addendum (besides silly note that from ".....temple..." reasons I didn't templed my Babelfish yet.......just not enough bloody time on my hands........and I'm spendin' my time on the net only when I'm not good for anything else) :

seems that every brick in temple is ,and must be , equal part of it. nothing more and nothing else . place accent on any word in "equal part of it" ;
just from this reasons sometimes wire from middle of pcb and shortie isn't crime ,but perfectly logical solution .


ok-I know that you know

:worship: :devilr:

ps. even if I like Niven, I usually prefer more Ursula's ways ......was I woman in previous life....or just wife ? :devilr:
 
GRollins said:

For some reason, I don't see peoples' PCB skills developing. Why, I'm not sure. Am I saying that everyone should be able to do their own boards? Not necessarily. But those who attempt it should attempt to improve their PBC abilities the same way that they put effort into learning the electronics angle. And yet, time after time after time, I see boards that are no better than the last board I saw from the same member. They're not improving. I'd like to see raised standards for circuit boards. Don't just go "Ooooh! Aaaah!" every time someone posts a picture of a board. Look at the confounded thing. No...really...I mean look at it. Does it look like it was made by a drunken monkey or a master at his craft? I see clever circuits here that make me think. I see pictures of metalwork that are models of competence. Then they pop the top and the bubble bursts. Ohmigod! You mean that marvelous chassis was built to contain that? Yuck! It's criminal...or should be.
Nelson has said that you should build a chassis as though it's a temple to contain your circuit.
I say, apply that same ethic to your circuit boards. Do not settle for "Well, it works, so I guess it's okay." Goad yourself. Make each circuit board better than its predecessor. Don't pat yourself on the back just because some other member went "Ooooh! Aaaah!" Their standards might not be high enough. Make yours higher. Do not accept mediocrity. We've got some pretty bright people here and there's no reason that they can't master the art of making printed circuit boards.
Excel, dammit!

Grey

Grey,
I'm in no way a pcb expert nor a seasoned designer, and I'm just trying to do things in such a way I won't be ashamed of my work looking at it one or two years later... I often fail at it, but I'm trying to learn from my mistakes.

In fact, PCB design seems to be the poor parent of DIY electronics. And it seems to be an international bug (or feature ?). Though I do no want to extrapolate to other countries, but as I constantly swap between this forum and french ones on the same topics, I see the same tendancy as you point out : Whoever comes with a nice design or idea, with such appealing audio characteristics that you spontaneously say "Hey, I want one !", the question araising almost immediately is "Is a PCB available ? Is there a group buy for it ?". Apparently, designing a PCB and having it etched and drilled seems to be the 13th Hercule's work. Even having a pdf file of the pcb is not sufficent for most of people to get the etched boards, there seem to be something supernatural finding a shop to do the job. (You know, the Pratchett's a-temporal shop wandering through multidimensionnal universes...) I happen to live in a small town in France, and even there, there's a local shop that does a very good job with my printouts... But it seems quite impossible to people living near Paris... But I'm losing focus ;)

In most of people's mind, PCB are only a way to go, just a board to plug parts on. But as J. Carr said somewhere, PCB is also a part of a design, and should be considered with the same care as choosing the right transistor in the right place. Designing a well laid out PCB supposes a minimal understanding of how the circuit works, where are the critical paths, where currents really flow, etc... . And I don't even speak of stray parasitics... So, as you mentionned, the first PCBs will be buggy, but I do not understand why the pcbs do not constantly improve along with the electronics knowledge of the author...

The major challenge is then to meet the electrical, mechanical and thermal requirements of the circuit, without speaking of the servicing aspect, and adding a somewhat artistic part, a personal touch that satisfies your aesthetic sensibility and makes you go further than you'd have gone if you'd only wanted this design to physically work. The devil often lies in details, but do some exorcism.

I spend a lot of time designing a single pcb, constantly modifying parts placement, conceptually rearranging the schematic from different points of view (electrical, thermal, mechanical, etc...), looking for a weighted compromise between appearance and physical requirements. And I stop and go to the real thing only when I deeply feel that nothing can be taken further... But the downside is that the "further" is constantly moving further :D

May be the "lot of time" stuff is the reason of the disaffection for PCB design. It of course takes a long time to tame pcb softwares and all their subtleties, and people often seem afraid of that point : too hard for me... But I found the softwares very useful to throw ideas on the "screen paper", cutting and pasting parts of the pcb to lay them out in another way while keeping the original, just to see if this damn idea is valid, not being ashamed to use the undo feature... While in progress, my screen is full of small portions of the pcb, mirrored, upside down, rotated, etc... Just critical building blocks, awaiting for the gathering... No autorouter needed, it is not aesthetical ;)

Just my two cents, for what they're worth
 
I find that I've redesigned my ZV7-T at least several dozen times. As the circuitry matures, so does the pcb artwork.

What would be exceedingly handy? How about a few "rules and guidelines" for the amateur electronics crafts-person.

So far I've gleaned:

1. The beauty of a ground plane for the signal portion of the board.

2. Logical placement of input and outputs connections; i.e. keep the wires away from components you'll be probing. If the wire is located in the center of a pcb, place it beneath the board. (I have no problem with a wire being in the center of my pcb if it's installed from underneath)

3. Symmetry and beauty in component and trace spacing. Mirror images are very handy in balanced circuits.

4. Proper through hole pin location and diameter. Software helps in this area, but so does looking up the part in question. Knowing which part to use is handy--pin spacing varies widely from one mfg to another, and one style to another.

5. Proper trace width.

To insist that someone "excel" without proper guidance is like punishing an infant for messing their diaper...........


So, let's get a list going.
:)
 
Things I hate on boards? Faston terminals and those small screw terminal blocks.

One of the biggest obstacles I see to doing some of the things that you propose is the max board size of the freeware and $125 versions of Eagle. Unlimited board capacity board CAD software is prohibitively expensive.

An old detective novel explained this point well:

"Is Florian here?"
"Florian died 20 years ago."
"Why does the sign still say Florian's?"
"Signs cost money."

So does PCB CAD software.

Bryan
 
Express PCB software is free, easy to use, and has an easy learning curve.

I'm not happy with small barrier blocks either, but am going to experiment with larger barrier strips mounted on my pcbs.

Seems to me that Express PCB will go as wide as 14" or so...... just checked, their max board size for prototyping is 14" wide by 12" in height. Their production run pcb size may be larger.
 
batee said:
Things I hate on boards? Faston terminals and those small screw terminal blocks.

Bryan


Actually, I rather like the screw block terminals. They are handy for removing boards from the chassis and reworking a board on the bench.

In my final designs I merely solder wires where the terminal blocks were. This seems like one area we could have flexibility since it's an easy choice to use them or not.

-David
 
I've been using larger terminal blocks more often in some of my power projects, too.

Dunno if they're audiophile approved (I can't imagine they are), but I agree the larger terminal blocks with #6 or #8 screws are practical for high-current connection to PCBs.

The more I use MOSFETs, the more I realize the need to include the ability to quickly disconnect wires from boards for servicing purposes. In non-audio applications, I also like the .156 spaced MTA connectors for power and the C-Grid connectors for signal purposes. Does anyone have experience with C-grid connections and (at least medium-fi) audio signals?

Bryan
 
Originally posted by batee One of the biggest obstacles I see to doing some of the things that you propose is the max board size of the freeware and $125 versions of Eagle. Unlimited board capacity board CAD software is prohibitively expensive.

If Windows, try Kicad : http://iut-tice.ujf-grenoble.fr/kicad/

If Linux, try pcb : http://pcb.sourceforge.net/ a part of gEda

Free, almost unlimited board sizes, a perfect fit for the diyer.
 
Hi David,

I was an avid user of the smaller barrier blocks, but discovered that they can loose contact with the FET or a cluster of wires (shove them out of the slot). You get these crazy sounding creaks and pops in the amplifier. Once, I went stark raving mad because of a voltage irregularity that couldn't be resolved no matter how hard I tightened the terminal screws--turned out that one of the screws wouldn't lock down on the component, even though it appeared to be tightened. It required removing and replacing the terminal block--not to mention removing all the hook-up wiring.

At least with terminal barrier strips, you can visually examine the contact. Another thing, since I wire the pcb to a terminal barrier strip anyway, why not save a step and include it on the pcb?

Hey Bryan, who mfgs C-grid connectors? I'm curious, and always looking to advance my pcbs.

John:)
 
C-grid SL are made by Molex/Waldom products. They're essentially a standard .1" grid jumper header with a plastic surround that allows the connector to lock together. They seem to work fine in microcontroller signaling applications. P144 of my Digikey catalog.

I have not tried these in any audio application other than PC CDROM audio.

Bryan
 
Thanks Cheff.

I'll check out KiCad while I'm waiting for the next Digikey order to arrive. I've been looking for a way to do some FET boards that're shaped like Nelson's Firstwatt boards. I'd love to have something that'll hold 6-12 FETs in parallel instead of the usual 3 FET boards that I'm seeing now. That would also let me spread the heat out over the whole heatsink instead of concentrating it into the smallest area possible in the name of "Mini". I don't want Mini. I'm an American - I want More.

I'll try to work in KiCad right after I finish replacing the surrounds on my Infinitys and building the next "temple" for my Aleph 30s. :smash:

Bryan
 
Originally posted by batee
One of the biggest obstacles I see to doing some of the things that you propose is the max board size of the freeware and $125 versions of Eagle. Unlimited board capacity board CAD software is prohibitively expensive.

Take a look at DipTrace. I recently switched to it from Eagle and instead of a limit on board size they limit by pin count, which I find a lot more flexible. For $125 there is a Non-Profit Licence that allows 1000 pins and 45 layers. I also find it a more intuitive software to use than Eagle. I guess the bottom line is that it was good enought that I bought it.

Cheers, Terry
 
My mother went through a phase wherein she fancied herself to be a whiz at furniture refinishing. She'd tear up stuff, then put this gawd-awful looking greenish-white finish on it. She told me it was called "distressed" furniture and solemnly assured me that it was all the rage.
My response was: Why beat up perfectly good furniture?
For years we ate off of this ugly-as-hell kitchen table. I was embarrassed to bring friends over. I confess that I never did understand the attraction.
Add to my list above:
I'm not entirely sure what the reason is--perhaps more of this one inch signal path nonsense--but some people seem to feel that they should have the mounting holes for resistors, caps, etc. right at the end of the body. This is foolish. By bending the leads right at the end of the body, you put undue stress on the junction where the lead attaches to the resistive element/foil/whatever. You're also depriving yourself of a bit of heatsinking, both when the part is soldered in and during operation.
I recommend leaving .1" (sometimes more) at each end. For example--I use the Vishay/Dale RN60 resistors (1/4W MILSPEC...1/2W in the real world).
*****
Yes, the RN55 are adequate for many applications, but then I'd have to stock two different wattages in every value. Sorry, ain't going to happen. The RN60s are more versatile and don't cost much, if any, more. Also note that there are those who feel that larger resistors sound better, due to the fact that they don't reach as high a temperature in marginal applications.
*****
While doing PCB artwork, I allot .6" per RN60 resistor. The body is .4", to which I add .1" on each end. I use a pair of needle nose pliers as a former when I bend the leads. All the resistors seat nicely and everything is tidy and uniform.
Something like ten or twenty percent of the "My New XXXXXX Doesn't Work" threads turn out to be dead resistors. The poster usually ends up saying something to the effect of, "Golly, I just can't understand why that resistor is bad. Maybe it was defective from the factory."
Or maybe they bent the leads too close to the body and broke the connection between the lead and the resistive element...

Grey

P.S.: Oh, and two more things:
--When you bend the leads, make sure the resistance nomenclature ends up on top. This isn't a problem if you use resistors with color bands, but if they've got writing, you'll want to be able to read it once it's in the board. I can't count the number of boards I've seen with the value of the part underneath the body. Dumb...very dumb. You have to unsolder the part if you want to check the value.
--Mount all parts so that the parts are readable from the same direction. Aside from the fact that most people can't read upside-down, it looks nicer.