Any interest in a standard PCB layout for the frontpanel components?

[sonicwarrior]

In the synthesizerforum.de there was the idea of creating a standard PCB layout

for a standard frontpanel for the MIDIbox SEQ.

Pros:

- No hassle with stripe boards/bread boards

- Standard frontpanel means cheaper case when doing

a groupbuy or something like that

Contras:

- No real individual layout possible even when the PCBs are splitted or something like that

- The layout cannot deal with all possible components. It should support

components that are easily available like the DTL-2 switches.

I know some people who are interested in the MIDIbox SEQ but are put off

by the complexity of having to deal with the frontpanel stuff.

Let me know what you think.

I know that a big group in the DIY-world argument that they want an individual

self-made design but what interest me is how many here would find a good standard

design ok.

And there is someone in Slovenia who can eventually make a standard case design

with silk screening and a professional look and even cheaper than an engraved panel:

http://www.prodigy-pro.com/forum/viewtopic.php?t=16021

(Downside: At least 10 people have to order such a case)

I'm currently waiting for 2 cases for my Gssl-Project to test the quality of Purusha's cases.

[Wilba]

It sounds like a good idea...

I've just finished my new MB-SID front panel PCB and I can give some advice.

It's possible to make a PCB footprint that can take the ALPS 16mm encoders or the smaller Bourns or ALPS ones... since the pins aren't overlapping you can add tracks to support the differences in pinouts as well.

Some people would be happy with TK's placement of the encoders (they aren't perfectly aligned with the "steps" on the 2x40 LCD) whereas some people (like me) may want to use smaller knobs and place the encoders exactly aligned with the "steps". One of the advantages of getting a PCB made is you aren't restricted to 0.1" (100 mils) alignment so you can align the encoders exactly to the steps. Either way you go on this, you still have to choose a "standard" 2x40 LCD size (they're not all the same!) so that the front panel cutouts are the right size, and the mount holes on the PCB are right.

Getting the right knobs is the biggest hassle, I've found... TK's design uses knobs from www.albs.de which are relatively cheap (I think 0,60 EUR?), and the front panel holes are bigger than the knob diameter so they poke through the panel. This I assume makes the construction easier, you can mount encoders, buttons and LEDs all on the one board. Basically if people are going to use a standard PCB and panel, they will also need the right knobs to suit. Same with buttons and button caps.

One final point: After seeing Rigo's ingenious use of JB Weld here, I plan to use this technique to mount my PCB to the panel and provide lots of support everywhere... I plan to use threaded metal spacers (looks like a 10mm tall M3 hex nut) and glue these to the panel and then mount the PCB to these. So maybe you don't need to drill mount holes in the front panel - people can use this trick to make mount points where they want them, and this might help accomodate slight differences in LCD dimensions, etc.

[moebius]

Hi.

I'm not a fan of vector or veroboards either, but actually MBSEQ is pretty straight forward wiring effort. The real fun only begins with matrixed frontpanel stuff.

I think the biggest problem with "a standard PCB layout" is the wasted PCB estate, because I'm guessing that these would have to be manufactured by a 3rd party?! It's costly to produce large PCBs mostly populated by bare copper ;)

(On the other hand, I can't see why this couldn't be done with one-sided boards and with large enough demand produced by Mikes Midibox Shop, http://www.mikes-elektronikseite.de/midiseite.htm)

Moebius

[sonicwarrior]

I like hot-melt adhesive more than "liquid" adhesive.

But I'm new to that and don't know how the quality compares.

The standard for encoders could be a problem as the ALPS and Bourns

are cheap and last only 10.000 whatever.

For the displays I planned to support the Reichelt "LCD 402A" (green) and "LCD 402A BL"

(blue). But they don't have exactly the same size. Unfortunately. :(

[DrBunsen]

How about modular boards for sections of the layout?  ie an 8x or 16x encoder and button board, and other boards for common but difficult to wire elements of the front panel, like the SID LED matrix.  With ribbon cable headers.  This would allow some custom layout but still take a lot of the tedium out.

[moebius]

I like hot-melt adhesive more than "liquid" adhesive.

But I'm new to that and don't know how the quality compares.

If You're referring to the hot-melt glue usually used with so-called hot glue guns - forget about it. It's great for securing wire bundles inside the chassis, but it won't work holding the panel stand-offs. "Crazy glue" really doesn't work either as it doesn't stick that well in bare metal to metal connections.

Any slowly drying liquid 2 component epoxy glue should work, these make extremely firm contact as long as left alone to dry after the positioning.

Moebius

[Twin-X]

I would like to add a comment about the cases from purusha.

one word "Awsome" today my cases arrived. This ultra proffesional stuff!.

[sonicwarrior]

How about modular boards for sections of the layout?

Yes, that was one idea but I wanted to wait until some people show interest in this PCBs.

I don't know what Mike needs as format. I'm using Abacom Sprint Layout 4 to design my PCBs.

This is a cheap and simple PCB designer software.

And I personally like Smash TV's PCB's much more (Epoxy, silk-screen, etc.).

[smashtv]

Hi guys

Talking about pro/legend/silkscreen PCBs here are several things to consider:

A single large PCB might cost less in the long run (one setup vs. several), but the various controls need different spacing from the panel (LEDs + encs on the same PCB can be done, but LCD + encs will leave ya with short shafts)

Separate PCBs make several setup charges, but allows for a little flexibility when its all mounted under the panel. 

Some things like the LED matrix are hard to layout on a single-sided PCB without a lot of wasted real estate and jumpers, but things like encoder + button boards could be single sided to save a little on overall cost.

That said if you guys will figure out an estimate of basic specs (how many boards needed per unit, how many square inches per board) I'll see what pricing I can get from the various fab houses I use.

Not volunteering to do the layout, but I'll help get them made if the price for pro fab is decent.  ;)

Best

Smash

[c0nsumer]

Not volunteering to do the layout, but I'll help get them made if the price for pro fab is decent.  ;)

Also, remember that any version of Eagle (assuming that is the CAD software used) which can do boards over 100mm x 160mm will cost a few hundred dollars. If I had the software I'd happily lay it out, but I'm a bit constrained size-wise.

That said, I personally really like the idea of a single PCB with a nice, standard facilities for everything on it. With some cut-outs or proper engineering with pushbuttons the displays could easily be integrated as well.

This is a project I'd love to do and would happily lay out, but I can't afford the software for it at this time. :\

-Steve

[tatapoum]

I've started to work in this way. If you're interested in my eagle file, I can post them.

4 encoders support. :

SID Oscillators, LFO and Filter switches and LEDs :

ludo

[sonicwarrior]

Also, remember that any version of Eagle (assuming that is the CAD software used) which can do boards over 100mm x 160mm will cost a few hundred dollars.

As I said:

I'm using Abacom Sprint Layout 4.

And as I'm the one who asked for an interest I will be the one who lays the stuff out.

The discussion is going into strange ways, I just wanted to know who is interested

to see if it's worth the effort. ???

[c0nsumer]

I'm using Abacom Sprint Layout 4.

And as I'm the one who asked for an interest I will be the one who lays the stuff out.

The discussion is going into strange ways, I just wanted to know who is interested

to see if it's worth the effort. ???

Ah, sorry. I'm speaking more figuratively if someone else were doing it. Personally, I think it'd be a great idea, provided you also have a pre-laid front panel (.FPD?) and BOM to make a standard item.

Personally, what I think would make it even better is if you were to fit it into a standard, off-the-shelf enclosure. If everything was mounted on the one PCB (which is how I envision it) you could do that, and put it all in a ~1" thick case. Maybe something like the <a href="http://www.pactecenclosures.com/Plastic-Enclosures/KE-Series.html">Pac-Tec KE-17 or KE-20</a>?

-Steve

[sonicwarrior]

You mean the whole MIDIbox SEQ on one PCB?

Personally, I don't like such big PCB's and they

increase shipping costs and decrease stability (break faster).

I like it the modular way. If you use the "sandwich" construction

you can use thin case designs anyway.

[c0nsumer]

You mean the whole MIDIbox SEQ on one PCB?

Yes. Isn't that what you were talking about? It'd be easy to put it all on one panel, and would cost the same to manufacture as the front panel PCB alone. There's plenty of room around the pots and LEDs and buttons to fit everything else.

Personally, I don't like such big PCB's and they

increase shipping costs and decrease stability (break faster).

I disagree strongly, and so do many commercial manufacturers. If the board is properly supported there's no reason it should break. And since shipping costs are fairly minimal when figured into the whole of the project, this is a rather moot point.

I like it the modular way. If you use the "sandwich" construction

you can use thin case designs anyway.

I understand if you do, but I'm just saying what I think would work best. Sandwhich-type construction requires either multiple boards with aligning connectors or cables.

You are already planning on laying out a massive PCB for almost all the IO bits. Why not also fit the core and other electronics in there? Orient all the connectors logically along one edge, lay everything out in a documented standard, and poof. An excellent, professional, single board MIDIbox SEQ.

If I had a copy of Eagle which would do boards that size I'd do one myself. I think it'd be a fun project. :)

-Steve

[sonicwarrior]

Yes. Isn't that what you were talking about?

No, I talked about PCB's for the frontpanel stuff, see the title of this thread.

For my taste a single PCB would be overpopulated with components.

The kind of thing where it's hard to trace errors.

And for my SEQ I want the AOUT and 8 gate outputs option.

I doubt this whole stuff can be reasonably fit into one whole large PCB

without creating a monster like the introspective 9090 PCB.

[c0nsumer]

And for my SEQ I want the AOUT and 8 gate outputs option.

I doubt this whole stuff can be reasonably fit into one whole large PCB

without creating a monster like the introspective 9090 PCB.

You're talking about the front of a 3U 19" rack. Figuring a 16" wide PCB, that's 72 square inches. That's an ENORMOUS PCB. If you were to use the KE-20 I linked to earlier you're talking about potentially even more room. Being mostly digital you'll get nowhere near the density of the 9090 boards. (Yes, I do have a set at home. They are amazingly dense, but still not difficult to work on.)

Even adding the AOUT with 8 gates wouldn't add much. It could easily be done with plenty of space left over.

Really, all you are talking about is a CORE, BANKSTICKs (these can be sprinkled anywhere the IIC lines and 5VDC run), 3x DINX4, simple 5V power supply (hell, use a simple switching one to cut down on heat), 1x DOUTX4, and LCD lines / brightness controls. I'd probably also throw a 74HC00 on there (basically, part of an LTC module) to give real, hardware MIDI LEDs. Although this isn't that important.

If I were doing this I would probably use standardized LCDs which will mount directly to (or stand just slightly in front or behind of) the PCB, similar to this. This would, of course, depend on the height of the rotary encoders and buttons and such, but could easily be varied as needed. Going off TK's design there would be plenty of room for the ICs and power around the controls -- even more so if you did it SMT. (This is harder, especially for the PIC, because it'd require either the builder to have a programmer or the PIC-supplier to have an interesting rig for programming them.)

-Steve

[sonicwarrior]

For the AOUT you need a +/- 12 V power supply.

For the 8 gate outpouts an additional single DOUT.

In the modular fashion people can choose to use the AOUT_LC

instead of the AOUT.

For the SEQ v3 the IIC MIDI module is needed further because

f**king Microchip is not able to fix the EUSART bug.

SMD is no option for me and would in my eyes scare off people.

And I would like to keep the PCB's as simple as possible

(no doube-sided, thru-plated PCB) so that everyone with the possibility

can easily etch it.

[c0nsumer]

For the AOUT you need a +/- 12 V power supply.

For the 8 gate outpouts an additional single DOUT.

Ah, yeah. Still not much more to add.

I had to add that on my SID implementation. Wasn't difficult to do. Just another pic + crystal + caps.

That I definitely agree with. I'm thinking of laying out a mostly-SMD MIDIbox FM (to make it tiny), but I don't expect too many other people to want to make such a thing. :D

I... don't think too many people would be etching them, as a blank board of that size is pretty expensive, and most people would have difficulty just printing a proper transparency for exposing it. Also, I don't know where one would find blank boards that wide... Then again, I might not be looking in the right places.

I wish you luck in doing that single sided. I think it'll be rather difficult.

-Steve

[sonicwarrior]

I... don't think too many people would be etching them, as a blank board of that size is pretty expensive, and most people would have difficulty just printing a proper transparency for exposing it. Also, I don't know where one would find blank boards that wide... Then again, I might not be looking in the right places.

Thanks, you're giving me an argument for modular PCB's. :P

As I said previously I want to do it the modular way.

Another advantage is that you can use some of the PCBs also for other similar

projects.

But it will take some time. I don't expect first showable results this year

because I have many other projects.

[c0nsumer]

Thanks, you're giving me an argument for modular PCB's. :P

As I said previously I want to do it the modular way.

Another advantage is that you can use some of the PCBs also for other similar

projects.

So you're talking about a modular front panel, not a single PCB for it? Sorry, I thought you were talking about a single front panel PCB.

What parts could you make reusable? Looking at the layout I just don't see what parts really carry over to other MIDIbox projects...

-Steve

[sonicwarrior]

I'm currently waiting for 2 cases for my Gssl-Project to test the quality of Purusha's cases.

Got them today.

Fantastic. Love them.

[sonicwarrior]

Man, this thread is really old and I had no time since then to work on this project.  :-[

Has anyone started a similar project in the meantime?

So you're talking about a modular front panel, not a single PCB for it? Sorry, I thought you were talking about a single front panel PCB.

As I read through the thread I was a bit scared how offensive my posts back here were.

And how not related and kind of mad.  ;)

I think I was not really sure if I should go the modular or the single PCB way.

What parts could you make reusable? Looking at the layout I just don't see what parts really carry over to other MIDIbox projects...

Yes, that makes sense. In the meantime I've got a bit more experience and would also go for one PCB for the frontpanel

and eventually even try to fit the whole SEQ on it (don't know, if that would work without SMD 'though) to reduce

the flat cable mess or layout a backend PCB with all the needed stuff on it.

[Echopraxia]

If this were to happen I would say modular pieces for each section of any midibox might make the most sense. You would also have to decied which type of tact switches will be used and encoders ect... I think though that having these would be great for people that are not good with vero board. Also the switches would need to be compatible with alot of differnet caps. I have never done pcb design so don't ask me :P Nice idea but I think everbody is probably still going to do there own thing. Unless its a wilba beast.....

[bugfight]

I've been thinking about modular control surface PCBs for a bit.  It would be useful, for example, to have 4 encoders with DIN shift register on board, with 10 pin dil headers for in and out like smashes DIN_X4 brd.  There would no doubt be different spacing needs but a small set of pcbs would be sufficient i think.  Could keep a list of designs in the wiki maybe, and as someone needs a different spacing, they could do the layout and add it to the list.  There are other obvious combinations, 8 buttons with leds for example 2x4 or 1x8 ...

Keeping the serial registers on the PCB makes interconnections far easier, and makes the PCBs truly modular since all connections are the same.  You can even mix DIN and DOUT by passing the SO or SI pin through when it isn't used.

In eagle, the schematics could be shared with different pcb layouts for different control footprints/spacing

Wise has done some of this already: http://www.midibox.org/forum/index.php?topic=6550.0

My first attempt at this was a 4x16 duo matrix with sparkfun buttons: http://www.midibox.org/forum/index.php?topic=10004.0