Jump to content

Mongrol's MidiboxSEQv4 Build


mongrol

Recommended Posts

Inspired by Hawkeye I thought I'd document and photograph my MidiSEQv4 build in the hope that it helps others.

I'm planning a fairly vanilla SEQv4 with Wilba's CS and 2 MIDI IO boards. I'll add the Raystar OLED displays instead of LCD. I intend to build a Ponoko acyrlic enclosure and will publish the SVG for once done.. I'll edit this post as I go.

#0 - BOM and Excitement
#1 - The Core - Storage for baby voles
#2 - The Displays - Trust datasheets
#3 - The MIDIIO's - Chill to the bood
#4 - The CS - Time is neither created or destroyed. Merely borrowed and given back.
#5, #6 - The Enclosure - Fat Tabs
#7 - The CS - Alps Encoder Edition
#8 - LED's - Aaah
#9 - Fitting and Finished - Dun dundun dundun

 

NOTE: This is not a build guide. It is a documented summary of my attempt to build a SEQv4. Some parts from the BOM below may, and indeed already have, changed as the journey continues.

 

Full BOM Summary

I'm in Australia where the AUD makes this quite expensive. I'll include costs in AUD and the original currency.

|   Qty | Component                       | Source |
|-------+---------------------------------+--------|
|    16 | Knobs                           | Mouser |
|    58 | Switches                        | Mouser |
|    58 | Switch cap                      | Mouser |
|     8 | Ceramic Cap 100nF=.1uF “104”    | Mouser |
|    12 | 10k 6 Pin SIL resistor network: | Mouser |
|    58 | 1N4148 (small signal diode):    | Mouser |
|    16 | Bicolor LED's                   | Mouser |
|     1 | USB Socket                      | Mouser |
|-------+---------------------------------+--------|
| Total | $120.45 AUD                     |        |

|   Qty | Component                      | Source   |
|-------+--------------------------------+----------|
|    12 | 220R resistors                 | Reichelt |
|     6 | 74HC165:                       | reichelt |
|     2 | 74HC595:                       | reichelt |
|     1 | Red 5mm LED                    | reichelt |
|    28 | Green 2mm LED                  | reichelt |
|     8 | 16-pin DIP IC socket:          | reichelt |
|     4 | 16 pin IDC connectors          | reichelt |
|     1 | 16 pin flat ribbon cable       | reichelt |
|     1 | 16 pin dual inline pin headers | reichelt |
|     2 | 10-pin ribbon wire             | reichelt |
|     6 | 10-pin crimpable IDC socket    | reichelt |
|     1 | 10 pin "box-header" angled     | reichelt |
|    17 | ALPS STEC12E encoders          | reichelt |
|     2 | Raystar OLED's                 | reichelt |
|-------+--------------------------------+----------|
| Total | $98 AUD                        | 69.07 EU |

From SmashTV's Midiboxshop

  • Ordered: 20/1/2016 - Processed: 26/1/2016 - Shipped:12/2/2016 - Received: 22/2/2016
SEQ CS board			$35.00
CORE STM32F4 Kit Board		$30.00
2xMIDI I/O module Kit Board	$46.00
Shipping and handling		$27.29
USD Total			$138.29
AUD Total			$209.88 (ouch!)

From RS Electronics (free delivery in Oz)

STM32F4Discovery Kit	$23.45 AUD

From TM Electronics

2xRaystar OLED 40x2 displays REC004002AYPP5N00000
EU Total	90.04EU
AUD Total	$144.86 AUD (double ouch!!!)

From Conrad

Posted to mumsie in UK who'll forward it on to Oz. Conrad wanted $80AUD to post a $4 part.

1x ALPS-863002-Rotary-Knob-With-Finger-Grooves (Datawheel)
UKP: 2.69

 

Build Status

It's time to turn this...

16020011.thumb.jpg.5784db7160aaf0cce0dcf

Into this...

mbseq_case30.thumb.jpg.6597f485024843a24

Let's do it!

Edited by mongrol
Update
Link to comment
Share on other sites

4 hours ago, EsotericLabs said:

 There must be some acrylic case designs lying around somewhere in the forum.

 

The most standard laser cut template I believe is this one from smokestack. It has bugs though and hasn't been updated. I'll be making a new one with interlocking slots round the back and sides with the only front edges using the "teeth" method. It'll be optimised for size as much as possible without modifying the boards beyond vanilla. 

 

Link to comment
Share on other sites

  • 2 weeks later...

When using the ALPS 12mm encoders, make sure you cross the center and the right pin (as seen from above). It takes about one minute per encoder and is the correct way of attaching them. You can live with DETENTED1 probably, without this mod, but it is best to do it correctly (for more info, see last posting in the phototutorial thread) :)

Many greets,

Peter

 

 

Link to comment
Share on other sites

On 26. Januar 2016 at 0:32 AM, mongrol said:

The most standard laser cut template I believe is this one from smokestack. It has bugs though and hasn't been updated. I'll be making a new one with interlocking slots round the back and sides with the only front edges using the "teeth" method. It'll be optimised for size as much as possible without modifying the boards beyond vanilla. 

 

That sounds just like what i am trying to come up with. If you'd share your files once you're done, that'd be great :)

Link to comment
Share on other sites

  • 3 weeks later...

#1 - The Core

  • Music: Aphex Twin - Drukqs
  • Time taken: 90 minutes

16020015.thumb.jpg.aeadfd88cfdcb22d67877

I'm roughly following Hawkeye's tutorial. Which directs me to the instructions are on the core page on the ucapps site. A very straightforward build.

The core is a storage unit for tiny plastic boxes. You can keep anything in it but the spikes will get in the way. Small mammals may complain.

Actually it's just a breakout for the STFM4 Discovery board.

P1080773.thumb.JPG.b5f89a341bd894954f5ea

Bootloader

Next I installed the bootloader. Reading the instructions isn't totally clear but I figured out I needed something to connect to the STLinkv2 debug program on the board. Being on linux I opted for QStlink2 on Github. This cloned and compiled no problem on my Debian Jessie box once I installed some lib-devs. You have to install some udev rules to run it under a normal user. Once done it detected the Core and I could upload the bootloader. It worked flawlessly and much less hassle than trying to get a Virtualbox up and running with Windows on it.

MIOS32

I hit a bug in MIOS32 trying to upload the SEQ project.hex file. This thread had a workaround.

SEQv4 application uploaded and appears ok. Hmmm, next up, MIDIIO or displays? DISPLAYS!!! I have to see things ON!

 

 

Edited by mongrol
grammar fixes
Link to comment
Share on other sites

#2 - The Display

Well that was both easy and hard. Making IDC cables is very easy. Simple cut the cable with scissors and, making sure you align the striped bit with the little triangle on the plug, poke it through the jaws till it's flush with the other side. Then squeeze it together. Make sure you do it evenly (a vice is good) as if you don't you may snap the side clips. So use a big set of pliers or something you can get even force. Then fold the cable over the top nice and tight and push on top locking thingy.

I made cables up and looked at the pinout on the OLED. Pin 1 was nearest to me so I put the cable on with the stripe at pin one. Plugged it into the Core and nothing. Hmm, stared and poked, replugged, stared at solder joints. Tested with meter and got funny readings. Then found a post about 3.3v and set lcd_type 0x02. I started playing around with that and managed to get one display up. No luck on the other. After hours of staring, poking, reflowing joints, reading schematics and testing I discovered the 5v OLED's should really be 5v after all. Who...would....have....thought?

Tada!

P1080774.thumb.JPG.3857b0025490586b3d707 

 

Link to comment
Share on other sites

#3 - The MIDIIO's

  • Music: Sparklehorse - Vixiedixiesubmarinetransmissionplot, Mantra - Shukyamuni Buddha (took a lead from Phatline and zoned)
  • Time Taken: 90 mins

P1080775.thumb.JPG.2e6585d6d826896af80d4

So easy I didn't even use a Hawkeye visual aid. The ucapps midiio page has all the placement information in it. One thing I will point out is the order is wrong. Do the IC sockets and headers before the big caps (height order y'know) then finally the MIDI sockets.

P1080776.thumb.JPG.caed7d1642f8e31e0217a

Next, the big one. BRING THE THUNDER!

Link to comment
Share on other sites

#3 - The Control Surface

Music: Tortoise, Atoms for Peace, Bert Jansch

The CS seemed a bit daunting at first until I found out about "top soldering". The CS is double sided with pads on both sides meaning you can solder on either side and get good joints. I took advantage of this wherever I could. Amazing speed up.

First we put the board up on standoffs high enough to let the wires of the components dangle without touching underneath. Following Hawkeyes Tut I dropped the diodes, resistors and caps in one go then soldered them all up, turned it over and snipped all the tails off. Top soldering is a magical time machine! The joints look nice and clean and it also makes for a lovely neat underside.

P1080777.thumb.JPG.95bfa8f52ee087754511e

Unfortunately that's where it stops. The IC sockets can't be top soldered so I used the tape method instead. If you don't want to use tape I sometimes use a bit of cardboard to cover them and hold them in place while I turn the board over. Same result.

P1080778.thumb.JPG.0079ecd7747362966b859

After that we're onto the switches. They all click in nicely and stay in place when you turn the board over. 232 joints later and we're nearly done.

But disaster striketh the mong! Alas I'm stuck! My boat has been cast upon rocks. I've been foiled by incompetence (probably my own). The box connector needed to talk to the Core is a 10 pin SMD part. I've been sent the wrong thing!!!

Well that answers the question I had when I first went through my delivery. :( I now await the opening of a local and dusty electronics store before I can move on.

P1080780.thumb.JPG.ad7b67782d950e6f51898

Link to comment
Share on other sites

Alas I'm struck down. I can't for the life of my work out the orientation and wiring for the IDC ribbon to the Core.

I was trying to do an angled IDC box header underneath but no matter how I make the cable or which ever way I put the box header in I always end up with the wrong pins.

I assuming S0 goes to S1, Rc goes to Rc etc? But I always end up with either S0 to S0 and Rc to Rc or S0 to S1 but with Rc going to VS instead. 

It's driving me mad.

Link to comment
Share on other sites

16 minutes ago, Shuriken said:

It's a straight connection. So just make sure that you have Pin 1 (red wire) coming from the Core on J8 to CS J1 go to Vs.

Yes! If you are missing that one connector and just can't wait to "fire her up", you can always solder in the required wires directly on the CS board (and add an IDC plug on the other end of the ribbon wire for the core, so you can still disconnect things). You can measure with a multimeter in connectivity/beeper mode from the core and solder in the wires on the appropriate pins of the cs... This should be done in a few minutes... :-)

Many greets!

Peter

Link to comment
Share on other sites

Should be mentioned that the ribbon connection is made on the underside of the SEQ PCB. Also that you don't HAVE to use a box header, regular 100mil pin headers are just fine. When looking at the IDC connector from a bird's eye view, the notch goes at the bottom and pin 1 is the lower left. Just match the left sides together as Shuriken says and you're golden.

(BTW it's SI and SO, for serial input and output. SO from the Core drives the DOUT chain, while SI to the Core carries the DIN data stream.)

Link to comment
Share on other sites

#4 - Enclosure Design

With all my switches on and 1 pot (and waiting on parts) I might as well get on with the enclosure.

Lots of people like to use the tooth method for acrylic boxes made either manually or with Boxmaker.

Boxmaker's pattern looks far too aggressive for me, I bet it grips on like a ugly steel bear.

boxmaker.jpg.b19a0a0e0f5bea1e626b06832ff

Others use the same method but reduce the tooth count to make it nicer. This works to a certain degree but reduces the stability and solidness of the casing. I had one of these around my LXR for a time. It was ok as I don't gig but it wouldn't last long.

lxr.jpg.8da827853e88450463bee96e10895ec7

All of these teeth designs use "key" screws in the sides to keep them together. Without them the side's would be pushed apart by the pressure from the top and bottom panels which are screwed together.

On the Axoloti forum I found this design. The teeth go inside slots in the top and bottom panels which are screwed together. If the teeth and slots are accurate enough (kerf accurate) then you don't need key screws in the sides. It also has nice rounded corners. I like this a lot and will attempt it with my SEQ.

10606466_162871650747027_160560703694351

I'm also going to attempt a couple of other improvements

  1. The only screws that show in the top panel will be the ones around the side holding the top and bottom panels together. The CS will be standoff'd to the bottom panel instead of the top to avoid any middle of the panel screws showing. To stop panel flex the CS will have 10mm standoffs on top holding up the panel. You just won't see them but they will be there making everything sturdy.
  2. I'm going to reduce the size by using the CS mounting holes (and 1 Display hole) for the case edge standoffs that hold the case together. The CS will be 1-2mm away from the case edges with no standalone standoffs around the side.

The display holes are a bit tricky to bottom mount but we'll see how we go.

I hate the thickness of the core. Would love to decimate that audio socket badly but that would mean making a non-standard case that others can't use.

Link to comment
Share on other sites

So the enclosure plan is almost finished but I'm just too scared to send it in. This is a $100 AUD print and I've found Inkscape to be a total disaster when it comes to accurate SVG design. Draw a box, set it's size. Go make a cup of tea and when you come back, the stupid things added on some random small number and moved things around tiny amounts. Horrid! I'll need to find some other Linux app to tighten things up before I send it in.

Link to comment
Share on other sites

If the added numbers are very small, don't worry too much, these might be floating point inaccuracies. It is important, that the drawing is precise to 0.05 - 0.1mm or so, the rest is quite irrelevant when lasering acrylics :-)

Many greets,

Peter

Link to comment
Share on other sites

I've found out that Inkscape adds the stroke width onto the shape width afterwards. This explains why when I look back at it there's an extra 0.100 on it. I'm lot more confident now I know what's going on. Turns out a good tip is to set the stroke to the size of the kerf, then you get an actual visual of the kerf cut in your model and can use it to adjust slots for fit.

Link to comment
Share on other sites

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.

Guest
Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

Loading...
×
×
  • Create New...