Honeymoon

When I received my Elektron OctaTrack (7/2015), I was excited to play, but cautious about what I was getting into. I knew just from having studied the manual and watching tutorials that this was going to be a deep device with a steep learning curve, and no obvious/default application.

I bought it assuming it would be a “swiss army knife” for all my sampling needs. In practice, I find that it’s bit more like a “erector set,” since most applications require not only some setup (on the device) but some planning (in your head) based on the (hopefully increasing) comfort level you have withs it’s architecture.

So, as much as I expected (and confirmed) that my RYTM would be a “quick and intuitive drum machine” for my right-brain, I knew that the OT would be a puzzle of lucid sampling” for my right brain.

What I did not expect is for the hardware to start failing…

Fool me Once…

After a month of light-but-steady use, the Sample A/B button stopped responding. This button not only controls recording from the first pair of 4 inputs, but it also is the ONLY way to get to the Recorder Setup menu in order to configure how sound would be sampled/recorded to a(ny) track, which is the first step in this device’s unique powers.

Octatrack copy
as you can see, most buttons not only serve a direct, hands-on purpose, but also (with the Function key) access some additional parameters. In this case, I could neither record from A/B, nor configure Setup on ANY recorder track.

The first step in confirming ANY problem is to perform a DIAGNOSTIC test.

  1. Enter the Early Startup Menu:
    1. power unit down
    2. hold the Function key
    3. power unit on
  2. Select Test Mode to test screen, hardware, and
Screen Shot 2017-04-20 at 2.43.00 PM.png

excerpt from OctaTrack user manual, summarizing the Early Startup and Test Mode screens.

As you push the first few trigger buttons, the OT will check the lights, screen, and confirm reception of EVERY button on the Front Panel user interface.

Sure enough, the Rec A/B button is not responsive.

By 8/26, I sent a question about arranging repair or ordering parts from info@elektron.se, and I put in a ticket (#9260) with Elektron Support through their (old) Customer Zone. They quoted $45/half-hour of bench time, and 2-4 weeks turn-around time.

Since I couldn’t claim a warrantee on an eBay purchase, and I realized the fix may be manageably small, I decided to attempt it myself.

 

so… TO THE BENCH !

Surgery #1 (2015 / 07 )

After some serious mediation, I realized I was ready to crack open and attempt (even rudimentary) otherwise-fully-functional machine that cost me over a grand. But, as I say…
“It’s no really yours until you void the warrantee”

…and even though I never had a warrantee on this, I must follow the spirit of the law.

I was impressed to find how easy and intuitive it was to take the OT apart.

 

IMG_20150925_215514511.jpg
The OctaTrack has a LOT of buttons, so the old ice-cube tray helped to sort all the parts (knobs, screws, etc) on the way in, so I could just reverse on the way out.

 

 

After removing the main hex-nuts framing the front-panel, the entire top lifts off easily, revealing the horsepower inside.

IMG_20150728_120425979.jpg
looking in from the front-left, we can see the audio, MIDI and CompactFlash slot, attached to the mother board, with ribbon-cables running up to various sections of the User Interface (UI) board where the switch in question resides.
IMG_20150728_120443735.jpg
looking at the other corner of the motherboard, we see an unoccupied row of pins. Wonder if those were programming the original board, or for some expansion (memory, etc) that was never implemented.

 

Once I got the board off, I probed and tested the solder-pins of the switch in question…

IMG_20150728_121941644.jpg
a quick continuity test proves that… YEP, the switch is ALWAYS open whether pushed or not.

..my initial diagnosis of the construction, button seating, and solder traces suggest that there was just a cold joint on the AB switch, so I just re-heated the pad with an iron, let it cool, making the switch return to normal functionality.

 

IMG_20150925_212708833_HDR.jpg
As we can see below, this go the switch working again, so that, when the the button was pressed, the switch closed (“shorted”) and worked properly.

 

Careful with re-assembly #1:

I got the unit back together, and it seemed… DEAD.

No need to panic, I just went re-disassembled, and found my problem.

 

IMG_20150728_123443240 copy.jpg
…when you re-assemble a unit such as this, make sure that the ribbon cables are re-connected with ALL the pins matching correctly.

As you can see above, I was too hasty to get the board back together, and mis-aligned the 6-pin (2×3) ribbon-connector block. I’ll bet that’s the set of wires that handles power/data/etc up to the screen, if not also the rest of the buttons and UI.

After booting the unit back up, I ran test mode, and confirmed the Rec AB button worked !

IMG_20150925_215911209_HDR.jpg
when in Test Mode, each button will confirm it was pressed by showing a unique numerical address between a combination of the 4 Page LEDs (right side), and one of the 16 Trig LEDs on the bottom row. (perhaps this is their 8-bit binary address). The Rec AB button (should) show Trig 4, and pages 1+3.

 

Having this buttons working allowed me to control Recording of inputs A/B, as well as enter the Recorder Setup menu (for ANY track recording behavior. Since sampling is a key feature of the OctaTrack, this repair was key to e back to working my way up the (STEEP) learning curve of this machine.

 

Surgery #2 (2016 / 8 )

The OT worked fine… for a few months. Then the same button failed.

Opening the unit back up revealed the same behavior, so I figured it may be some failing inside the switch, so I looked into replacing the switch itself.

Thankfully, Elektron’s parts department will not only sell superficial parts (knob caps) but also electrically connected parts (i.e. the switch I probably needed to res-older), so I paid $22 to paypal@elektronmusicmachines.com to receive 5 replacement switches.

The switches (SPST momentary panel mount, which seem to be from the Marquardt Series 1840 product line) came 2 weeks later.

IMG_20170501_104138469.jpg
Add the Elektron switches to my (ever-growing, ever messy) collection of parts drawers.

 

In order to desolder and remove this switch, I had to remove the UI board from the front metal front panel by loosening the Torx connections…

IMG_20150728_122259444.jpg
after (careful) disconnection of the ribbons, a torx bit is needed to remove the UI board form the front panel
IMG_20150728_121524654.jpg
The the brass nuts seem fixed to the front panel, which really helps the “all-one-piece” and “built-like-a-tank” feel of the construction. I now better understand how mounting the board on flexible/shock-absorbant (plastic, rubber) standoffs gives a give “spongey” feel to all buttons and knobs on other gear (ahem, Roland, Korg, etc).

 

IMG_20150728_121532881.jpg
I gave the naked board a quick dusting.

This leaves the main UI board with its buttons and LEDs exposed, so I highly recommend any face-down handling of a circuit board be done on a towel or closed-cell foam pad.

IMG_20150728_124636435.jpg
now to inspect the offending Rec A/B switch: it seems mounted just as solidly as its “Rec C/D” switch. By the way, beautiful labeling, Elektron !

I chose a soft towel, and got to work clipping off the leads to desolder and remove the suspect switch.

IMG_20150925_213353707.jpg

IMG_20150925_214744687_HDR.jpg
and IN comes the brand new switch. Note slot-notch that holds the red button cap.

After putting in the new switch, I closed everything up, and (since I put the stand-offs on correctly this time), it worked right away.

IMG_20150925_215856834_HDR.jpg
another “success light” in Test Mode !

Unfortunately, after this surgery the button failed after only a another month of use.

 

period of reflection and regrouping.

…as one can imagine, to have repeated failure (let alone inconsistent access) to a button that is critical to setting up and controlling the sample-recording action of what many argue is the most powerful sampler on market is… to put it lightly, discouraging.

So, this OT spent the better part of 2016 turned off on a stand while I regrouped and did some more research on fixing soldering. I didn’t want to have potentially damaging “learners’s mistakes” on the circuit board of this device.

 

Also, I bough a copy of the OctaEdit program from the magnificent Rusty O’hara. I was blown away at how easy it made it to explore and sketch within the OT’s functionality, and (among so many other things) helped me work around a broken button to play and learn with the other facets of this machine !

surgery #3 (2017 / 3 )

In the middle of 2016, I started more regularly attending The Hacktory in Philadelphia, and got to do some touch-up solder work with their very nice iron (especially nice compared to the crummy Weller 40W iron I’d been using before).

 

IMG_20170413_205253474.jpg
Dave at The Hacktory  in Philadelphia was kind enough to let me use his top-line solder station, which includes an iron with built-in vacuum, and a heat-rework gun to remelt and re-fuse solder to boards.

 

So, with a a combination of some cool-down time away from the device, and the confidence of having access to better tools and expertise, I went in again.

IMG_20170413_201802884-.jpg

here we go again, this time at the well-equipped and spacious desk of The Hacktory.

By this time, I suspected that the problem was not the switch, and not the solder joint, but the trace leading off of it.

IMG_20170413_203656620.jpg

first we melt and scrape away the old solder joint from the (replacement) switch

 

IMG_20170511_205603913.jpg
testing continuity of the trace that leads from the switches solder pad to whatever chip processes that switches electrical state.

Doing a continuity test with the solder trace removed showed that the connection was intermittent. This suggests that there might be a small gap or fissure in the trace.

Upon closer inspection, I found that there seemed to be a small crack or chip right where the PCB’s trace meets the solder-pad. This would explain it. Since my previous surgeries failed after re-soldering the pad, this suggest that the solder was not making an adequate connection, and would fail after moderate stress or time.

I didn’t manage to get a picture of it (my cell phone cannot pull focus close enough (and I couldn’t get it to show through a magnifying glass), but I figured I’d found the culprit, and decided to make a wire-bridge.

IMG_20170511_210517086_HDR.jpg
Here, I used a single strand of 28 Gage wrapping wire to reach from the solder pad of the switch’s throw to the trace. Steady hands and 4 attempts seemed to stick.

 

IMG_20170511_210640864.jpg
Careful to use nips to cut off the wire slack without ripping off the concave fillet of the bridge.
IMG_20170511_211109099_HDR.jpg
…and we’re OFF.

 

Hopefully, the third time is a charm.

IMG_20170413_204826427.jpg
For the third (and hopefully last) time, ran in test mode, and the button worked.

 

Holding ? (2017 / 5)

It’s been about a month since the last (successful) surgery. Given my free-er schedule this spring, I’ve been able to much more time (hours of usage) on the OT than I had previous lapses.

…and you know what they say about mechanical switches:

"it's not the age, it's the mile-age..."

Keeping my button-mashing fingers crossed that this will hold indefinitely and I’ll be able to build and test proofs-of-concept for some ideas I’ve had while studying through the manual and OctaEdit‘s visual virtualization of the system.

Stay tuned !

If you have any questions or comments, I’d love to hear ’em !

 

 

Advertisements