fv.log(msgdmp)

Chemistry: Detecting the Presence of Copper Ions

Since I'm currently in the process of finding out how to best remove copper from used up sodium persulfate I needed a way to detect the presence (or absence) of copper.

Reclaiming Copper from Sodium Persulfate #4: 6B Pencil Lead Anode

As previously noted I've decided to give softer sketch pencil leads a try.

It looks like the electrodes are more efficient and a bit more stable. I've also noticed that the solution seemed to have taken on a lighter blue color and the copper cathode seems to have gotten thicker. In principle it seems to work! However, the thicker anode too started to disintegrate after a few hours at 2 to 2.1 V.

Reclaiming Copper from Sodium Persulfate #3: Bigger Graphite Anode

I tried scaling up the graphite anode by connecting several sketch pencil leads together.

I've done this using single strands of stranded copper wire to weave and knot the individual pencil leads together. Once tightly woven into place I've finalized everything by adding generous amounts of solder.

With regards to being able to use higher voltages this was no help (of course) and the pencil leads still disintegrated.

What I've learned:

Fellow hacker reloc pointed out that the softer a pencil lead, the more graphite it contains. I've picked 2H leads for my tests so far and will switch to 6B leads (the softest I could find at short notice).

What is going into solution depends on the voltage applied (potential) and which materials of lower potential are still there (if available, the lowest energy bond breaks first, even with high voltages).

Reclaiming Copper from Sodium Persulfate #2: Stainless Steel Anode

I got my hands on some pieces of stainless steel to replace the graphite anode.

Reader's Digest version: the mess got bigger. Green is the new blue.

What I've learned:

Stainless steel doesn't seem to work well, the solution quickly turns green.

Reclaiming Copper from Sodium Persulfate #1: The Naive Approach

At shackspace we're somehow in the business of collecting sizable amounts of used up etching solution. We're using Sodium Persulfate which on its own is also used as a bleach, detergent or disinfectant. However, with plenty of copper ions in the used up solution it's a very potent poison and you should not by any means pour it down your drain, ever.

A quick search on how to get rid of the stuff yielded Marc Schaffer's page (German) who explains his electrolysis setup using a 12 V 900 mA Photocell, a copper cathode and an anode made from either stainless steel, coal or platinum.

Platinum was no option because of its high price. I didn't have any stainless steel that I could verify to be stainless steel at hand, so I opted for coal. The closest thing to coal I could think of was graphite. And since I'm a lazy person, I simply got some sketch pencil leads. All the core, none of the casing :)

The result of the first experiment can be seen in the pictures above:

The pencil lead disintegrated

A huge mess

What I've learned:

12 V won't work. Anything much higher than 2 V will cause the lead to disintegrate

Electrolysis should start at around 1.8 V, that's when bubbles (Oxygen) start to form at the anode

Graphite particles can be filtered out quite easily using a piece of paper towel

Teardown: Water Flow-Sensor

Recently the shackspace lasercutter started to act up a bit. Turns out that I've probably put a bit too much copper sulfate into the cooling water and it started to react with the aluminium connector pieces which then corroded and closed up on themselves.

However, after having just cleaned them, the lasercutter complained again that there was no flow. So this time around I suspected the flow sensor itself. Turns out I was wrong (it was the just recently cleaned connectors, again) but I took apart and cleaned the flow sensor anyway.

The sensor itself is genius in its simple design. It has a single moving part: a small flap that blocks flow in one direction but lets it pass with close to zero force in the other. The flap position is then detected, I'm going out on a limb here, most likely using a hall sensor or something similar.

What I also found is a piece of horrible engineering. One side had a connector screwed in, secured with Teflon tape and a small O-ring. The O-ring was pressed to the casing by the face of the screw-in connector tube. So as soon as you start fastening the connector, the O-ring will inadvertently end up being squeezed into the tube and bunch up there, restricting flow and doing actually a worse job with water proving than if it simply wasn't there at all.

Project Fronius Repair #5: No More Smoke!

After I posted a call for help to the shackspace mailinglist, asking for someone who owns or knows someone who owns a similar welding machine to take a peek at, chris suggested to just give Fronius a call. He even went one step further and actually called Fronius Germany!

From Fronius Germany he got a digital copy of the manual, troubleshooting guide and a high-level wiring diagram. While the latter did help in identifying the two stray wires, it didn't help much in figuring out where they should be connected to.

Fronius Germany was also not able to give out any more details circuit diagrams or repair tips, since they only handle replacements.

This didn't stop chris. He called up Fronius Austria and ended up talking to a very kind technician who offered to identify blown parts if we send him a photo (the resistor I couldn't ID turned out to be 33 Ohm 5 W) and in the end sent a photo himself showing how to connect the two stray wires!

I already had a hunch that one wire had to be connected in an H-Bridge fashion since it was part of a pair, both connected to a transformer. The Fronius technician verified that hunch.

Cautiously turning on the machine for the first time after repairing bits and pieces here and there I was half expecting it to blow up again. Nothing of the sort happened! The machine booted up nicely and I could toy around with the settings on the front panel.

So far I've not had any of the torches or the water cooling system connected. I'll do that once I've got someone with me who knows how the machine is supposed to operate to verify all is in order :)

shackspace Infrastructure: OR4 Room Rearrangement

shackspace recently got a full-body scanner donated and we needed some space to set it up. For this we needed more space in room OR4.

shackspace Infrastructure: Spray-Etch Tank Heater Fix

The last time bronsen and I wanted to use the spray-etching machine the heating element broke in a spectacular fashion (smoke, blown fuses, everything).

I've ordered a new heater (thanks Timm for pitching in) and now only hat to somehow mount it in the tank. This wasn't straight forward since the old mount broke as well.

shackspace Infrastructure: Laser Exhaust Fix

The shackspace laser cutter had a bit of a suction problem. The low pressure side of the big air pump connecting the pump to the cutter was soft and unstable and fell in on itself when the pump was running.

I've replaced the tube with a big flexible clothing dryer tube and got an angle and reduction piece to connect the lower diameter high pressure side to the exhaust going out of the building.

With help of dritter I then mounted the pump itself on the underside of the laser cutter table to clean up the mess under the table a bit.

Project INFRA #5: Light Barrier Breakout

A few weeks ago I decided it's time to build a small breakout board for the EE-SW1070. I also decided to try out OSH-Park to order the boards.

Automatic Drip Irrigation - Off-Grid-Version

After I've set up the first version a few days ago, I've now arrived at stage two.

Here's the full shopping list to date:

Drip irrigation system (Gardena 1265-20)

12V 35Ah car battery (Energizer ELB1330)

Power inverter (Filmer 36206)

Trickle charger (Black & Decker BDV090)

Timer (Goobay)

The trickle charger, power inverter and battery are all hooked up in parallel. If the balcony door is open I can connect the charger to mains power and leave it sitting there for as long as I want, it's designed for exactly that use case.

The timer triggers the drip irrigation pump a few times during the day (five times for now). Using this setup the battery should last for up to seven hours of continuous pumping (ignoring loss and consumption of the inverter and timer). Considering the pump is powered for five minutes of which one minute is actual pumping of water and that five times per day, the battery will outlast the water supply by at least two orders of magnitude ;)

The first measurements are in!

The amount of water used for a single irrigation run is around 1.3 to 1.4 liters. Pretty much exactly what it said in the manual.

Except for the Wall'o'Strawberries all this water is lost as in consumed.

The system uses three distribution adapters with various throughput (15, 30, 60 milliliters/run). Of those, the one with the highest throughput (60 ml/run) is used for the strawberries.

The strawberry wall is actually designed to be a water circulation system since most of the water you put in goes right through (once the soil is damp enough). So I would expect most of the water to drip through after a few minutes.

Going by what the manual says, the total amount of water used for just the strawberries is 8*60 ml which adds up to 480 ml. 35 minutes after an irrigation run, the water collected from the strawberry wall ends up to be approximately 400 ml. Since it didn't stop to drip entirely and I didn't measure the actual in-flow, I consider the numbers to add up well enough.

Work in Progress: uc-basics Workshop - Larson Scanner #2

I've spent last Monday together with dojoe working on finalizing and testing the teaching samples' code and breadboard setup for the microcontroller basics workshop.

Project Rigol D1082E Repair #2

It is done! Thanks to a spare encoder I got from dop3j0e my scope is usable again. I've encountered the same issues as already reported on the Dangerous Prototypes forums:

The rotation direction is inverted.

The new encoder is longer by about 3 mm or so.

The replacement encoder has 20 pulses per full rotation compared to just 12 of the old one.

The latter really isn't a bug, it's a feature. I can scroll way quicker using the 20 pulses ;)

With regards to the inverted operation, I'll probably just let it be. So far it's not annoying me enough to disassemble the scope again and apply the fix.

Automatic drip irrigation for the balcony garden

The balcony at kombinat23 sports a nice small urban gardening effort. Today I've upgraded it to be semi autonomous using an automatic drip irrigation system made by Gardena.

Since this system is marketed as a "save your potted plants from dying while you party at the beach for a week" system (read: vacation helper), it turns on only once per day. This isn't enough for an outdoor garden that has a) bigger pots and b) way more evaporation losses. Therefore a timer switch is used to make the Gardena electronics think its a new day five times per day. Works like a charm!

There's still a few problems on the TODO list though.

I'd like to run the watering cycle more than five times per day, I need a bigger reservoir. Actually, big enough so I can run it for at least a weekend, better a whole week unattended.

There's no mains power outlet on the balcony. I have to either use a car battery and inverter, install a power outlet myself, or build something way more awesome to go right through the window without drilling a single hole. For the time being, the timer only runs as long as the balcony door is open (which is most of the time right now).

Project Rigol D1082E Repair #1

My trusty Rigol D1082E DSO recently showed some wear and tear with the timebase encoder knob. I could easily enough turn it, but the time setting didn't seem to change at all or jump around like crazy.

I also don't seem to be the only one with issues here, as I found at least one other guy which also had to replace his encoders.

I'm also running into similar "problems". I can't seem to find a 12-click encoder knob, only 20-clicks per revolution. However, that's very minor, since I tend to scroll a lot with the timebase knob, so this should be more of a feature than a bug ;)