Saturday, 14 March 2015

Since returning to work full time I have really quite neglected this blog, maybe its not really the format that I need, because sometimes I just want to share a link, but today I  found something that I just had to post. Is this the world's largest Raspberry Pi Cluster?

Sunday, 18 May 2014

Oh dear what I have I done!

I haven't posted much since going back to work, bu I wanted to show a few people my latest purchase, its a very VERY poorly Juno 106. It was £225 on e-bay and this is what the text with it said
"
I bought this a little while back on Ebay as a restoration project, but I have since got hold of another one, so have decided to sell this one on. When I bought this there was almost no information on it's condition, other than that it had be stored in a garage for a few years and had started to show signs of rust on the casing, it wasn't powering up either to be able to test. 

I won this and had it sent direct to Synth Repair Services in Stoke-On-Trent, where I send all of my kit for service or repair. The guys at SRS fixed the power and opened and tested the machine for operation. They ran a 'soak test' and determined that only 3 of the 6 voice chips work. They also stated that there is no guarantee on how those 3 would last and recommended replacing the complete set of VCF/VCA & Waveform Convertor Chips.

Here is a run down of what was recommended and required to bring this back to full working order:
1no. New Set of VCF/VCA and Waveform Convertor Chips New Bender Wheel
1no. New Battery
2no. New Slider Knobs
1no. New IEC Mains Inlet (power is currently working, or was when last switched on for test - but it is strongly recommended to replace this - as with the other items mentioned above)
1no. Replacement Front Panel - Original (optional ) "

So I have got my work cut out then.The work order will be  
  1. Service the power supply before ANYTHING else. If its dodgy it could wreck the synth.
  2. Turn it on and make a list of what does and doesn't work
  3. Take LOADS of photos of the inside.
  4. Strip it down to parts
  5. Remove contaminants from the PCBs
  6. Get the case shot blasted, and powder coated white.
  7. .... I dunno lets see whats wrong with it first :-)















Sunday, 1 December 2013

The Soundlab Diaries - Part 3 - Build disaster.

Green Omeg pots on the SoundLab panel

Problem 1

Two months into the build I noticed that occasionally VCO1 would "warble" when I touched the coarse tuning pot(entiometer). At first I thought it was a just a dodgy component that had been damaged as I worked, but closer inspection showed that several of them did it.
I put a post on the Electro-Music forum to see if others had the same problem, and they did, it seems that the little green Omeg pots I used are well known for going wrong after a few months.
These pots are sealed so there is no way they can be cleaned or fixed, and as it seemed they were all going to fail eventually I didn't have much choice, if I wanted a reliable instrument they would all have to be changed.
Alpha Pot - Note the metal shaft cannot be pushed into the body.

Problem 2

There are no other pots on the market the same size as the ones I have used, so I can't just buy a set of replacements and wire them in.
The panel is already pretty tight for space so anything larger would not fit.
The only affordable, smaller,  ones are clones of the tiny American "Alpha" pot. People on the forum reported good results with them but they require a 7mm fixing hole, much smaller than the 11mm drilled in the panel for the green Omeg ones, so they won't fit in easily either.

Possible Solutions

If time was not an issue I would simply make a new panel with smaller holes, but that would mean a complete strip down and re-wire, which would set me back months, so I have broken one of my personal rules and  ordered a new set of pots without knowing exactly how I am going to fit them.This normally ends up being completely the wrong thing to do and I regret it, but on this occasion I can't see any other way through.
There is nothing on the market to make these fit so I am going to have to make something. I have had similar problems before and in the past have tried these ideas.
  1. Reduce the size of the holes by filling them in with plastic metal and re-drilling, but in this case that would mean a complete strip down and would take ages. 
  2. Stick on a thin skin of aluminum, or plastic laminate, on the front panel and drill new smaller holes in that, but again that would mean a complete strip down.
  3. Rings made from rubber tubing slipped over post shafts to make them a snug fit, but I don't have any the right diameter.
I have just found some nitril o-rings that are 7mm internal diameter and 2mm thick which would give it an external diameter of 11mm... I think this might work and have just ordered some.
Finger crossed.


Update.. arghh.. I have just found some plastic pipe that fits perfectly...

The Soundlab Diaries - Part 2 - First stage build.

Front panel, half way through construction.
The final design was based arund a Music From Outer Space SoundLab PCB. This is designed to be a very simple switched synthesizer for making sound effects,  but it always screamed out to me as being capable of much more.
It is built on 72h (14.4") 3U (Eurorack) format panel, and the lettering was printed on a laser printer, laminated, and glued to the metal before trimming. It is the first time I have tried this method, and it produces a neat easily useable result, but nobody is going to be fooled that it is a professional job.
The sellotape to protect the corners and the Stig sticker are not staying :-)

Modifications from original


Behind the front panel, with main PCB and extension board
Like I said this little beast is capable of more than its original design and I could not resist building in the following modifications.
  • Routing switches changed for 4mm sockets,
  • LFO - Rate in fast mode reduced to 30-0.05 Hz, by increasing C14 to 22nf
  • LFO - LED on triangle output
  • LFO - Inverted triangle output added
  • AR - Threshold level on external input
  • AR - LED
  • AR - Inverted AR output added
  • AR - Attack/On phase output added to provide additional control gate.
  • VCO - VCO1 upgraded to be exactly the same as VCO2
  • VCO - Trigger input via a capacitor so they trigger on any square wave
  • VCO - Triangle output added (Not tested yet)
  • VCF - Control front end modified for logarithmic response (Not tested yet)
  • VCA - Control front end modified for logarithmic response (Not tested yet)
  • Control inputs on VCOs, audio and control on VCF and VCA have one direct socket (input 1) with no attenuator and at least one socket with an attenuator
  • All outputs via 1K resistor to eliminate problems with shorts, and multiple outputs can be directly connected together.
The circuitry for these additions and modifications are on a couple of fiber glass matrix boards, one at each end or the main panel, which are held in place by PCB brackets mounted behind potentiometers

Progress

Today, 1st December 2013,  the LFO, AR are completely working, the VCOs are working but not of the modifications have been implemented yet. The VCF and VCA have not been started.
Well that was what I thought...

Sunday, 27 October 2013

The Soundlab Diaries - Part 1 - General Design

I have always wanted to build a modular synthesizer, but components were expensive, racks were the size of a large bookcase and it would have taken years to build even the simplest system, so like many other things, I never got round to it.
I had almost given up on the idea when in 2007 I stumbled across a website showing a range of modules built using the new Eurorack, format. These were much smaller, simpler, and used readily available components if I bought one from completed modules it would me more than I would normally pay for a car, including tax and insurance. But there were also kits, and better still some manufacturers would sell you just the blank PCB.
I spent weeks researching the various options and eventually came up with a clear idea of how I wanted my system.

1 - Eurorack panel size format

Ironically Eurorack panel sizes were originally defined in imperial measurements and are all 5.25" (3U) high with widths in multiples of 0.2" This is small enough to be easily made by hand, and smaller modules can even be made from lengths of aluminum strip. Best of all I already had enough parts to make two double height 19" racks.

This skelton rack from 4ms wuld be enough to get started, but you will soon run out of space and need something bigger. The bus here is made on a ribbon cable, but most use a PCB fixed across the back of the rack.

 2 - Eurorack bus

Eurorack bus connector
All eurorack modules use a standard 2x8 connector, as shown above. The pin spacing is 0.1" so it is easy to use on pin board and veroboard.
The only thing I don't like is that it doesn't require a keyed connector so its easy to plug in round the wrong way, potentially destroying your module.
The IDC ribbon connectors do have a key block on one side so I will be using keyed connectors on the other half, just to make sure.
Bus cable being used to connect a module to the back plane. Note the back plane connectors do not have a key so the convention is colored stripe down, but it would be easy to wire a connector up wrong and blow up your module.

3 - Connectors

4mm Banana plugs. Note this type has a hole so that plugs can be stacked up so you don't need "multiple" modules.

Between modules in the same case I am going to use 4mm Banana plugs. This may seem like and odd choice as most Eurorack modules use 3.5mm jacks, but I might want to use this live and my experiences of these have not be good.
4mm bananas are almost bullet proof and if the worse came to the worse you can always strip some bell wire and stuff it in the sockets.
Between cases I will be using standard 1/4" jacks, they have their own earth, so are shielded, and are robust enough not to worry about. I use these for audio connections so already have a good supply.

4 - Build standard

I might want to use this live, so it has to be reliable and be able to take a few knocks. I used to work in the Avionics industry where even prototypes were built with an eye to military specification, and anything less now seems like sloppy work.
For my music gear this means
  • All standard PCB components are soldered through the board and secured on the other side with solder, so any strain is pulling the track towards the board and not lifting it. (Except SMD)
  • If I am adding a component and there is no spare hole than I will fit a vero pin and solder to that. This might mean lifting out an existing component from a hole, fitting a pin, then both components are wrapped round it.
  • Flying leads on a PCB will be as short as possible and either soldered to a pin, or if under the board soldered onto a pad with a component lead, so again any movement of vibration can't lift it.
  • No components in flying leads. Any vibration or movement in the wire could cause the component lead to repeatedly bend where it enters the body, leading to stress and potential failure.
  • Panel wiring will be neatly laid out and fixed down. Laying things out neatly makes it easier to check your progress and find any faults that might occur later. Fixing the wires down, reduces movement at the solder joints and thus stress, making it far more reliable.

Tuesday, 22 October 2013

Tested : Ctrl + Shift + T

From Tested "Our Favorite Keyboard Shortcuts by Norman Chan By far our favorite keyboard shortcut ever. Used in any modern web browser, Ctrl + Shift + T (in either Windows or OS X) lets you restore your last browser tab that you closed. Yes, you can turn back internet time! In Chrome, this lets you restore up to 10 tabs. We love this shortcut so much, we even made a video about it."
OK so its not actually a video, just a series of photographs, but several of these were pretty obscure so check them out :-)

Sunday, 6 October 2013

Raspberry Tank

"The “Raspberry Tank” is a project to create an autonomous vehicle with a web interface, using a cheap hobby RC tank and a Raspberry Pi computer."
The scary thing is in thatthe only think you would need to scale up for this to control a full sized tank is the mtor control mechanism. The Pi, the sensors, the software would all be the same.
I must look on e-bay to see if anybody is selling an old tank. :-)
For details of the smaller one go to http://raspberrytank.ianrenton.com/

Wednesday, 2 October 2013

A Power Supply & Self Powered USB Hub for Raspberry Pi

This is another of those, "this is so simple and so clever, why didn't i think of it" moments.
Like many of us Bharath Bhushan was slightly miffed at the fact that you need separate supplies for a raspberry Pi and its peripherals.
His solution was both simple and elegant. He simply chopped open his USB hub, and tapped off a power feed to go directly to his Pi.
I like the clever way he cut his existing USB cable, the yellow one in the photo, using one end for the power supply and the other to extended the hub so it could more easily reach the Pi.
For an detailed explanation and full construction details go to http://bharath.lohray.com/weblog/a-power-supply-self-powered-usb-hub-for-raspberry-pi/#.UkwOQFOnepF
[Via Make]

Unfortunately the USB hub I was using did not seem to have a separate supply regulator and so the raw 9V goes straight into a large blob on the PCB. 
Its going to be a bit of a gamble if your hub is suitable, and I don't think you can easily check before buying as I can't see the people in PCB World being terribly understanding if you start taking their kit apart in the middle of the shop.
My hub was glued together, you can't just unscrew it, so I had to carefully cut round the seam with a craft knife and then gently lever it apart. Its very easy in this situation to (a) cut yourself (b) cut into the circuitry inside.
If possible hold the Hub in place with a clamp so it doesn't shift, which makes it much easier to work on accurately, and wear gloves.
Once you start to cut through it is normally possibly to gently prize it apart, slowly, without extra cutting, although you might have to work on the case to get it back together again.

Make: Visualize Your Heartbeat With This Homemade Pulse Sensor

Here is a clever little project from those awfully nice people over at the make website.
It s a front end sensor for measuring your pulse rate with an arduino, raspberry pi or any other board with GPIO access.
The principle is simple, illuminate the inside of your thumb with an LED, and measure how much light comes back. It works because there is an artery in your thumb that reflects back some of the light, but as the blood pulses the diameter of the artery increases so it reflects more light.
The Op-Amp circuit detects these changes in light level, amplifies them and passes then as a clean digital signal onto the processor board.
The whole thing can easily be built on vero-board or some other pre-printed generic PCB, although in practice if you were just making it out of curiosity I can't see any reason why it wouldn't work on a breadboard.
It only uses readily available components, although its not clear why they use a quad op-amp and then don't use half the circuit, perhaps it was just the way their PCB was made. Personally I would try a dual op-amp to save space and few pence.
For full details, and a video,  go to http://makezine.com/video/visualize-your-heartbeat-with-this-homemade-pulse-sensor/

Friday, 27 September 2013

Rock, Paper, Scissors, Lizard, Spock

This confused the hell out of me at first, but seeing it like this makes it easy to remember.
Arrange the objects in a circle, in the order of the title, Rock, Paper, Scissors, Lizard, Spock.
Each item beats the next one item on the circular list. Is defeated by the one after that, defeats the next and defeated by the last.
I have no idea what the hand movements are for lizard and Spock!