Simple Optical Compressor for bass guitar – construction

An extremely simple yet capable compressor circuit is published by Elliott Sound Products, here:

Compressor Mk I

I built the circuit exactly as-is (in an Altoids tin). 

Care was taken to optically couple the bulb to the photocell (LDR), using a perspex rod, wrapped in smooth tin-foil.  The perspex came from eBay:

as did the miniature “Grain of Wheat” bulbs:

This was put together, and sealed in heatshrink tubing, to make it fairly lightproof from external light sources.  The rest of the components came from (free postage) or from my junk box.

Here’s the final assembly.  The green wire provides grounding (shielding) for the tin-foil, as per the original article description.

Here’s the finished article.  The Altoids tin was really too small, so I had to cut notches in the lid, which rather spoiled the look.

The duck tape on the bottom of the box provides insulation.  The whole tin is effectively coupled to the amp “earth” via the jack socket outer rings.

I soon discovered the following:

  • Using jack-sockets to carry high-current, low-impedance amp output signals is a very dumb idea.  I knew this before, I just kind-of forgot it in the urge to get everything to fit in an Altoids tin.  Even with the amp output muted (but powered), I got a fat blue spark when unplugging the box.
  • The “grain of wheat” bulbs don’t really get into their proper operating range of brightness when you have an efficient speaker cabinet (like a one) – even when the variable resistor is turned right up.
My amp has an FX send / return loop, so I was using that for the input side (on the left).  The two speaker sockets (on the right) were there in case a “pass thru” was needed, although again my amp has two speaker-outs, so no need to use that in practice.
I was a bit concerned about the way the tin was earthing together the FX send/return loop with the “-” side of the amp output.  It worked fine for the two amps I tried, but what would happen if it was connected to an amp in bridge-mode where the “+” and “-” sides were both driven?  Clearly not ideal.

Compressor Mk II

 Mk II was going to be built in a proper case, with a Speakon plug on a trailing lead (so it simply cannot be plugged in wrongly), and would use a high-efficiency LED with suitable dropper resistors instead of the bulb (so as to be more effective at low amp output levels).

It would also have electrical isolation between the amp out / speaker side, and the delicate FX send/return loop (which will be used for a shielding / ground connection for the case).

Here it is:

The Speakon cable enters through a grommet, is double tie-wrapped to the floor of the case, so even your clumsy drummer friend shouldn’t damage it.

The variable resistor is a 5 watt, 500 ohm type which I found on eBay:

You’ll notice a pair of 47 ohm fixed resistors, in parallel, on the green input wire next to the variable resistor.  In parallel that makes around 24 ohms.  The purpose of that, is to prevent the last little bit of the variable resistor’s track dissipating a lot of power.  In fact it’s hardly necessary as the LEDs use less power than the bulbs.  I used two small resistors in parallel because they have a low power rating (I think 1/4 watt) and didn’t want them to get too hot.

Because LEDs can “flash” much faster than a bulb can, and only respond to current in one direction, we need a “rectifier and smoother” circuit.  This is what the small diode and big  capacitor does.

The capacitor is 470uF, 25v (from the junk box).  Voltage rating is important, as your amp can produce quite high peak voltages.  Make sure you use at least a 25v rated one, higher if you can – preferably 63v.

The output of the capacitor feeds two separate LED circuits.  The first one has 2x 220 ohm in parallel (effectively 110 ohm) feeding the high-efficiency LED that is inside the shrink wrap.  Because the LED light output is very directional, I did not use the perspex rod – I just put the LED right up against the front face of the LDR.

The second circuit goes through 470 ohms to a small front-panel LED.  The idea being that you can see what the compressor is doing from the outside.

The final part of the circuit is the same as the Elliott original.  A 33k resistor couples the input side (amp:  FX SEND) to the output side (amp:  FX RETURN), forming a potential divider with the LDR.

With the lid off in the dark, you can see both LEDs are working together.

The case was sprayed a non-Ashdown blue, and a nice quality collet knob from CPC/Farnell added.

The final circuit seems to work well – the component values will need some tweaking to suit your amp / cabinet combination.  If you want a circuit that is less sensitive to amp/cab variations then you might want to look at the other Elliott pages.

Circuit diagram:

Free download in PDF
and ExpressPCB .sch format (R-click / Save As) .. software available here

Sony Vaio VPCEB4L1EWI – Heatsink clean and refit

This time, it’s a Sony Vaio VPCEB4L1EWI, with i3-380M CPU.  About two years old.  Running Prime95 stress test, the cores were reaching 85 deg C and the fan was rather noisy.
1.  Strip off the back plate, undoing all the visible screws, taking careful note of where they all go back – there are several different sizes and types.  The back plate then snaps off – be careful with the small white lugs on the upper side here.  One came off despite me taking care. 
Always wear an anti-static wrist strap, connected to a metal part of the laptop (e.g. the USB connector) when working.

2.  Here is the heatsink / heat pipe / fan assembly removed.  The grey heatsink paste had set quite hard.
3.  Clean up the CPU top surface carefully with meths, a cloth, and cotton buds. 
 4.  Clean all the old paste off the heatsink.
5.  To do a really good job, “flat” the heatsink down with a diamond file, cleaning the copper dust off regularly.  I kept applying a few drops of meths as a non-greasy lubricant while working.
After a few minutes you should be able to get a near mirror-finish on the heatsink.  Don’t over do it – copper is a very soft material.
7.  Apply a new heatsink paste (I used Arctic MX-4), and re-assemble the machine.
Measuring the core temperature with “CPUtemp” and running “Prime95” stress test, the peak core temp reached 68 deg C instead of 85, and the fan ran much quieter as a result.

Building a quick & easy website .. pros and cons

Just finished the first draft of a website for my wife’s home tutoring business.  I used the free web site builder tools from, and they’re not bad.  A few things we came across were a bit frustrating however:

  • Once you’ve chosen your template, the background image (in our case, of the coloured pencils) seems to be unchangeable
  • There’s no “undo” button, so use the “Save and Publish” option frequently to avoid losing work
  • When you create a new menu item (on the top row, in our case), it throws away work you’ve just done on your current page, so save it first!
  • Working with images is somewhat frustrating in terms of positioning, resolution etc

There are also good things about the sitebuilder:  The sites look good on iPhone and Android devices as well as on a PC / Mac.

I also noticed that when running this auto-generated page through W3C Validator, it finds 5 “errors”.   Hmmm.

I looked around for help pages, or a forum on this website builder too, to no avail.  Anyway I should probably be using something “proper” like WordPress, Joomla or even good old Dreamweaver.  Thoughts /comments appreciated.

Improving an Ashdown Little Giant bass amplifier

I have an Ashdown Little Giant amp.  Several reviewers have mentioned that this amp is not at loud as it should be (e.g. here).

It all depends on whether the output levels coming from your bass are enough to drive the output stage fully.  I have an active bass, but even with the input set on max, to get a decent output level I need the output volume control set to around 3 o’clock.

While there are workarounds available, e.g. use an effects send/return through a pedal which has some gain, I wanted a simple ‘no extra boxes’ solution to go with my lightweight cab.

A nice chap called Wilfried Klaas has come up with what I consider to be the perfect solution to this – adding an extra small preamp (gain) stage inside the amp itself. His web page (in German) describes this improvement.

Wilfried was very helpful and supplied me a ready-made amp board for around €15.00.  This was fairly straightforward to fit.  I’ve recorded the process of testing the amp, fitting the board, and re-testing the amp.

Incidentally – my amp, purchased off Ebay, started off life as a green-front, LG350.  However something strange had been done to it, the fan was wired to run continuously off the 12v supply, rather than being connected to the variable-speed fan output.  Eventually the power module failed, and it went back to Ashdown, who replaced it with a Powersoft Digimod 1000 module.  Only one side is connected and used, so I have effectively an “LG500”.

1.  Measure the amp’s existing output

First, I hooked up the amp’s instrument input to a sine-wave source.  I used a 1kHz, 0dB digital source which you can get hold of from here.  This was loaded as a .wav file onto an IPod.  The output from the iPod was checked on a scope; above 90% volume there were some signs of clipping (presumably the iPod’s headphone output stage).  This input level was just enough to light up the ‘peak’ light on the LG’s input stage.

The amp output was connected to a 4 ohm dummy load, which I made as follows – two 8.2 ohm resistors on big heatsinks, connected in parallel, as shown here.

For the actual test, these were placed on a metal tray, on a wooden board, as they will get hot if you run the amp into the loads for more than a few seconds.

With a sine wave input, the amp produced a maximum output of around +/- 60 volts peak, shown here:

 The RMS power in a sine wave is

    (V^2 / R) / (sqrt(2))  =  (3600 / 4) / 1.41 = 636 watts

After a few seconds,  some kind of power limiter kicks in, presumably in the power amp module itself, and the output falls to 44 volts peak, which is

     (1936 / 4) / 1.41 = 343 watts.

So the amp can produce more than the rated power (for short periods), the problem is down the lack of gain.

2.  Fitting the “Wilfried Klass” buffer stage

Firstly, switch the amp off and unplug the mains lead, leaving the load connected to dissipate the internal power, and take the top off the amp case.  The cooling fan has a 3-pin connector which looks very similar to PC-type cooling fans, this is easy to remove.

Inside of Little Giant amp
The white wires on the bottom left carry mains electricity, and even the +/- 60v output from the amp could deliver a shock, so always work with the power off.
The module requires + / – 12v DC power and an earth, which are obtained from the green connector following Wilfried’s instructions.
Power connector On Digimod 1000 module

The next step is to find and break into the grey ribbon cable that goes from front to back. , on the right side of the amp.

The third wire from the bottom carries the preamp signal.   This needs to be picked out of the ribbon and cut into.  I used the scope at this point to check that the sine wave signal was indeed on pin 3

before I cut into it.   I broke into the cable carefully using a craft knife:

The buffer circuit is then connected up, with the input coming from the front panel, and the output going towards the back panel:

 I then powered up and tested the amp was working, but this time maximum output was reached (before onset of clipping) at a lower setting on the output knob.  The limiter behaviour is the same as before.

Finally, the buffer circuit was fitted into the insulating sleeve supplied by Wilfried (red colour in the picture below).  I used double-sided sticky tape to secure this to the inside of the case, and tie-wraps to keep the wiring in place.

Having just used the amp in a band rehearsal, I would say it’s a great improvement – performance volume levels can be reached, while there’s still some ‘reserve’ on the output knob.  There is no appreciable increase in the level of hiss or hum from the amp (it’s pretty quiet both before and after the mod).

So thanks to Wilfried for making an OK amp (available for reasonable money on EBay) into a great little amp.

Helping my son revise for a maths test .. most of the questions seemed a bit dull and unlikely to catch a boy’s imagination, so I made this one up for him.

A crazy sniper shoots a high velocity bullet vertically upwards.  The muzzle exit velocity was previously measured on a firing range, using an accurate timing device, to be 980m/s.

Use the simple equations of motion (known as the SUVAT Equations .. available on thickipedia here) to calculate

A) the time the bullet takes to reach the top of its trajectory
B) how high it goes
C) bonus mark : could the gun have shot a hole through the Red Bull Stratos balloon just before Felix jumped?

Sony Vaio VPCF13M0E – CPU heatsink clean and re-fit

My work laptop (a Sony Vaio VPCF13M0E) had become very noisy, with the fan running full speed as soon as it did any actual work.  A bit of googling found that this was a common problem.  The original heatsink paste becomes dried up, and doesn’t conduct heat well.

Here’s some pictures of the steps I took to fix it.  Obviously ONLY ATTEMPT THIS if you know what you’re doing, and the machine is out of warranty.

Time taken: 2 to 3 hours total. 
You will need:

  • Small cross-head screwdriver
  • Anti-static wrist strap
  • Heatsink paste
  • Methylated spirits
  • Cotton buds, kitchen roll
  • Somewhere safe to put all those tiny screws – I recommend a large blank sheet of paper (with a drawing of the back of the machine) to put the screws down onto

1.  Remove the battery, and all the screws on the back plate.  Keep careful note (with a drawing of the back of the machine) where all the screws go.  There are several different kinds, and a few little tricks.

Remove the screws – on the back
– inside the battery compartment
– near the battery connector
– covering the RAM (middle plate)
– covering the hard disk (lower left)
– holding the hard disk in.

The optical disk drive can be pulled out of the machine, revealing three small screws (on the left).

I counted 24 screws in total.

Next is the tricky bit.  The plastic back plate can be gently lifted off, taking care to not stress the USB, SATA, ethernet and VGA sockets on the right of this picture.  Nearly there!

2.  Wear an anti-static wrist-strap for this next stage.  You can see mine is clipped onto the metal part lower-right.

The CPU and GPU heatpipe plates are held on by springy metal strips and four screws each.  Take these off and keep them separately.  Remove the small fan connector from the motherboard.

Once it’s free to move, ease the plates away very gently from the top of the CPU and GPU chips.  Some dried paste will be left behind on each one.

3.  Clean all the old grey gunk off the plates with cotton buds and methylated spirit.  Very carefully clean around the CPU and GPU chips with cotton buds/meths, making sure not to damage any of the small surface-mount components.

4.  Examine the state of the heatpipe.  There should be no leaks or other damage.

I found that the copper heat plates that go onto the CPU and GPU, were not very flat!  You can see the ridge marks where it has a concave bit.

A bit of work with a diamond file got them to be much flatter (this picture was taken about half way through the flatting work). Only do this if you are confident and skilful with this kind of work; you should still get OK but not ideal results if you just clean (but don’t flatten) the heat plate.

Take care not to remove too much material!  Keep cleaning both the file and the work with methylated spirit to avoid build-up of copper particles that could leave deep scratches.

4. Finally you should have everything clean and ready to put back together.

Apply a thin coat of good quality heatsink compound (I used Arctic Silver).  The amount I’ve put on here is plenty – any more will just get squeezed out and be wasted.

5.  I gave the heat plate springs a gentle tweak to make sure they were pushing the plates down. I also gently cleaned the fan blades with a paint brush and hoover to remove dust.  There wasn’t much dust build-up on the actual heat sink blades.

6.  Re-fit the heatpipe carefully.  It’s best to work slowly and fit screws in diagonal opposites i.e.

1  4
3  2

7. Fit the back of the laptop back on, and re-fit all the parts and screws.

8.  Finally – proof that it’s definitely worked is that your laptop should run quieter and better.  If you want to give it a good workout, I would recommend downloading and running some Cosmology@home tasks.

You can see with all CPUs running at 100%, the cores are at a max. of 68C which is fine.


SATA drives and Acer T135

More on the SATA / Acer T135 saga.

Having got the PC to “see” the drive in the BIOS, it should be a simple matter of restoring the system volume onto this drive (or cloning it from the old drive), making it a primary, active partition (so it’s bootable into Windows), and off we go, yes?

So I created 1 large partition for the system and restored onto it. Did this work? No. Not in these Acer machines.

According to boot.ini (which the wonderful StorageCraft ShadowProtect software can read easily), partition 2 was set as the boot volume. And I only had 1 partition – so better change it.
So I edited it to be partition 1, and tried again. No luck.

I recall that the machine was supplied with the drive having two partitions –

  • Partition 1, around 2.8 Gbytes, for automatic ‘recovery’ of the machine
  • Parition 2, the usual Windows system volume.

The recovery process is an Acer thing. Basically on boot, you hold down a key (Alt+F10) and the machine goes into auto-recovery mode. This basically re-initialises the system partition (2) from a factory-created archive held on partition 1. That’s what the 2.8 Gbytes is used for. It provides a degree of safety for uses who trash up their machines so they won’t boot, and I suppose it’s cheaper than supplying Windows installation media.

I’m not a great fan of this kind of auto-recovery (after all, it will trash all your programs and data, for sake of getting Windows re-installed). I prefer to use StorageCraft ShadowProtect with an external USB drive, as then I can get the machine back exactly to “how it was” the last time backed up.

Out of a mix of desperation and curiosity, I tried holding down Alt+F10 as the machine booted (off the new disk, which did not have any recovery software on it, only a clean system image). Voila! it booted into Windows, off my new partition 1. So this told me two things:

  1. I could restore the machine, somehow.
  2. The Acer BIOS somehow seems to be ignoring boot.ini, and doing its own thing, based on this Alt+F10 key (or lack of it).
  3. The Alt+F10 key forces a boot from partition 1. It works, but it’s a bit inconvenient for everyday use.

Another attempt

Now knowing this about the Acer machine, I tried again, but this time

  • Small initial partition (around 2.8 GB), which I put a copy of the Acer restore stuff onto from the original disk
  • Large system partition (the rest of the disk) for Windows etc.

The first was copied from the original Acer-supplied hard disk. The second came from my ShadowProtect backup.

Still no luck. It would not boot into Windows. But with Alt+F10 held down, it would boot into the Acer Recovery software on Partition 1.

Yet Another Attempt
Finally, in desperation – I thought – why not run this Acer recovery software, and at least see if it can make the machine (normally) bootable, off a freshly restored partition 2. So I let it restore, effectively over-writing partition 2 with its virgin install of Windows + the Acer utilities etc. that came with the machine. (No CD-ROMs are supplied as standard).

This resulted in a machine that actually boots up (although missing a lot of stuff, obviously). So – here’s the trick – that Acer Restore feature is actually necessary for a machine rebuild – in a totally non-obvious way. boot.ini looked just the same, before and after the restore.

I wonder what the Acer Restore does to the disk. It does some magic that is not obvious.

So … having proved that was possible, I then restored the ShadowProtect image back over partition 2. And that finally worked! I had a machine that booted normally, and had everything back where we wanted it.

I still don’t know why the machine is so fickle with its boot-up, and why it ignores Windows’ own boot.ini.


  • If you have an Acer T135, that 2.8G of “stuff” on the Restore partition is essential to the continued use of the machine, if you replace hard disks. Don’t just think that having valid Windows install media and a valid key will be enough to get you out of trouble.
  • Acer T135 has its own boot control which ignores Windows.
  • Don’t forget to jumper new SATA drives down to the older standard, if they can be detected.

and finally

  • Don’t buy machines with proprietary motherboards and BIOSs if you can help it! Sooner or later they will drive you mad.

Recommended tools and parts for this kind of work:

  • Patience
  • Anti-static wrist strap
  • Screws for holding the new drive in
  • SATA cable (usually not supplied with ‘OEM’ drive packages)
  • Jumper (again, usually not supplied)
  • Tweezers for fitting the jumper
  • Partitioning / formatting / backup / restore software, such as Norton Ghost, Acronis TrueImage, Partition Magic. My personal favourite at the moment is Shadowcraft StorageProtect 3.2 which works fast and well in my experience. It comes with a bootable CD-ROM image that means you can work with machines even when the hard disks are unbootable. Review of the software here

Acer T135 with and new SATA drives – a snag

Just came across a nasty little snag with SATA disk drives. I thought I’d write it up, in case you run into something similar.

One of our family computers is an Acer Aspire T135, around 3 years old now. The 80GB hard drive is getting rather full and slow. Time for an upgrade.

Step 1 – check what the existing PC can do.
It has a spare SATA connection on the motherboard, so that seems like the best way to go.
There is plenty of space in the case for a new drive.
There is a spare SATA-compatible power connector (different to the old-style 4-pin Molex’s) dangling in the case.
I’ll need to get a new SATA data cable for the drive.

Step 2 – plan how to use the drive
My initial thought is to use the new drive for the OS and programs, and the old drive as scratch area and – if it makes things run faster – the windows swap file.

Step 3 – choose a drive
A quick look at drive specs etc. suggests a 500GB drive as the best price point at the moment (given that the machine does not have to handle huge amounts of data usually).

A quick check on the Acer website suggests no particular compatibility requirements for second drives (then again, it doesn’t really say a lot about their machines anyway).

Western Digital do a nice looking “Green Power” model which claims to save power. As family PCs get left on all the time, that seems like a good idea.

I got mine from Novatech as they are local (

Step 4 – install the drive
Short work with a screwdriver and the new drive was ready to run.
Boot up, check in the BIOS that the drive was recognised … nothing, only the existing drive is found.

I tried various things, including:

  • Swap the existing and new drive data cables
  • Swap the existing and new drive power cables
  • Try with another SATA drive I had available (an older 160GB Maxtor). Ah-ha! That works!

So why does the new drive not work?

Step 5 – fault-finding .. it’s NOT the BIOS …

  • Maybe the BIOS is too old to read 500GB disks. So I flashed the BIOS with a newer version, directly from Gigabyte’s website, rather than from Acer’s website.
  • The motherboard is a K8VM800. But it’s labelled as the K8VM800MAE on the motherboard itself. Gigabyte list two versions – Rev. 1 and Rev. 2 – but nothing about the MAE. According to the chips fitted it’s closest to Rev. 1. (looking at the Ethernet chip type).
  • Incidentally, flashing BIOSs on computers like the Acer T135 that lack a floppy drive is somewhat time-consuming and troublesome. It hasa nice integrated flash card reader though … so … After a bit of Googling, and making a DOS-bootable disk on a Flash SD card, which worked fine on one of my other machines, I gave up, yanked a floppy drive out of another machine, and used that instead …
  • It flashed OK (but with a warning message), and it ran the computer OK, but it kept giving checksum errors on startup. Not such a good idea. Revert back to the original BIOS (which I had kept).

Step 6 – try another 500GB drive

As it happens, I have a 500GB SATA drive (different brand – Samsung) in another machine. So – let’s try that.
It works! So there isn’t a capacity issue. Must be something weird with the drive.

Step 7 – has anyone else hit this before?
After some tricky Googling to avoid the hundreds of old review articles on T135’s, many that have a different spec, I found a forum that suggested that:

  • SATA has different speeds – 150MB/sec and 300 MB/sec
  • While everything is supposed to auto-adjust, it might not

A quick check on the drive manufacturer’s website pulls up a datasheet, and — guess what — there are some jumpers to force the drive down to the slower speed.

It’s that OPT1 jumper (pins 5-6 on the drive family I’m using).

Now the BIOS sees the drive. But the original drive seems to be taking ages to boot up in to Windows (despite not being changed at all). More on this tomorrow.

Happy Solstice (yeah, right).

I received a “Happy Solstice” message from someone close to me. While wishing them no ill whatsoever, I felt like sharing my own thoughts on the matter.

As a fan of Western reductionist scientific principles I was drawn to musing on the possible reason why the sun might not ‘rise’. I came up with:

1. The sun stopped shining.

2. A large, opaque extra-terrestrial object (such as an Arcturan Mega-Goat) suddenly parked its hairy ass somewhere between the Sun and Earth.

3. The Earth stopped spinning.

Other than that, well, yup, the Sun will generally rise as usual.

On (1) – the bad news is, this will happen sooner or later. But there’s probably a billion years to go. As the Sun runs out of hydrogen fuel and starts to consume heavier elements, the most popular scenario is that it will gradually expand and swallow all the inner planets (including Earth). Bummer.

On (2) – many of your friends who have chemically prepared for the Solstice might worry about this. Goats are usually friendly, but are omnivores and have large appetites. I would recommend further research to establish whether goats prefer red planets to greeny bluey ones.

On (3) – Newton’s Laws suggest that if, for some reason, the earth did suddenly stop rotating (and therefore the Sun not rise one morning), you would most certainly know about it, even if you were in bed at the time:

  • Earth rotates Eastwards (i.e. towards the “rising” sun) at approx. 1000 miles per hour at the Equator. About 600 miles per hour at European latitudes.
  • You are rotating at the same speed, snug in your bed.
  • Earth stops. (This is unlikely as it would require a huge force to be applied to something that’s very large, very heavy, and basically 99% molten rock so quite squishy and hard to get hold of). A large meteorite might just do it.
  • You, your bed, your house, and everything else that’s not tied down initially keeps on moving at 600 miles per hour to the East.

This is likely to be quite noisy and uncomfortable, and so people would (briefly) wake up.

You can therefore banish any lingering spiritual guilt you may feel if you are not standing in some damp field with other beardy weirdies, swaying and singing out of tune.

I suggest that you pass this tip on to your earthy friends, so that they don’t have to get out of bed each year specially just to check. On the other hand it’s mostly harmless and probably quite good fun, and you might even get lucky afterwards …

I’ve been playing about with KPL (Kids’ Programming Language), my first little program is a Celestial Mechanics simulator – try it out here.