The amplifier is based on the 12AU7 valve (part number ECC82 in Europe). The schematic came from here, it's a nice kit, but lacked a power supply and the layout wasn't quite what we needed for kits in TinkerSoc. I added a LDO 12v regulated power supply, an input volume control pot and kept the design single layered (with one jump). The final schematic can be viewed here
A couple of things needed to be debugged on the prototype, I'd accidenty created a high impedance short between the gnd plane and the AC input, but thankfully I popped a scope onto the power supply before installing the valve on the board, so it was caught early and rectified. To stop this happening again I've excluded the ground plane from the AC input section of the PCB.
The board can be run of an AC or DC power supply. If you know you're going to use a DC power supply you can connect the positive and negative up acordingly and omit the rectifier, and possibly decrease the size of the resevoir cap.
There was also an issue in that I incorrectly put a choke into the output of the voltage regulator to eliminate transients. This had far too much impedance and caused a massive voltage drop on the circuit (so it didn't work) and got incredibly hot. I removed this and put a jump in it's place, and the board worked perfectly.
There is some setup to do on these before connecting headphones. The two trimmer pots must be adjusted to get the right bias. You set this by adjusting the 50k trimmer until the source is at half the suply (drain) voltage. In our case a smidge over 6V. You should check this a few times after the device has warmed up to make sure everything is settled in.
Check the current on your headphone jack between L and GND; and R and GND are both less than 10mV.
I found that with the mains supply I had hooked up to the board that the LDO regulator (VR1) was getting very hot - so it might be worth considering finding a small heatsink for it. If you hear sizzling when you touch it, it's probably too hot.
We designed a nice simple case as well. I'm still a little concerned that it might need some extra ventilation, but it should work for now.
The top of the case needs to be made of very flexible ply wood of some description, we used aeroply. It's a little expensive, but no prohibitively so. It was all laser cut, and is simple enough to be nice and cheap
We also wood stained it to look pretty.
Building the case is quite easy, you bend the top part over the sides, gluing in place, and mask up the whole box until the glue on the top has set. Then you remove the bottom and put the tube amp and connectors in.
These things work out at a fairly reasonable price. By far the most expensive component is the tube, which we found online for £9 inc P+P. The rest of the components came to about £15 from Farnell, and we got the PCB for free from our University's electronics department. The PCB is very easy to manufacture by hand using home etching techniques, as it is single sided and my routing uses very generous trace sizes which are more than big enough.
As far as performance goes, it amplifies - I'm not really sure how much, but acording to a more audio savvy friend it sounds very good. The 12v reg put out 12.47v, which is just over what I'd like, but is still fine for the tube and within 5% tolerance. A better linear voltage regulator would have fixed this, as the prototype has a cheapy one from ebay. We got a very good DC offset at 1.2mV, typical is 3mV.
The PCB files are for EagleCAD, and can be downloaded here. Just stick it in your EagleCAD projects folder.
The AutoCAD case design can be downloaded here.
Licensed as CC-BY-NC 3.0, 2013