There is an update to this article here.
As I mentioned in my post on the vagaries of phantom power, I have been looking at designs for a compact self-contained box to connect an upright bass to the PA. The bass pick-up in question is a piezo jobbie integrated into the bridge like the Fishman BP-100. Needless to say Fishman recommends a suitable pre-amp if using a cable longer than 10 feet (3m). Since starting the project, I mentioned it to another friend who has purchased an inexpensive pickup for her violin – similar to the one pictured below.
The tone/volume unit is passive – there is no amplifier and no battery – so again, major tone-suckage has been experienced.
Whilst trawling the net looking for inspiration, I came across Scott Helmke’s “Mint Box Buffer” project and the Altoids Piezo DI box was born. Let’s cut straight to the chase and show a photo of the finished item:
Not much to see from the outside. There’s a 3.5mm jack for the input, an illuminated switch to select battery or phantom power and a 6.35mm (TRS) jack for the balanced/unbalanced output. Let’s have a look inside:
There’s a single-sided PCB with a JFET buffer and an op-amp balanced driver. The unit may be powered from battery or phantom power. For best performance the unit should be connected to to a P48 (i.e. 48Vd.c.) phantom-powered microphone input on a mixer but if necessary it can be battery powered and plugged into a balanced or unbalanced line input. The reason there is a 3.5mm jack (rather than 1/4″) on the input is that players with small acoustic instruments (such as violin, mandolin or ukulele) will appreciate the smaller jack plug on the end of their passive pickup. In order to use the unit with a microphone input a special output cable will need to be made with a 3 pole TRS plug on one end and a male 3-pin XLR on the other end. Assuming the player keeps the DI box about their person, the special cable will only need to be about 2m or 6 feet long as – once it gets to the floor – it can be plugged into a conventional mic lead of any suitable length.
Here’s the schematic:
C1 is for RF immunity. R1 and C2 form a low-pass filter with a corner frequency of 34kHz again to (hopefully) eliminate RF interference. C3 blocks DC from the instrument. The input impedance is essentially set by bias resistor R2 at 10M ohm. Q1 forms a source follower with slightly less than unity voltage gain. Q2 forms a constant current source to bias Q1. R3 provides negative feedback to Q2 to improve its temperature stability and reduce the value of the constant current from Q2’s Idss to some smaller value (dependant on the specific JFET used for Q2). U1 forms a balanced driver. There is provision to add gain at the non-inverting stage of U1.
For no (unity) gain (recommended):
- R5 short circuit (fit wire link in place of R5)
- R6 omitted
Alternatively, voltage gain = 1 + Rf/Rg.
- suggest R5 at 33k
- R6 at 10k
to give a voltage gain of 4.3 or +12.7dB.
The choice of op-amp is entirely up to you. I used an LM833 but a TL072 or RC4558 will also work. Note that the NE5532 will not work with phantom power because it requires too much current. Of the op-amps I tried (all of the above and from a test involving only one sample of each device), the TL072 requires marginally less current than the LM833 and therefore supply voltage will be higher and the driver will have more headroom. However the LM833 may well sound better than the TL072.
D1, D2, D3, D4, R11 and R12 serve to protect the op-amp output stage.
The battery is only connected when a plug is inserted in the output jack. If the switch is pushed in, the LED will illuminate to remind you that the unit is operating on battery power. Finally, there is 400uF of power buffering so the unit will take a few seconds to “warm up” if it has not been used for a while and the capacitors are fully discharged.
The grounding scheme is important and should be followed carefully. A small wire should be soldered from the chassis connection on the pcb to the Altoids tin (which should have a sheet of thin cardboard in the bottom to prevent short circuits). Also, I drilled through and soldered a brass M3 nut to the inside of the box so that the lid can be screwed shut.
Here are the PCB details:
Note that the PCB layout uses 1/8W (0.125W) size resistors except for R15 which is a 1/4W size. At the time of writing all components are available from Rapid Electronics in the UK.
Note that the presence of R3 in the circuit means that you shouldn’t need to match Q1 and Q2. However, matching is highly recommended and you should select for similar Idss of about 2-3mA. The sample with the higher Idss should be used for Q1 and the lower for Q2. If your Idss values are substantially higher then you may need to alter the value of R3 to bring the quiescent current down to a suitably low value.
Errata – Added 18 March 2012
It turns out that I read the datasheet for the 1/4″ jack socket incorrectly and because the pinout of the jack socket is symmetric, I didn’t notice on the prototype. Consequently the PCB layout APDV2DS07 inverts the phase of the signal. If you are concerned about absolute phase (I’m not), you can work around this by crossing over the two links on the PCB.