Project files: Single-supply experiments…

After a short delay caused by me finishing the HF-lineamp writeup first, here are the promised project files for the single-supply minipre and the railsplitter mentioned in the previous post about my USB-C experiments. Although these two designs are pretty different, I’ve decided to group them together because they also have some things in common that would otherwise be a bit repetitive.

What is it?
PCBs for my two single-supply circuits described here. Although these are designed originally for powering via USB-C, they will of course work with any single-voltage supply within the limits given below.

The minipre can be used both as a preamp and as a headamp, although especially difficult headphone loads might be a problem for the TLE2426 (not tested). The rail-splitter can be used for powering other preamp/headamp designs intended for symmetrical supplies. Since it offers more power even some smaller class-A circuits should be feasible to run off it.

How big are the boards?
The amp board measures 2.0” x 2.25” (app. 51 x 57 mm.) and the rail splitter measures 1.15” x 2.3” (app. 29 x 58 mm.)

What is the status of the boards?
The rail splitter board is version 1.1. I’ve changed the silkscreen of the board to reflect the real current limit of the board (see below), but that’s all.
The amp board is called version 1.5 because the basic design is lifted from version 1.5 of the dual-supply original. Apart from adding the virtual ground circuit no changes were made and the board worked first time.

Does it use any special/expensive/hard-to-find parts?
Not really. The ICs may not be available everywhere, but there are lots of credible suppliers so no matter where you live in the world you should be able to source the parts.

Anything else I need to know?

• On both boards the noise reduction capacitor for the TLE2426 has a dual footprint, either a 5mm MKT film cap, or a 1206 X5R/X7R SMD cap mounted on the bottom of the board. I use the MKT cap on my board but it’s mainly for cosmetic reasons, I doubt it makes a practical difference.
• The rail splitter can in principle supply 250mA continuous output current (= the output current limit of the BUF634A buffer). I’d originally put a limit of 150mA on the board because I thought I saw that somewhere as the practical limitation for the standard SOIC-version (without a power-pad). However, a quick look at the datasheet doesn’t seem to support that view, especially when the buffer is in low-BW mode as it is here. You’ll need to keep an eye on the temperature of the buffer though because it is on the underside of the PCB and thus not as easily able to dissipate heat.
• The TLE2426 on the opamp board can supply around 25mA max. If you want to use the board as a preamp then I don’t think you need to worry much about this and I would recommend the OPA1656 opamp as a starting point. If you want to drive headphones (hard) then it might be worth thinking about the idle current of the opamp and the headroom you have available. In that case the OPA1642 and OPA1692 opamps might be worth investigating, because they have very low supply currents.
• The rail splitter has a maximum input voltage of 30-36V. A standard opamp will work up to 30V, but with a higher-voltage capable opamp (such as the OP07 or OP27) the limit will be 36V dictated by the buffer. If you have older BUF634 versions (and not the newer “A”-type), then note that they are also only rated up to 30V.
• The max. input voltage for the preamp is 40V or the max. voltage for the opamp you use, whichever is lower.
• On both boards the LED is in front of the TLE2426. This means that the current draw from the LED is not loading the railsplitter circuit but also that the LED resistor sees the full input voltage. For USB-C voltages this is not an issue, but if you have ambitions to get near the max. input voltage that the boards will tolerate then be sure that you also calculate the power dissipation in the LED-resistor.
• Note that for both of these designs that you are creating a “virtual” ground. If you use the boards on their own (or the railsplitter with a normal amplifier design) that is not likely to be an issue. However, if you start combining multiple boards with different supplies in the same box you need to be a little careful as to which “ground” is connected to what in order to avoid hum and noise problems.

Downloads:
Download design files here

Related information:
For the minipre board this is basically the good old “CMoy” circuit and as such there are plenty of opportunities to learn about how the circuit works should you be so inclined 🙂 I can absolutely recommend the old but still excellent articles from Headwize/Head-fi member Tangent here and here. Tangent’s pages also have a ton of other useful information and although the site isn’t updated any more (and it’s quite old) there’s still plenty of good stuff.

If you are more technically inclined then probably the best resource is the “Opamp Applications Handbook” from Analog Devices and edited by Walt Jung.

For the railsplitter board, be sure to check out Neal’s original build version that inspired this as well as Tangent’s original “research article” on virtual grounds which is still live at the time of writing this post.

Note: Always read the “intro post” for additional important information about my designs.