RE: An isolated low noise +/-5V DC power supply

strb — Mon, 04 Nov 2024 21:06:05 GMT

Very interesting, thanks for sharing.
It's amazing what can be achieved with a humble 7805

RE: An isolated low noise +/-5V DC power supply

scottiebabe — Sat, 02 Nov 2024 21:49:25 GMT

Neat, thanks for sharing.

I have one of the ADUM3160 USB Isolator dongles that works well enough.

I should try out the newer ADUM3166 which supports high-speed USB at 480 Mbps.

RE: An isolated low noise +/-5V DC power supply

DAB — Sat, 02 Nov 2024 17:44:54 GMT

Very interesting and well written blog.

Well done.

RE: An isolated low noise +/-5V DC power supply

shabaz — Sat, 02 Nov 2024 15:19:22 GMT

Hi Michael,

This is a very useful blog!

It's been bugging me, how to measure noise on a higher-voltage supply (up to say 24V), where I feel uncomfortable with AC coupling in case the supply is powered on/off when the 'scope is set to a very low volts/div.

I was wondering if I should use an external capacitor, followed by a TVS, although then all that would need to be shielded too.
In my environment, the MXO 4 picks up noise (at about 17 MHz I believe from memory; don't know where it is coming from) even when using double-braided (RG316DS) coax, so I had to move to semi-rigid coax to remove that.

There are interesting "R-core" transformer (they are available on AliExpress), and I've measured the capacitance between primary and secondary to be 100 pF on them (versus 200-300 on normal E-core transformers) but you make a good point, that loads will have their own filters for HF.

It's impressive that the MXO 4 can be used for such low-noise measurements.

A side project I've been working on is a preamplifier (not specifically for PSU noise measurements, but perhaps useful for that) for 10 Hz to 100 kHz. I found a design in an old magazine article, but some of the parts are obsolete, so I've had to pick alternatives. I've entered the schematic into KiCad and prepared the relevant footprints, but have yet to start the PCB layout (mainly because I wasn't sure if I wanted to proceed with this or not; it's quite an excessive low-noise and low-distortion design presumably for testing audio circuits; but could be fun to experiment with one day). (I have simulated bits of it, especially the areas where I had to choose different parts).