Recovery and restoration service:frontea online,corp.

Recovery and restoration service:frontea online,corp.

[Abyss Audio] Step‑Style CC Dual Power Supply Optimized for Startup Testing | PopoCLiPS BR DIY

‹ 2026/06/30 ›

Good evening. This time, we’re building a power‑supply circuit.


Yes, yes — that’s right.

Because of what happened in PopoCFA Clarity Prototype (2), I needed to reflect and build a dual‑rail power supply with current control (step‑style CC) optimized for startup testing. Sweat



Here’s the back side of the completed enclosure.


It has two boards installed — one is an older rectifier board I made for dual‑rail supply.

There was space, so I added it as a bonus; I won’t touch that circuit this time.


(AC‑AC adapters without center taps that can generate ± rails… truly precious.)


Purpose

The goal is to build a power supply that provides safety and convenience specifically for startup testing.


Based on the failures during the Clarity prototype, the objectives are:


  • Prevent component burnout due to mis‑wiring or overcurrent
  • Allow simultaneous monitoring of positive and negative rails for easy error detection


“Why not just use a commercial bench supply?”


Well…

“Dual‑rail supplies that show both rails simultaneously are expensive, big, and heavy.”

Sure, compact ones probably exist, but considering convenience… yeah. Sweat


Also, for my projects, the currents I want to observe during startup are usually small — around 10–60 mA.

Clarity is the same.

I don’t want to drag out a huge, heavy bench supply every time.


Target Specs

  • Voltage: ±6–24 V, as regulated as possible
  • Current: step‑limited up to ~100 mA
  • Monitoring: simultaneous visual display of positive and negative voltage & current
  • Operation: live switching of current limits across 5–6 steps (10–100 mA)


If it can do this, I’ll be satisfied.


Current‑Limit Circuit

The first‑stage power source is a commercial adapter or a homemade supply.

So this time, we build the current‑limit stage that follows it.


V1 and V2 are external supplies — even batteries would work.

Q1/Q2 are current sources; Q3/Q4 are current sensing; R3/R4 define the actual current limit.


D is for biasing, C is for input/output decoupling and HF noise removal (C1–C4), and C5/C6 are auxiliary stabilizing caps.


The transistors used in the simulation are the actual parts I have on hand.

(Parts list omitted — please check the schematic.)


Implementation Board

The circuit isn’t very complex, and this is a one‑off tool, so I’ll use a universal board.


However, for PopoCLiPS, I already know I’ll need to pull wires not only from IN/OUT but also mid‑board for meters and switches.


So yes… as usual…


A wiring hell awaits.

Therefore, even though it’s a universal board, I’m using KiCad!

(Otherwise my brain melts and I make mistakes… Sweat)


Here’s the board layout.

Pretty clean, and extraction points are easy to define. ^^

Simulating a universal board in a PCB tool is surprisingly tough. Sweat


Meter Circuit

Next, the meter circuit.


I thought about using a millivolt meter, but the ones available are expensive and bulky.


So — let’s use Akizuki’s affordable VU meter!

I chose the DE‑1434, a tiny toy‑like unit.

The catalog looked decent, and it even has a small trim screw.



Here’s the actual unit.

The box shows two AC connection examples.

(Pretty good for a budget VU meter.)


This time, I’ll use it as both a voltmeter and ammeter, so I’ll remove the printed panel and redesign the scale. ^^


Meter Calculations

The DE‑1434 movement is 500 µA, 650 Ω.

We want a scale of 20 V max and 100 mA max.


So:

Full‑scale voltage (FV) = 0.0005 A × 650 Ω = 0.325 V

Required series resistance (TR) = 20 V / 0.0005 A = 40,000 Ω


Thus:

Voltage‑meter resistor = 40,000 − 650 ≈ 39 kΩ  

Current‑meter shunt = 0.325 V / 0.1 A = 3.25 Ω (≈3.3 Ω)


So we insert 39 kΩ for voltage sensing and 3.3 Ω for current sensing.



I also made an LTspice simulation for this.


Before that, I needed to create symbols for the switch and VU meter.


Because the switch toggles between voltage and current modes, I needed ON/OFF states — so the model ended up with an unnecessary V3 supply.

But as you can see, ±20 V simulation works exactly as expected. ^^


SPDT Switch Model

.subckt SPDT IN OUTA OUTB CTRL

EINV CTRL2 0 VALUE {5 - V(CTRL)}

S_A IN OUTA CTRL2 0 SWA

S_B IN OUTB CTRL 0 SWB

.model SWA SW(Ron=0.01 Roff=1Meg Vt=2.5 Vh=0)

.model SWB SW(Ron=0.01 Roff=1Meg Vt=2.5 Vh=0)

.ends SPDT


DE‑1434 Model

.subckt DE-1434 P N

Rcoil P X 650

Bmeter X N I = { V(P,N) / 650 }

.ends DE-1434


Enclosure Design

Once the circuit and board are ready, the rest is enclosure and wiring diagrams — 99% complete.


Because I added an AC/DC converter as a bonus, wiring becomes quite complex.

To avoid brain meltdown, I split the diagrams into two systems.


Current‑Limit Selector System

Uses only a 2‑pole 6‑position rotary switch.

OUT is split into supply and sensing paths, connected to the shortest return path on the board.


R3/R4 define the current limit

I_limit ≈ 0.6 V / R3 (positive)

I_limit ≈ 0.6 V / R4 (negative)


Using:


60 Ω / 40 Ω / 20 Ω / 12 Ω / 8.2 Ω / 6.8 Ω

→ approx. 10 / 15 / 30 / 50 / 70 / 90 mA


Nice balance for 6‑step switching.


“But I don’t have such weird resistor values.”


Correct — I don’t either. Cry


So we make them. ^^;


6.8 ≒ 10|22

8.2 ≒ 10|47

12 ≒ 22|27

20 ≒ 10 10

40 ≒ 30 10

60 ≒ 30|30


(| = series, = parallel)


This gives smooth switching from 10–90 mA — perfect for live testing. ^^


Power and Meter System

Quite mixed. Sweat


Components:

  • DPDT ON‑ON ×2
  • DPDT ON‑OFF‑ON ×1
  • DE‑1434 ×2
  • Connectors


I tried hard, but if V/A switching is done via DPDT, the shunt resistor stays energized… Sweat



“Well, fine — this way output stability is better anyway.”


Implementation and Measurement

As shown in the first photo, wiring is dense as usual, but the build turned out nicely.

Now for actual measurement. ^^


Using a ±16 V supply (Chinese unit).

(Positive/negative differ by ~1 V — that’s just how this adapter is.)


Testing around 30 mA with a 500 Ω load:


10 mA → OK

15 mA → OK

30 mA → OK

50 mA → limiter activates → OK


Great. ^^


Voltage side:

Meter reads slightly high, but adjusting resistors endlessly is pointless — I tuned the meter trim instead.


Current side:

Also good — shared meter, but current accuracy is quite nice.


PopoCLiPS BR Completed!

Now I can finally perform detailed startup testing for PopoCFA Clarity.

Excellent.


By the way, the replaced meter panel ended up with V/A switching reversed relative to the scale… but I’ll accept it. Sweat


Next time, we can finally move on to the PopoCFA Clarity Release Candidate.


Stay tuned♪