TankYou rev2

A simple 4-line linear power supply bank (9V DC) connected to the mains.

Caution

SAFETY WARNING: This project involves working with mains voltage (230V AC). Lethal voltages are present in the circuit. Do not attempt to build or test this device unless you are experienced in handling high-voltage electronics. Always ensure the device is unplugged before touching any internal component and that the primary section is properly insulated.

Specifications

• Power source from mains
• 2 galvanically isolated ground domains
• 4 Regulated 9V DC Lines (2 per isolated ground)
• 200mA max per line (Center-Negative polarity)
• Heavy-duty PVC enclosure (from an electrical junction box)
• An old-fashioned external look

Design

The design intentionally prioritizes simplicity, availability of common components, and practical reliability over maximum electrical performance... Moreover the exposed aesthetic identity of the encapsulated transformer and the 4 giant and so beautiful caps.

Schematic

Circuit Analysis

The circuit is organized in three stages: transformer and rectification, filtering, and voltage regulation.

Transformer and rectification:
Notes: The transformer is rated at 9 VAC nominal, but the unloaded secondary voltage measures approximately 13 VAC, which is typical for small EI transformers.

The mains transformer provides two independent secondary windings to achieve galvanic isolation between the two ground domains. Each winding is rectified independently by bridge rectifiers, producing the following $DC$ voltage before filtering:

$V_{DC} = V_{AC} \times \sqrt{2} - 2 \times V_D$

Assuming $V_D=0.7V$ and a measured secondary voltage of $13 VAC$:

$V_{DC} = 13 \times \sqrt{2} - 1.4 \approx 16.9V$

Filtering:
Each rectified rail is filtered by bulk electrolytic capacitors ($C = 4700\mu F$).
The ripple voltage under maximum load is calculated as:

$V_{ripple} = \frac{I_{load}}{2 \times f \times C}$

At maximum combined load for each isolated domain ($I_{load} = 400mA$ total per domain, $f=50Hz$):

$V_{ripple} = \frac{0.4}{2 \times 50 \times 4700 \times 10^{-6}} \approx 0.85V_{pp}$

Voltage Regulation:
Four standard $7809$ linear regulators produce the four regulated $9V$ output rails. Under an actual full rated load of $200mA$ per line, the transformer exhibits a typical voltage regulation drop, settling the regulator input voltage ($V_{IN}$) at $\approx 13.3V \div 13.5V$ (as documented in the Test Log). This dynamic adjustment yields a real static headroom of:

$V_{headroom} = V_{IN} - V_{out} = 13.3V - 9.1V \approx 4.2V$

which is safely above the $7809$ internal dropout voltage ($\approx 2.0V$) and the calculated ripple, ensuring stable regulation.

Power dissipation:
Using the real-world measured $V_{IN} \approx 13.3V$ under the maximum nominal current ($200mA$):

$P_{U_{MAX}} = (V_{IN} - V_{out}) \times I_{MAX} = (13.3 - 9.1) \times 0.2 = 0.84W$

The $78xx$ regulators in TO-220 package have a thermal resistance $\Theta_{jc} = 5 K/W$. All four regulators are mounted on a shared aluminum heatsink with an estimated thermal resistance ($\Theta_{hs}$) between $5K/W$ and $7K/W$. With $T_A = 30^\circ C$ and $\Theta_{cs} \approx 1$:

$T_{hs} = T_A + 4 \times P_{U_{MAX}} \times \Theta_{hs} = 30 + 3.36 \times 7 \approx 55^\circ C$
$T_j = T_{hs} + P_{U_{MAX}} \times (\Theta_{jc} + \Theta_{cs}) \approx 55 + 5 \approx 60^\circ C$

Junction temperature remains well within the absolute maximum rating ($125^\circ C$).

Protection:
Each regulator is protected by an antiparallel diode connected between OUT and IN. This safeguards the regulation stage against reverse currents from inductive or capacitive loads during power-down cycles.

Implementation and Test

PCB Layout

The circuit was designed with ExpressPCB.

Note

The circuit is housed in a standard PVC electrical junction box. Because the enclosure is fully made of insulating plastic material, a chassis earth connection is not strictly required for external user safety. However, inside the box, all AC mains paths are kept physically separated from the low-voltage DC sections.

Calibration Procedure

No calibration required. Output voltages are determined by the fixed internal reference of the $7809$.

Test Log

Load and Dropout Test

Line 1:

Load (mA)	$V_{OUT}$ (V)	$V_{IN} regulator$ (V)
0	9.2	16.8
50	9.1	15.9
100	9.1	15.7
150	9.1	15.3
200	9.1	13.3
500	8.9	10.8

Line 2:

Load (mA)	$V_{OUT}$ (V)	$V_{IN} regulator$ (V)
0	9.1	16.8
50	9.1	15.9
100	9.1	15.7
150	9.0	14.0
200	9.0	13.3
500	8.8	10.8

Line 3:

Load (mA)	$V_{OUT}$ (V)	$V_{IN} regulator$ (V)
0	9.2	16.7
50	9.2	15.6
100	9.0	14.6
150	9.0	14.1
200	9.0	13.5
500	8.9	10.9

Line 4:

Load (mA)	$V_{OUT}$ (V)	$V_{IN} regulator$ (V)
0	9.2	16.7
50	9.2	15.6
100	9.2	14.6
150	9.1	14.1
200	9.1	13.5
500	9.0	10.9

Thermal Dissipation Test

Due to test bench limitations (availability of two dummy loads: one resistive dummy load and one MOSFET active dummy load), the test was conducted by drawing a continuous maximum load of $200mA$ simultaneously on 2 lines (one per isolated domain: Line 1 and Line 4) for a duration of 60 minutes, at a stable ambient temperature ($T_{amb} = 30^\circ C$):

Elapsed (minutes)	$T_{hs} (^\circ C)$
0	30
15	37
30	39
45	40
60	40

• Domain 1, Line 1: $V_{IN} \approx 13.3V$, dropping $13.3V - 9.1V = 4.2V$ across regulator U1. Power dissipation: $P = 4.2V \times 0.2A = 0.84W$.
• Domain 2, Line 4: $V_{IN} \approx 13.5V$, dropping $13.5V - 9.1V = 4.4V$ across regulator U4. Power dissipation: $P = 4.4V \times 0.2A = 0.88W$.

Total power dissipation delivered to the shared heatsink:

$P_{tot} = 0.84 + 0.88 = 1.72W$

Using the thermal Ohm's law, the actual thermal resistance of the external heatsink ($\Theta_{hs}$) is experimentally derived as:

$\Theta_{hs} = \frac{\Delta T}{P_{tot}} = \frac{T_{hs} - T_{amb}}{P_{tot}} = \frac{40^\circ C - 30^\circ C}{1.72W} \approx 5.8 , K/W$

Output Noise and Ripple Test

The output AC ripple and noise were monitored under a full $200mA$ resistive load using an FNIRSI 1014D digital oscilloscope (AC coupling). The output waveform shows no periodic $100 , Hz$ ripple component above the instrument's noise floor (~50mV/div minimum scale).

Testing the supply output connected in series with a $1\mu F$ polyester capacitor directly into a high-gain guitar amplifier at maximum volume, reveals only broadband thermal noise with no tonal AC components, confirms the absence of audible hum or buzz.

Conclusions

Results: The unit ensures good regulation within the full nominal operating scale ($0 \div 200mA$), which is within standard audio/pedalboard requirements.

Suggestions: Add a slow-blow fuse on the transformer primary side to improve safety against internal shorts.

About & License

Author: Alessandro Fraschetti (gom9000).
Technical Notes: The hardware design was supported by ExpressPCB and the custom expresspcb-goslib libraries.
License: This project is licensed under the MIT License.