This is the second blog post of my RoadTest review of the AIM-TTI QPX750SP power supply: AIM-TTI Bench Power Supply

 

Previous blog posts:

AIM-TTI QPX750SP RoadTest | A Non-Conventional Power Supply Review

 

As I mentioned in the first blog post, the plan for this RoadTest is to test the AIM-TTI QPX750SP against a diverse collection of power supplies, usually used to power DIY projects.

 

In this blog post we will go over some a series of experiments involving passive loads.

 

1. Introduction

 

A passive load (or maybe "static load") is an electrical load consisting from passive components: resistors, capacitors and inductors.

 

We can categorize passive loads in 3 main categories:

  • resistive loads - heating elements, etc.
  • inductive loads - DC motors, inductors, etc.
  • capacitive loads - capacitors

(resistive, capacitive and inductive components can also be combined arbitrarily)

 

When a DC voltage is applied to circuit consisting of passive loads, the circuit is supposed to reach a stable state after some time.

 

For this reason, I decided to focus on measuring the output ripple and noise of the tested power supplies. Ripple is supposed to be AC component of the output voltage produces by the power supply. On top of this noise caused internal of external factors can also be present on the output.

 

Along this I also looked at the behaviour at powered on and power off events.

 

For this test power supplies were tested in constant voltage (CV) mode (where applicable), with current limit set to values that are not exceeded.

 

The accuracy of voltages produced by the power supplies was just briefly checked to be in spec (~ ±0.05V for bench supplies, ±0.5V for others).

 

The measurements were using an Keysight MSOX3034TMSOX3034T oscilloscope and handheld digital multimeters. I tried to probe the output signal of the supplies, as close to the output terminals as possible. I didn't bothered to try to eliminate all the possible noise sources, so noise / ripple measurements are just indicative rather than exact.

 

2. Resistive Loads

 

I stared the experiments with resistive loads, mostly heating elements salvaged from broken equipment:

 

 

Voltage

(V)

Resistance (Ω)
Current (A) | Power (W)

4.20 Ω


19.00 Ω
(36 Ω + 41 Ω)

36.00 Ω

93.00 Ω

3.30.792.59xx0.090.30xx
51.195.95xx0.140.69xx
122.8634.290.637.580.334.00xx
163.8160.950.8413.470.447.11xx
19xx1.0019.000.5310.03xx
48xx2.53121.261.3364.000.5224.77

 

I tried came up with a series of voltages that mostly cover the output voltages / voltage ranges of the tested power supplies. Than, I selected suitable resistive loads for each voltage. The power supplies were then tested with these combinations of voltage an load.

 

For each combination of the following was measured:

  • Ripple and Noise - RMS and peak-to-peak values, measured using the Power Application of the MSOX3034TMSOX3034T

       (for the Ripple measurements I used the RMS value; the peak-to-peak value is a good indication of the noise present on the output signal)

  • Power On and Power Off behaviour - Turn On and Turn Off time

 

A screenshot of the waveform displayed on the oscilloscope was saved for each test.

 

Aim-TTS QPX750SP

 

I started the testing with the subject of this road test, the Aim-TTS QPX750SP. These are the values I got:

Aim-TTI
QPX750SP

Output Ripple
(mV)

Turn On

Time

Turn Off

Time

Waveform

rmsp-pmsecmsec

3.3 V

4.2 Ω2.2923.0~ 150~ 120
36 Ω1.5518.0~ 150~ 120

5 V

4.2 Ω1.7823.0~ 120~ 70
36 Ω1.5819.0~ 130~ 150

12 V

4.2 Ω1.7821.0~ 200~ 35
19 Ω1.6718.0~ 160~ 140
36 Ω1.6819.0~ 160~ 130

16 V

4.2 Ω1.7719.0~ 175~ 25
19 Ω1.6619.0~ 175~ 175
36 Ω1.7419.0~ 175~ 175

19 V

19 Ω1.5919.0~ 175~ 160
36 Ω1.6819.0~ 175~ 175

48 V

19 Ω1.6418.0~ 175~ 60
36 Ω1.6819.0~ 175~ 150
93 Ω1.6820.0~ 150~ 150

The output ripple is pretty low, around 1.5 - 2.3 mV for all voltages / loads. The average peak-to-peak noise is 18 - 23 mV.

 

 

RD Tech DPH5005

The tests were continued with the RD Tech DPH5005, the low cost bench-style power supply. The Aim-TTS QPX750SP was used to provide an 24V / 10A input to the power supply.

 

RD Tech
DPH5005

Output Ripple
(mV)

Turn On

Time

Turn Off

Time

Waveform

rmsp-pmsecmsec

3.3 V

4.2 Ω4.80109.9~ 200~ 10
36 Ω3.7093.9~ 150~ 50

5 V

4.2 Ω6.62113.5~ 180~ 5
36 Ω4.5792.9~ 125~ 60

12 V

4.2 Ω15.15152.0~ 200~ 15
19 Ω11.90115.9~ 225~ 60
36 Ω8.70112.7~ 225~ 110

16 V

4.2 Ω21.14214.5~ 200~ 15
19 Ω14.87124.1~ 225~ 65
36 Ω10.86121.5~ 200~ 100

19 V

19 Ω16.56132.4~ 200~ 75
36 Ω12.28119.6~ 225~ 125

48 V

19 Ω9.35163.4~ 220~ 80
36 Ω8.11153.4~ 225~ 110
93 Ω7.44115.0~ 250~ 325

The output ripple of this supply is significantly higher, around 4 - 21 mV depending on the load. The output signal is relatively noisy, with average peak-to-peak values of 90 - 210 mV.

 

ATX Power Supply

 

The next victim was the old ATX power supply which was transformed into a "bench supply".

 

The Element 14 system allows a maximum of 110 images per blog posts. Please find the rest of waveform images in the following blog post AIM-TTI QPX750SP RoadTest | Static Load Test - Waveforms

 

ATX

Output Ripple
(mV)

Turn On

Time

Turn Off

Time

Waveform

rmsp-pmsecmsec

3.3 V

4.2 Ω4.7595.8~ 25~ 100

Link to waveforms...

36 Ω3.7598.3~ 25~ 120Link to waveforms...

5 V

4.2 Ω11.62109.6~ 25~ 150Link to waveforms...
36 Ω3.3381.9~ 25~ 220Link to waveforms...

12 V

4.2 Ω18.69166.1~ 25~ 125Link to waveforms...
19 Ω9.5691.2~ 25~ 250Link to waveforms...
36 Ω6.82100.5~ 25~ 350Link to waveforms...

 

 

The ATX power supply produced output ripple and noise levels similar to the RD Tech. However, the accuracy on the 12V rail was pretty poor, with voltage regulated to ~11.3V, instead of 12V.

 

LTE 12V Power Brick

Next was a 12 V / 5 A power brick produced by LTE:

 

LTE
12V

Output Ripple
(mV)

Turn On

Time

Turn Off

Time

Waveform

rmsp-pmsecmsec

12 V

4.2 Ω43.54663.6~ 10~ 35Link to waveforms...
19 Ω22.55591.7~ 20~ 200Link to waveforms...
36 Ω24.24982.0~ 20~ 300Link to waveforms...

 

The peak-to-peak output noise values are all over the place. I think, is the worst from all the power supplies tested. This is quite unfortunate given that the power supply is part of a relatively high value Xilinx ZCU104 development kit.

 

Mean Well 12V Power Adapter

 

Next was a 12V power adapter from Mean Well.

 

Mean Well
12V

Output Ripple
(mV)

Turn On

Time

Turn Off

Time

Waveform

rmsp-pmsecmsec

12 V

4.2 Ω12.66178.3~ 10~ 50Link to waveforms...
19 Ω6.05178.3~ 10~ 250Link to waveforms...
36 Ω4.44176.4~ 10~ 350Link to waveforms...

 

The output ripple values are pretty good, with a bit of noise on the output.

 

Mean Well (NXP) 12V Power Brick

 

I continued the tests with a 12V power brick produced by Mean Well for an NXP dev kit.

 

Mean Well
/ NXP
12V

Output Ripple
(mV)

Turn On

Time

Turn Off

Time

Waveform

rmsp-pmsecmsec

12 V

4.2 Ω31.01331.0~ 50~ 100Link to waveforms...
19 Ω29.29472.6~ 60~ 400Link to waveforms...
36 Ω21.02468.7~ 60

> 500

Link to waveforms...

 

Output ripple and noise are "average", bit noisy.

 

IBM 16V Laptop Charger

 

Next was an old 16V laptop "charger" from IBM.

 

IBM

Output Ripple
(mV)

Turn On

Time

Turn Off

Time

Waveform

rmsp-pmsecmsec

16 V

4.2 Ω63.49590.51~ 25~ 35Link to waveforms...
19 Ω26.93248.3~ 10~ 150Link to waveforms...
36 Ω18.73248.3~ 10~ 225Link to waveforms...

 

The output ripple and noise are a bit high on the IBM. Maybe because it's quite old.

 

ASUS 19V Laptop Charger


I continued with ASUS 19V laptop charger.

ASUS
19V

Output Ripple
(mV)

Turn On

Time

Turn Off

Time

Waveform

rmsp-pmsecmsec

19 V

19 Ω27.24373.7~ 15~ 150Link to waveforms...
36 Ω20.48320.6~ 15~ 250Link to waveforms...

 

Nothing too interesting here.

 

LiteON 19V Power Brick

 

Next was a LiteON 19V power brick from a NVDIA Jetson dev kit.

 

LiteON
19V

Output Ripple
(mV)

Turn On

Time

Turn Off

Time

Waveform

rmsp-pmsecmsec

19 V

19 Ω35.59327.8~ 50~ 80Link to waveforms...
36 Ω25.79316.7~ 40~ 175Link to waveforms...

 

The power supply produced similar results to the Asus.

 

ChipKit UNO 32

The next victim (literally, this time) was a ChipKIT Uno32.  The plan was o power it from a 12V jack style supply.

 

Oppsie!

Looks like the 3.3V regulator released the magic smoke.

 

After some investigation, I found that a jumper was set to bypass the 5V regulator, so the 3.3V regulator ended up being powered with 12 V. It supports a maximum of 6 V. Yeah, user error!

 

Arduino MKRFOX 1200

Then, I continued with an other Arduino board the MKRFOX 1200, this time powered from USB.

 

LiPo Battery
4S / 15.2V

Output Ripple
(mV)

Turn On

Time

Turn Off

Time

Waveform

rmsp-pmsecmsec

3.3 V

19 Ω4.0291.1~ 5~ 10Link to waveforms...
36 Ω4.0395.9~ 5~ 10Link to waveforms...

5 V
(*USB)

19 Ω5.7798.0~ 5~ 15Link to waveforms...
36 Ω5.36105.1~ 5~ 25Link to waveforms...

The output ripple and noise are OK. The 5V is probably coming directly from the USB.

 

LiPo Battery

I finished the tests with a LiPo batery:

 

LiPo Battery
4S / 15.2V

Output Ripple
(mV)

Turn On

Time

Turn Off

Time

Waveform

rmsp-pmsecmsec

15.2 V

4.2 Ω1.4433.5xxLink to waveforms...
19 Ω1.4231.9xxLink to waveforms...
36 Ω1.3125.4xxLink to waveforms...

 

The battery produced the cleanest output, with the lowest output ripple and noise. This is somewhat expected, as batteries are known to have a low noise output.

 

3. Inductive Loads

 

Next, I continued the experiments with inductive loads.

 

The initial plan was to do some tests with a salvaged 220V DC motor ran at 48 V. As the motor was a bit weak at 48 V, I decided to also throw in a small 12 V motor I use for PCB drilling:

 

As the most of the above brick / adapter power supplies are not really designed for inductive loads, I decided to continue the tests just with the two "bench" style supplies. I also used a set of reverse protection diodes as protection.

 

These are the result I got from the tests:

 

Aim-TTI
QPX750SP

Output Ripple
(mV)
Turn On TimeTurn Off Time

Waveform

rmsp-pmsecmsec
220V Motor
@48V
5.51198.8~ 175~ 150

Link to waveforms...

12V Motor45.14

308.8

~ 150~ 1100Link to waveforms...

 

 

RD Tech
DPH5005

Output Ripple
(mV)
Turn On TimeTurn Off Time

Waveform

rmsp-pmsecmsec
220V Motor
@48V
10.94106.2~ 200~ 2250Link to waveforms...
12V Motor35.52343.3~ 200~ 1800Link to waveforms...

 

Surprisingly, the two power supplies produce pretty similar result in terms of output ripple and noise.

 

On the other hand, we can observe an interesting difference on the turn off test. When turned off, the Aim-TTI seems to discharge the output a bit quicker compared to the RDTech.

 

4. Capacitive Loads

 

I also wanted to do some experiments with capacitive loads. As there are not many real world examples of pure capacitive loads, I used a couple of electrolytic capacitors for testing:

The total capacitance is about 2000uF.

 

I didn't measured output ripple, as capacitors just acts as no load after charged. I captured however the behaviour on Turn On and Off events.

Aim-TTI
QPX750SP

Turn On TimeTurn Off Time

Waveform

msecmsec
~ 2000 uF @ 24V~ 150~ 150Link to waveforms...

 

 

RD Tech
DPH5005

Turn On TimeTurn Off Time

Waveform

msecmsec
~ 2000 uF @ 24V~ 200~ 5000Link to waveforms...

 

Here we can also observe the more aggressive discharge behaviour of the Aim-TTI.

 

5. Conclusions

 

The result of the conducted experiments were not too surprising. The AIM-TTI QPX750SP produced the best result in almost all tests, voltage and power ranges.

 

The other "bench" / "lab" style power supply the, the RD Tech DPH5005 is a bit worst compared to the AIM-TTI. But, given its price (~ $50), I think is well in the acceptable level.

 

The "brick" / "adapter" style power supplies produced varying results in terms of output ripple and noise.

 

Arduino style boards are an option to power low power electronics. The downside of these is the lack of protections, lab supplies usually have.

 

The next blog post will include some experiments with dynamic loads. We will also take a look at what protection mechanisms lab supplies usually have.

 

Hope you enjoyed this update!