Atten APS3003S/3005S Variable Power Supply – Review and Teardown

The Atten APS 3003S/3005S is an entry-level power supply designed for use by hobbyist and by students in school labs. The two models, the APS3003S and APS 3005S are almost exactly the same, but there are slight design differences and specs between the pair. The 3003S features a maximum current output of 3A and the fan is constantly on. The 3005S model features a maximum current output of 5A and the fan is controlled by a thermostat, which monitors the temperatures of the heat sinks from the two power transistors and discussed later on. This series of power supplies is usually seen sold on eBay for a relatively cheap and fair price of around about $100, which means that it is easily obtainable for both hobbyist and school use. It can output 30 Volts at 3 or 5 Amps depending on the model and it does also have some very promising specs. Every lab needs a proper power supply for many purposes such as testing electronic circuits, charging batteries and for just burning things up :D. I’ve been needing a proper power supply for my lab for a while now and after searching for a bit I decided to buy this power supply because I was unable to find another power supply in the same price range with the specs that this unit offered. Now after receiving and using the power supply for a bit, I am able to provide a full review and a small teardown of the supply.

Feel and looks.

Here is an image of the power supply, front, side, and back. The instrument measures 240X150X150mm and weights 4 kg, which is reasonable, but I wouldn’t carry it around with me everywhere. The footprint of the instrument on the workbench is very nice as it is basically a rectangular prism it would fit very nicely with the other instrument on the workbench. Another thing to take into account is that since the ventilation is on the sides and on the back, you are easily able to stack your other test equipment on the flat top of this power supply, but be sure to leave some space and the sides and at the back of it for ventilation.

As spotted by Electro Elvis, there is a strange 3A fuse on the back of the instrument even though it is able to provide a full 5A. Upon checking the fuse at the back I can confirm that the fuse is 3A, I currently have no problem with this as I have yet to come for a need of 5A of current. The design of the device is relatively simple and the looks are nothing to write home about. The power supply casing is made from some kind of metal with the front panel being made from plastic. It doesn’t really feel that rugged and strong. If you were to drop it, the casing would most likely bend and get a dip in it and the plastic may crack or break. I decided not to test this just for the sake of still having a good-looking case after this review and teardown. The casing itself did not fit perfectly onto the base of the power supply and was a bit wobbly. The knobs on the front are relatively good, but there is nothing special about them. You can see the PCB through the display – a feature which I am not particularly concerned about, only the fact that it would of been done better. The output terminals are great! They are screw on and also accept banana plugs, so you can just put a normal wire into them if you don’t have any banana plugs. The final thing to comment about is the button, which turns the device on, sometimes when you press it the spring inside makes a horrible noise which echoes the sound inside the metal casing of the unit, not too pleasing. Afterwards I began to disassemble the unit to see just what was inside…

The teardown.

First step in tearing this down was to take the cover off. The cover was held by a 9 Phillips (+) screws. 4 on each side and 1 on the back of the power supply. Upon unscrewing them it didn’t really feel like they put some effort into keeping this thing together, some screws screwed off really easily and others a little bit harder, but overall they did not hold well. The screws were screwed into the metal casing and I think that they could of put a bit more effort into holding the supply together by investing in some bolts. Below is an image of the casing; it basically is just a piece of bent metal, nothing special.

Upon opening the power supply and removing the case the next thing was to remove the two supports that held the power supply together. While removing I noticed that on the front side the screw had bolts on them and not the back, I really don’t understand the logic of that. After that the front and the back got really wobbly, this tells me this it was not very well designed.

Now it was time for the analysis of the inside electronics of the device. Referring to the image above, the first thing that I noticed was the massive transformer! It took about 1/3 of the entire space of the unit. The transformer has multiple inputs and outputs, I noticed that the serial number on top of it has 3005S written in the middle of the other part of the number, this means that this is probably some kind of custom made transformer just for this specific unit.The transformer feels and looks very solid and of good quality. The next thing that I noticed is that the unit has 2 relays and 2 very large capacitors as seen in the image below.

The caps are both rated 4700uF for 63V branded “KDC” and the relay is “Tianbo” brand. I also noticed that the ground wire ran through the one side of the transformer and went out of the other side and onto the ground socket, which is located at the front of the unit. This effectively grounds the transformer in case of a fault. The image above is one side and the image below is the opposite side.

The connectors on the unit were coated with some weird form of glue which didn’t seem to be set very well and still soft. I then began to attempt to unscrew the boards off the base of the unit. I only succeeded in doing so for the front panel meter, since the connectors (coated in weird glue) were basically falling off the PCB and I was unable to remove a connection from them without breaking them. I decided not to risk it. I noticed that one of the transistors has been bent and screwed onto the panel PCB. I did not like the look of this very much as firstly the device could give off heat and affect surrounding components and the traces of the PCB. And secondly with vibration it could very well fall off. The voltage and current outputs are controlled by 4 potentiometers, 2 for coarse and 2 for fine. This is alright, although you cannot spin it all the way around and it can get nudged while the device is off and change the voltage. This may possibly destroy the device or the circuit that the supply is connected to.

Below are some images of the removal of the panel PCB.

An interesting solution was used to save money for the display of the panel. There is simply a slightly dark screen which just goes straight to the displays, it slightly covers the PCB and the trim pots on it, however they are still visible. I am not particularly too picky about this feature as I tend to ignore it, however I do think that it could’ve been done better.

With the panel is back on, I can confirm that the PCB is visible.

The next part I go to analyse is the thermal design of the device. On the unit there are 4 heat sinks and one fan. 3 heat sinks are fairly small and are attached to TO-220 package transistors, the large heat sink which is near the fan has two TO-3 package 2N3055 power transistors attached to it. The fan is activated by a KSD-01F thermostat attached to the heat sink – a feature which the 3003S model does not include.

During the inspection I noticed a very bad case of thermal design. The images below illustrate how two of the heat sinks are located near a capacitor and a transistor, the capacitor doesn’t really matter that much, however the transistor does. As the heat off the heat sink comes off it will affect the transistor and cause “thermal runaway” – a characteristic that should be given attention to when designing circuits.

Overall the component connections are good, and I have not seen any dry joints however some of the components have been soldered onto and have been bent a little bit, also not to mention that the different connectors are a little loose. The PCB design is good but nothing special. The choice of components and brand names is decent, and I cannot comment much about it. Thermal design is alright, however I am not very happy with the heat sinks near the transistors. The next bit of this review is we need to see how this performs.


The specs on the unit look very promising and there are shown below.

Rated Output Voltage
0-30V ( Adjustable )

Rated Output Current
APS 3005S : 0-5.00 A (Adjustable)

Voltage Accuracy
± 0.5% Rdg +2bit

Current Accuracy
± 0.5% Rdg +2bit

Ripple And Noise
≤ 3mvrms (Test Conditions 30V / 5A)

Load Stability
≤ 0.3% (Test Conditions 30V / 5A)

The unit also has current limiting, but not the type you may expect. When it hits the max set current it drops the voltage in order to keep the current at that set point. Basically rendering the current feature useless, besides for safety purposes. Another annoying feature of the current is the fact that the set current is not shown, you have to short the two outputs together in order to set the max current. When the unit heats up from the high current it turns the fan on, the fan is incredibly loud and I am probably going to change it soon. The voltage fine knob goes up to 2.9V and the coarse knob go up to 29.5V, making the maximum voltage, when used together, 32.4V – a very generous voltage. The knobs are a bit “touchy” and you have to play around with them for a bit to get the right voltage and current. The panel meters are affected by outside circuits, so if you connect a battery to the output terminals it will display the battery voltage if the output terminal is smaller than the battery voltage. The same goes for current. As seen in the above specs it states that the voltage accuracy and current accuracy’s on the panel meters are ± 0.5%. The panel meters display the voltage in steps of 100mV and the current in steps of 10mA, by my testing they are not very accurate and not to the ± 0.5% stated as shown below.

The relays kick in at 6.8V, 15V and finally at 22.1V. After testing the relays I began to test the rest of the specs such as the noise, overshoot and under load conditions.

Please note that this may not be an accurate representation of the actual noise from the device as when this was tested the oscilloscope was not floating.

Here is the noise at 5V

A considerable amount of noise with some large spikes at 5V, peak to peak value is 10.9mV.

The 30V seem to be even worse with the peak to peak value being 16.8mV.

Overshoot at 5V, very little, will not damage anything connected.

Overshoot at 30V, decent amount, not damaging. However little negative voltage spike for some reason.

Overshoot at 30V with a 10K load resistor across the terminals, nothing special, but a bit better than without load (above).

Changing the voltage from 30V to 20V seems to be very stable.

A very interesting result occurs when the unit is switched off and then on. When turned off there is a massive voltage spike as shown above. I decided to investigate this a bit more, by moving the waveform down so we can see just by how much the voltage overshoots.

54.8V WHAAAAT! THATS A 24.8V OVERSHOOT. INCREDIBLE! WILL DEFIANTLY KILL SOMETHING CONNECTED! Unbelievable…  After testing it is present on every voltage range, here is the closeup of the noise at 5V.

At 5v the noise still seems to be present, it lasts about 720us. As I have experience this slight pulse is not enough to destroy a circuit or a device connected, unless its really sensitive to sudden voltage changes.

The verdict.

A decent power supply which is easily repairable due to the internal construction and easily obtainable through hole parts. Will definitely suit beginner hobbyist and school labs, however I would not recommend that professionals use this. This is a power supply for just the sake of having one, the performance could definitely be improved, but for this price range it is good. If I was a bit more experienced with $100 I could of built a better power supply than this, and I would recommend building one if you have the knowledge and skills to do so. It has all the features that you want in a basic power supply, variable voltage, current limit and ground terminals. I would not go onto high current and voltage applications with this supply, but it definitely is suitable for low voltage and low current usage in beginner hobbyist and school labs. Overall I’m disappointed with this power supply and I believe that there are many improvements that could be made to it. Once again you get what you pay for.

Features: 6/10
Performance: 6/10
Design: 6/10
Build Quality: 5.5/10
Value for Money: 7.5/10
Overall: 31/50

– Mint Electronics.


, , , , ,

  1. #1 by Ig Saturation on July 10, 2012 - 12:44 am

    Nice review and photos, mint. But with the voltage spike is completely unacceptable for any purpose. The final rating, IMHO is 0/10, its useless for most purposes in electronics unless that spike is an artifact.

  2. #2 by Ig Saturation on July 10, 2012 - 12:48 am

    PS, that noise you see is not representative of that coming from the supply if you connected the probes of the Rigol to the PSU. It will pick up ground loop noise or coupled ground noise, since the scope is at earth ground. You must float the scope, not recommended, use a differential probe or a portable battery operated floating scope.

  3. #3 by mintelectronics on July 10, 2012 - 8:27 am

    Hey Saturation!
    Perhaps I agree with you about the voltage spikes, however the test was run at 30V only. Upon testing this furthermore at 15V and then 0V it appears to still be visible on the scope and still has around about the same overshoot voltage (24V), when I tested this to see whenever it was an artifact or not, which I thought was highly possible since none of my circuits connected died, the led that I hooked up to the output terminals did flash for a very tiny amount of time. I guess that the overshoot voltage is present there, however it is over so quickly that nothing connected gets destroyed. Thanks for pointing out to me about the noise, I added a note to the post saying so 🙂

  4. #4 by Monkeh on July 17, 2012 - 12:31 am

    The 3A fuse is for the AC input, not the DC output. 3A at 240VAC is 720W. Which is 24A at 30V.

    • #5 by mintelectronics on July 17, 2012 - 5:38 pm

      Hello Monkeh!
      Are you sure this is correct? I always thought fuses were always current rated ignoring the voltage rating whenever it is AC or DC.

      • #6 by Monkeh on July 17, 2012 - 10:18 pm

        Yes, I’m quite sure. Ohm’s law. 3A at 240V is not the same as 3A at 30V.

      • #7 by mintelectronics on July 18, 2012 - 10:42 pm

        Seems like your reasoning is correct. Then the next thing that puzzles me is why they have a fuse for 3A. Even though I doubt that there are 24A delivered to the internal construction, the transformer is probably some current limiting one. Thanks for pointing that out for me about the fuse, much appreciated! 😀

  5. #8 by Monkeh on July 18, 2012 - 11:37 pm

    Because 3A is an appropriate fuse for the 240VAC input?

  6. #9 by Monkeh on July 19, 2012 - 1:35 am

    And the only limiting that transformer is going to do is when it melts. That’s a very ordinary, cheap EI-lamination transformer.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

%d bloggers like this: