One of the most basic pieces of test equipment in the arsenal of an Electrical Engineer is the humble benchtop power supply. While I have gotten away with using old plugpacks, AT/ATX power supplies as surrogates, they have never had the flexibility of continuously variable voltage and current limiting and this sometimes complicates servicing and testing – a short circuit could easily and quickly liberate heaps of smoke.
As part of my PhD, I needed a benchtop power supply for my experiments. Of course, proper supplies from test equipment companies like GWInstek command a fair premium – so I thought I’d be cheap and buy one from Jaycar. Big mistake.
The unit in question was the MP3087 (linear) dual rail tracking laboratory power supply. It is by no means a laboratory power supply. The first unit had a second rail that failed to deliver its rated power as it constantly chattered its relays when asked to deliver more than 0.7A. This was returned, and a replacement took two weeks to be sent – and was lost by the courier. A fresh unit was dispatched and this one was faulty out of the box with the left rail stuck at 30v and no relay action as if it had a welded relay. And there was a screw securing the transformer inside that was loose and almost falling out.
This cost a fair amount of time for my PhD – my supervisors weren’t happy, and neither was I. At this stage, we needed to obtain something which could meet the requirement of at least having two rails of up to 30v, adjustable current limit of at least 3A per rail. Surveying the equipment suppliers showed that the GWInstek supplies had a four week lead time, so I went elsewhere.
This resulted in sourcing two Manson HCS-3102’s at a total cost around $350 delivered for both units. Manson is a Hong Kong company, and they have been manufacturing switching power supplies for Dick Smith and Jaycar for a while and generally work fairly well. In fact, I already own a Manson supply, branded Dick Smith Electronics, 3-30v at 20A (non current limited) and generally they’re “good enough“.
The unit arrived in a plain cardboard box with a small label on the side identifying the power supply contained within. It is nice to see that this has C-tick approval as well.
The box is generously filled with goodies – a CD containing the software, individualized test report, manual, USB cable, IEC power cable and remote control connector.
A detailed look at the test reports seems that they take great care to qualify each and every unit coming out of their factory:
The power supply is absolutely tiny compared to what I expected – here’s its size compared to a DDR DIMM:
The front has an old-fashioned physical toggle power switch which controls the primary. There are two rotary encoders, for changing voltage and current which operate by push toggle between fine (0.1v/A steps) and coarse (1v/1A) steps.The front panel AUX output can output the whole current rating of the supply, but has no binding posts. There’s LEDs to indicate the use of rear control (remote or USB) and CV/CC modes. The front panel meters toggle between displaying the set-voltage and set-current when adjusting, and the measured voltage and current when not adjusting.
There is a vent for air only on one side, and the front panel clips into a tiny slot in the side as well which will make taking the lid off a bit of a challenge:
The rear of the supply shows the 80mm fan with chromed grille, made by Sunon (decent quality) and the mode switches, remote control connector, USB and main output connectors, and fused IEC input.
Specifications are provided in the brochure, but my summary is:
- 1-36v output (adjustable to 0.1v steps) with 50mV load regulation (10-100%) and 20mV line regulation (90-264VAC) with floating terminals.
- Ripple and Noise: 5mV RMS, 50mV p-p, 30mA current noise.
- 0-5A current limited output (adjustable to 0.1A steps) with 100mA load regulation (10-90% rated volts), 50mA line regulation (90-264VAC)
- Dual 3-digit LED display of voltage and current +/- 0.2% + 3 counts
- Switch-Mode design with universal 100-240v 50/60Hz AC input at 82.9% efficiency.
- Switching Frequency 100-120khz.
- Thermostatically controlled fan (0-100%).
- Over-voltage, Over-load, Over-temperature protections.
- 200x90x208mm size with 2.4kg weight.
- Rotary Encoder Knob adjustment with push-toggle between fine and coarse adjust.
- Remote control (on/off, voltage and current limits) by 0-5V control signals and over USB with provided software and command set.
- Supports cyclic test program development with provided software.
- 3 preset voltage-current combinations available.
- Front panel AUX output and rear panel MAIN output with binding posts.
The complete manual can be found on Manson’s site here. The output voltage steps are a bit “coarse”, but the use of rotary encoder allows for precise changes without concerns about worn out pots or nudging pots carefully. The rotary encoder itself is clicky, and I think it’s not likely it’s an optical one, so it could probably wear out prematurely (but no worse than potentiometers). It’s not a big issue since this supply offers USB and remote programmability – so even if the front panel is completely borked, it’s still great. The main outputs being on the back is a bit unusual – and the binding posts are without holes (as is a “Manson specialty”). Presets make it quick to recall settings – say 5v/500mA for powering USB devices. The low weight and thermostatic fan is a blessing – no longer is there a continuous hum from the cooling fan – it only comes on when needed. The high efficiency makes this even less likely to be the case. But it is a switch-mode supply, so the noise from the supply could upset sensitive devices like radios.
I’m sure you’ll agree, it’s not really quite on par with the expensive “fully programmable” power supplies that are complicated enough to be wave generators – but at under $180AUD delivered, it really offers a good range of features and looks to be quite suitable for the hobbyist.
Inspecting the Internals
It’s not my sort of review if it doesn’t at least take a peek under the covers, so lets take off the eight Phillips head screws and get inside – no warranty seals whatsoever. For those following at home – do NOT try this at home – at least, not without a lot of caution since the caps could be charged with a lethal charge. Unfortunately, the cover needed a bit of prying to separate itself from the front plastic cover – but here’s the beast in its glory.
Near the rear IEC connector, we can see what looks to be a passive PFC and line conditioning hardware. There are large primary and secondary caps scattered around – as well as one near the rear main terminal outputs. Interesting is just how many transformers there are. Wow.
Staring at it from the secondary side – we can see neatly tied up wiring looms. One connects the rear preset switches to the front panel where the controls are, another which connects the rear terminals bypassed by a large electrolytic and two smaller ceramic caps/ Ferrite bead suppression is used for both front and back outputs – definitely some attention to EMI here. The small USB serial converter is visible in the bottom as an add-on.
Here’s the view from the primary side – and oh no. CapXon capacitors. Eugh. They’re a bit questionable and have been considered a poor quality cap. All electrolytics in this supply are CapXon – that’s the price one pays for cheap supplies. Aside from that, it’s clear the front panel switch controls the primary AC – which then enters a hefty choke and bridge rectifier on the right side of the image. At least it doesn’t look like the cheaper ATX PSUs where even the chokes are omitted from the board and the components are populated at every-which angle.
Looking at the front panel board – this is basically where all the control takes place. The firmware is Rev 2.4 – and it’s running on an Atmel ATMEGA64A which is kinda like a “smaller” cousin to the Arduino Mega’s 2560. There are two variable pots and a push button inside – maybe this is for calibration and test purposes but I wasn’t going to break my supply finding out.
It was hard to get a look at most of the semiconductor devices on it owing to the tight construction of the power supply – but we see on the right most, an ON Semiconductor device (which is good quality). The other two, I thought were Texas Instruments, until I looked at the shape carefully. That’s not the shape of Texas! It’s a logo with an ‘i’ in it, and as far as I can tell, it’s Vishay Siliconix which is also a good brand! So it looks like these guys know the importance of using properly rated, branded semiconductors, but not capacitors. Puzzling.
Using the Supply
While the specs are very good, the real test is how it performs when it’s being used. I’ll try not to say too much – lets go with a simple pro and cons breakdown.
- High efficiency – less heat, less power waste.
- Low weight – easy to move for those without permanent lab spaces.
- Thermostatic fan – no need for annoying fan noise all the time.
- Good voltage range and current range – 36v is nice and high, and 5A is better than 3A that you would expect from these types of supply, stability seems pretty good, with microprocessor control and built-in self testing. The fact it has current limiting in this price range is quite nice, and you’re not bundled into buying two rails before you need it. And when you do, the floating terminals allow you to series-up the supplies if you need to.
- Presets – easy way to fix certain parameters to avoid errors and damage to equipment when used as a permanent supply.
- Remote and USB control – great to have an alternative to using the front panel which allows for more sophisticated control possibilities.
- Protection mechanisms – this has extensive protection mechanisms so it’s unlikely it will hurt itself.
- No skimping on the accessories in the box – it’s quite annoying when other manufacturers omit cables.
- Build Quality – the metal casing, and the plastic front panel feels solid enough and the internals are neat. Individual test reports provided give confidence that they have decent QC procedures from the factory.
- Price – definitely a big plus.
- Front panel ergonomics – it’s fiddly having the push toggle between fine and coarse settings. Sometimes you don’t know till you’ve changed the output, and the LED dim/bright indications are not that clear due to some level of flicker in the display. Better keying it in remotely if you want to be sure. Also the sharing of the same displays to show the set point and the present reading is a bit annoying. The rotary encoders may have a limited lifetime.
- Granularity of settings – while it is a “lab” supply, the voltage steps and current steps at 0.1v and 0.01A are a bit imprecise. Of course, one will have to pay more to get more, but at least an analog supply can “in theory” be regulated to an infinite number of voltage steps. The other thing is that the supply does not go down to 0v – so working below 1v isn’t really a possibility (although it’s rare that one would need such a low voltage).
- Memory – the power supply will start up at your last settings which is fine, but you have to power it up first, check the value, power down and attach load (or use an external switch) since there’s no way to disable the output from the front panel. It would be so easy if they added a push button like in some of their other models. You can do it in the software and by command over USB.
- CapXon capacitors – these aren’t exactly a great brand, and are marginal quality. Depending on how much you use the supply, and how demanding you are in regards to voltage ripple, this might not be a problem for the intended lifetime, but replacing these won’t be that cheap of an exercise.
- Noise – due to the switching frequency which is low, it seems that there are some subharmonics from the supply which become audible in certain voltage ranges as an extremely high pitched whistling/whining which varies according to load. Better supplies are perfectly inaudible – this one needs a bit more refinement when it comes to this.
- No spare fuses – while you shouldn’t blow a primary fuse under all operating circumstances, most other manufacturers do include one, two or three spare fuses just in case. Nothing worse than having equipment sitting and waiting for the right fuse to repair it.
- Self test routine – every time the unit powers up, it takes a good 10 seconds or so to run its self tests. Switching the output on and off by using the toggle switch which switches the primary is a tedious exercise that really shouldn’t be needed. Fit an output on/off switch!
Playing with the Software
Now we get to the fun part. Remember when I said this was programmable? Well, lets take a look at that now. The USB connection is implemented with a Silicon Labs CP210x USB to Serial Converter. Drivers are provided on the CD, however, more updated drivers are available at Silicon Lab’s website which may resolve compatibility issues.
You then need to install the HCS version 0.9 software from the CD (not available on their website, curiously enough) which has an outdated wizard and looks like it was made for Windows 95. Running it under Windows 7 64-bit edition, I needed to run the HCS tool with Administrator rights and with a COM port number between 1-4.
I was greeted with being able to control the supply with their software. While it is rudimentary, for simple tasks, it definitely is useful.
The settings tab allows us to select the COM port, the data sampling time (from 1 second upwards) and control the upper voltage and current limit selectable by the front panel. This can be quite a good “safety” feature in case someone starts turning the knobs too much. The left hand side allows you to instantaneously commit a new voltage and current setting to the supply, while showing the current measured voltage and current, and multiplying them for power. There’s also the output on/off display which, when clicked, allows you to toggle the output to off. So the supply is capable – but the front panel just doesn’t have the button.
One of the useful features is to make a timed cycle program. If you’re testing something and you need to change voltages and current limits or turn things on and off at intervals, this allows you to set 20 entries for voltage and current combinations, with a time of up to 99 minutes 59 seconds for that combination, and a number of repetitions count.
Sure, 20 entries can be quite limiting, and the lack of “on-condition” flow-chart style command is a bit annoying – but this program is version 0.9. Unfortunately, I don’t expect Manson to do much more on it – so making the most of the supply will be your responsibility!
The next tab allows you to change the rear panel presets for the power supply. This allows you to bypass the annoying front panel procedure of pushing knobs etc etc. Great!
The more interesting part is the Data Logging feature, which allows you to sample the front panel readings up to once a second and plot it.
In this case, I left the sampling at 3 second intervals – but you can see me playing with changing voltage levels, then topping the voltage out, and then bringing it down to 1v. You can literally see the caps discharging right before your eyes.
To make it a bit more interesting, I changed the sampling time to one second, and changed the scales. I attached an 8cm case fan to the outputs and stepped from 1v to 10v in 0.5v steps at 10 second intervals and you can see it plot the measured voltage, current and computed power. Quite neat.
Of course, the voltages during transitions can be somewhat rough – that’s just the nature of how the beast works.
But here’s the most interesting part – the fact that the supply has a very simple TEXT based protocol and it is fully detailed in the manual. I’ve copied it here just to give you an idea how SIMPLE it really is.
Just opening Hyperterminal and keying in the commands gives you the expected results. You can thus write Matlab scripts, C programs etc to overcome limitations in the software provided and make the most of the programmability! How neat is that. One thing you do have to be aware of if you’re commanding it by hand is that the device doesn’t echo your characters back – so you may see no characters appear, but it will respond correctly. The other thing is that replies have <CR> but no <LF> which may cause Hyperterminal (or whatever terminal program you’re using) to overwrite responses due to the multiple lined response nature (usually a numeric string<CR>OK). So twiddle your terminal emulator settings to append <LF> to <CR>.
What we have here is a fairly inexpensive power supply with a lot of good features to tempt hobbyists. The programmability of the supply isn’t fully exploited by the provided software, but it isn’t really a difficult thing to improve upon it if necessary. The ergonomics of the supply is complicated by the shared displays, coarse adjustment steps of 0.1v/0.01A (possibly improvable through the use of the rear analog control port) and dual mode coarse/fine dials with a lack of output on-off switching on the front panel. But by far, the supply seems to be let down by mediocre quality software and capacitors and a whine/whistle. This happens on both supplies, but maybe I am being a bit too picky. After all, it’s a supply I’m very happy to have in my lab because it really does almost everything at a price I can afford. For those who aren’t using it day in and out or at high loads continually, it’s unlikely that the capacitors will be a problem anyway, and at this price, you could probably afford to pick up another anyway. Almost all other brands will have you paying more, for less.