Dp832a Manual

[Giordano Thibault]

Electronicgadgets need power. And if the device in question doesn't come witha power supply or it's a repair job and you suspect that the PSU is part of theproblem, then a lab power supply comes into play. I've had a VoltcraftPS405-pro power supply for ages and a Manson NSP2050 switching power supply isalso hiding somewhere. Both are OK, but I felt the urge to treat myself to anew one that is properly programmable and offers remote control and a betteruser experience.

Siglent and Rigol fall into the "Respectable Chinese Vendors" category, whileOwon is known more as el cheapo supplier (although I found some very positivereviews of the PSUs mentioned above). UNI-T's reputation seems to be mixed:Their multimeters are pretty popular but the community doesn't seem to likeUNI-T as well as Rigol or Siglent. Nevertheless, after hours of studying datasheets, manuals and reviews, UNI-T's offerings appealed to me most:

The device makes a solid impression. The keys are made of rubber but have aclearly noticeable click to them. The knob clicks nicely and precisely. The displayhas glare, but that doesn't really bother me.

The display always indicates whether we are in constant voltage (CV) orconstant current mode (CC). In addition, all three channels have an LED abovethe respective sockets, which lights up in green (CV) or red (CC), as is commonon traditional power supplies. I think that's great, because it makes it somuch easier to see when the current limitation takes effect on a channel.

When channels 1 and 2 are configured to be in parallel or in series, thedisplay prominently shows a small picture of the appropriate wiring in therespective configuration. What appears to be missing is a classical trackingmode in which ch1 and ch2 are tied together and controlled jointly. You can kindof get that in serial mode if you use all four output sockets but the displaywill only show the added voltage and, of course, both voltages are identical inthis mode. I guess a true tracking mode could be added in firmware so maybethere is hope to get that in the future.

Now let's put some load on it. So we set the power supply to 20V and theelectronic load to 50W. After 10 minutes I switched off the load, but thepower supply fan hardly seemed any louder. That was probably not enough load toreally put stress on it.

It seems to be decent, but the installation is very large and takes forever. It also seemed to me that the power supply reacts to the software rather sluggishly.But that that may be due to the overhead from the virtual windows.

When you want to control more complex scenarios it's often easier to write ascript than fiddle with vendor software. The device has three ports you can useto connect it to a computer: USB, RS232 and ethernet. All three will give you thecapability to send SCPI commands. A programming (SCPI command) manual isavailable from the manufacturer's web site.

We already played with the Windows software above, but I prefer LINUX and would love to use Python for scripting. So I looked around and discoveredPyVISA. That looks quite promising. So let's install it:

Unfortunately, I could not find a firmware update function in the userinterface or the manual. And there are no firmware files on the UNI-T web sitefor the instrument. That's a pitty because I think getting firmware updates fora $600 power supply isn't asking too much.

Excellent! I did let the support people know that I am very happy with this andI would be even more excited if this information and firmware files wereavailable from their web site. According to their reply, UNI-T is planning toact on this some time in August. Wouldn't that be cool?

In order to prevent wasting too much energy in the voltage regulator, lab powersupplies have multiple taps at different voltages and choose an appropriate onedepending on the set output voltage. So let's try and find out how many thereare and at which voltages. For that, I started at 0V and slowly cranked up thevoltage until I could hear the telltale click of a relay.

The exact switching voltages differ depending on where you are coming from. Ifyou are increasing the voltage, the switch will be at a slightly different pointthan when turning down the voltage (hysteresis). And that makes a lot of sensebecause you don't want to get your power supply in a situation where it keepsswitching just because you are operating close to a tap voltage.

First, we'll check if there is any voltage present at the sockets when thechannels are switched off. The above mentioned thread on EEVblog forum says there are small negative voltages present. Let's check:

I see no overshoot in channels 1 and 2 while there is a little bit of it onchannel 3 at lower voltages. Channels 1 and 2 also show some struggling to getto the target voltage in some cases. I have no idea what is going on there.However, I don't know if that is the power supply's fault or if the electronicload has a part in it. So I'll repeat the experiment using a power resistor(12Ω, 100W) instead of the electronic load.

Next, we'll test how well the supply can handle changing loads. I programmed theelectronic load to alternate between 1A and 4.5A every 2 seconds (transientmode). The supply is set to 30V. I didn't really see much on the oscilloscope soI switched to the best multimeter I have (Siglent SDM3065X; aperture 0.5 PLC).

Good power supplies have a current limiting mode so that you can set amaximum current and the device will quickly lower the output voltage when thelimit is exceeded until we are back at the current limit. That way, we canprotect our circuit from over current damage in case it is faulty and tries todraw too much power. The faster our power supply manages to get back to normal,the better.

For testing this feature, I set the current limit to a certain value and ran itat a low current (load 1) before suddenly increasing the load above the currentlimit (load 2). On the oscilloscope, we observe the voltage (Ch1, yellow) andthe current (Ch2, magenta, via a Hantek CC-65 current probe).

Now we set the current limiter to 20mA and the voltage to full throttle (30V). Turnon the channel and see what happens. The power supply immediately goes into CC modeand the LED glows nicely at 1.9V, 20mA.

In addition to current limiting, the device has an over current protection modein which it will inactivate the channel after the maximum current was exceeded.To test it, we will use the same setup as above and set both limits to 500mA.

For this experiment, we will set various different voltages and loads (ET5410A+)without a current limit. I am using a new Siglent SDM3065X multimeter tomeasure the actual output. As this is a pretty precise measurement, I turnedon all instruments involved and allowed them to sit and equilibrate for aboutan hour at 22.2 C (Fluke 87V with a k-type probe).

Now, we plot the deviation (in mV) of the set/readback voltages (V.set /V.psu) against the measured voltage (V.dmm). The green lines represent thespecification limits and the dots are color coded by the current reported bythe PSU (A.psu)

In order to analyze current accuracy, we will take measurements at variousvoltage/current settings while setting the electronic load to 6A so that itwill always try to consume more power than the limit allows, thus putting thePSU into cc mode.

The term Ripple refers to the periodic AC signal that originates from the50/60Hz line frequency (or the internal switching frequency in case of aswitching supply). Noise is all the remaining junk that comes on top of it.

Not good. Apparently my probe is enough of an antenna to pick up quite someripple and noise out of thin air. And, of course, the scope itself also hassome noise. All in all we already have 2mVpp which is the spec limit of thePSU... So I connected a BNC-adapter to the probe and plugged that into a BNCto Banana adapter which is then plugged into the scope. If we are lucky, thathas enough shielding to get rid of most of the crap.

At the settings above, ripple&noise are drowning in the huge signals. Note thefactor 1000 difference in sensitivity between Ch1/Ch2 and the difference trace.The probes just receive too much crap out of thin air and the adc simplydoesn't resolve the noise at these settings. Increasing sensitivity on thechannels doesn't help either, because then they go off the screen (clipping)and there is nothing left for the math function to work with. Maybe I can bodgetogether some kind of shielding, later.

So at this point I give up and admit that I am not sufficiently equipped (orknowledgable) to really quantify ripple & noise of the supply any better.I guess I'd need a really low noise differential amplifier...

The power supply is easy to operate and has many useful features. All aspectsof performance that I tested look really good. The PSU is easy to program viabuttons and menus and remote control via SCPI was easy to set up. I was goingto complain about the lack of firmware updates but as I said above that hasbeen fixed. It would be nice if a future firmware update added tracking mode tothe feature list. And compared to the Rigol DP832A, the manual of the UNI-Tdevice is not that great but I managed to figure everything out without muchproblems.

I recently used relays and the digital I/Os of the DP832 to control the polarity of the outputs. The circuit needs an external 24VDC supply to activate the relays. This makes it a bit impractical to use. I also plan to design a circuit that would automatically calibrate the DP832 also using relays.

The USB host port could be a possible solution. It would be interesting to see if it can supply enough power to activate the relays. Unfortunately the DP800 Series user manual does not provide enough information about the type of the USB port. I assume it is a normal USB 2.0 port capable of supplying 0.5A of current.

I tested the port with 600mA for more than an hour without any interruptions. The voltage was 4.55V but I am not sure that it was not the voltage drop on the cable. In any case we should be able to get 2.5W out of the port continuously. That should be enough for at least 10 5V micro-relays.

3a8082e126