UNI-T UDB3000S Series: UDP3305S / UDP3305S-E

[gabrieLL]

I am on the same boat at this time. Can't decide between Rigol DP832 and UDP3305S-E. 

[Pfriemler]

You know about all the downsides the RIGOL has? If you can live with them, you don't have to worry. But be beware of them, as it will be your tool, not ours.
I went ashore again - I am expecting my UNI-T at home tomorrow and may then be able to provide a few more arguments.

[gabrieLL]

Haven't found anything specific against RIGOL so far. Only thing bothers me at them moment is the <3A and monochrome display and possibly a loud fan , higher price +80€.  Could you elaborate some more about the downsides of the RIGOL ? Thank you

my understanding of the rigol is :
DP832 is on the market for +-10years which is a good thing and still sells. HW issues were ironed out with board revisions, SW too. Components inside should be proven to work reliably. Lots of documentation , community , support , higher chance of self repair success or aftersales repair at all.

UNI-T
Nothing mentioned above. Only that it is nicer, more up to date. <5A , color display, numpad.
- Not sure about the internals. There is zero in-depth teardown of the UDP3305.

[faktorqm]

Hey gabrieLL, check the attached file. It's a comparison (excel sheet) with all the interesting parameters of each model of the power supplies in the same price level - quality. Please ignore the colors.

I will go for the UNI-T because I like it very much.

As you can see in the excel, on Rigol PS the ground between the channel 2 and 3 is shared, (which I found it unforgivable, but that's my personal opinion) and the voltage accuracy is better on the UNI-T S-E model. The current accuracy is slighty better (0.15 vs 0.2), the rigol is 3A per channel, UNIT is 5A. For me, it has more value the current rather tan voltage, again, my personal opinion. As last but not least, UNIT has a fourth channel 5A 2A fixed which personally, I really appreciate.

Regards!

[Pfriemler]

My first impression is excellent, not as big as I feared, with a permanently running fan, but much quieter than even the latest Korad KA3005P (the Welectron edition with the green-red display), good voltage accuracy even under full load, almost intuitive operation. I really like the calculator-like numpad (not the Rigol ring one). I was attracted by the series/parallel switching at the push of a button, the "rules" (switch off when output current is below xy) etc. and the easy-to-use additional functions (programming a voltage ramp etc. or quickly testing 100 on/off switches is quick without having to start the computer) or the built-in recorder that @thePhil described here.
Firmware 1.10 bootloader 1.06 is installed. I wonder why UNI-T does not publish the 1.11 Tony Albus offers here.
Yes, it may be that the Rigols' firmware and hardware have matured over time. There are definitely more users and therefore more potential advisors in the event of a fault. I hope the UNI-T users stay in touch here in case something happens to us after the first two years 

[Pfriemler]

Hmmpf. Looks like I will not sell my KORAD KA3005P ...

First: Set to 5 volts 10 mA and shorten the outputs with a 1R. This will result in spikes - the more output capacity, the more the curve is wide.

This is a comparison between my KORAD KA300P (dark purple), UNI-T's CH2 (light purple) and CH3 (green). Nearly 5A spikes. The KORAD is way faster than the UNI-T.

To get an idea how much the effective output capacity is: 10V shortened with 1R, KORAD vs UNI-T@CH1 vs a 10 volts charged 450 µF (measured capacity):

Damned. Did I cry about the 100 µF at the OWON ODP? I did.

So, if I attach an LED to an active output >10V, I may kill it? Checked it with an 230 mA LED: remarkable flashes with the UNI-T, barely noticable with the KORAD.


Second:
off > on behaviour 5V 3A with no load. As stated in @thePhils blog, the UNI-T power channels are rather slow with a small spike in the beginning, whereas Channel 3 is much faster with an app. 10% overshoot.


Well, the KORAD shows remarkable stairs ... no stairs at all if the current limit is set to max (5.1A).

Now we try 3.3V @ 0,1A:

Look at the beaty overshoot the KORAD (purple) makes. And the UNI-T does, too (yellow).

From the frying pan into the fire?

edit2: I forgot:

... Channel 2 outputs a voltage of -243 mV when turned off. In parallel mode this drops to 127 mV, possibly because Ch1 takes some of it. Ch1 is at -11 mV.

Nothing here, looks very fine: The CH1 and CH2 output voltage @OFF is both -2.2 mV with my UNI-T UDP3305S-E, -1.7mV if switched on and set to 0.00V

[faktorqm]

Good test! In which real life scenario that could be a problem...

a) the overshoot for a very short time period? 10% overshoot you stated in your post.
b) The active fuse time: In the first millisecond all three are below 50% of the voltage (Depending on the board, 1ms could be an eternity) but it looks like not a bad time. Isn't it?

The rise time but ch2 could be lower ch3 is behaving like it should. Thank you! Regards!

[Pfriemler]

Overshoots when switching a channel on is always a problem with cheapo regulated psu, as far as I experienced. A Basetech BT-305 peaks up to 4.8 volts when switching on to (preset) 3.3 volts with no load connected - if you set the current limit down to 0.2A, the overshoot is gone, although it takes about 100 ms to load the (from my point of view horrible big) 660 µF output damping capacitor.
Strange: Both Korad and UNI-T CH1+2 do not tend to any overshoot with no current limitation, but do when I lower the current limit. That's what I really do not expect: "be safe with low current limits" is not true here. But they are moderate here, I can live with that.

Active and fast current limitation may save boards with failures under test. I did not examine the behavoiur on a short circuit (or heavy load) when switching on the channel - here it shows if you plug in or (mechanical) switch on a device connected to an active output.
Even if the output stages reacts in microseconds, the energy stored in the output capacitors may blow something to heaven, thus a very low output capacity is important. I do not see any on the KORAD circuitry (although I think ther is some), my Voltcraft DPS2010 has something ~50 µF, the Basetec mentioned above 2x330 µF, the OWONs ODP board photos let me assume ~100 µF (although I have no idea how they would do in my test).
Looking to the curves in the first shot: Peak current is limited only by the 1R resistor. Whe have a full unlimited current of 5 amps for ~600 (CH1) and ~200 (CH3) microseconds before the current limitation kicks in, the discharge of a (presumed) output capacitor begins after that - but almost no delay at the Korad. Due to the decreasing, Korad and UNI-T CH1/CH2 may have a similar output capacity (pic2). Compared to the green cap discharge curve, it may be something ~100-200 µF. CH3 decreases much faster, but puts more even more energy to the DUT as the KORAD does.

Overall: It is not that bad. But I am impressed how fast(er) the KORAD acts.

It would be very interesting to see how a Siglent or Rigol react here: Simply set em to 5V 10mA, put a 1 ohms resistor on the active output and look how the voltage decreases.

[dreamcat4]

in the realm of power tools, and other industrial tools we have the concept of 'soft start' feature. so why cannot be possible a choice here for the on/off behaviour as a setting? for a slower ramp up of the voltage?

because for some certain endurance, and power cycling stress tests, then you do wants to go straight / full on. then full off very rapidly. or to bring up and switch on a circuit 'fast enough' for a clean boot or initialization of a digital circuit.

however at other times... then you might be testing some sensitive component. lets take driving a fet as an example. but it could be any sensitive things such as leds or whatever unprotected circuit. then we want the option for a slower ramping up (the voltage). not to overshoot. and with the cc on

this seems like reasonable feature to be asking of a bench top lab psus? given that not everybody can afford something too much expensive / higher quality.

[Pfriemler]

Do not misunderstand me: Of course it's very easy to setup a smooth voltage ramp, and even without that: If you know what the device does, you can set it to your needs.
Connecting a LED to an active channel will show the spikes as short flashes, but if I connect the LED to the inactive channel (set to 5V 10mA) and switch the channel on/off, there are no visible spikes - this works very fine as it should.
I will check that later and put the pic in here..

I think (real) professional PSU behave much better with that. But again: connecting an LED to an active KORAD set on 30V (!) 10mA let the LED flash barely noticeable. And it's really cheap, compared to the known brands.

[dreamcat4]

ah ok then! thanks so much for clarifying much appreciated 

[BillyO]

Wow, I wasn't aware of that. Good to know. I don't think that's even legal.

Not sure what it's like on your side of the pond, but here in Canada there are statutory warranties.  I believe it is a full year on new goods.  If the manufacturer does not have a means to handle the warranty claims, it falls back on the retailer.  If the retailer is off-shore all bets are off.  I think you might find it's pretty similar throughout the 1st world.  A good manufacturer will honor their equipment warranties regardless.  Korad is such a company.   They don't have a service center in North America, but will do everything they can to help you out.  UNI-T on the other hand will just ignore you.  I have personal experience with both of these companies.  A power supply from Korad and a Oscilloscope and a benchtop DMM.  I will never buy anything from UNI-T again that has a value over $50.  Both the UNI-T scope and DMM sit as reminders in the back of my basement, dead and abandoned by the company that made them.

Siglent and Rigol are in a different league.  Both have service centers around the world and offer great support.


Edit:
I notice UNI-T now have a location in NA and EU:

Uni-Trend Technology US INC.
3171 Mercer Ave Suite 104 Bellingham WA, 98225

Uni-Trend Technology EU GmbH
Affinger Str. 12, 86167, Augsburg (Munich area in Germany)


Well, maybe things have improved since the dark ages.  We need to get some success stories then.

Edit #2:
Just sent the NA location a general query.  We'll see if they respond.

Just an update on this.  I sent the NA location 3 separate queries requesting service information and I got no reply whatsoever.   So apparently hanging a hand painted shingle on an empty unit and taking a picture is what UNI-T think is all they need to do.  Buyer beware.

[faktorqm]

Thank you very much for your answer. Very clarifying. I hope anyone with the Rigol 832A or Siglent SPD3303X(-E) can post the same test to compare the behaviour.

I think I will end buying the UNI-T, at the end of the day, it has a lot of things I appreciate and it's affordable for me. Maybe not the best one but what I want is to don't buy a problematic device.

Regarding support, I really don't care I had a lot of problems with several different devices and always they wash their hands. Bah. Regards!

[Veteran68]

I have a SPD3303X-E (upgraded to SPD3303X) and can try to create the same test tonight if I get time.

[Pfriemler]

I have a SPD3303X-E (upgraded to SPD3303X) and can try to create the same test tonight if I get time.

Please, do!

Did some futher checks ... as mentioned above: connected a LED, set to 5V 20mA, switch channel on - absolutely no overshoot of current.

Next issue: draining external sources. I did check this with the KORAD KA3005P and the UNI-T channels, coupled in parallel with cables.

Korad 5V 200mA, UNI-T CH1 5V 200mA, both channels on: Korad reads 0mA, UNI-T 10mA. (UNI-T's output voltage is mV higher than the KORAD's)
If you increase voltage on UNI-T, KORAD voltage reading follows, current remains 10mA. Looks like the KORAD is draining 10mA constantly, also when ouput is switched off.
If you increase voltage on KORAD, UNI-T CH1 loads the Korad with 154mA, CH3 with 15mA. Voltage Readings show actual values. BTW: CH3 voltage reading limit is 7.38 volts.

I do not like this.. This is not the behaviour of an 1-quadrant psu.
So, putting a diode in series when charging batterys seems to be necessary like with the RIGOLs, otherwise the battery will be discharged. And we do not know if the UNI-T will bear this a long time.

So, the next task to check SIGLENT: couple both channels - do they load each other when set to different voltages, and how much?

[faktorqm]

I have a SPD3303X-E (upgraded to SPD3303X) and can try to create the same test tonight if I get time.

That would be great sir! I you can do the same test it would be great. We don't have a rush, so take it easy! And thanks for your time in advance!

[Pfriemler]

Next issue: draining external sources. ... UNI-T CH1 loads the Korad with 154mA, CH3 with 15mA...

Just to clarify this: It is not that easy...

So I did a general look how my power supply units act when an external voltage is applied to the output terminals, regarding if it's under/above the settings, channel output on or off.
First: "PEMENOL DC-DC USB Step Up/Down Modul" - a small chinese board with USB-in, QC/PD support or simply 4-12V, out adjustable 1..30V 0..2A max 15 watts (2023 ~15 EUR)
wether on or off, depending on the setting of the output voltage - I assume it's simply the output divider for setting the voltage ... acts like a resistor 4,7k(5V), 6,5k(10V), 8,2k(20V), linear behaviour,
maximum 6,44 mA in 5V setting with 30V external, ~200 mW in total. May damage the (small) resistor.

Second: "MATRIX MPS3206", cheapo switching PSU (2022 ~80 EUR). On or off, it acts like a resistor of 2k, drawing a maximum of 15 mA @ 30V = 0,5 watts. Nothing spectacular!
No difference if output is switched on or off.

Third: "KORAD KA3005P", well known. As with the "MATRIX",similar to a 480R resistor, i.e. 20,6mA@10V=206mW, slightly more current at 20V: 49mA@20V=0,98W, ~400R. Did not test above to avoid internal damages.
No difference if output is switched on or off.

Fourth: "Voltcraft DPS 2010PFC", similar to "McVoice SNNT-2010pc" etc, switching psu with linear regulater (~3V overhead)
Well, this drains much more as I expected. Starting with ~100mA when external voltage is 0.1V above the setting, I stopped examing with currents >1A. If set to 20V, putting in 20.7 volts, it draws 443mA, resulting in >9 watts dissipating power.
I did not check if there's any limitation. No wonder if I killed it a few months ago.
If output is switched off (relay), there's only a 10k resistor (attached by me to discharge the 47µ output capacitor)

Now let's come to the UNI-T UDP3305S-E Channel 1. In a few words: it depends on various things.
a) with output OFF, it drains ~45mA whenever the external voltage is lower than the setting of the channel
b1) with output ON or OFF, it drains between 52 and 157 mA whenever the external voltage is higher than the setting of the channel
e.g. apply 10v to the channel set to 12V, while switched off -> 45mA is drawn from external. Now set the channel to 9 volts -> 154 mA.

b2) current rises slowly from 150 to a maximum of 157@11.4 volts, then decreases down to 52mA@16.7 volts, then rises to a max of 60mA@33 volts.
Thus, the comsumed power has two maximums at 11.4 and 33 volts and do never exceed 2 watts.
That looks safe, so I assume that there's no risk to damage anything. Well done, UNI-T!

Last but not least: While experimenting with CH3 (remember: 6.2 vmax), I fired a crowbar inside at 10.6 volts. Nothing happened (current limitation).
So beware: Do NEVER apply any battery or something like that with >10V to CH3!

[thm_w]

From what I remember the Rigol had a ~20mA constant current load on its output, which is somewhat necessary as you need to discharge the output caps once you disable or reduce the voltage.

Either way, I would not use it to charge a large battery. You can get a Riden RD6006 or similar which has a dedicated battery charging function (will show you total Wh/mAh), output relay, and fuse. Though their prices have come up a bit in the last few years.

[Pfriemler]

Of course it makes sense to use dedicated battery charger, it's kind of overkill to use a laboratory PSU for that. I have something in that style, including balancer function etc, fine. But I thought of the UNI-T for some very special things which my small charger do not can. Although the UNI-T has no dedicated battery charge function, you can record values and calculate later. Or you may charge partly, in intervals, stop below a special current - which is easy to achieve with the "Monitor" inside (switch of channel if current is below x).
Well, I keep that in mind. Maybe I construct an relay output breaker, using the digital I/O on backside: you can program them to change level depending on the output state. It's even a workaround that the UNI-T makes some alert instead of switching off the channel.

[audiotubes]

Dave has a nice video on recharing Lions with a programmable PS.

[Pfriemler]

So, basically it's well known why you should not use lab psu for charging batterys -  and how to set if we have to use 'em. There was nothing new to me. Just in the comments to Daves video

a user reclaimed a damaged RIGOl DP832 after charging a 26v lithium battery pack (and with 12V nothing happens) and linked to a RIGOL document: https://www.batronix.com/files/Rigol/Labornetzteile/DP8_DP1_ActiveLoads.pdf
But yeah, the RIGOLS are known for those damages. I will not try to blow up my UNI-T in the same way, although I think it would be safe somehow. I think the RIGOL will draw much more than 20mA in that case.

[Pfriemler]

Had a deeper look @ RIDEN ... hm, not only a dedicated battery charging function as @thm_w said, but a separate relay switched battery charger output. Nice to know. A 6012P is added to the wishlist.


Next issue: I am struggling with network access. Something very stupid is going on here. I use (Germany's top model) AVM hardware, the UNI-T is coupled to a WLAN repeater ("1200") in a (so called) mesh to my 6591 router. Both DHCP and manual IP settings seem to work, net scans with 2 computers and mobile phone do find it, pings are fast, an open port 111 (rpcbind) is reported. Although, the UNI-T "Instrument application" do not find the device via network (USB works fine). Both computers are running stupid simple Windows 10/11 with defender only. No firewall setting popped up. Any hints, folkss?

and here's the solution (afterwards I remembered Tony Albus' video where he did so):

If I open the scan on my MS Surface Pro Tablets, I will see the upper dialog with the options USB / LAN /RS232.
You have to expand the window downwards to make an input line visible. Here you have to type in the exact IPv4 address of the UNI-T
Now scan - and tadaaa! It's there.
Once you did so, next time you won't have to expand the windows, as long as the software remembers the IP address of the UNI-T (here, at least).

So stupid! I even installed newer NI-VISA runtime versions on the third PC, but you won't have to do so, it works with the V17 as well.

[Pfriemler]

Another topic: the trigger settings. I experimented yesterday with strange results. Looks pretty like "work in progress". Hmpf.

The UNI-T has four digital I/O lines backside (3.3V level). They are marked as "G 4 3 2 1" on the housing, referred as D3, D2, D1 and D0.
All triggers are disabled if the UNI-T is switched off  ... and, unlike any active "monitor", there are no signs in the status line to show if a trigger is enabled.

They can act as input - The H/L line levels are named as "High Level" (3.3V) and "Low Level") 
You may short them to ground, resulting in a ~0.7mA current, so you can use it - as an example - to switch ouputs on/off/toggle by falling edge/rising edge/low level/high level, with a so called "Sensitivity", which seems to be a delay only for debouncing. Funny: set the trigger line to toggle the output at low levels and shortcut the line - and the output is blinking in different speeds, according to the "sensitivity" setting. Nevertheless: the first task, an external all-off-button, is very easy - simply attach a button or switch and define the trigger to set ouput off on all channels with high sensitivity and bob's your uncle, so to say... 
But: If you expect this to work after a reboot ... no, it's disabled. 


Now to the output mode. The H/L line levels now are named as "Positive" (3.3V) and "Negative" (0.0V). 
If you short the output to ground, a current of ~6mA (H level) and 0.7 mA (L level) is measured.
It looks like there are hardware flip-flops triggered by some conditions. 
Very stupid: there is simply one condition per output. So, if you switch the UNI-T on and find the ouput at H level (3.3V), you can trigger it to L let's say with the condition "output CH1 is off" to realize let's say an external relay trigger. Yes, it works - once. Now the output is at L level and you have no chance to get it to H again. You can go to the trigger menu, switch the condition to "auto" and change now the output level manually in the next field, that's all. I did not find any reaction in "Auto" condition when using the UNI-T. Do not look at the manual, it's useless here.

So, to set a trigger line to simply follow the output state of a channel is not possible, if I not missed any other setting. 

[faktorqm]

OK I finally get decided and purchased 3305S at eleshop.eu. Once it arrives I will start with some tests and let you know. Regards!

[ArcticGeek]

For those that are interested in one of these Power Supplies Uni-T currently has them on sale (20% off) if you are in USA if you order directly from Uni-T.  Perhaps the same is true for other countries as well.

https://uni-trendus.com/collections/dc-power-supplies

In the US, that puts the price of the 3305S-E at $295 and the 3305S at $415.  You will have to pay shipping charges for the 3305S-E but the 3305S will ship free since the order is over $399.  That puts the price of these power supplies at a very attractive price point.

I just ordered the 3305S for myself - I can't wait to try it out.

Best,