UNI-T UDB3000S Series: UDP3305S / UDP3305S-E

[Self Bias]

These seem to be relatively new: https://instruments.uni-trend.com/EU-EN/powersupplies/udp3000s.html

I'm looking for a new bench PSU and stumbled across this one: It has more current than both the Rigol DP932 and the Siglent SPD3303, over the Siglent it also has a numpad (which I favor). Compared to the GW Instek GPP4342 it lacks a channel and the load sinking capability, but I don't need that anyway. I'd rather have something with more current.

I was curious enough to order the S version. I'll take some measurements when it's here. I'm especially wondering if they chose a linear regulator for pseudo channel 4, if so it'd be useful for USB powered test equipment like differential probes.

Did anyone else try them yet?

[thm_w]

Channel 4 is just a generic USB 5V/2A output, nothing special, probably not linear, +/- 0.25V accuracy. Still useful though.

Channel 1 and 2 are both 5A, so could be paralleled for up to 10A. Also seems to have internal relays for parallel/series. That is impressive.
For some reason load regulation in series tracking mode is not great (<300mV). Otherwise the specs look very good, essentially copy paste of the rigol. Weight is about the same, 10kg.

https://unitrend.oss-cn-hongkong.aliyuncs.com/20230301/1677654502688937.pdf

Price:
UDP3305S-E $370
UDP3305S $500

Hardware might be the same inside both.

[Self Bias]

I have received the unit (haha) today, but I won't be able to test much today. First impressions:

+ The 2A rating on the USB port is correct. It shuts off hard at 2.2A and recovers instantly once the load is removed.
+ Noise seems pretty low, even on the USB port. It might actually be a linear supply for that "Channel".
+ Set and Readback accuracy on the voltages is pretty good. Set voltages are off by mostly less than 2 mV, Readback is around the same.
+ I did not see an overshoot when activating the output with a load present or when adding a load and having current limiting kick in.
+ The user interface is great. There's a few typos here and there ("Templet" = "Template") and language selection is possible in two spots for some reason. But the overall workflow is intuitive.

- 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, Ch3 at 0.03 mV . Changing settings doesn't seem to have an effect. The short-circuit current in this state is -0.5 mA. I'll see if that's still there tomorrow, it's not something I like to see. The maximum allowed reverse voltage for many logical gates is 0.5V, so it's about half of that. I don't know about more sensitive low voltage circuits though.

All in all I'm not fully happy yet. Pseudo-Channel 4 could at least have an on-off-switch, I already missed that when connecting a USB connector for measurement. The negative voltage when turned off is at least an annoyance. There's also no relay click when switching outputs, just when chhanging ranges, which indicates that the outputs are permanently connected.

I went to the Embedded World, an exhibition that is held this week. I spoke to the representatives from Siglent and GW Instek about their selection of power supplies, specifically mentioning that I just ordered the UNI-T for its 5A/Ch output rating. The responses were:
- Siglent: We have a new model planned for this year that will feature a higher output current, but we don't know when it will be released.
- GW Instek: It's on their list. I filled out a form to specifically request that and possibly bump up the priority.

Both said that multiple customers came with that specific request. So maybe UNI-T will soon have some competition.

[BillyO]

UNI-T are famous for having zero support.

If you have no problems with it then it's fine, but if you have trouble you are generally out of luck.

People with good to great support are Siglent, Rigol and Korad.

I have been on the sad side of UNI-T support twice.  Now my policy is buy nothing from them over $50.

[thephil]

Thanks for sharing your observations! Did you get to play with it some more in the meantime?

I stumbled upon this instrument and from a distance it looks quite sweet in comparison to some contenders like the Rigol DP832, the Siglent SDP3305x and the Owon ODP3063. I really like that it has a proper number pad which many competitors lack. The user interface looks promising and (in contrast to the Owon) the color coding seems consistent between case and display. So I am very tempted to get one but I'm reluctant because there are very few reports online at this point. But maybe I'll have to be among the guinea pigs...

Phil

[Self Bias]

I decided to keep it as the negative voltage issue when turned off seems to be common, even among more expensive PSUs.

Automation works like a charm, I haven't seen any odd behaviour when controlling it remotely via USB. I'll see if I can verify the spec sheet sometime soon, but at least Ch1 and Ch2 seem to work fine. I didn't use Ch3 much so far, but I have no reason to believe it doesn't.

[Self Bias]

I went ahead and tested all the channels. TL;DR: It is well within spec with a max error of 4.4 mV and 1.2 mA. For those who enjoy details:

Test Setup
I used a Keysight 34465A that is out of calibration (> 2 years since last cal), but I have no reason to significantly doubt its results.
It is set to automatic range for both voltage and current measurement, current measurement was done on the 10A socket. Voltage measurement is with no load, current measurement is with the DMM shorting the PSU.

Everything is controlled by a Python script that knows about the PSU's datasheet specs and runs a voltage/current curve with 25 points. The lowest value is the range's resolution, which also taken into account for each step. For each step, minimum and maximum allowable deviation is calculated and logged to a CSV file along with the actual deviation.

Terminology & Units
"Set" is is the comparison between the voltage set in the PSU and the actual voltage measured by the DMM. For current, it is the current limit set vs. the actual current measured.
"Readback" is the comparison between the actual voltage/current measured by the DMM and the voltage/current reported by the PSU.

In the attached CSV files, the column "Set" is in Volts/Amperes. Everything labelled "Diff" is in mV/mA. This is to make scaling easier when plotting.

Results
Channel 1 Max Set Deviation: -1.3 mV
Channel 1 Max Readback Deviation: +1.3 mV
Channel 2 Max Set Deviation: -4.4 mV
Channel 2 Max Readback Deviation: +4.4 mV
Channel 3 Max Set Deviation: +0.6 mV
Channel 3 Max Readback Deviation: -0.6 mV

Channel 1 Max Set Deviation: +1.2 mA
Channel 1 Max Readback Deviation: -1.2 mA
Channel 2 Max Set Deviation: +0.5 mA
Channel 2 Max Readback Deviation: -0.5 mA
Channel 3 Max Set Deviation:  -0.5 mA
Channel 3 Max Readback Deviation: +0.5 mA

Conclusion
It is obvious that the readback is lying: It always reported the same value that was set, which leads to the deviations being exactly the same with inverted sign. Since the magnitude is so low, I'm personally okay with it thouggh. I wonder if the deviations will differ in magnitude when the unit starts to drift over time.
I did not measure the accuracy of the voltage readout under load. The unit doesn't have a sense input though, which is a shame, and it seemed to loose a few millivolts at max current due to connector and trace resistances.

I had done a short measurement of transient response under sudden load changes. Sadly I didn't take scope screenshots of them, but they looked okay to me. If someone is really interested, I can do those again.

[hsorbo]

Anybody knows how the fan noise is on this compared to Rigol?

[iAleks]

This device works much quieter, I had a dp832 before, and I sold it, now I have this uni-t, and it completely suits me in everything. I would buy 2 of these uni-t power supplies instead of one dp832.

[thephil]

In the meantime, I have bought a UDP3305S and spent some time playing with it. overall I am very happy with the instrument. I tested it against some of it's specs as far as I could with my means and it did quite well. I also had a conversation with UNI-T support and they eventually provided a firmware image and instructions to perform the update.

If you are interested, I documented my experience in my blog (I'm not a pro so be gentle. However I'll be happy to receive constructive feedback):
http://techbotch.org/blog/lab-power-supply/index_en.html

So, I didn't regret my purchase.

[iAleks]

Hi! Сan you give a link to the firmware. 1.10

[thephil]

Hi,

Yes, I got the following url for downloading the firmware image (v1.10) https://25628395.fs1.hubspotusercontent-eu1.net/hubfs/25628395/Instrument%20%20FIRMWARE/UDP3305S%20Update.rar. I don't know if/how long the link will last. As described in my blog post(http://techbotch.org/blog/lab-power-supply/index_en.htm), the update worked fine on my UDP3305S. I have no Idea if the same image is also valid for the UDP3305S-E or if that would brick the device.

• Unpack the *.rar file to get the firmware image UDP3305S.bin.
• Put the image into the root folder of a FAT32 formatted usb thumb drive. Do not change the filename.
• Turn off the power supply.
• Insert the usb drive into the usb port on the back of the power supply.
• Press and hold the UTILITY button and turn on the instrument.
• Now the update process starts.
• Once the update has finished turn the supply off and on again.
• Done.

[audiotubes]

Thanks for your web page on the UDP3305S, this was interesting.

Why did you upgrade the firmware? Was it to solve some problem(s) you found, or just to have the latest?

It looks like one of my usual shops has this for 600 euros. If anybody has a reliable source in EU which is considerably cheaper please let me know.

I will probably get one of these in the next weeks.

Thank you!

[thephil]

No – I didn't have any issues. I just wanted to make sure I would be able to benefit from future updates once bugs have been found and fixed or new features are implemented. That's why I got in touch with support. Let's face it – every software has bugs and if I pay 600€ for a PSU I want to be able to benefit form future fixes. So once I learned there is a new firmware, I wanted it even if I don't know what was changed at this point ;-)

[audiotubes]

No – I didn't have any issues. I just wanted to make sure I would be able to benefit form future updates once bugs have been found and fixed or new features are implemented. That's why I got in touch with support. Let's face it – every software has bugs and if I pay 600€ for a PSU I want to be able to benefit form future fixes. So once I learned there is a new firmware, I wanted it even if I don't know what was changed at this point ;-)

Thanks. I'm a software guy so I am an intentional late adopter. Let the other guys break and after the company fixes their fixes, then I'll try it

[thephil]

BTW: Does anyone here have a UDP3305S-E? It would be interesting to compare the two and try to find out if they use different hardware or if the difference is only in firmware.

[iAleks]

I have UDP3305S-E, but in this weekend il take UDP3305S, I want to take them apart and compare, and make S-E version too S version

[thephil]

Cool! Would also be interesting if you could get the 1.10 Firmware image for the S-E from UNI-T and then compare if it's actually identical to the S image we already have. Could be identical or could contain the model number in the image somewhere...

[iAleks]

First, I upgraded the S-E version to the S version, using the ST link programmer (STM32 ST-LINK Utility) and the serial flash copy from S version, just copy firmware from STM32H750 and serial flash. The S version was later, it has first version of the PWR PSB, all other boards are identical.

DO NOT program my 2U4 24CL04.BIN too you without backup! its calibration data! For upgrade you dont need this file!!
Here some photo and firmware S version
https://drive.google.com/drive/folders/162RZpVvgZ5Ns_vbXwYi4y9uXyf6soXJO?usp=sharing

[thephil]

Very cool! Thanks for sharing. Apart from the PCBs looking pretty much identical, did you verify that the upgraded S-E now actually honors the 1mV steps? I'd expect it to but you never know...
And good to know about the calibration data – that may pave the way to an unauthorized calibration procedure in the long run.

[iAleks]

Oh, yes it work not correctly on S-E,
set           measure     show on displey
10.000     10.000        9.999
10.001     10.000        9.999
10.002     10.002        10.002
10.003     10.002        10.002
10.004     10.004        10.003
10.005     10.004        10.003
10.006     10.006        10.005
10.007     10.006        10.005
10.008     10.008        10.007
10.009     10.008        10.007
10.010     10.010        10.009

[thephil]

Well – I think it mostly did work: An unmodified S-E would only allows for 10mV steps while you seem to be getting 2mV steps which is certainly an improvement. You may just need a voltmeter with more resolution to see the difference in the 6th digit for the 1mV steps...

[thephil]

@audiotubes: Seems like it's cheaper in croatia: https://www.chipoteka.hr/alati/izvori-i-napajanja/laboratorijski-izvor-2x32v-2x5a-5v-udp3305s-uni-trend-9150044680.

Disclaimer: I have never bought from that shop so I wouldn't know if it's real or a scam...

[iAleks]

Okay, now i see, S version have on all channel ADC 16bit (ADS1118IDGSR) and 16bit DAC (DAC8562), S-E version have ADC 16bit (ADS1118IDGSR) and 14bit DAC (DAC8162).
Il try order DAC8562, after place new ic it need calibration.. but im try it.

[audiotubes]

@audiotubes: Seems like it's cheaper in croatia: https://www.chipoteka.hr/alati/izvori-i-napajanja/laboratorijski-izvor-2x32v-2x5a-5v-udp3305s-uni-trend-9150044680.

Disclaimer: I have never bought from that shop so I wouldn't know if it's real or a scam...

Thanks, I'll have a look at that in more detail. I wonder if they didn't include VAT in the price shown...