Seeking PSU advice: Rohde & Schwarz NGE103B COM3 or Keysight E36313A.

[RBBVNL9]

Hi,

I am looking for a decent power supply, and it most important features should include three channels, parallel mode, digital IO, and a convenient user interface. Initial candidates are the Rohde & Schwarz NGE103B COM3 and the Keysight E36313A. Both are in their second hardware iteration and had several firmware update, and are from reputable suppliers, so you’d expect stable units where teething issues are fixed.

Since I liked the NGE103B’s small compact factor and identical channels, and the E36313A seems out of stock everywhere, I purchased the first. I also am a very satisfied user of a R&S RTB2004, which also made me choose for R&S.

But my experiences with the NGE103B power supply are far from satisfying. In particular, the prominently advertised parallel mode does not seem to work well at all (see a more detailed, initial impressions review below, which is longer that I was anticipating to write…). Wondering what I should do… Should I keep the NGE103B and get used to its issues? If someone can show me I’m wrong in my assessment, I’d be happy to keep it. Or return it and get the E36313A instead? Or even another unit/brand?

Happy to receive views and recommendations!

*** initial impressions review of Rohde &* Schwarz NGE103B ****

The parallel mode is a very centrally promoted feature of the NGE103B. However, in practice, it does not seem to work well at all. The current supplied by the individual channels is very unevenly spread. Some channels are fully loaded while others have plenty of power left. After a while, the fully loaded channel jumps into protection mode, turns off, causing other channels to overload, which in turn off as well, etc. The user manual says that one has to carefully try to increase or decrease voltages of the individual channels to find a balance in current supply, but I never managed to do so. Moreover, I want to be able to focus on my projects, not on how to keep the PSU up and running. The manual also says “If you wish to distribute the load to multiple channels, it is recommended to set the current limit of the channel that is to supply the main current to a fraction of the current.” But that is besides the point: the purpose of the parallel mode is to deliver more current than a single channel can! Also, the fact that there is no combined screen reading for the combined parallel (or serial) channels (like other manufacturers do), or internal relays to wire these modes, is quite a disadvantage. But the parallel mode not working at all is a real showstopper for me! 

The UI is far from intuitive. Not having dedicated channel on/off buttons, results in lots of keystrokes. Every time the rotary dial goes into live mode (which is basically very often), most other keys cannot be used for the next 7 seconds. The user interface is also inconsistent across the various menus. Sometimes the direction pad is used to increment and decrease values and select digits, but at other times it is not (like in the EasyArb mode, making entering up to 128 value points a painstaking process). The push/dial also behaves differently, depending on in which menu you are. Tracking mode is far straightforward. It does not give the same values to the selected channels, but only applies the same relative changes. Also, each time you change from voltage to current setting, you have to re-initialize and configure the tracking mode again. Finally, before I my purchase, I had noted there was not a numerical keypad, but only now I realize how much I miss it on this unit.

The remote control of the device with the WEB browser (via the optional ethernet and optional LAN interfaces) is extremely limited. Unlike many other devices, you don’t get the front panel with all its readings and controls, but a page that allows you to enter a few parameters only (basically voltage and current). More in general, given the list price of the three options that are available for this PSU (IO, Ethernet and LAN ), what you get is very poor.

The user manual seems to be a last-minute job. Page 8: “When operating the instrument, several alternative methods may be available to perform the same task. In this case, the procedure using the touchscreen is described. Any elements that can be activated by touching can also be clicked using an additionally connected mouse.” I can assure you, however, that this power supply has no touch screen, neither supports a mouse. Perhaps a copy-paste from another manual? Many functionalities are poorly described, and sometimes the important answers hidden somewhere elsewhere in the text (for instance, the fact that the EasyArb configuration can be saved). The service manual, reportedly available after registration on R&S’s GLORIS information system, is missing there.

On the positive site, the device is compact and quiet (the fan cannot be heard unless under great load). But I’d have preferred the fan to be on the side or the back, not the top. Because now one cannot stack other devices on top. The unit has feet for stacking, and the cover of the manual shows a nice picture of two units standings on top of each other, but inside, the manual says that “one should not obstruct the ventilation holes (which effectively is what you do if you place another unit on top of it), and that in rack use, at least 1HE empty distance above the unit needs to be respected. Having to stand freely party undoes the advantage of the unit’s small size.


*** So, happy, again, to receive recommendations on whether to keep this NGE103B or switch!

[Fixpoint]

Admittedly, the NGE103B is not perfect, but what you are saying implies that you didn't pay enough attention. The NGE103B is not at fault here.

The current supplied by the individual channels is very unevenly spread. Some channels are fully loaded while others have plenty of power left.

What you are experiencing is normal and not a design flaw of the instrument. The NGE103B is an entry-level instrument and does not feature regulated load sharing. If you need something like that, you must shop explicitly for this feature. In any case, this is something that you can know in advance before buying, so I don't consider it an error on the side of the instrument.

This does not mean that it is impossible to make use of parallel channels -- you just need to know what you are doing. You can't just naively connect everything in parallel and expect it to work as if you only had 1 channel instead of 3.

Don't blame the manual. The manual authors expect you to know what you are doing, they don't give you a whole lecture in electronics. I myself am no expert, I'm just a hobbyist -- so I know that I have to learn before buying something!

The manual also says “If you wish to distribute the load to multiple channels, it is recommended to set the current limit of the channel that is to supply the main current to a fraction of the current.” But that is besides the point: the purpose of the parallel mode is to deliver more current than a single channel can!

The idea is the following: Suppose you need up to 7A at 5V. Then, you should set all channels to 5V and limit each of them to 2.4A, so that the device is forced to spread the 7A more evenly.

Yes, in case you only use 3A one of the channels could deliver 2.4A and another 0.6A. In order to solve this problem, you must take measures yourself. For example, it is important that you match your wires to your load very closely (length, gauge, connectors) because the voltage drop among them influences the load sharing. You want a certain voltage drop in order to give the channels a chance to balance themselves. Using voltage drop for self-regulation is called "droop sharing".

Also, you might want to use power diodes in order to protect the channels against reverse current, depending on what you are doing. A PSU can be damaged if you have reverse voltage. This fact is especially important when your load is inductive.

The UI is far from intuitive. Not having dedicated channel on/off buttons

What?? Of course it has those buttons, and they are clearly labeled and prominently placed! Just have a look.

results in lots of keystrokes. Every time the rotary dial goes into live mode (which is basically very often), most other keys cannot be used for the next 7 seconds.

Shouldn't the live mode quit when you press the dial? I personally have the PSU configured such that it is ALWAYS in live mode and I don't have any problems.

there was not a numerical keypad, but only now I realize how much I miss it on this unit.

Since you can edit each digit of the settings individually, you can make quick adjustments of the settings without a numerical keypad. Many adjustments are even quicker this way than with a numerical keypad. You should make use of that feature.

Perhaps a copy-paste from another manual?

The NGE comes from the NG line of new PSUs that all share the same basic UX and UI concepts. The manuals are basically the same.

But I’d have preferred the fan to be on the side or the back, not the top. Because now one cannot stack other devices on top.

There is still enough space between the unit and the next one in the stack, this shouldn't be any problem.

For a change, I can tell you a REAL nuisance with the NGE103B: It's fan control algorithm is not very good. After running for some time (it does not matter whether there is a load or not), it enters a control loop that starts and stops the fan periodically even when there is no load (at room temperature). R&S have told me that they have this problem in their backlog, however I don't know whether/when they will fix it.

[RBBVNL9]

Dear Fixpoint,

Thank you for your extensive comments, appreciate it. I learned several things from what you wrote.

But on some other, major issues, my experiences / observations are, however, not the same as yours.

The NGE103B is an entry-level instrument and does not feature regulated load sharing.

Your comment implies that it was unreasonable of me to expect, in this price class, a parallel mode in a PSU that would not require the user to manually have to achieve load balancing for the device to work. To find out whether that was true, I now searched more specifically for parallel mode implementations of devices at the same or lower price level. A video of the Keysight E36313A, the main alternative I considered, with a similar list price of the NGE with similar options, suggest that that PSU works precisely as per my expectations, with no requirements for the user to balance loads across channels. Also has relays that set up the internal wiring. Checking a bit further, the Keithley Model 2231A-30-3 (considerably cheaper than the NGE103B) also works this way.  Heck, even the much, much cheaper Siglent SPD3303X-E performs automatic load sharing between its channels in parallel mode (see the following this video at 41:55 - https://www.youtube.com/watch?v=MwSO9cu5c1o. Also this one has the internal relays to set up parallel wiring.

The idea is the following: Suppose you need up to 7A at 5V. Then, you should set all channels to 5V and limit each of them to 2.4A, so that the device is forced to spread the 7A more evenly.

So, I just performed a test following your line of reasoning. Each channel is advertised to deliver up 32V, up to 3 amps, where the total power per channel needs not to surpass 33.6 Watt. So, at 12V, the max. channel load is 2.8 amps (which is indeed exactly what the user interface shows us when 12V is selected).

So I connected a load that at 12 Volt draws 7.35 amps. At 7.35/3=2.45 amps per channel, that is below the max. rating. I set each channel at 12V, and, in line with your suggestion, I then set the current for each channel lower than the max rating, at 2.6 amps. When I turn the outputs on, and initially all three channels are still nicely 10% below their maximum load rating (<30 of a maximum of 33.6 Watt). But one channel gives up after 25 seconds (turns into OPP mode) and subsequently the other channels jump into current limiting mode and drop from 12 to 5.4 Volts.

Again, we are clearly still in the devices specs here...

I'd be really interested to see whether you (or someone else) could replicate this simple test, to see whether its just my unit that keeps turning of channels while used within the specifications, or whether you find the same. For the load, for convenience, I just took four halogen 12V light bulbs....

I must say that low(er) loads, the current limiting method you suggest does work, whereas using the method listed first by R&S in their manual (“By increasing the voltage slightly, the load distribution can be manipulated”) I could not get stable channel load balancing at all and channels kept turning off even at lower loads.

what you are saying implies that you didn't pay enough attention.

I understand that, depending from the perspective you come from, this 1200 Euro power supply is an entry-level instrument. But if a PSU is prominently advertised to be able to deliver up to 9 amps (within the limitations of the overall channel power), then the user should be able to trust on it that it does so, in a reasonably uncomplicated manner, and following the information in the user manual. Not that channels turn of after 30 seconds. I want the device to meet the specifications as advertised.

I don't think that is a matter of not me not paying enough attention.

Since you can edit each digit of the settings individually, you can make quick adjustments of the settings without a numerical keypad. Many adjustments are even quicker this way than with a numerical keypad. You should make use of that feature.

Yes, the digit entry feature is useful. But this entry mode is not available when you most need it - for instance if want to make a number of points in the TrueArb table. No idea why this entry mode does not work here, but the only way to populate this table is painstakingly dialling in all voltages, currents and times with the rotery dial. If you know a way how to activate the direction keys in the TrueArb table mode, then please share with us how that is done!

The UI is far from intuitive. Not having dedicated channel on/off buttons
>>> What?? Of course it has those buttons, and they are clearly labeled and prominently placed! Just have a look.

I really don’t see dedicated channel on/off buttons. There are three buttons, labelled “Ch 1”, “Ch 2” and “Ch 3” respectively, but dependent on the state the device is in, these buttons serve to select channels for voltage setting, for current setting, for tracking selection, for "fuse" settings, or for channel on/off. So these are not dedicated channel on/off buttons.

Concerning the manual, well, it may be well true that this device is closely related to other devices that R&S makes. But this particular manual is for the NPE100B series devices only. So I find it strange that it refers to the touch screen or mouse controls, while the NPE100B series certainly has neither of these.

There is still enough space between the unit and the next one in the stack, this shouldn't be any problem.

I would be happy if I this was the case. But on this subject, the R&S user manual (p. 19) states “The R&S NGE100B is fan-cooled and must be installed with sufficient space on the top to ensure a free flow of air. Required minimum distance: 1 rack unit (RU).” One rack unit (generally known as 1HE) equals ~4.4 cm. The distance between stacked NGE103B devices is only around 1.0 cm so way below the required 4.4 cm.

For a change, I can tell you a REAL nuisance with the NGE103B: It's fan control algorithm is not very good. After running for some time (it does not matter whether there is a load or not), it enters a control loop that starts and stops the fan periodically even when there is no load (at room temperature). R&S have told me that they have this problem in their backlog, however I don't know whether/when they will fix it.

Thanks for mention this - even if I'm looking for solutions, no more problems I now left the unit on for a most of the day, without any load, and indeed it gets into a loop of turning the fan on and off. 

[Fixpoint]

Hello RBBVNL9. Maybe my reply was too harsh. I just wanted to defend the instrument from criticism that I felt was coming from an angle that appeared unfair to me.

Heck, even the much, much cheaper Siglent SPD3303X-E performs automatic load sharing between its channels in parallel mode

I understand your frustration, but this is a pretty fundamental thing: R&S is one of the world's "first-rate manufacturers", and when you buy from them, this can have DRAWBACKS instead of advantages. Why?

Because of the way premium manufacturers work. They will charge you a premium for even their most basic entry-level instrument and will still provide you with much less features than the (non-premium) competition. Rule of thumb: A premium manufacturer doesn't charge you for features but for reliability and support. (Also, they take pride in who they are and often don't manufacture in China but in other places that are more expensive.) That means, when you buy a 1200 USD premium instrument, you are likely to get less features than with a 600 USD non-premium instrument. That's how it is: don't buy premium if you don't have VERY deep pockets. Otherwise, you might get less than you thought.

With premium manufacturers, you might even find that NONE of their devices have all the features you want, not even the most expensive ones. That's because they expect you to buy specialized equipment for each and every task because their target audience are companies and professional laboratories, not hobbyists. Some of them just won't give you an instrument that does everything. Of course, the non-premium manufactures will -- and for a fraction of the price.

Another example: Gossen multimeters cost approx. 600 or 700 USD but "only" have 12000 counts, are pretty slow and clunky as hell. As far as I know, you can get a 300,000 counts meter that is much faster for a fraction of the price (Brymen, I think). However, the Gossen is built like a tank and comes with a detailed calibration certificate. Also, it is very difficult to connect it the wrong way. Even the feel and sound of its rotary dial is a special experience. Reliability is king, that's what you pay for.

I want the device to meet the specifications as advertised.

Yes, of course. I never used the NGE103B with 9A, so I don't have experience with that. I can only say how I personally think: I accept that these specs are the extremes and if I REALLY needed 9A in a real-world application, I would use an instrument with at least 200W. Of course, you don't need to think the same way -- it is just my experience (also with other devices, not just lab equipment) that the high end of any spec almost never represents an area where you would actually want to use the device.

For example, if you needed to measure a voltage of approx. 1000V, you wouldn't want to use a multimeter that can measure up to 1000V. Of course, your region of interest would be within spec, mathematically spoken, but practically, you would probably have more fun with a higher-spec'ed instrument.

In my opinion, the NGE103B is not suited for applications that require 9A. At least for me, that's ok. Your mileage may vary.

Now the final question, out of interest: What happens when you disable OPP?

[drwho9437]

The biggest weaken for me of the NGE103B was the noise/ripple specs. An E36312A has 350 uVrms on all 3 channels. The NGE103B has 1500 uVrms.

The best things about the NGE103B are the size and that all the outputs are the same and go to a healthy 32 V.

I have used a lot of power supplies including endless HP ones, Tek, R&S, and more. I was looking at all these supplies recent, and although I haven't gotten it yet to test it I put an order in for a GPP-4323.

This supple the ripple is weirdly measured only to 1 MHz instead of 20 but on the 32 V rails it is at 350 uVrms. Given it probably isn't white and it has more low frequency content given it is a linear design inside I think, I'm counting on there being nothing interesting raising the noise levels between 1-20 MHz. Even if it ends up being a bit worse than I would guess I think it will be better than the NGE103B at 1/2 the price with an extra channel although not all are matched and the outputs for the 4 channel version ended up being on the left first because I guess they didn't want to move the layout for the other channels. There is almost always some digital service in everything I make so the two more dirty rails are useful for motors and digital while the bipolar low noise for the juicy analog low noise stuff. Of course it depends on what you are using it for. I mostly put it on order to replace a very loud Tek supply.

If you can handle some fan noise the best thing to do for power supplies is normally just to go buy used ones. If you don't need programmable outputs you can get very fine triples HPs for less than 100 dollars used. Singles for 50 dollars. I might have continued on that route but most older ones don't let you see the current on N channels at once and some old triples have no compliance at all on rails which is non-optimal for certain prototyping. 500 dollars for a PSU though given people on here only buy 1000 dollar scopes is probably too much. Anyone who feels they have to stay in the 1000 dollar scope budget should try to find a decent used power supply I think. Fully programmable are available for 200 normally, they are just loud and have more ripple than the newest ones.

[RBBVNL9]

Dear Fixpoint,

I appreciate your second set of comments, and you are right in many things you write. And I was also pretty direct in what I wrote (but do still stand by those words).

With a premium manufacturer, you may indeed often pay premium, even for a more limited feature set. Having said that, when it comes to the parallel mode implementation we were discussing, the trouble-free, ‘automatic’ load balancing is also found on Keysight and Keithley entry-level devices, which I think are to be considered premium manufactures like R&S. My quick survey of the devices I considered to buy suggests that the NGE is really an outlier in not having ‘automatic’ load balancing, but we would have to look at many more devices (and actually use them) before coming to conclusions.

So,do I really need 9 amp? Honestly, not that often.

Should I buy the NGE103B if I really need 9 amps? You are right, probably not.

Thing is that if I buy a new test device, I really want to get well acquainted to it, use its various feature, make sure I understand its functioning, particularities, ins and outs, and limitations. With the NGE103B this has been such a bumpy ride, so many things that felt disappointing when I actually put my hands on it.

It feels so different from my Rohde & Schwarz RTB2004 oscilloscope, I get the impression that the NGE is really in another league, as if it’s not even from the same maker. The RTB oscilloscope is satisfactory to me in almost any dimension, it’s a pleasure to use, well designed, pleasant and consistent in UI, etc. (As a matter of fact, it was the R&S service that let me down with the RTB when it took them several months to replace a broken probe they supplied with it, but that is another story.)

So, bottom line is the following: (1) Can the NGE perform most of my use cases? Answer: technically probably yes, even though it seems not to comply with all its specifications. (2) Will I be a happy, satisfied user? Not sure I will.

Wo what to do? The Keysight E36313A seems to be a more attractive option to me now (I know that in parallel mode it delivers ‘only’ 4 amps, but at least the mode seems well designed). Buts it’s not available before October or so. There may be other options too, from Keithley, BK Precision, AIM-TTI, EA, and others, or I may even look at lower end options from Siglent, among others. May be less attractive in some dimensions, more attractive in others. A colleague of mine here is scholar in (human) decision theory, perhaps should have a talk with him ;-)

[RBBVNL9]

Dear drwho9437,

although I haven't gotten it yet to test it I put an order in for a GPP-4323.

The GW Instek GPP-4323 had not been on my radar yet, but at its price point, to certainly seems to model to look at. Please do share your experiences (ripple, but also otherwise) once you get your unit!

[Wolfgang]

Ever looked at a Rigol DP832A (or DP832, which is upgradeable) ?
Almost as good, for a small fraction of the price. Parallelizable too.

[Fixpoint]

Keysight and Keithley entry-level devices, which I think are to be considered premium manufactures like R&S

I think Keysight and Keithley are different from R&S in that they indeed offer top-of-the-line instruments but also have a non-premium segment which is something R&S don't do. If I remember correctly the cheaper models are just branded with their names but are manufactured by some Asian company, at least for Keithley. Correct me if I'm wrong.

And of course I also would appreciate if R&S cared a little bit more for the NGE series, including regulated load sharing. But for me, the fan problem is the most annoying thing and I find it pretty ridiculous. And that is something you cannot know in advance, so it is definitely not ok. It is very annoying when this control loop goes on and on for hours.

I think when an R&S-branded PSU is left alone without any load at room temperature, then the fan must be OFF and must remain OFF, that's it, no questions asked. I have the feeling that in this instance, R&S just didn't care. It seems that they thought "well, the NGE is mainly for schools and universities, it just doesn't matter".

It feels so different from my Rohde & Schwarz RTB2004 oscilloscope

I also use the RTB2004, so apparently we have the same gear on our benches. And yes, the RTB2004 appears to be better engineered, however it also has some software bugs I again and again come across (UI glitches), and I don't like the dented dials (smooth scaling and zooming is not possible with dented dials). And, well, I think everybody hates the glossy display, I think that's just a marketing gimmick so that it looks nice on product shots.

what to do?

Because this is not only a technical question but also an emotional one (nobody wants to feel ripped off), it's not so easy

From the fact that you are considering the Keysight with only 4A parallel current, I can infer that that you don't really need parallel mode, because otherwise you would say that 4A is not enough! 4A is actually almost nothing That in turn means that you could well go with the NGE (which, by the way, should be able to handle 6A parallel current without problems, I assume).

Keep in mind that parallel current sharing can always be an issue, also with regulated PSUs (device health issues in case you make a mistake and regulation can suffer), so if you REALLY needed 6A, 7A or 8A at least occasionally, then you definitely should buy a PSU that supports at least 10A per channel. That will provide you with superior performance. Remember that parallel mode is just an emergency solution for the case that you unexpectedly need more current than usual, but it is not a (good) solution for your everyday work.

Did you try to turn OPP off?

[Fred27]

Some interesting views. I've got a E36312A but like the look of the NGE103B - mainly because of the size (although I see it's actually bigger than I though, only really being less tall that the Keysight) and because I really like my RTM3000 'scope.

The one thing I'm not keen on with the E36312A series is that it draws 10W when "off" - i.e. powered off using the soft button on the front. There's no physical switch on the device. Other than that I'm pretty happy with it.

[RBBVNL9]

I also use the RTB2004, so apparently we have the same gear on our benches. And yes, the RTB2004 appears to be better engineered, however it also has some software bugs I again and again come across (UI glitches)

Yes, at times I run into glitches on the RTB too, and occasionally parts of the UI become unresponsive altogether and I need to reset the whole device (booting the instrument while pressing and holding the 'preset' key). Still, taking all together, this instrument is a pleasure to work with and I have not regretted my purchase one single second.

Did you try to turn OPP off?

No, I did not, probably I'm hesitant to do so (/having to do so) because I assume its there to protect the device. But may dig into this.

I think Keysight and Keithley are different from R&S in that they indeed offer top-of-the-line instruments but also have a non-premium segment which is something R&S don't do. If I remember correctly the cheaper models are just branded with their names but are manufactured by some Asian company, at least for Keithley. Correct me if I'm wrong.

I think this is indeed true for the entry-level models of Keithley. As for the Keysight E36313A, I have the impression this is really designed by them, and did not see other, very similar looking devices from others that indicate possible rebranding. But it's always hard to be sure.

Keep in mind that parallel current sharing can always be an issue, also with regulated PSUs (device health issues in case you make a mistake and regulation can suffer), so if you REALLY needed 6A, 7A or 8A at least occasionally, then you definitely should buy a PSU that supports at least 10A per channel.

Agree. I don't think I need >4 A that often , and if I would, I should shop indeed for something that has higher currents on a single channel.

Having said that, a parallel mode seems to be a good and cost-effective trade-off between having the flexibility of multiple channels on a device, and occasionally to have higher currents. In that sense, the parallel mode is something that suits my intended use pattern. But it just feels wrong when such a mode is advertised, but does not work properly.

[RBBVNL9]

And of course I also would appreciate if R&S cared a little bit more for the NGE series,

Trying to put this into something constructive, I’d like to submit a firmware/device improvement list for the NGE103B. Happy if others add to this!

Will start with ‘low hanging fruit’, then mention some more radical improvements.


Firmware improvements (compared to FW Version 1.54)

1.   Fix the fan issue, so it does not get into a loop of on/off while the unit is not having any load.
2.   Make the UI consistent across all menus and all modes of operation. For instance: in the TrueArb menu, make sure the direction pads set the voltage and current the same way as it does in other modes. Also, make sure the dial and its push-button work consistently across menus.
3.   Implement a proper load balancing for the parallel mode. Really should not be that difficult, I must assume (comparable power supplies from premium manufacturers, as well as lower-cost power supplies, most often have such load balancing).
4.   Implement a tracking mode where all tracked channels take the same value (and not only chance their relative values).
5.   Implement a decent remote web server functionality. Especially important if you make customers pay steep license fees to have ethernet or WiFi functionality. Either emulate the full screen (preferable option). Or implement different user interface (at risk this is a subset, and users may disagree what important things should be in the subset). At least provide web interface controls for the functions that require a lot of user input data, such as TrueArb.
6.   Provide a setting to set system time, or take system time from a web server. This way, files (e.g. screenshots) are saved with correct time stamps.

Documentation and marketing

1.   Reorganize the User Manual by putting the information more clearly together. For instance, for the TrueArb, include information of its working as well on how to store the TrueArb settings all together in one section
2.   Remove text from the User Manual that is not related to this device (for instance the text on touch panel and mouse control)
3.   Include reference drawings/schematics for the digital IO pins (how to hook up switches, electronics and/or relays).
4.   Make sure that the devices sold indeed meet the specifications (32V, 9A, up to 33.5Watt/channel) in a reasonable, stable way. If specifications are not met, then remove those claims from the marketing. (My own unit is not able to get to 90% of the stated specifications, as explained above.)
5.   Should the parallel mode stay as it is, then provide consistent and good information on its usage. Do not use conflicting language (at one points explain the user has to load balance by very small variations of the voltage settings, at other points explaining the user needs to do so by changing current limit settings).
6.   Be consistent in usage requirements. For instance, do not show a picture of stacked units or provide stacking feet, but then write in the manual that minimal required vertical clearance is 4.4 cm.
7.   If you announce that a service manual is available after registering to R&S GLORIS, then do make it available there.

Hardware improvements (limited to relatively small changes only)

1.   Provide dedicated channel on/off buttons. If there is not enough size for additional buttons, consider to move M1 to M5 to the system menu (these are not needed that often, and you may even provide, say, 10 memory slots this way). Then put dedicated channel on/off in their place – after all, these buttons get to be used all of the time. Making the buttons light up in the same channel colour codes as on the front panel and the display (similar to many oscilloscopes) would be great.
2.   Have relays that set the wiring for serial and parallel mode.



 

[RBBVNL9]

It seems that they thought "well, the NGE is mainly for schools and universities, it just doesn't matter".

Interesting comment... While the device I was discussing here was bought by myself, my regular job is at a technical university, working as a full professor.

And - believe it or not - at universities, we do care about things like UI, usability (including noise levels), and power consumption. In fact, topics like sustainability and usability are more prominent in our own research and teaching then they ever have been before. 

So, if anyone is listening at R&S (or other firms, for that matter), please do listen to user comments and feedback...

[Hydron]

Just noticed this for the GW Instek:

"The CH1/CH2 of the GPP-Series are designed with the load function. A single power supply can set
one channel as the power output, and one channel for the load function to consume the power of the
DUT so as to meet the basic charging and discharging test requirements for battery. Channel 1 and
channel 2 not only provide 32V/3A power output, but also feature built-in maximum 32V constant
voltage load (CV), maximum 3.2A constant current load (CC) and maximum 1k constant resistance Ω
load (CR) function."

Has anyone had any experience using this supply? Certainly sounds like it's worth looking at given the load feature. No sense connection unfortunately, and may have to be careful not to get the stupid safety version without binding posts though (see TME's pictures of it vs the manufacturer site). It's also not obvious how to get it with a LAN port.

[2N3055]

You should have bought TTi PSU..

https://www.aimtti.com/product-category/dc-power-supplies

R&S entry level equipment is not worth 1/3 of money they ask for it...

[jjoonathan]

I love my GPP-4323 supply!

[Fixpoint]

R&S entry level equipment is not worth 1/3 of money they ask for it...

That, of course, is not true. The NGE is a definitely a very good and solid device, it protects your circuits very reliably from mistakes, and the UX of the new NG line is really very good: very good color scheme, very good use of space, all the important info is right where it should be, and the UI reacts to input in an intuitive way. The NG line follows the established best practices and avoids common pitfalls that other manufacturers have in their UIs. Also, there are no UI glitches in the current firmware version (1.54), at least not to my knowledge. Yes, the Arb function does not have a very efficient UI, but that is no surprise without a touch screen.

The only problematic things are software-related: the fan control (they didn't really implement that, the control loop is extremely annoying) and the missing load balancing in parallel mode.

[2N3055]

R&S entry level equipment is not worth 1/3 of money they ask for it...

That, of course, is not true. The NGE is a definitely a very good and solid device, it protects your circuits very reliably from mistakes, and the UX of the new NG line is really very good: very good color scheme, very good use of space, all the important info is right where it should be, and the UI reacts to input in an intuitive way. The NG line follows the established best practices and avoids common pitfalls that other manufacturers have in their UIs. Also, there are no UI glitches in the current firmware version (1.54), at least not to my knowledge. Yes, the Arb function does not have a very efficient UI, but that is no surprise without a touch screen.

The only problematic things are software-related: the fan control (they didn't really implement that, the control loop is extremely annoying) and the missing load balancing in parallel mode.

With all due respect to you, your opinion (and even your national pride  ),  when all those "minor" niggles happen in chinese equipment than it's "reason why that crap is cheap"...
And if I buy PSU, I didn't buy graphic tablet with video game.. I don't care how fancy graphics look..

I have 2 cheap no name PSU that work nicely and reliably in parallel mode.  They sell this R&S 1500€ + PSU and advertise parallel mode.
And then when it doesn't work, they tell you they didn't really lie, because small print in datasheet says "it works with caveats".
I simply hate equipment from top notch manufacturers that is specified by lawyers, not engineers.

Don't get me wrong, once you cross into midrange and high end R&S is world class and well worth the money.
I'm not happy about entry level equipment from any premium tier manufacturer. They are all either provide inferior specced equipment (carefully and deliberately designed not to compete with their own higher end equipment) or they relabel chinese products and charge twice fro the label.

It just so happens that that second tier manufacturers are making entry level equipment that is very good quality and features for the price, because what is "cheap stuff for simpletons" for premium manufacturers , is second tier manufacturers top equipment and their bread and butter.. They simply put more effort into it. It has to be good for them to prove themselves.

So that is the source of my opinion.

With best regards,
Sinisa

[drwho9437]

I love my GPP-4323 supply!

I actually watched your video before I ordered mine. Channel 4 is first because the dual power supply has the nicest two outputs already and they didn't want to change some of the wiring I guess. You would think they could relabel it but alas. I can see why on the dual they wanted the two channels centered on the front.

If you have the ability to measure the ripple accurately I'd be interested.

[Hexley]

Another satisfied owner of the GW Instek GPP-4323 here.

The small size and quiet fan are great features, as are the four isolated channels and the active-load capability.

BTW, there is a profile written for the GPP-4323 in HKJ's TestController software, so you can do remote control. The LAN interface option is inexpensive and nice to have.

One note -- be sure to update to the current firmware. Earlier versions (such as 1.10) had an issue with a glitch at power-on for Channel 3. Now fixed.

[Fixpoint]

I made a simple ripple & noise measurement for the NGE103B. I cannot design a professional test setup, so I would be interested in results that are produced by someone who is qualified.

The measurements are at two vastly different time scales! 10 µs/div (5V and 10V, for noise judgement) and 50 ms/div (10V, so that the ripple becomes visible).

[jjoonathan]

Yeah, that could be. It's even crazier on the GPD-3303S, where the order is "CH2 CH1  FIXED," although I suppose that too could be the result of  optioning.

As for ripple, it's <150uV RMS by my simplistic measurements:

[Rich@RohdeScopesUSA]

Hi,

I am looking for a decent power supply, and it most important features should include three channels, parallel mode, digital IO, and a convenient user interface. Initial candidates are the Rohde & Schwarz NGE103B COM3 and the Keysight E36313A. Both are in their second hardware iteration and had several firmware update, and are from reputable suppliers, so you’d expect stable units where teething issues are fixed.

Since I liked the NGE103B’s small compact factor and identical channels, and the E36313A seems out of stock everywhere, I purchased the first. I also am a very satisfied user of a R&S RTB2004, which also made me choose for R&S.

But my experiences with the NGE103B power supply are far from satisfying. In particular, the prominently advertised parallel mode does not seem to work well at all (see a more detailed, initial impressions review below, which is longer that I was anticipating to write…). Wondering what I should do… Should I keep the NGE103B and get used to its issues? If someone can show me I’m wrong in my assessment, I’d be happy to keep it. Or return it and get the E36313A instead? Or even another unit/brand?

Happy to receive views and recommendations!

*** initial impressions review of Rohde &* Schwarz NGE103B ****

The parallel mode is a very centrally promoted feature of the NGE103B. However, in practice, it does not seem to work well at all. The current supplied by the individual channels is very unevenly spread. Some channels are fully loaded while others have plenty of power left. After a while, the fully loaded channel jumps into protection mode, turns off, causing other channels to overload, which in turn off as well, etc. The user manual says that one has to carefully try to increase or decrease voltages of the individual channels to find a balance in current supply, but I never managed to do so. Moreover, I want to be able to focus on my projects, not on how to keep the PSU up and running. The manual also says “If you wish to distribute the load to multiple channels, it is recommended to set the current limit of the channel that is to supply the main current to a fraction of the current.” But that is besides the point: the purpose of the parallel mode is to deliver more current than a single channel can! Also, the fact that there is no combined screen reading for the combined parallel (or serial) channels (like other manufacturers do), or internal relays to wire these modes, is quite a disadvantage. But the parallel mode not working at all is a real showstopper for me! 

The UI is far from intuitive. Not having dedicated channel on/off buttons, results in lots of keystrokes. Every time the rotary dial goes into live mode (which is basically very often), most other keys cannot be used for the next 7 seconds. The user interface is also inconsistent across the various menus. Sometimes the direction pad is used to increment and decrease values and select digits, but at other times it is not (like in the EasyArb mode, making entering up to 128 value points a painstaking process). The push/dial also behaves differently, depending on in which menu you are. Tracking mode is far straightforward. It does not give the same values to the selected channels, but only applies the same relative changes. Also, each time you change from voltage to current setting, you have to re-initialize and configure the tracking mode again. Finally, before I my purchase, I had noted there was not a numerical keypad, but only now I realize how much I miss it on this unit.

The remote control of the device with the WEB browser (via the optional ethernet and optional LAN interfaces) is extremely limited. Unlike many other devices, you don’t get the front panel with all its readings and controls, but a page that allows you to enter a few parameters only (basically voltage and current). More in general, given the list price of the three options that are available for this PSU (IO, Ethernet and LAN ), what you get is very poor.

The user manual seems to be a last-minute job. Page 8: “When operating the instrument, several alternative methods may be available to perform the same task. In this case, the procedure using the touchscreen is described. Any elements that can be activated by touching can also be clicked using an additionally connected mouse.” I can assure you, however, that this power supply has no touch screen, neither supports a mouse. Perhaps a copy-paste from another manual? Many functionalities are poorly described, and sometimes the important answers hidden somewhere elsewhere in the text (for instance, the fact that the EasyArb configuration can be saved). The service manual, reportedly available after registration on R&S’s GLORIS information system, is missing there.

On the positive site, the device is compact and quiet (the fan cannot be heard unless under great load). But I’d have preferred the fan to be on the side or the back, not the top. Because now one cannot stack other devices on top. The unit has feet for stacking, and the cover of the manual shows a nice picture of two units standings on top of each other, but inside, the manual says that “one should not obstruct the ventilation holes (which effectively is what you do if you place another unit on top of it), and that in rack use, at least 1HE empty distance above the unit needs to be respected. Having to stand freely party undoes the advantage of the unit’s small size.


*** So, happy, again, to receive recommendations on whether to keep this NGE103B or switch!

Hi RBBVNL9 - just saw your PM.  Sorry, I've been on vacation this past week.  I'll talk with the team in Germany and get back to you.

-Rich

[drwho9437]

Yeah, that could be. It's even crazier on the GPD-3303S, where the order is "CH2 CH1  FIXED," although I suppose that too could be the result of  optioning.

As for ripple, it's <150uV RMS by my simplistic measurements:

Thanks, that's about 2-3x better than the spec. The standard is to measure it to 20 MHz due to classic scope bandwidths and this is what Keysight, Tek, Rigol and others all use, but GW uses 1 MHz just as you did here. If it is white it can only about about 3x worse than what you report which is still better than the specs on the E3x Keysight series which is good. Its worth noting to others who might be reading this that the other channels are not as good noise wise in the datasheet, but I think they really serve as AUX supplies and the ones you tested are the real ones you would run sensitive stuff off of.

Thanks again for the quick test. I am not suppose to get a unit for another month it looks like. Pretty long lead.

[maginnovision]

Can you turn the 4323's display off? Or does it have a screensaver/auto off like the Rigol DP832?