RoadTest: Become a Tester of the R&S® NGC103 DC Power Supply
Author: Fred27
Creation date:
Evaluation Type: Power Supplies
Did you receive all parts the manufacturer stated would be included in the package?: True
What other parts do you consider comparable to this product?: The Keysight E36312A
What were the biggest problems encountered?: No problems, although the small screen meant it wasn;t possible to see the actual output and the settings simultaneously
Detailed Review:
I decided that my approach to road testing the NGC103 would be to start with usability. They say that you only get one chance at a first impression, and I thought that I'd give this impressively spec'd power supply a chance to do just that.
What I did was to just plug in in, stubbornly refuse to read any manuals, and just see how intuitive it was to do what we probably all what to do with a power supply - dial in the voltage, add a sensible current limit, and go. I genuinely did this "blind" without reading even the quick start guide, although I must admit I took a couple of takes recording the video.
You may want to just jump straight into watching the video, but in case you prefer reading to watching, then I can summarise my findings. It's is of course fairly easy to use. I'd hope it would be, but there were a couple of points where I had to think just a little bit before pressing buttons.
After that initial stumblings around the very basics, I decided to delve a little deeper into using the power supply. There are 2 main ways that the output of a power supply is limited. It can either be in Constant Voltage mode where the output voltage is being regulated. This is probably where most power supplies spend 95% of their time. Alternatively it can be in Constant Current mode, where the output has hit a limit on the current that is being supplied and the voltage will drop to ensure that this current limit is not exceeded. As demonstrated in the video, you don't really chose which mode the power supply shoudl be in, you set sensible limits on both and whichever is hit first.. well that's what mode you're in.
The NGC103 also has three levels of protection which can be consider more as safeguards that active limits. Ideally you wouldn't be hitting these, and if you do it's probably a sign that something has gone wrong and the power supply has stepped in to save you and whatever device you're powering.
The NGC103 is the three-channel option in the NGC100 range and has 3 x 32V 3A channels. Whilst it might seem that the NGC101 has the advantage of a 10A output, and the NGC102 a coupel of 5A outputs, this shouldn't put the NGC103 at a disadvantage. You see, those 3A channels can be stacked in parallel to combine those power ratings. It can also stack the channels in series to turn 3 x 32V into 1 x 96V.
My test was to stack the channels in parallel and for this the ability to have the channels track and provide the same output voltage is obviously a requirement. You really don't want to wire two channels of a different voltage in parallel.
This test seemed simple enough, and it was indeed fairly easy to ensure two channels had the same output voltage, However, when I put this into practice and tried to get more than 3A for combined channels, things didn't quire go as expected. It seemed that as one channel hit it's current limit, it dropped into Constant Current mode and the voltage began to sag. This seemed to happen even if wired in parallel to another channel at the same nominal voltage. You can see this happening in the attached video. I'm assuming that perhaps I'm doing something wrong. Maybe if you can see my mistake you'll let me know. Anyway. nothing bad happened. It just seemed that combining channels in parallel didn't quite work as planned.
As the current supplied by a power supply goes up, the resistance of the supply leads - and more importantly the voltage drop across them - becomes significant. The way this is solved is by measuring the voltage at the load rather than out the power supply output and compensating for any drop. As the current drawn by the supply's measurement circuits are small, the voltage drop across these conductors is indeed small enough to ignore.
Whilst the single-channel NGC103 has sense inputs on trhe front panel, the NGC102 and NGC103 don't have room and the connection is via a small Wago 713 series connector on the back.
I found this a little fiddly to use, but as it's Wago it is good quality and can be replaced if you ever lose it - unlike a custom connector. It didnt help that the NCG100 Series manual said:
"Insert shielded wire to the pluggable terminal block. Tightened all the wires with the screw on the pluggable terminal block."
There are no screw terminals, but the datasheet available from Farnell / Newark explains that you need to insert a small flat blade screwdriver down the side to release the wire clamp and insert the wire. I'd advise using a small guage wire and that single-core is probably easier.
Using the Sense funtion however is very easy and works as you'd expect. Just connect the wires near to you load and the Sense function is automatically enabled and the voltage at the load is more accurate to that set on the power supply.
I did try silly things like hot-plugging (which works) and connecting only one sense wire (positive is all you need to enable the function). They didn't cause a problem but are obviously not recommended.
Sequencing the channels was fairly easy. To get ot the appropriate menu, you need to press [Adv]anced and then select Sequence. For each channel that you want to sequence, you need to set ACT.CHx to "On" and set a delays time. Once you've done this for all the channels you want to use, you select "Activate" if the feature is to be initiated by the Master On/Off button or Triggered if it is to be initiated by an external trigger.
To start the sequence, you need to have Mast On/Off set to "Off" and confusingly, you need to have one or more of the channel's individual On/Off buttons selected. If you don't have one selected then nothing will happen when pressng Master On. If it's correctly configured then all the sequenced channels will go off and the switch on once the programmed delay time has been reached.
The EasyArb function of the NCG103 was also fairly easy to use. As I've done with most things on theis RoadTest, I started by ignoring the manual and seeing if it was obvious how to use the feature. With a couple of quirks, I'd say it was.
For some reason, the Arb function has its own main panel button, but it can also be accessed via the menu as with the Sequence and Ramp features. For each channel that you want to create a seuqnce for, you need to activate the sequence and then edit the "waypoints". You have a maximum of 512, btu I kept it to just 4 for a demonstration.
It was a little clunky editing this table of values on the device. For larger or more complex sequences it would make sense to edit this on a PC using HMExplorer or as a CSV file and then upload.
One of the first things I do with any device is check for updated firmware or software. (When using Linux, "sudo apt update && sudo apt upgrade -y" is second nature.) So, I decide to see if there was an update for this power supply. It seems that the usual R&S firmware links nw need you to log in using an account. No problem, I'm sure I created one using my personal gmail address... Oh. Apparently these sort of addresses are no longer permitted. I neede to create an accoutn with my work email - even though I'm not doing this review on behalf of my employer. Even then I had to jump through a couple of hoops to have my new account "upgraded" to allow firmware downloads. I'm not sure exactly why this was necessary, but it does seem to follow a trend of test equipment manufacturers going more "corporate only". I believe Keysight also no longer "support" personal purchases of equipment.
Anyway, odd thought it was it didn't take too long to get approved and I applied a minor patch for "SCPI command stability optimization". A similar process was also required to download the user manual and the HMExplorer software.
It's wasn't too hard to setup control of the NGC103 over the network. I connected an Ethernet cable and was able to see the assigned IP adderess using the Setup / Interface / Ethernet / Parameter menu on the device. Once I knew the IP it was a case of opening up a web page served by the power supply. I found out afterwards I could also have used the default hostname and go straight to http://R-NGC103-000000.local. To be honest there isn't that much the web interface needs to do other than show some basic info and provide an SCPI terminal for entering commands. I suppose I was expecting something more elaborate such as that provided on the RTM3004 oscilloscope but to be honest that wouldn't really by necessary unless you were in a classroom and wanted to show the display to a room full of students. Taking screenshots of the display was easy enough but I don't believe there is any way to get a continuous live view of it.
| {gallery}LAN control |
|---|
|
IMAGE TITLE: Device page |
|
IMAGE TITLE: Settings page |
|
IMAGE TITLE: Security page |
|
IMAGE TITLE: Terminal page (SCPI) |
I also found that connecting via HMExplorer over IP was easy too. (The name is a bit of a throwback to the Hameg days!) All I had to do was change the port used from the default of 50000 to 5025 to match the hardware and off I went. Strangely, it seemed that once I selected the NGC103 the only option available was also an SCPI terminal.
Whatever I tried, I couldn't get the device connected over USB - either as VCP or TMC. Th link for driver in manual is incorrect and links to https://www.rohde-schwarz.com/product/ngf201 . Once I found them, it seems that none of the available drivers at https://www.rohde-schwarz.com/product/ngc100 work with VCP. I ended up following a link to ni.com for NI VISA drivers but needed another account. I decided that I didn't need this enough to continuem so stopped there. To be honest, all I really wanted was a way to create screenshots and I discovered that a long press on Save/Recall button saved a screenshot to a USB stick attached to the front panel.
I was looking forward to evaulating the specs of the NGC103. I've got a matchingRohde and Schwarz 'scope that is up to the job of checking the ripple and noise from the outputs. I got a programmable load especially so I could program in some load changes and see how the NGC103 manages load regulation and how quickly it recovers from sudden changes in demand. However, as I started on this I hit a bit of a problem. It wasn't a problem with the power supply. It wasn't even a problem with the equipment. The problem seemed to be with my surroundings.
No matter what I tried, I seemed to be plagued with some much electrical background noise that it made any attempts to measure these things a bit meaningless. Rohde & Schwarz quote < 6mV peak-to-peak and <1mV RMS of noise from the supply and I was getting more than this just measuring a short! I tried a few things - disconnecting as much equipment as I could, changing my measurement setp and even going down the bottom of the garden, I got some good tips from other E14 members here, but nothing made enough difference for my measurements to be useful.
In the end, I decided that the spec sheets are very unlikely to be incorrect and that I'd be adding very little to professinoal measurements by adding my own amateur ones to the mix. Also, perhaps my fellow roadtester may have covered this adequately. i decided there wasn't much to be gained from pressing on with this.
If you're in the market for a power supply in this market segment then I can thoroughly recommend the NGC100 range. This is a great 3 channel supply but the NGC101 and NGC102 would do a fine job if you don't need as many outputs I loved the fact that there was a physical power switch and it didn't have too much vampire power draw when not being used. I thought it was a nice, compact addition to any workbench and certainly doesn't take up as much space as my broadly equivalent Keysight E36312A. The features such as sequencing and EasyArb are really useful if you're trying to either protect or (quite the opposite) deliberately stress your powered device.
I only had 2 minor niggles with the device. Firstly, the small size and screen meant it was just a little clunkier to use than a device with a bigger screen and permanent keypad. Secondly, I couldn't seem to connect 2 outputs in parallel to get more than 3A on one output, Perhaps this is my mistake. If so, please let me know and I'll re-test this.
Top Comments
I don't see anything wrong in the video? Tracking doesn't force the output to stay at given value. Tracking just sets the voltage *limit* and the current *limit* to set values.
What happens is…