EEVblog 1402 - Rohde & Schwarz NGA100 PSU Teardown + GIVEAWAY

[Astrodev]

The overshoot is definitely a potential problem but it would be interesting to see how the  CV vs CC affects any transients on the output when used at the upper end of the power envelope, as this has a lot of potential to really do some damage to the circuit being powered. I have also seen some bench power supplies get damaged under these conditions at the edge of the envelope.

It would be interesting to know how these power supplies cope under these heavy load conditions.

[timgiles]

I could do with a decent power supply - so would be great to be part of the giveaway :-)

[Stuart Coyle]

How many power supplies do I need? The answer is, of course, always one more than I already have.

[Kean]

With the rear ouput and seprate rear sense terminals / cables, I am slightly missing a way to switch between front and rear. The drive terminals may be in parallel, but the sense part would need some switching (front, rear with 2 wires and rear with 4 wires).

Maybe the relay at 12:25 in the video is for switching between front/back sensing?  It does look a bit beefy for that requirement, but wouldn't you need two relays to select between isolated, series, and parallel modes?

Was there any other relays for series/parallel mode switching?  With isolated channels and 250V DC max to chassis on both channels, I guess they could be using transistors for switching into series and parallel modes.

[Kleinstein]

The large relay is likely for the series / parallel switching. It should need 2 relays for this. Switching with transistors is tricky and would be odd.

Switching the sense lines would be more practical to do with with electronic switches or maybe reed relays. One side (usually the positive side with the usual floating regulator) could use a simple CMOS switch.

[ElectricGuy]

I'm in... like flin!!! 

[f4eru]

I have a good old linear 1kW R&S NGRE "STROMVERSORGUNGSGERÄT" like this one but in 10A 100V version, so i'll pass on the giveaway
Here you even can select if you want current overshoot or not, there's a button for it and that's really useful

R&S cannot pretend they did not know about current overshoot... if they do it switchable since 40 years



Nice and reliable stuff.
When I opened it up, I saw that the voltage ref. is a neon tube
None of this digital panel with non selectable overshoot rubbish

[rf-design]

Now you can see that an agile code developed over distributed sites for a target RTOS samples a grey frontend knob and allow two different ramp functions 

[mrprecision]

I would like to win that power supply to make a review and a comparison to the siglent power supply.

[nctnico]

R&S cannot pretend they did not know about current overshoot... if they do it switchable since 40 years

No. It is to select control loop behaviour for loads with a large or small capacitance. It is not uncommon on high-end / high-precission power supplies.

[Doctorandus_P]

That overshoot is simply unacceptable for a power supply in that price range.
Especially the voltage spikes when it is in constant current mode  and the set voltage is adjusted.
 

It's also very big for a 2x 40W power supply.

I do wonder a bit about all those secondary transformer taps, and how they are switched and there seems to be a black relay stuffed in a corner.

All together I'm not impressed.

[Kleinstein]

The 40 W (per channel) power limit is odd for a linear power supply. I would normally expect more like a stepped function like 2 A up to 40 V and 1 A when higher, depending on the transformer taps used. There may be some regulations that make is easier if only specified to 40 W.

The transformer wires seem to come in colored pairs, so no split coils with 3 wires. Still there are quite a few wires. The expected votlages would be something like +-10 V (so 2x2 wires) for the regulator and 50/120 V (4 wires) for the main ouput, the 2 large filter caps. There may be an addition 5 V for the regulator / ADC/DAC part, though this may also start from the higher supply rail.
With 2 darlingtons on the heat sink I would expect them to be in "series" with a diode to bypass the upper one if the lower tap is sufficient. So the cross over between 2 voltage rails can be electronic. The Keithley SMUs and some audio amplifiers use a similar configuration, though with push pull.
There is also some smaller power transitor / FET, that may be used for some switching, maybe limiting the voltage charging the fitler caps. Without this the worst case heat dissipation would still be in the 100 W range (2 A into a short), which would be quite a bit of powert for s single 150 W rated darlington.   

[Bouteille]

Hi there,

yet another greedy message to win a power supply with high resolution metering...
Would love to reproduce the "self mixing interferometry" experiment with laser diode (from this other great guy on YT)

See you and thanks for all you do !!

[SoiioS]

I’m in
Nice power supply
 

[mrcole]

Count me in!

[RBBVNL9]

Thanks Dave, nice video!

Concerning the front panel connector type: this NGA device can get to quite high voltages - especially with 100V channels in series - which may explain the choice for this safety connector. Maybe less convenient than regular banana sockets, but safe.

I suspect that one of the reasons R&S upgraded their older NGE100 (with smaller grey binding posts with black and red rings) series to the NGE100 B series was also to offer safer connectors (large blue and red binding posts). Not many other changes were implemented, if any.

(And I’d like to be in for the draw ;-)

[cortex_m0]

This would not be one of my top choices for a power supply. My first WTF moment was when Dave powered it on and it took 20 seconds to start. It really didn't get better from there.

I appreciate the min/max function, but it's useless if it includes the charging of the output caps at power-up, or at least needs an easy to access clear key.

The absolute accuracy on the current readout doesn't seem good enough to justify all those digits on the display. IOW, I can't claim to replace a bench ammeter with the internal current readout on this thing. The extra effort R&S went to to produce "10uA/1uA" resolution was wasted, compared to a model that shows 1mA resolution.

It looks like a good electrical/mechanical design, but poor software carries the day again.

[radioing]

As for the CC overshoot:
I'm pretty sure that the problem of this PS with voltage overshooting in CC mode occurs due to the fact that the voltage feedback loop is faster than current feedback loop plus the transformer winding switching (or some pre-regulator?; it's 100 V PS). Therefore, after the PS starts, the transformer winding switch is set according to by user adjusted voltage (e.g. 15 V) and voltage feedback loop is pushing the PS to required 15 V. But, with some delay, the current feedback loop starts to decrease voltage to reach the adjusted current limit 2 amps. As the voltage is going down, the control logic switches transformer winding to lower voltage and then current feedback does the final job. It is visible at Dave's osci screen that after the winding switch the current feedback must increase the voltage to reach required current.
With 3.3 V adjusted, the winding switching does not happen and only the slower current loop is the reason that the PS goes to 3.3 volts first and then down to 2.x volts (due to the delayed current feedback) to reach 2 A current.
The pulse train generated by Dave is the winding switching effect also.
It could be interesting to adjust 100 V and see how many times the winding is switched (how many section is there) to reach 2.x volts for 2 amps at the output (but it is out of the PS power limit so the adjusted current must be probably decreased to PS accepted value for this test).

[MBY]

I think I could live with the glitches

Please include me in the draw, thank you.

I'm in

I could do with a decent power supply - so would be great to be part of the giveaway :-)

I'm in... like flin!!! 

It occurred to me that this is the wrong thread for the giveaway! Post your "I'm in" once in this thread instead: https://www.eevblog.com/forum/contests/giveaway-rohde-schwarz-nga102-psu/msg3602896/

[madsbinger]

Thanks for the tear down and always good videos to learn from. I have learned a lot from you and keep learning. I would like to win, so I have something to use home in my private hobby lab.

[Prithul0218]

Count me in please! My Ruideng supply isn't good enough
Thanks.

[Brumby]

GIVEAWAY IS OVER.

Draw was made 19th July

(Besides, this was the wrong thread for the giveaway.)

[greencardigan]

There was firmware released a few days after the video was released.

V02.006 - https://www.rohde-schwarz.com/au/firmware/nga100/

The first improvement listed in the documentation is current overshoot.

Dave do you still have the other unit to test the new firmware?

[Pinkus]

I am in too.

[Refrigerator]

I am in too.

No you're not