EEVblog #549 - Rigol DP832 Lab Power Supply Followup

An update on the Rigol DP832 lab power supply fix.

EEVblog Main Web Site:
EEVblog Amazon Store:
Donations:
Projects:
Electronics Info Wiki:

[TehGordonFreeman]

EEVBlog, the only YouTuber I know who says 'A very quick video...' at the beginning of a 20 minute video... But somehow keeps me around for the entire time anyway...

[CamronLocke]

I'm 14 and I learn so much from EEVblog I've learned ALL the basic parts to a circuit from you Dave! Greetings from North carolina!

[sharedknowledge]

A Chinese engineer I worked with said what a pcb LOOKS like is the most important criteria for Chinese management. They apparently don't usually care about much else, including the thermal design, as long as it looks "high quality" and similar to competing products. He said Chinese managers are rarely well educated in engineering and focused much more on the marketability of products, matching the most obvious features with the competition even if the features don't fully work, and they don't fix known problems unless they become a significant financial issue. If a product is bad enough it's sold off, rebadged If needed, and often marketed on eBay with essentially no usable warranty. They usually recover their investment even when the products are grossly flawed. Which, in turn, gives them little incentive to properly engineer anything.

[jurgentraude]

Great job! And for RIGOL I have to say they are listening and try to improve! Thumps up! Credit to you Dave.

[chrispychickin]

I noticed a few minor layout differences other than the lm317 you mention:
1> top right corner, different layout and one less small electro (between the 3 large ones, and the power devices above them)
2> a few minor component selection/layout differences on the bottom of the board, check the far right, and around the regulator on the left.

[avidrandomer]

This was a common problem with some power supplies I was using back as a student. If we had a circuit with multiple input voltages but only 1 ground plane, we were always told to short the grounds on the supply together.

[jelberth]

Hi Got my DP832 PSU today. I did buy it mostly based on the reviews here and thought most of the bugs would be resolved by now. But the OVP is a big FAIL. When I put a 12V lamp on output 1 and set the OVP on 12.5V. Then I put the output on 11.6V and turn the knob to jump from 11.6 to 12.6. The output first jumps to 12.6 stays there for 330ms and then turns off the output. But if I do the same with a jump from 11.6 to 21.6 the output stays on for a bit more than 500ms and then turns off. That is way too long and can fry your very expensive electronics just before turning off.
So basically the OVP is only useful when the increment steps are small.
The negative pulse on the 5V at switch on is still present.
I have software version 1.14 and that is the latest one.
They could simply resolve this in software by not allowing the voltage setting to increase over the OVP value wile in constant voltage mode. Instead they measure the voltage on the output and if that one goes above the OVP value is shuts the output of. That routine is slow so the output stays on for a long time. If OVP is on that routine should get a high priority.
This is poorly designed software.
Maybe a new followup?

[zuccasnow]

What about the overshooting voltage issue when switching on the channel? We should not forget that. Dave took apart this device for investigating this problem...

[ericvieira]

7:19: Two components between the new heat sink and the left edge (inverter + power resistor?) have been rotated 90° (lower left side of heat sink)...

[panzerschrekIOI]

I freaking like those videos. Keep up the good work Dave.

[Jobucko]

"Just a quick video" 20 minutes.  I love

[benjaminpajk]

Emona Instruments established by Alfred Brežnik. Nice to see Slovenian roots Down under. By the way, Emona was a roman city, right where Ljubljana - today's capital is.

[jonka]

This all makes me think that Rigol's management system is poor from the top down. Ok, so some young designer was told to deal with the 5v supply and made a poor set of choices. However as Dave asks how did this pass muster and even get past first base? I wonder if any changes have happened at Rigol's management and product development stages.
I also wonder what (if anything) would have happened if Dave had not flagged this to a global audience. Good on you Dave.

[FireMouseHQ]

So it maxed out at around 80 degrees now. But what if it gets mounted into a rack with a higher ambient over 40 instead of the normal 21 room temperature?

[TradieTrev]

Sounds like a trap for young players! Well spotted Dave and nice follow up!

[tubical]

Dave, usually i agree on how you put a light on this time...IMHO this is *not* a "...killer precision power supply". (check it at 20:14) To be honest: NO

A real -IMHO- killer PS would have:

1) Isolated and/or insulated grounds (=read individually) may with a switch to put them together if wanted but not by default.

2) No cap in the output after the regulator.

3) No resistors in ground path at all !!

[volodymyrzakolodyazhny]

Why they haven't just put that LM317 onto that medium-size plate radiator (in the center)? It could be easier and cheaper.

[Darkfuturee]

What about an update to your's power supply designs?

[redtails]

9:40 you do have pretty high mains to have upwards of 245-250v.. Then again, these sorts of machines will have quite some tolerances I can imagine. Having a trannie makes a device way less flexible, though. You'll see that switching PSUs can easily handle like 70-250v without any differences in output voltage. If this was Photoninduction's channel, he would certainly have hung this PSU on a variac to see if the issue was the higher voltage.. Like maybe this machine was made for 230v max and they're just stretching the limits for the Australian market. Anyway, even without variac, Dave's investigation is nice

[ronaldlijs]

Thanks Dave, still not great. I was planning on buying one of these, won't now.