Comparing Power Supplies

[hcoverlambda]

So I just got a refurbed Rigol DP832. Had an eBay cheapy for a long time (Mastech HY6003D seems to be rebranded a million times). I'm comparing the voltage of the new and old power supplies and the Rigol looks much noisier than the Mastech. I'm not sure if what I'm seeing is good or bad and my google-fu is failing me. Can anyone point me in the right direction?

BTW, when I received this unit, output one was stuck in critical mode. Cracked open the case and two cables were disconnected and the pins bent (See the pics here https://www.facebook.com/media/set/?set=a.1675093016069677.1073741839.100007069846938&type=3). O_o Anyways, that has made me more concerned that this unit might have issues.

[tautech]

Welcome to the forum.

Is your Ch 2 probe set to 1x ?

[hcoverlambda]

Thx man!

Naw, both set to 10x. Good idea tho, I didn't even think about that.

[max666]

Looks way out of spec indeed:

Low Ripple and Noise: <350 uVrms/2mVpp

[hcoverlambda]

Aaah, cool didn't even think to look at those specs. Thx!

[rx8pilot]

How are you probing it? The probe should be directly on the posts and keep the ground loop small or use a spring for the ground. Check to see if you get the same results on Ch1 and Ch2 probes.

Limit BW to 20Mhz as it specifies in the data sheet. I don't know where they measure the noise spec for the data sheet or what conditions the PSU is under at the time - voltage and current. I suspect that the measurement is taken at the output stage across a capacitor. Leads will act as an antenna and pickup all the available energy on your bench.

[nugglix]

Hi!

Can't complain about my DP832.

Regarding the correct way to measure the output of power supplies, there might be some ideas hidden in:
   http://www.eevblog.com/2014/03/21/eevblog-594-how-to-measure-power-supply-ripple-noise/

Cheers

[hcoverlambda]

Awesome! Thx for the great suggestions and vid, will try/check them out.

[max666]

How are you probing it? The probe should be directly on the posts and keep the ground loop small or use a spring for the ground. Check to see if you get the same results on Ch1 and Ch2 probes.

Limit BW to 20Mhz as it specifies in the data sheet. I don't know where they measure the noise spec for the data sheet or what conditions the PSU is under at the time - voltage and current. I suspect that the measurement is taken at the output stage across a capacitor. Leads will act as an antenna and pickup all the available energy on your bench.

I agree.
Common mode noise can be a bitch too, Dave explained it in one of his videos, don't know if it 's in EEVBlog 594.
But since the Mastech, which I assume is connected the same way, doesn't show excessive noise, I'm inclined to say there is something wrong with the Rigol.
Edit: You could connect a power resistor or an incandescent light bulb and measure again under load, maybe that sheds some addition light on possible failure modes of the power supply.

[hcoverlambda]

Wanted to do a quick follow up. Watched the vid (Exactly what I was looking for BTW, thx!), and as a quick sanity check adjusted the scope from DC coupling/10x/No BW limit to AC coupling/1x/20 Mhz BW limit, also connected directly to the power supply and got some very different results. I need to dig deeper into that vid but looks like just those small tweaks helped get a much clearer picture. Another thing I noticed about the Mastech is it's not at all stable. Was jumping around like a bull at a rodeo. In fact it took me a bit before it settled down enough where I could take a screenshot. So it looks like like the Rigol (Blue) is at least in the ballpark of the specs and the Mastech (Yellow) is worse (as I would have expected). Going to try the cabling Dave suggested in the vid and see if it gets me to his measurement of the DP832 (Last pic).

Anyways, thx for all your suggestions!!

[hcoverlambda]

BTW, just powered back up to zoom in a bit and the Mastech is Vpp has gotten worse. O_o Not sure why that happened but here is a 5ns time scale.

[tautech]

Basic scope use lessons required.

Questions
What frequency ripple would you expect? Set the timebase for that.
What amplitude of ripple would you expect? Set the V/div for that.
Would you expext or be interested in MHz of ripple? No. Turn the BW limit on.
Still have some jitter of varible peaks? Turn Averaging on.

That should get you some readings you can make sense of. 

[amyk]

You may also want to check for bad caps.

[hcoverlambda]

Great, thx!

[ConKbot]

Youre picking up FM broadcast radio. Notice how once you get away from the trigger edge the traces dont line up? Thats the modulation of the signal

The rigol has a higher impedance at 100 MHz than the mastech.  This could be something as simple as the location of the output capacitor on the rigol vs mastech.  I.e. ive seen some cheapie supplies where the output cap is soldered almost right to the output terminals (plus they like to use big caps on cheap supplies, which is frowned upon for supplies with a constant current mode)  I cant seem to find a definitive picture quickly for the 832 output cap location. 

[hcoverlambda]

Aaahhh, would never have guessed that, thx! I remember Dave saying in the vid that the ground wires on the probes are like antennas, so makes sense. Just ordered some coax cables and bnc/banana adapters, will be interesting to see the results when using those.

[Stupid Beard]

I cant seem to find a definitive picture quickly for the 832 output cap location.

If I remember right it's 1000uF right at the banana plugs. That's from memory so could be inaccurate.

[hcoverlambda]

I cant seem to find a definitive picture quickly for the 832 output cap location.

If I remember right it's 1000uF right at the banana plugs. That's from memory so could be inaccurate.

Wow, spot on!  https://www.youtube.com/watch?v=s4rV20ulsm0&feature=youtu.be&t=19m43s