Agilent U3606A (DMM + DC): Noise on DC output

[anotherlin]

Hi everyone,

I just purchased an Agilent U3606A DMM + DC combo.
It is a 5.5 digit DMM + a 30V/1A-8V/3A DC supply.
Both functions can operate simultaneously and independently.

The DMM seems to work just fine. I've measured a few things around and compared results with my handheld Fluke 179.

The DC power seems problematic. The output is correct. I've tried various voltages and currents, using my fluke to measure them.
They were always spot on. So I decided to plug the output to my scope to check out the "cleanness" of the DC.

Problem: it is extremely noisy, see the attached screenshot. I'm testing using 0.5V. And ripple is basically 200mV peak-to-peak.
That's way out of spec (0.01% + 3mv).

I've tried switching the scope to AC coupling with a higher "speed". And I'm have some form of sinusoidale waveform at 500KHz.
I've tried to check on the internet, seems like I may not be alone : http://www.home.agilent.com/owc_discussions/thread.jspa?messageID=98850

Is my brand new U3606A faulty ? Should I contact Agilent ?

I would definitely appreciate some comments or suggestions.

Best regards,

[alm]

From your description it appears that the unit is faulty, but I would repeat the same measurement with the probe tip (and ground lead) both connected to the negative output terminal to exclude any common mode problems. I would definitely talk to Agilent tech support and see if they have any further troubleshooting instructions. It doesn't happen to have a noise / sine wave generator feature that is accidentally turned on, does it?

[bsgd]

Hello,

I have this unit, but never really used it. It is in storage somewhere.

I did quickly tested it and found an issue: when powered on with the probes connected to the DMM jacks, the power on self test would take ages to finish and then the unit would reset the stored setting to factory default or some other random values and start misbehaving.

Can you try and check if your unit has the same problem?

Mine would misbehave completely when powered on with probes connected. Did you try to do the same test powering it without probes?

[anotherlin]

The unit is brand new. I received it this thursday.

I've always turned it on with probes unconnected. And it powers-up fast.
I can give it a try with probes on, but I don't think this should be problem (why would be ?).

The DC supply can work as squarewave generator. For my 0,5V test, I didn't turn it on, of course.
But I've tested it (the squarewave gen), and it is working. Frequency, duty cycle, can be set. And it all works fine.
Except that I end up with a squarewave with the same kind of noise.

I'll post a picture of that tomorrow.

[bsgd]

anotherlin,

My unit was also brand new, new in box. For some strange reason, whenever I have the probes connected, it would misbehave, either giving me false readings, or loosing all stored data, etc.

I dont know why is that. Heres is a way that you could test yours:

1- Power it on and set it to read DC Volts, autoranging on.
2- Turn it off and connect probes to DMM jacks
3- Turn it on and see what happens.

Im not sure this is somehow related to your problem, as you never turned it on with probes connected. All I know is that I certainly have some weird issues with my unit, and thats why I dont use it.

I will try to locate and test for noise. I never did that.

[bsgd]

I have tested my 3606 and if yours is defective, mine is too.

I get the exact same noise, no matter whether I use S1 or S2. IMO this unit has some firmware bug and is not up to Agilent quality standard.

So far I have found two issues with this unit: the power on selft tested with leads connected and the high PS noise. Not good.

[bsgd]

Me again.

I just noticed Agilent released a firmware upgrade. I went ahead and upgraded my unit. The process went well, but at the end, guess what? My unit died completely.

Well, good thing I dont really need this. I am sure this is not a good product... shame on Agilent.

[T4P]

Contact agilent and they will help.

[anotherlin]

Ok, I've done some further tests.

So I've tried to power-on with test leads, no problem whatsoever. 
Selftest runs for about a second and everything works fine.
I've even gone into the "utility" menu and manually launch a self test, no problem.

The DC supply can act as a squarewave generator, so I've given it a try.
I set it to 1.5V (squarewave peak-peak from 0V to 1.5V), at 480Hz (around 2ms period), with a 25% duty cycle.
As you can see on the attached picture, except for the noise, it works fine.

I've checked by feeding the output to the DMM input, readings all fine. And double checked with my Fluke handheld.

I'm puzzled

[anotherlin]

One more picture.
This is with a 0.5V DC from the U3606A.
As you can see, the sinusoid has a 2us period (500Hz).

[ThongLin]

Please be informed the U3606A Ripple & Noise is specs from 20Hz to 1MHz. In the U3606A User’s Guide page 166 a differential amplifier is recommended to connect the source output before connect to Oscilloscope for Ripple & Noise measurement. The basic idea of using a differential amplifier is to filter noise above 1MHz and also amplify the U3606A output for more accurate measurement. Due to the x10 amplification from the differential amplifier the result should be divided by 10 in order to obtain the actual value.

U3606A User’s Guide download URL
http://cp.literature.agilent.com/litweb/pdf/U3606-90013.pdf


If you need technical support for your highlighted U3606A unit issue, please contact your nearest Agilent Service Center or submit your technical support request via Agilent Technical Support link below. I hope this is help.

https://www.home.agilent.com/agilent/editorial.jspx?cc=US&lc=eng&ckey=1536179-1-eng&nid=-536902435.899262&id=1536179-1-eng

[saturation]

A common error is to stick a bench scope into an floating DC PSU and see large ripple and other noise.  You will pick up common mode voltages from various sources between the isolated ground of the PSU and the earthed scope ground.

The easiest way to test the output of a supply is to float your scope, which means its safest as a battery powered scope.  If you still see ripple/noise with a battery powered scope, you indeed have something to investigate.

Otherwise, you're left with this.

http://www.eetimes.com/design/power-management-design/4012479/Power-Tip-6-Accurately-Measuring-Power-Supply-Ripple

[anotherlin]

Yes, I was aware that scope measure from earth ground and DC supply output is floating. That's why I wanted to post on the forum before actually going to check with Agilent.

I don't have a differential amplifier. But yet, I think I can do a test by using a 9V battery powered opamp in a differential amp circuit, with DC supply as input of opamp, and measure the output. This should reduce much of the common mode.

Yet, I'm wondering, what would the usefulness of a DC supply if it has lot of common mode noise on its output. My circuits using it (the supply), would work well, but how can I then measure it easily ?

[saturation]

Hi!, now IIRC:

Common mode noise maybe an issue if your circuit design is earth grounded and your design requires noise isolation better than the PSU can provide, switching PSU noise is ~ in mV, which I think the 3606A is.  If you were to design low noise circuits, this PSU could be a problem that's why for general use, a cheapo linear supply is far quieter and applicable for most design work.

To measure the PSU noise properly, you have to remove noise introduced by the scope.  It may only exist when a 1Mohm+capacitance load is placed between the negative input terminal and earth ground.  You could try an isolation transformer for the PSU and measure.  Isolation transformers still have active chassis grounds, and your PSU will likely have the same.  You can still pick up noise coupled by unintended capacitance from the jacks to the chassis or the extend leads of the probe, particularly from RF sources, radio signals and wifi is a good source of contamination.

The best and easiest way to measure noise is with a fully floating portable scope even a cheapo like a Velleman; its cheaper than a differential probe, which is your other alternative.   Having a portable cheapo scope to test line voltage circuits and other designs best done with isolated inputs does come in handy, because the inputs to DSO can only take so much differential voltage between its inputs and ground.  Sacrificing the Velleman, worse case, is less painful.

Yet, I'm wondering, what would the usefulness of a DC supply if it has lot of common mode noise on its output. My circuits using it (the supply), would work well, but how can I then measure it easily ?

[anotherlin]

Just one more question :

Some Agilent supplies (like the U8001 or most of their E36xx series) have three connectors : plus, minus (common), and earth ground.

How about I ground the common (minus out) with earth ground ? This way my scope and the supply both use earth ground as reference ?

Mine (U3606A) doesn't have earth ground up on front of unit, but I can use a cable to do so.

[saturation]

A PSU without a ground pin will float, its the better kind and less noisy.  If you short the - pin to ground, the negative pin is now referenced to earth.  However, this does create a true ground loop impressed on the circuit under power 

Just one more question :

Some Agilent supplies (like the U8001 or most of their E36xx series) have three connectors : plus, minus (common), and earth ground.

How about I ground the common (minus out) with earth ground ? This way my scope and the supply both use earth ground as reference ?

Mine (U3606A) doesn't have earth ground up on front of unit, but I can use a cable to do so.

[anotherlin]

I've done some further testing this weekend.

So I don't have a differential amp, but it is easy to make a very basic one using an opamp  (see wikipedia)  and two 9V batteries.
That's basically what I've done using a OPA2134 and 4x1.5kOhm resistors (unity gain), so Vout = VpositivePSU - VnegativePSU.

And indeed, the 0.200mv 500KHwnoise is common mode. With this differential opamp buffer, the noise is gone.
So my U3606A may be working just fine after all. Yet it is not much of any use for analog application (guitar preamp).

However, with a direct connection, no load, direct to oscilloscope probe, the output is "clean" for about a second, and then noise is there.
So I think my unit is actually faulty.

I'll try to post you some oscilloscope screen shots.
And I'll contact Agilent monday.

[Isamun]

Would be nice to hear some updates on these issues. Maybe someone has bought it recently and has some new insight (no pun intended)? Looks like a handy piece of kit.

[jcoigny]

I have the exact same issue with all 3 of my U3606A units.  I have finally found a way to improve it quite a lot.  I use the rear terminals for my power connections to the rack its installed in.  Basically I have to connect a wire to the earth ground screw near the terminal block to earth ground on the terminal block.  Then I connect a shunt on the terminal block between earth ground and S-.  This gets my noise down to sub 10mV range.  It is of course not as clean as other supplies I have plugged into the racks power input but its fine for this application. 

I was really frustrated by the 200mV rms noise I was experiencing with this supply.  While in America it worked pretty well using my standard 3 prong power cord however after shipping these racks to Taiwan the problem really cropped up.  They only have 2 prong power and no earth ground runs through the buildings here.  My solution probably works a whole lot better when earth ground is brought through the power cord of the meter.

Anyway I hope this helps someone, if there is anyone still using these things anymore.

[nfmax]

I checked my U3606A the other day. I get only about 50mV pk-pk, as seen on my 200MHz MSOX2024A, all of which appears to be common mode. Everything on the bench has a mains earth taken to a common point (i.e. it all runs off one socket)