QA401 Right Channel Noise Problem?

I just finished building a pre amp and decided to get some measurements from my QA401. All seems well, except the right channel is showing significantly more noise than the left.

After much swapping of cables, I think the QA401 may be at fault. The following screen shot was taken with the left channel QA401 output driving the left channel of the preamp. The left channel preamp output was split using a Y splitter and is driving both the left and right channel of the QA401 inputs. All single ended connections using terminators for the QA401 inputs.

I swap cables at the QA401 input and the high SNR remains on the right channel. I swap terminators – same thing.

I’m an infrequent user of the QA401. Using the latest software. Cockpit error or does the QA401 have a problem? All suggestions welcome.

Hi @Buckskin, at first blush it looks like something is up with the QA401. Can you do a File->Reset Startup Defaults and then select +6 dBV input and, place terminators on all 4 inputs, make sure both channels are enabled and then share the plot?

If the left and right channel are different, try swapping terminators, and also try rocking the terminators to see if perhaps they aren’t making contact? The plot you show looks like the right channel isn’t getting terminated. If either the + or - or both are open then it will revert to the higher noise floor (which is basically the noise from the 100K input impedance).

Thanks, Matt

Matt,

Many thanks for the response. I really appreciate it.

I’ve reset to startup, selected +6dBV and terminated all four inputs. Enabled both channels and got the plot shown below. No amount of wiggling or swapping changes anything. Time to take a peek inside??

Hi @Buckskin I think you should send it back for checkup. Something is definitely up. Please send a short mail to support at quantasylum dotcom and we’ll send a return label and figure this out.

Thanks, Matt

Email sent. Thanks for the help.

Sounds like you guys have this all solved… But, here is my experience today:

Bought QA401 a year or two ago, never have really used it until today. Got it out, used my laptop, no USB hub, laptop power supply, no other connections…

My right channel was showing a similar 20dB higher noise floor… Similar to yours. Mine had a 20dB bump at 60Hz though, yours looks less than 10dB higher at 60Hz… So in classic audio response to a 60Hz, started playing with grounding… The thing that solved it was a 24ga wire stuck in the USB jack and connected to ground (wall outlet).
Now I have L and R at the same noise floor, and no more 60Hz.

I figured there was a posting on this forum about how to properly ground, or not to ground… then found this thread.

What is the official stance on grounding when using a floating system such as a laptop?

Bob

" 24ga wire stuck in the USB jack and connected to ground (wall outlet)."

Huh??

only tried for about 5 minutes… This worked, so went back to the task at hand. This outside USB ground seems to be close to the same thing as a chassis cover screw.

I am sure this is not the official procedure… So looking for any official guidance on it.

Hi @pricom and welcome! The wire in the USB jack shouldn’t be needed. The ground on the USB side is completely isolated from the ground (common) on the audio side. That is, your DVM should report an open from the BNC shells (outer conductor) to the USB shell.

On a normal PC, the USB shell on the PC is at ground, and thus the USB shell on the QA401 is at ground too. But if you use a hub (which usually has a pseudo isolated supply or a laptop) it should show solid and comparable noise if both inputs are fully shorted with shorting blocks.

Can you share a plot of what you are seeing with the grounding wire? Thanks

Since you are interested, will dig in a bit and report back the findings.

Ok, some simple experiments. Reminder, this is a Laptop connected to wall power (US 115VAC), USB cable straight to QA401, no hub. My first test was to check shield (not ground) continuity on a sampling of ASB A/B cables. Some show very low resistance (under 20 Ohms), a couple are close to 0 ohms, and some are higher than 2K Ohms. So assuming a user provided USB A to B cable has continuity in the shield (not ground) is not good.

With nothing connected, and no caps on the QA401 inputs, it looks pretty good and L/R chans are close enough to each other FFT wise… This is with no ground on the USB jack on the QA401.

So the first test that is worth noting:

  • DUT connected to QA401 Output and not powered on, QA401 inputs open (no caps)
  • Laptop power supply attached
  • USB Cable shows continuity between shields

image

Please note the 60Hz peaks, and the Right Chan has more noise than Left. When I first saw this, as reported a few days ago, the right chan noise was more like 20dB higher than Left, but today it was closer.

Next test was adding shorting caps to the QA401 inputs:

image

Close to the same thing

Next test was adding ground wire to USB jack on back of QA401 and attached to cover ground screw of wall outlet.

image

60Hz is gone, and the L/R levels are together now.

For verification, now I disconnected the ground wire from the USB jack on the QA401

image

Back to the same thing, more or less.

Then next, disconnect the laptop power supply from mains power (it is a grounded power supply, but the output is floating). No USB Jack ground wire on the QA401

image

I have not looked at the inside of the QA401 to see how USB shield is treated compared to USB ground compared to circuit ground compared to audio ground compared to input BNC ground. For me, the ground wire solved the issue of the day, so I can go on to other work.

Bob

Hi @pricom, thanks for the measurements this is very clear now. The USB ground and shell are tied together in the QA401. And then USB ground/shell should show an open between USB shell and BNC shell. Can you verify you see an open between USB shell and BNC shell? I suspect it is.

The fact that your spectrum cleans up when the laptop supply is removed suggests that the isolated laptop supply has its “ground” swinging wildly, which means the ground entry into the QA401 is also swinging wildly, and that is radiating and being picked up by the analog.

A lot of folks thinks a laptop brick provides clean isolation. It provides isolation, but it’s usually not clean. If you laptop brick has a ground on the input, then it’s not generally isolated (you might see 1 kohm from output ground to the “chassis ground” (3rd plug–the round one). But these generally have a cleaner output ground.

But if your laptop supply has a two-wire inlet, then its output is mostly isolated (probably >1Mohm) but the “ground” could be swinging 10V or so.

The fact that the earth ground, when connected to the QA401 USB shell, kills the 60 Hz when the laptop is plugged in suggests your laptop ground is one of the ones that swings wildly. And you could probably clean it up also by touching the earth ground to the “ground” inlet of the laptop while it’s being charged.

If you have a scope, the scope probe ground is generally tied to earth ground through the scope’s internal wiring (you can verify with DVM). So, you can touch the scope probe tip to the laptop charger output ground and see what you see. Below is the “ground” measured on a Samsung 45W USB-C laptop charger.

There’s 8Vpp of 60 Hz on this. And if this is charging your laptop, then there’s 8Vpp of 60 Hz on the logic-side “ground” of the QA401.

Do you have a recommendation for a suitable wall wart to power a USB hub? Presumably it would need a three prong input based on your analysis above…

Just preparing ahead of time…

You should be able to connect ground anyplace–a mic input on the PC, HDMI output shell. It’s a pretty common problem on PCs where PC mic audio is rife with hum when plugged in. A lot of folks complain their PC mic has a hum when on AC power (HERE).

A good place to pick up a legit earth ground on smaller PCs (small fanless types that mount on back of LCD panel) is in the HDMI as many monitors still have an earth ground. But some have made plugs specifically to allow you add an earth ground to your laptop. See HERE

And here’s a hub that is grounded. Pretty pricey though (HERE)

Thanks very much for the suggestions.

I’m using an iMac here running Windows 7 though Boot Camp. The computer itself has a grounded power cable. But, almost every USB hub will use a wall wart that presumably has the isolation issue you described above.

I’ll experiment more to see what effects I’m seeing and what possible solutions work and don’t work. Will report back.

Here’s my QA401, inputs shorted, with no hub and with this hub:

Moving various components around didn’t make any difference. Yeah, there’s really only one tone, but it’s there.

I should’ve added that I’m using a better than average USB cable between the computer and the hub (a nicely shielded WireWorld sample I had sitting in a drawer) and a better than average WireWorld cable between the hub and the QA401. The latter is the same cable I used between the computer and the QA401 for the no hub test. The hub stayed plugged in for this test, both to the computer and to the AC adaptor.

I was happy to find this thread to seek help with a similar issue but after some tinkering I narrowed this down to a broken BNC 50Ohm terminator on the ‘-’ connector of one of the channels, it checks open with a multimeter. With the input shorted using a trusted paper clip both channels are now matching, phew.

What is the recommended method to terminate the other input side for unbalanced measurements? A 50 Ohm BNC terminator or just a short? Or perhaps using central terminals of the +/- BNC connectors is yet better as it should reject common mode noise?

Hi gvl. At this link I have found useful tips to improve the measures by reducing the noise that is introduced by the setup of the measures themselves

Interesting presentation. @matt do you have any thoughts on the approaches mentioned in it?