buried in the noise floor but measurable

going over some units that have a hum, you can measure it and it's well buried in the noise floor of the unit but we can measure it. Since it's not audible is it worth even worrying about?

Hum is generally a noise that could be engineered out with better layout discipline or audio signal integrity management. Hiss on the other hand is often intrinsic  to devices used so can be minimized but not completely eliminate.

How much hum is acceptable is subjective and arbitrary... how much time and patience, and money do you have to tweak the design?

When repairing an existing design, re-engineering is not often an option.

JR

PS: Hum in a guitar pick-up is hard to eliminate, hum inside a metal chassis should be easier.

Thanks for the insight John.

pucho812 said:

going over some units that have a hum, you can measure it and it's well buried in the noise floor of the unit but we can measure it.

Click to expand...

How much?

Since it's not audible is it worth even worrying about?

Click to expand...

As JR says, it depends. Hum maybe there from day one, so you'll have to live with it, but it may also be a sign that something is deteriorating (lytic caps), so it may be worth investigating. Also hum may come from a newly installed adjacent piece of equipment radiating interference. That is also worth looking at.

It's an off the self PSU, brand new that was being considered for some design stuff. It is external but is a brand new thing...

pucho812 said:

It's an off the self PSU, brand new that was being considered for some design stuff. It is external but is a brand new thing...

Click to expand...

The plot thickens...
Please characterize hum, is it acoustic noise emanating from it or is it electric hum measurable at the output? Is it 60 or 120 Hz? How is it measurable? Directly or via some equipment that it powers?

it's 60 cycle and on the output, measured via an ap1

pucho812 said:

it's 60 cycle and on the output, measured via an ap1

Click to expand...

Is it the indication given by the measurement panel on the AP1, Or is it what you see on a 'scope (connected at the "Processed signal" output)?
What level? What is the nominal DC voltage?

Pure hum is easily audible far below random hiss.

Pure hum is totally avoidable (if you care enough). Run it on batteries in the woods, no hum.

Put in the studio, actual operating conditions, and use your ears. Will the customer notice?

Will have to do the tests again on the AP1....

Not trying to induce a bad case of gear envy, but an AP2 that has DSP can allow you to use averaging techniques to separate correlated noise sources from random noise sources. Essentially, taking the average of a number of FFTs of the residual from (for example) a THD measurement allows you to push the noise floor down, much as the ear does. This works because a truly random signal, when added to another truly random signal, will only increase in level by 3dB, whereas a non-random signal added to another similar signal, like mains hum or a spurious distortion product, will increase in level by 6dB.

Stated differently, averaging two FFTs will push the noise floor down 3dB, but present non-random interference at the same level. Averaging four will push the noise floor down 6dB, and so on. Basically, you get 3dB improvement for each doubling of the number of averages you take, or 3dB * log2(# of averages). This allows you to dig into the noise floor and accurately separate correlated noise/interference/distortion products from uncorrelated (i.e. truly random) noise.

While this is simple to do with an AP2, I'm sure this could be done with a computer and a sound card, set up to take FFTs of its input and average them in the complex domain. This also might be done by recording some signal and then doing post-processing to do the averaging with some clever software along the lines of Matlab.