Question about oscilloscopes for audio

The terminology is a little misleading here. It's useful to digitize audio signals to 16 bits -- and your ear can certainly tell the difference between

8 bits and 16 -- but for *viewing* a signal you probably aren't going to get much useful information out of more than 8 bits unless you have a device (newer digital scopes, a spectrum analyzer, etc.) that can display, e.g., log(input signal) (in other words, can display the input signal in dB).

For a headphone amplifier you'll be able to perform design & troubleshooting just fine with an 8 bit scope. If you're building a very high accuracy amplifier, however, you'll need something better once it's finished to measure, e.g., total harmonic distortion.

Note that many (even most) low cost "16 bits" oscilloscopes will not actually achieve 16 bits of resolution without a *huge* amount of averaging -- 16 bits is over 100dB dynamic range, which doesn't go together with "wideband" and "cheap."

---Joel

Ahh, I see. But isn't total harmonic distortion always an important factor in audio design? Personally, I don't mind a little distortion as long as I can still enjoy the audio.

Thanks, Joel!

This is what works for me. A USB audio interface unit for your PC, a good one is the M-Audio Audiophile -USB found here for $150.

This will give you 2 input channels and 2 output channels of 24 bit- 96KHz performance, fully isolated. Then get a software package to interpret and display the data on your PC in scope or spectrum formats. Here is a good one for $100. This setup will output sine waves with harmonics down >90db and has an input noise floor of -110db at 1.0KHz and will do a full frequency response with one keystroke in

Harry, your posting sort of baffled me, I had it in the newsreader, then decided to erase it and convince myself I had never seen it :-)

There are several reason for wanting to do that, first I started wondering 'What headphone has 110dB range, my Sensheiser HD201 does not dampen for example room (PC fan) noise enough to get to that, given distortion at the upper end'. Then why all the expensive stuff, while every PC has a 16 bits soundcard, and programs that are free like 'oscope' (Linux) exist, also there are free fft packages.

Realy I have done audio with a normal analog scope and may still prefer it to a digital one in some cases. As to 16 bits scopes, OK, lets say most sensitive range is 10mV (1mV for a good analog audio scope existed too) per division. So if 10 divisions on the graticule, 100mV, and 65536 steps (2^16), the lowest signal would be 1.525 MICRO Volt. Now maybe somebody here can make a headphone amp with that low noise, but I cannot.

So that is why I wanted to forget about what you wrote, but something in me looked it up in google (sorry if google reformat my text in weird ways), and as I am very old, maybe I have been missing some important new audio topics.

Hell you can get a good idea about distortion simply by substraction output from input. And since mp3 nobody hears anything anyways. There were days of HiFi and .000001 % (;-)) harmonic and crossover distortion, but since the appearance of the 10 cc PC speaker I am not sure it makes a difference.

So, for what it is worth: And you can take a metal screw, drill it in the skull, wind some coper wire over it, bias with a magnet.... just like chewing cookies. (This last sentence was a joke OK A JOKE). Pfff.

So please correct my simplistic audio views if you find thse flwaed. But in my days audio sounded a lot better then mp3 on PC speakers.

24 bit, 48 bit or whatever...

8 bits is just about ok for a scope I gues, just about the thicknes of the trace line I gues, although if it has zoom and other functions this makes it a bit limited.

But if you realy need to do much more than just look at the basic shape then maybe you need something more than a scope.

Colin =^.^=

Its all Johnny Cash's fault! After all, he got it "One piece at a time", instead of all in one piece. ;-)

--
Service to my country? Been there, Done that, and I\'ve got my DD214 to
prove it.
Member of DAV #85.

Michael A. Terrell
Central Florida

Memories! And singers had to 'know' how to talk (spell) before beeing sicked against audience.

Have fun

Stanislaw

"PPP

** Groper alert !

** For audio design and * troubleshooting * work - what you need is a standard * ANALOGUE * scope.

Typically dual trace with 20 MHz vertical bandwidth or more.

Are none available in your lab ???

Using either 8 bit digital scopes or a PC soundcard as makeshift audio scope is plain dumb - the former suffer from poor image quality and aliasing while the latter has insufficient bandwidth for what you intend doing.

....... Phil

I am in complete agreement with Phil on this. An Analogue scope with

20MHz bandwidth can be had for relative peanuts and is indispensible for this sort of work.

Cheers

PeteS

Aliasing is a thing of the past now that virtually all DSO's (even the low end ones) use real-time sampling (usually >10 times the bandwidth) instead of repetitive sampling.

But yeah, agreed, a regular analog scope is what you need here.

Dave :)

This "higher accuracy" is so that software can make FFT out of the signals that accurately show signals 90dB below the max (for showing low levels of harmonic distortion), neccesary for testing high-quality audio systems. FFT's from 8-bit scopes can show at most 40 or 50 dB below a max signal.

The truth is, there were pretty good headphone amplifiers made before there were 16-bit analog-to-digital converters. But if you want to test your amp for the lowest distortion using the easiest and most convenient way, you should use one of these:

• An oscillocope with 16-bit vertical input and FFT capability (I'm not familiar with any 16-bit models, and like in your lab, all the ones I've used are 8-bit).

• The M-Audio or similar interface and a PC with appropriate software such as
  formatting link

• A hardware system made just for the purpose such as an Audio Precision system
  formatting link

There are other, analog ways to measure distortion, but these systems make it about as easy as hooking things up and pushing a button.

On the other hand, you likely won't need these to make your amp. Just put it together and if there's no obvious distortion (in your ear or shown on an 8-bit or analog scope), then you almost surely did it right.

I still have problems, on rare occasions, getting a modern DSO to behave due to aliasing issues. Usually turn on the analog scope to figure out what I'm looking at then set up the DSO accordingly.

For audio, noise, and odd digital problems analog scopes are the way to go. If you need to measure distortion, use a decent spectrum or distortion analyzer.

--
Mark

I have and use analog scopes but also have and use combination analog and digital scopes. I have an old Phillips PM3350A and often recommend the Hameg combination scopes with CRT's. HM507. I think they deleted a cheaper model with less bandwidth.

More bandwiths is useless, but more samples is good, as well as at least 12 bit vertical resolution minimum if you have to use digital.

greg

greg

go

a

Design

scope

At a customers last week, had to use their new, Tek TDS3024 'scope. Must admit to being somewhat leery due to a previous time waste with a HP model. Yep!, right on cue the damned thing refused to display a 100Hz, 15V, 20uS signal. It looks like a girls handbag and about as much use :) john

--
Posted via a free Usenet account from http://www.teranews.com

Sorry if my post baffled you, I am known to do that. A post by Ben Bradley in this thread said best what I was trying to relate. You do need a scope for design of audio equipment but for final test and verification your client my need data in the frequency domain. Using a PC based sound card normally has high audio and EMI noise levels. The USB sound card that I proposed is at the end of a 5 meter cable and electrically isolated from the host PC. This allows gain and distortion measurements from any node in your DUT. The -110dbu noise levels are nice even if you only require -50dbu. The software I suggested controls output waveforms and crunches input data yielding great printable curves of the DUT's performance. YMMV Harry

The myth continues through many generations that HP knew how to build oscilloscopes.

Their last decent scope was the 130C. It has been stadily downhill ever since.

--
Many thanks,

Don Lancaster                          voice phone: (928)428-4073
Synergetics   3860 West First Street   Box 809 Thatcher, AZ 85552
rss: http://www.tinaja.com/whtnu.xml   email: don@tinaja.com

Please visit my GURU\'s LAIR web site at http://www.tinaja.com

There is a special button on the scope just for such occasions, it's called "Autoset". Students today make extensive use of it ;-)

The new Agilent ones are brilliant. The older mixed-signal 54620 series or the newer 6000 series are just so easy and intuitive to use, and the

6000 series performs as good as an analog scope in most respects. I've lost count of the number of times people have come to borrow my Agilent because they couldn't figure out how to drive their Lecroy or Tek.

Dave :)

The rest of their stuff is bad, too. Their all-in-one printers with hundreds of megabytes (Yes, I am not kidding) of driver bloatware are horror. Won't print B/W fax without installing a new colour cartridge first (which by the way was not empty anyway as I only printed BW faxes).

A bit OT, but had to get it off the chest.

SioL

[snippage]

In John's case, he was looking at a 0.2% duty cycle signal. That's hard for a digi scope to deal with when looking at a full cycle or two.

--
Mark