Something that has been exercising my mind lately is building a "control unit" for the test bench. It started off, simply enough, as an I/O device for computer-based audio testing. But then I got to thinking about integrating my hardware test equipment. Something to simplify connecting the various items of test equipment to devices under test (DUTs), including breadboards. Something to let me listen to and see, analyse and measure the various signals with a minimum of messing around. Here's my current thinking…
It would be in a standard 2U rack box, and sit at the bottom of my other rack gear on the shelf at the back of the test bench.
Starting at the left of the panel, there would be a Source selector for each of the two channels feeding the DUT. Options would be:
· Computer soundcard output (from software signal generators such as Visual Analyser, REW, Rightmark, etc)
· Low distortion oscillator (existing hardware)
· Function generator (partially built hardware)
· Spare(s) for future developments (who knows?)
· Other channel (to make sending exactly the same thing to both channels easy)
The source selectors would feed a pair of line drivers via preset level controls (to calibrate the max levels) and variable front panel controls, for easy adjustment. The line drivers would take the signal up to say +20dBv to allow for overload and headroom checking of DUTs. Stepped attenuators on the front panel would take the levels down from +20 to –60dB in 10dB steps for microphone and other low level input preamp testing. There would be a final " minus infinity" position for noise measurements (no voltage but same output impedance). Output RCA connectors on the front panel would feed the DUT.
Then we come to the input areas. There are three.
Firstly we have a pair of BNC inputs feeding high impedance attenuators and buffers designed to work with standard oscilloscope probes. The probes will be used for getting into the middle of the DUT or breadboard. There have been a number of suitable buffer circuits published over the years. The usual thing is to have something like x1, x10 and x100 attenuation, and maybe x10 gain. One interesting option would be to use the Nuova Elettronica approach which connects to the computer via USB rather than soundcard. Because of the high impedance, these signals will be fine to look at, but can be expected to be a bit noisy for in-depth analysis. (Or am I being too fussy?)
Secondly, there would be another pair of RCA connectors for the outputs of the DUT. This direct connection will be used for serious measurements of noise, distortion etc, so needs to not contribute significant degradation. I guess some gain or attenuation may be needed here too to facilitate testing at elevated and lower levels? To save panel space, perhaps this attenuation could be controlled by the same rotary switches as used for the probes?
The microphone input would lead to a low noise balanced mic amp with phantom power, for measurement mics. Not sure if one or two channels, but probably go for two, if only to allow relative phase checks? Question – stepped attenuator on front panel, to help keep calibrations valid? Heh heh, or the probe switches again?
These feed the Analyse selector switches, which would have these options:
· Source (output of the Source switch, i.e. whatever's being sent to the DUT)
· Probes, for probing about inside stuff
· Output from DUT
· Microphone(s)
· Spare(s) for future developments
· Other channel
These Selectors feed outputs to the hardware Noise and Distortion meter.
They also feed a two position Monitor selector switch, along with the analysis output of the N&D set (ie the leftovers after filtering out the fundamental). The Monitor selector feeds the computer sound card input, the hardware Digital Storage Oscilloscope and the monitor speaker amplifier.
There might be an External Trigger selector for the hardware oscilloscope, with choices Source L, Source R, mains. This would keep the DSO triggered even when nulling out signals.
Finally, we have a Stereo/Left/Right/Mono/Off selector, and a volume and balance control. They feed a stereo amp in the unit, say 2 x 40W, which drives the test bench loudspeakers. A headphone socket switches speakers off if headphones are used.
So, in a typical setup, I could choose between the signal sources, and send one or more to the DUT. I can set the max level to anywhere between –60 and +20dBv, or zero, and run calibrated (full level) or variable. And I can measure, see and audition that level on any of my measuring tools.
I can probe around inside the DUT with the probes, or select its output, or a combination. I can compare input to output. I can also input from microphone(s), in which case I'd monitor on headphones. I can measure, see and listen to either channel output, or listen to and view the noise and distortion products, or analyse them on the computer FFT. My hardware oscilloscope can view along with the rest, and be triggered by the source or mains to look for hum. I can readily set up for full analytics using Rightmark and other software.
Is it a plan? Is it unnecessarily complex? Or am I overlooking something important? What would you do?
Terry
It would be in a standard 2U rack box, and sit at the bottom of my other rack gear on the shelf at the back of the test bench.
Starting at the left of the panel, there would be a Source selector for each of the two channels feeding the DUT. Options would be:
· Computer soundcard output (from software signal generators such as Visual Analyser, REW, Rightmark, etc)
· Low distortion oscillator (existing hardware)
· Function generator (partially built hardware)
· Spare(s) for future developments (who knows?)
· Other channel (to make sending exactly the same thing to both channels easy)
The source selectors would feed a pair of line drivers via preset level controls (to calibrate the max levels) and variable front panel controls, for easy adjustment. The line drivers would take the signal up to say +20dBv to allow for overload and headroom checking of DUTs. Stepped attenuators on the front panel would take the levels down from +20 to –60dB in 10dB steps for microphone and other low level input preamp testing. There would be a final " minus infinity" position for noise measurements (no voltage but same output impedance). Output RCA connectors on the front panel would feed the DUT.
Then we come to the input areas. There are three.
Firstly we have a pair of BNC inputs feeding high impedance attenuators and buffers designed to work with standard oscilloscope probes. The probes will be used for getting into the middle of the DUT or breadboard. There have been a number of suitable buffer circuits published over the years. The usual thing is to have something like x1, x10 and x100 attenuation, and maybe x10 gain. One interesting option would be to use the Nuova Elettronica approach which connects to the computer via USB rather than soundcard. Because of the high impedance, these signals will be fine to look at, but can be expected to be a bit noisy for in-depth analysis. (Or am I being too fussy?)
Secondly, there would be another pair of RCA connectors for the outputs of the DUT. This direct connection will be used for serious measurements of noise, distortion etc, so needs to not contribute significant degradation. I guess some gain or attenuation may be needed here too to facilitate testing at elevated and lower levels? To save panel space, perhaps this attenuation could be controlled by the same rotary switches as used for the probes?
The microphone input would lead to a low noise balanced mic amp with phantom power, for measurement mics. Not sure if one or two channels, but probably go for two, if only to allow relative phase checks? Question – stepped attenuator on front panel, to help keep calibrations valid? Heh heh, or the probe switches again?
These feed the Analyse selector switches, which would have these options:
· Source (output of the Source switch, i.e. whatever's being sent to the DUT)
· Probes, for probing about inside stuff
· Output from DUT
· Microphone(s)
· Spare(s) for future developments
· Other channel
These Selectors feed outputs to the hardware Noise and Distortion meter.
They also feed a two position Monitor selector switch, along with the analysis output of the N&D set (ie the leftovers after filtering out the fundamental). The Monitor selector feeds the computer sound card input, the hardware Digital Storage Oscilloscope and the monitor speaker amplifier.
There might be an External Trigger selector for the hardware oscilloscope, with choices Source L, Source R, mains. This would keep the DSO triggered even when nulling out signals.
Finally, we have a Stereo/Left/Right/Mono/Off selector, and a volume and balance control. They feed a stereo amp in the unit, say 2 x 40W, which drives the test bench loudspeakers. A headphone socket switches speakers off if headphones are used.
So, in a typical setup, I could choose between the signal sources, and send one or more to the DUT. I can set the max level to anywhere between –60 and +20dBv, or zero, and run calibrated (full level) or variable. And I can measure, see and audition that level on any of my measuring tools.
I can probe around inside the DUT with the probes, or select its output, or a combination. I can compare input to output. I can also input from microphone(s), in which case I'd monitor on headphones. I can measure, see and listen to either channel output, or listen to and view the noise and distortion products, or analyse them on the computer FFT. My hardware oscilloscope can view along with the rest, and be triggered by the source or mains to look for hum. I can readily set up for full analytics using Rightmark and other software.
Is it a plan? Is it unnecessarily complex? Or am I overlooking something important? What would you do?
Terry
You've seen Pete Millett's soundcard interface, I suppose?
Soundcard Interface
I've seen a number a variations on this built around the country, they work well.
Soundcard Interface
I've seen a number a variations on this built around the country, they work well.
No, hadn't seen that, thanks (remember I'm very new 'round here!).
More grist to the mill. Problem is, what am I to do with all this grist?
Terry
More grist to the mill. Problem is, what am I to do with all this grist?
Terry
depends on how you value your time - a new or used Audio Precision unit will likely do more, better than you could put together, program, debug, calibrate in years of full time of labor
they really aren't charging too much for their decades of experience in design and in the debugging, verification, robustness that comes from supporting units in customers hands over that time
they really aren't charging too much for their decades of experience in design and in the debugging, verification, robustness that comes from supporting units in customers hands over that time
Ah yes, but where's the DIY in that? 🙂
You're probably right. Unfortunately though, second-hand Audio Precisions are a bit light on the ground around here. I live in a small coastal village in the former British penal colony known as New South Wales. The nearest one on Ebay at the moment is in Korea!
Terry
Make flour, bake bread!
There are a number of folks on here who have made front ends for soundcard measurements, including Pete. Maybe we can get them to chime in?
Good idea to look around and see what others have done. Here's a few so far:
Millet
Marsden
xoscope
Nuova Elettronica
dpscope
(interesting - seems to be DC coupled and wide bandwidth)
Opencircuits
We can build on that list....
Terry
Millet
Marsden
xoscope
Nuova Elettronica
dpscope
(interesting - seems to be DC coupled and wide bandwidth)
Opencircuits
We can build on that list....
Terry
Last edited:
This is interesting too:
Turn PC Sound Card into a High Speed Sampling Oscilloscope
There are sure enough brains out there wrestling around this topic!
Terry
Turn PC Sound Card into a High Speed Sampling Oscilloscope
There are sure enough brains out there wrestling around this topic!
Terry
Well, I've stopped wondering, and started doing! I've built a test mock-up of the switching aspects of the proposed control unit, in a plastic jiffy box. That will enable me to test the idea before proceeding. Here's a copy of the front panel layout. The words around the edges are labels for connectors on the sides. The blue circles are where the knobs go. Left channel at top, right below them. The left-most knobs determine what source is sent to the Device under Test. The middle knobs determine what is sent to the various items of measurement equipment. The right hand knob determines what you hear on the test-bench monitor speakers. The two switches allow you to monitor the signal from the DUT or the analysis output of the N&D set.
So far, it seems promising. I'll run with it for a while and if it seems to do the job, build the real thing.
Terry
So far, it seems promising. I'll run with it for a while and if it seems to do the job, build the real thing.
Terry
- Status
- Not open for further replies.