My interest is in building, designing/prototyping interconnect cables.
Of most importance, I need to build a fair number of interconnects and need to develop a test to ensure basic quality of termination, beyond signal continuity as tested with a voltmeter. Things like cold solder joints which cause noise but still allow signal to pass. Other flaws in connectors, wire, etc. Also interested in measuring effects of shielding on EMI and RFI noise attenuation, linearity of cable design.
I read the thread on purchasing used oscilloscopes.
http://www.diyaudio.com/forums/parts/104684-buying-oscilloscope.html
Based on the threads recommendations I have decided on picking up a reconditioned/calibrated/90 day warranty Tektronix 265B from a reputable source that I have been dealing with for the past 5 years.
Question 1:
For basic production problems. As opposed to listening to every pair, is there a good testing technique that can be employed? Possibly using a function generator, or other device.
This same company has a number of options for gently used/calibrated function generators.
At the high end, they highly recommend the HP33120A, ~ $1000:
The Agilent 33120A Function / Arbitrary Waveform Generator uses direct digital-synthesis techniques to create a stable, accurate output signal for clean, low-distortion sine waves. It also gives you fast rise and fall-time square wave, and linear ramp waveforms down to 100µHz
• 15 MHz sine and square wave outputs
• Waveforms, Standard: Sine, triangle, square, triangle, ramp, noise, sin(x)/x, exponential rise and fall, cardiac, dc volts
• Waveforms, Arbitrary: length 8-16000 points, 12 bits (incl. sign) amplitude resolution with 40 MSa/s Sample Rate
They also have many units under $500.
Question 2:
If I am going to do testing as well as cable research/prototyping, will the HP33120A be a good investment? Any good suggestions on function generators to complement the Tektronix 465B? I am looking for something that generates the signal with low distortion.
What about something like a less expensive Wavetek unit, e.g. the Wavetek 134 which can do
"sine, square, triangular, and sawtooth waveforms with selectable and variable outputs over a dynamic frequency range of 0.2 Hz to 2 MHz. Can be operated as a triggered or gated CW generator and as a continuous or triggered sweep generator. Sweep time of 10 µsec to 100 sec, up to 1000:1 with calibrated control."
http://www.diyaudio.com/forums/parts/28662-advice-function-sweep-generator.html
Thanks in advance for any suggestions.
Of most importance, I need to build a fair number of interconnects and need to develop a test to ensure basic quality of termination, beyond signal continuity as tested with a voltmeter. Things like cold solder joints which cause noise but still allow signal to pass. Other flaws in connectors, wire, etc. Also interested in measuring effects of shielding on EMI and RFI noise attenuation, linearity of cable design.
I read the thread on purchasing used oscilloscopes.
http://www.diyaudio.com/forums/parts/104684-buying-oscilloscope.html
Based on the threads recommendations I have decided on picking up a reconditioned/calibrated/90 day warranty Tektronix 265B from a reputable source that I have been dealing with for the past 5 years.
Question 1:
For basic production problems. As opposed to listening to every pair, is there a good testing technique that can be employed? Possibly using a function generator, or other device.
This same company has a number of options for gently used/calibrated function generators.
At the high end, they highly recommend the HP33120A, ~ $1000:
The Agilent 33120A Function / Arbitrary Waveform Generator uses direct digital-synthesis techniques to create a stable, accurate output signal for clean, low-distortion sine waves. It also gives you fast rise and fall-time square wave, and linear ramp waveforms down to 100µHz
• 15 MHz sine and square wave outputs
• Waveforms, Standard: Sine, triangle, square, triangle, ramp, noise, sin(x)/x, exponential rise and fall, cardiac, dc volts
• Waveforms, Arbitrary: length 8-16000 points, 12 bits (incl. sign) amplitude resolution with 40 MSa/s Sample Rate
They also have many units under $500.
Question 2:
If I am going to do testing as well as cable research/prototyping, will the HP33120A be a good investment? Any good suggestions on function generators to complement the Tektronix 465B? I am looking for something that generates the signal with low distortion.
What about something like a less expensive Wavetek unit, e.g. the Wavetek 134 which can do
"sine, square, triangular, and sawtooth waveforms with selectable and variable outputs over a dynamic frequency range of 0.2 Hz to 2 MHz. Can be operated as a triggered or gated CW generator and as a continuous or triggered sweep generator. Sweep time of 10 µsec to 100 sec, up to 1000:1 with calibrated control."
http://www.diyaudio.com/forums/parts/28662-advice-function-sweep-generator.html
Thanks in advance for any suggestions.
In all my years of working in this field I have never had use for a function generator. Sine wave, yes. No other waveforms have been needed. A good quality audio generator is necessary if you are measuring distortion, but of course in that case you need a distortion measuring device. If you want to rely on your eyes on the screen for distortion, almost any generator will do, from the Heath and EICO and Knight on up. Or you can build one easily.
Oscilloscopes need to have whatever bandwidth you intend to use. If you want to track signals in a 10.7 MHz strip, of course you need bandwidth greater than that. If not, then something with response beyond, say, 100 kHz is adequate. If you intend to work with Class D and up amplifiers, you will definitely need more sophisticated equipment.
High sensitivity on a 'scope is handy, say 5 mV/div. You need probes too. No need for delaying sweep, storage, other bells and whistles. (Except as mentioned above for switching type amplifiers and power supplies.)
Measuring power output of amplifiers requires 8 Ohm power resistors. You should have a variable autotransformer to control the power input to the amplifiers, and it's doubly useful for applying low voltages after a repair job. And an audio voltmeter to get more accurate results than the 'scope can provide.
Oscilloscopes need to have whatever bandwidth you intend to use. If you want to track signals in a 10.7 MHz strip, of course you need bandwidth greater than that. If not, then something with response beyond, say, 100 kHz is adequate. If you intend to work with Class D and up amplifiers, you will definitely need more sophisticated equipment.
High sensitivity on a 'scope is handy, say 5 mV/div. You need probes too. No need for delaying sweep, storage, other bells and whistles. (Except as mentioned above for switching type amplifiers and power supplies.)
Measuring power output of amplifiers requires 8 Ohm power resistors. You should have a variable autotransformer to control the power input to the amplifiers, and it's doubly useful for applying low voltages after a repair job. And an audio voltmeter to get more accurate results than the 'scope can provide.
In looking for defects I would be looking at visual differences between a reference signal from a known good cable and that from the defective cable. In case of resolution would 2mV/div be better than 5mV/div? the Tektronix 465B is 5mV/div. The Tektronix 2465B is 2mV/div and appears to be more $.
audio express had an article a while back 3-4 months, concerning conectors. The author suplied a circuit for evaluating the various types of conectors. The author is here somewhere
ah I found it, its post 7644 in thred
I don't believe cables make a difference, any input?
ah I found it, its post 7644 in thred
I don't believe cables make a difference, any input?
With most audio work, cables matter very little. If you have a very long run, are in a corrosive environment, or are working with second rate equipment, well maybe.
The best tests for cables are to measure capacitance, resistance, and insulation integrity. You don't need an oscilloscope for that.
If you have an open cable and want to repair it, measure the capacitance at each end and it will tell you where the break is.
The best tests for cables are to measure capacitance, resistance, and insulation integrity. You don't need an oscilloscope for that.
If you have an open cable and want to repair it, measure the capacitance at each end and it will tell you where the break is.
The HP3312A is the older version of the 33120A. The older function generator is a good piece of gear and can be had for relatively little cash. The 33120A is awesome! Very low THD on the sine wave and fast square wave.
I think for cables, though, the best test method is a TDR (Time Domain Reflectometry) measurement. All that's required is a fast oscilloscope and a step generator with an insanely fast rise time.
~Tom
I think for cables, though, the best test method is a TDR (Time Domain Reflectometry) measurement. All that's required is a fast oscilloscope and a step generator with an insanely fast rise time.
~Tom
Could you provide a link to this article? I looked on the audioXpress web site and wasn't sure where to look. Or is on another site?
Can TDR be used to detect defects in short cable runs (1m) that are not technical short circuits, e.g. cold solder joints, areas of deformed insulation, etc. Wikipedia indicates that TDR is primarily for testing breaks in long cable runs but are also used to detect wire taps.
Time-domain reflectometer - Wikipedia, the free encyclopedia
TDR Tutorial - Intro to Time Domain Reflectometry
TDR Circuit
Also can the HP3312A and HP33120A serve as the step generator, with insanely fast rise time?
I am a diyAudio newbie so please excuse the trivial questions.
It was in the November 2009 Issue, page 22 of AudioXpress magazine.
"Distortion Meter" by Ed Simon (who is a member of our forum).
If I were going to design and test cables, a function generator and scope would be about last on my list of things to buy. I'd want a milliohm meter and something to measure impedance, a vector voltmeter at the low end, or a network analyzer at the high end. Or a decent impedance analyzer. I think TDR can work to a meter or less. Even a decent CGRL (LCR) bridge would be useful.
Conrad
Conrad
speedskater beat me to it. Looks like on the audioXpress web site you will have to buy the magazine. They have some of the past articles available to look at online but not that one. I don't know who has the copyright, Simon or AudioeXpess. In the post I mentioned above Simon had some comments a few posts later maybe you can PM him.
The 33120A is quite good, especially if you need/want the arbitrary waveform capability. I have loaded pink and brown noise waveforms, as well as an emulation of an S/PDIF signal carrying a 1 kHz sine wave onto mine. The RS-232 port is valuable here if you don't have GPIB/IEEE488.
The 33120A sine wave is specified to have 0.04% THD, and that is also what I measure it to be - not bad but not great. Rise and fall time on square waves and other steps is not good, making high frequency square waves look more like a flattened sin wave. (It looks just like a Besel filter step response, because it is). So to answer your question directly, no it definitely can't do a step with insanely fast rise time.
It uses direct digital synthesis at a fixed frequency: 40 MHz. So it basically has a waveform (sin, triangle, arbitrary) of up to 16000 points stored in RAM, and it outputs samples from there at 40 MS/s. The output frequency determines the rate at which the instrument increments samples. At low output frequencies, samples are repeated, sometimes many times. At high output frequencies, samples are skipped, sometimes most samples of the waveform are skipped. There is no interpolation or proper resampling, so this can result in some additional waveform distortion creeping in, even after the reconstruction filter.
Other features are nice to have: logarithmic frequency sweep, and even modulation. The burst and AM modulation enables you to create tone-burst signals, useful for example in testing transient responses of drivers or rooms, or testing the dynamic headroom of an amplifier. You can also create signals to test for intermodulation distortion.
Download and read the manual; it is a wealth of information on the abilities and limitations of this instrument.
The 33120A sine wave is specified to have 0.04% THD, and that is also what I measure it to be - not bad but not great. Rise and fall time on square waves and other steps is not good, making high frequency square waves look more like a flattened sin wave. (It looks just like a Besel filter step response, because it is). So to answer your question directly, no it definitely can't do a step with insanely fast rise time.
It uses direct digital synthesis at a fixed frequency: 40 MHz. So it basically has a waveform (sin, triangle, arbitrary) of up to 16000 points stored in RAM, and it outputs samples from there at 40 MS/s. The output frequency determines the rate at which the instrument increments samples. At low output frequencies, samples are repeated, sometimes many times. At high output frequencies, samples are skipped, sometimes most samples of the waveform are skipped. There is no interpolation or proper resampling, so this can result in some additional waveform distortion creeping in, even after the reconstruction filter.
Other features are nice to have: logarithmic frequency sweep, and even modulation. The burst and AM modulation enables you to create tone-burst signals, useful for example in testing transient responses of drivers or rooms, or testing the dynamic headroom of an amplifier. You can also create signals to test for intermodulation distortion.
Download and read the manual; it is a wealth of information on the abilities and limitations of this instrument.
- Status
- Not open for further replies.