The title for this blog started back in the early 70's with an article in Audio magazine about an amplifier made by the Lirpa company. The frequency response of the amplifier was listed as "DC to Daylight". I am, by nature an analog geek, and i seem to collect lots of random bits of information on anything analog, quite literally from DC to Daylight. If i don't have it, i have probably seen it somewhere, and give me enough time and i can usually retrieve it. I have a few focused interests, such as audio amplifier design and RF design. So this blog will sometimes contain something random, but hopefully useful.
wouldn't an AC ohmmeter be handy?
Posted 24th July 2009 at 04:49 AM by unclejed613
it would be nice to be able to measure impedance of an AC circuit as easily as measuring DC resistance, wouldn't it? there are many times in the audio business, when i would like to find out what the impedance of something is at a particular frequency, and i usually have to resort to some indirect method. an impedance bridge is a piece of test equipment that measures impedance, but such a beast is expensive, and is an indirect method, and a bit more complicated than putting a pair of test leads across a device and reading a number. i would like to be able to put a pair of test leads across a device and know it's impedance as a direct measurement. of course there would be one added step, dialing in the frequency i want to measure at, but i can live with that one extra step. so how to go about it?
a DC DMM ohmmeter circuit is a simple device. you have a digital voltmeter across a pair of probes, and a DC current source set to a known current. if the current were for instance 1mA, then the voltmeter would directly read 1V/kΩ, so a reading of 0.6V would be a measurement of 600Ω. simple.... i like simple, so how to do it for AC? could it be that simple? i think so. i have a function generator with a sine wave output, and a true RMS voltmeter, so how do i get a known AC current, when my FG has a voltage output? a series resistor introduces error, especially once the impedance being measured is more than a few percent of the resistor's value, so a series resistor is not the answer here.
it pays to browse the Application Notes of IC manufacturers like Analog Devices, TI, and National Semiconductor. i've learned a huge amount over the years from app notes, more in fact than one could learn in school. if you have a creative mind, you can see something almost insignificant in an app note and solve a huge problem with it (like using LEDs as opto sensors. somebody saw a little blurb about it in one of TI's app notes back in the 70's, did some experiments with it, and now it's common practice). so, while thinking about how to make an AC current source, i remembered seeing a "bipolar" current source. of course, this current source was shown with a DC input, because nobody in their right mind would ever need an AC "constant current" source (AC constant current source IS a bit of an oxymoron, isn't it?). so i did a search and found a whole app note by NS, about the Howland current source. it turned out to be EXACTLY what i was looking for, but from the text, not hardly a mention of using it for AC. so i loaded up LTSpice, drew the schematic and ran the sim using an AC source instead of DC. perfect, with 1V AC going in and a 1kΩ scaling resistor i got 1mA out, no matter what the load resistance was. stage 2 was actually building it expecting some unexpected error that would disqualify the circuit as unusable. no such error ocurred, it worked exactly like the sim (which if you have any experience building what you have run on a sim, is the exception rather than the rule). so now i have a working current source, and the scale factor is easy to set. with 1Vrms in, and a 1k scaling resistor i get 1mA out. with a 100Ω resistor i get 10mA out. i can literally take a DVM and my device, dial in a frequency, and do a direct measurement of impedance. so, shown below is an excerpt from NatSemi's AN-1515, and the circuit i used for my current source. i used a TI TL2072 op amp instead, and replaced the input voltage source with my function generator (the -2072 is a vastly improved -072).
a DC DMM ohmmeter circuit is a simple device. you have a digital voltmeter across a pair of probes, and a DC current source set to a known current. if the current were for instance 1mA, then the voltmeter would directly read 1V/kΩ, so a reading of 0.6V would be a measurement of 600Ω. simple.... i like simple, so how to do it for AC? could it be that simple? i think so. i have a function generator with a sine wave output, and a true RMS voltmeter, so how do i get a known AC current, when my FG has a voltage output? a series resistor introduces error, especially once the impedance being measured is more than a few percent of the resistor's value, so a series resistor is not the answer here.
it pays to browse the Application Notes of IC manufacturers like Analog Devices, TI, and National Semiconductor. i've learned a huge amount over the years from app notes, more in fact than one could learn in school. if you have a creative mind, you can see something almost insignificant in an app note and solve a huge problem with it (like using LEDs as opto sensors. somebody saw a little blurb about it in one of TI's app notes back in the 70's, did some experiments with it, and now it's common practice). so, while thinking about how to make an AC current source, i remembered seeing a "bipolar" current source. of course, this current source was shown with a DC input, because nobody in their right mind would ever need an AC "constant current" source (AC constant current source IS a bit of an oxymoron, isn't it?). so i did a search and found a whole app note by NS, about the Howland current source. it turned out to be EXACTLY what i was looking for, but from the text, not hardly a mention of using it for AC. so i loaded up LTSpice, drew the schematic and ran the sim using an AC source instead of DC. perfect, with 1V AC going in and a 1kΩ scaling resistor i got 1mA out, no matter what the load resistance was. stage 2 was actually building it expecting some unexpected error that would disqualify the circuit as unusable. no such error ocurred, it worked exactly like the sim (which if you have any experience building what you have run on a sim, is the exception rather than the rule). so now i have a working current source, and the scale factor is easy to set. with 1Vrms in, and a 1k scaling resistor i get 1mA out. with a 100Ω resistor i get 10mA out. i can literally take a DVM and my device, dial in a frequency, and do a direct measurement of impedance. so, shown below is an excerpt from NatSemi's AN-1515, and the circuit i used for my current source. i used a TI TL2072 op amp instead, and replaced the input voltage source with my function generator (the -2072 is a vastly improved -072).
Total Comments 0
