A trans-conductance amplifier can be approximated using an audio transformer, such as one of those "line matching" devices using in PA systems.
You use it backward - drive the low Z winding with the amp, connect the high Z winding to the speaker. The trick is to place a resistor between than amp output and the low Z winding, to convert voltage to current. Then the output current is simply Amplifier V / Resistor R, divided by the turns ratio of the transformer.
A pretty "cement half A to half B" way to explore the effect, versus building an entire amp. Wouldnt sell, because of the significant insertion loss - and every kid with a fat wallet wants full power, right?
You use it backward - drive the low Z winding with the amp, connect the high Z winding to the speaker. The trick is to place a resistor between than amp output and the low Z winding, to convert voltage to current. Then the output current is simply Amplifier V / Resistor R, divided by the turns ratio of the transformer.
A pretty "cement half A to half B" way to explore the effect, versus building an entire amp. Wouldnt sell, because of the significant insertion loss - and every kid with a fat wallet wants full power, right?
@arivel , I hope this helps.
A Transconductance or Current Source amplifier behaves in this way: A Voltage input produces a output current proportional to that input. When used with a speaker that has a flat impedance curve, it behaves somewhat like a voltage amplifier, however any change in the speaker's impedance will produce a change in Voltage at the speaker to maintain current proportional to the input of the amplifier.
An example of this would be: some speakers have an impedance dip in the area of a crossover between two drivers. The Transconductance amplifier would show a Voltage dip in this region. Another example would be: most speakers are inductive and show an increase in impedance at high frequencies. The Transconductance amplifier would show an increase in Voltage at these high frequencies.
A Transconductance or Current Source amplifier behaves in this way: A Voltage input produces a output current proportional to that input. When used with a speaker that has a flat impedance curve, it behaves somewhat like a voltage amplifier, however any change in the speaker's impedance will produce a change in Voltage at the speaker to maintain current proportional to the input of the amplifier.
An example of this would be: some speakers have an impedance dip in the area of a crossover between two drivers. The Transconductance amplifier would show a Voltage dip in this region. Another example would be: most speakers are inductive and show an increase in impedance at high frequencies. The Transconductance amplifier would show an increase in Voltage at these high frequencies.
To say it another way, the loudspeakers impedance curve becomes part of the frequency response.
If the impedance rises, so does the output.
So you are either looking for a loudspeaker with flattish impedance over its bandwidth, or one that benefits from a bit of “EQ”. My first experience with a current amp (actually had variable output impedance from voltage amp thru to current amp, and some loudspeakers definitely benefitted from turning up their Rout. In the room, there was a just right setting. This amp pre-dated Joe’s (and in private correpspondence, Joe learned some stuff from Daniel’s amp).
dave
If the impedance rises, so does the output.
So you are either looking for a loudspeaker with flattish impedance over its bandwidth, or one that benefits from a bit of “EQ”. My first experience with a current amp (actually had variable output impedance from voltage amp thru to current amp, and some loudspeakers definitely benefitted from turning up their Rout. In the room, there was a just right setting. This amp pre-dated Joe’s (and in private correpspondence, Joe learned some stuff from Daniel’s amp).
dave
Hi Arivel
There is some truth in saying that a current source can be modeled as a voltage source followed by a value of resistance. Except it is not a resistor, it is an impedance that has an equivalent value in Ohm. This is difficult to digest by some.
The truth is that a current source that has a high output impedance relies on feedback to achieve the effect, often able to get hundreds of Ohm, yet there is no resistor that impedance the amplifier to go full output and clipping like any normal amplifier does. This is how I see it (my real experience) and if anybody has a different view, then they are free to express it.
One trick I have done is to take a Class D amplifier capable of very high output voltage and put a 18-22 Ohm series resistor. The voltage swing is severely limited and I think I only got about 20-25 Watt out of it and a very hot resistor. It gave something close to a current source and you could indeed hear an improvement. Then I met Menno Vanderveen at ETF17 and he had a modified Hypex Class D amplifier with a 3-position switch, that achieved, without any series resistor, a near-zero Ohm, then 3 Ohm and thrid position 18 Ohm. Most preferred when it was set to 18 Ohm, even if the bass was somewhat fuller but surprisingly not out of control, but the music that was chosen probably helped here.
BTW, if you EQ the current of the amplifier flat (adjusted for the speaker load used), then output impedance no longer matters. Since the alignment also does not change with output impedance, this effectively puts an end to the so-called 'damping factor' or the ability of the amplifier to have some kind of stopping power (again the trolls can't see that). No commercial speakers do this kind of EW, only a few DIY designs (ahem, by me) do. Not only locks in the alignment, but also the crossover to not budge, no matter the output impedance of the amplifier.
So does that help?
Hi Dave
Note that I am discussing EQ'ing the current, which is of course not what you say re EQ of the response. EQ the current and the speaker becomes immune to the o/put Z.
There is some truth in saying that a current source can be modeled as a voltage source followed by a value of resistance. Except it is not a resistor, it is an impedance that has an equivalent value in Ohm. This is difficult to digest by some.
The truth is that a current source that has a high output impedance relies on feedback to achieve the effect, often able to get hundreds of Ohm, yet there is no resistor that impedance the amplifier to go full output and clipping like any normal amplifier does. This is how I see it (my real experience) and if anybody has a different view, then they are free to express it.
One trick I have done is to take a Class D amplifier capable of very high output voltage and put a 18-22 Ohm series resistor. The voltage swing is severely limited and I think I only got about 20-25 Watt out of it and a very hot resistor. It gave something close to a current source and you could indeed hear an improvement. Then I met Menno Vanderveen at ETF17 and he had a modified Hypex Class D amplifier with a 3-position switch, that achieved, without any series resistor, a near-zero Ohm, then 3 Ohm and thrid position 18 Ohm. Most preferred when it was set to 18 Ohm, even if the bass was somewhat fuller but surprisingly not out of control, but the music that was chosen probably helped here.
BTW, if you EQ the current of the amplifier flat (adjusted for the speaker load used), then output impedance no longer matters. Since the alignment also does not change with output impedance, this effectively puts an end to the so-called 'damping factor' or the ability of the amplifier to have some kind of stopping power (again the trolls can't see that). No commercial speakers do this kind of EW, only a few DIY designs (ahem, by me) do. Not only locks in the alignment, but also the crossover to not budge, no matter the output impedance of the amplifier.
So does that help?
Hi Dave
Note that I am discussing EQ'ing the current, which is of course not what you say re EQ of the response. EQ the current and the speaker becomes immune to the o/put Z.
A
The image in Joe’s opening post #1, I believe it’s the amp schematic, doesn’t show for some reason.
One time I forgot to pay the cable bill, they cut me off. I paid the bill and like magic I was back on line.
Hello All,
Remember that a transconductance amplifier (AKA a current source amplifier) is a voltage controlled current output amplifier.
A transconductance amplifier has a differential voltage input, a + input and a - input, the difference voltage is amplified. In terms of voltage input the it is a variable gain amplifier. The "normal" voltage input goes to one input and feedback voltage goes to the other input. The difference between the two voltages is amplified.
The feedback voltage is sampled across a sensing resistor in series with the load (driver). As the impedance across the driver varies the current also varies and the voltage across the current sensing resistor is feedback at the second voltage input of the transconductance amplifier.
Do not let anyone tell you that current is in charge. Voltage controls the output current.
There is not a large hot resistor at the output of the amplifier.
Thanks DT
Hello All,
Looking at post number 1 the Lm3875T chip amplifier has gone obsolete while we are waiting for test results. Interesting fact; the Lm3875T chip amplifier is a Voltage input Voltage output amp.
Joe has promised test results for a Scanspeak plain Jane 6 ½ inch mid-woofer. I have a pair on my bench. It seems like a good idea to test the concepts with a driver that is not someone’s darling. We do not want to offend anyone in the process of testing.
I went shopping for a Chip Amplifier Evaluation Module at mouser. You know an amplifier with differential inputs similar to the now obsolete LM3875T. The world has moved away from Class A/B so I purchased a Class D amplifier with differential inputs like the LM3875T. We do not want to be testing someone’s darling amplifier either.
I purchased a TPA3255EVM from Mouser. With the differential inputs it can be setup as either Voltage amplifier or “current amplifier”. “current amplifier” is really a misconception. With current/voltage feedback from the current sense resistor it is really a variable voltage output amplifier that compensates for the variable impedance of the driver load.
Output impedance is not really the concept; it is a combination of output voltage and current. A transconductance amplifier juices up the output voltage to maintain the output current into a variable impedance load .
Thanks DT
Looking at post number 1 the Lm3875T chip amplifier has gone obsolete while we are waiting for test results. Interesting fact; the Lm3875T chip amplifier is a Voltage input Voltage output amp.
Joe has promised test results for a Scanspeak plain Jane 6 ½ inch mid-woofer. I have a pair on my bench. It seems like a good idea to test the concepts with a driver that is not someone’s darling. We do not want to offend anyone in the process of testing.
I went shopping for a Chip Amplifier Evaluation Module at mouser. You know an amplifier with differential inputs similar to the now obsolete LM3875T. The world has moved away from Class A/B so I purchased a Class D amplifier with differential inputs like the LM3875T. We do not want to be testing someone’s darling amplifier either.
I purchased a TPA3255EVM from Mouser. With the differential inputs it can be setup as either Voltage amplifier or “current amplifier”. “current amplifier” is really a misconception. With current/voltage feedback from the current sense resistor it is really a variable voltage output amplifier that compensates for the variable impedance of the driver load.
Output impedance is not really the concept; it is a combination of output voltage and current. A transconductance amplifier juices up the output voltage to maintain the output current into a variable impedance load .
Thanks DT
Last edited:
Current output amplifier is an accurate and relevant description of the action of the amplifier and its feedback loop as far as the load we care about is concerned. The amplifier's inherent output impedance is less relevant, as far as the final outcome, but it's fine to describe that process as well.
As for the TPA3255, how do you plan on setting it up for current drive? The feedback network is built into the IC, and is pre-filter.
And as I post that, there are 4 threads on the main page about TPA3255 with post filter feedback, and one includes a link to a paper from TI on the topic. Nothing on achieving current drive but at least it's something.
Hello All,
Accurate and relevant are not measurable words. On the other hand variable/controllable output voltage and voltage gain is measurable. Variable current does not come into play until there is a connected load. Calling it current output is a confusing distraction from the function of the feedback loop.
By design the LM3875T is a voltage input and voltage output amplifier.
To avoid the confusion of the conversation of feedback in Class-D amplifiers I went out and found some LM3875T chip amplifiers to purchase.
Let us not go down that rabbit hole today. There is nothing about current drive in the LM3875T documentation either.
The TPA3255 is a project for another day. I do have a new in the box AP AUX-0040 filter to test drive.
Thanks DT
In terms of output impedance this may be of interest:
Negative output impedance:
https://circuitcellar.com/resources/quickbits/negative-impedance-converter/
Last edited: