Multiple segments as a resonant mode, yes.. but not multiple harmonically related frequencies..
As a secondary effect, yes (though it's typically controlled). This reminds me how important is it to be specific with this.
It appears to be part of the underlying confusion that breakup itself is a linear phenomenon.
I have built and use four amplifiers in current drive after modification of TDA2005 amplifiers. And one with TDA2003. Its the "Basic current feedback" method as described by Esa Merilainen:
https://www.current-drive.info/9
It is stated that the bass with current drive would become fat and undefined but on classical fullrange drivers without crossover I cannot find this to be true.
Fullrange drivers without any crossover only profit from this type of configuration - everything sounds better with this type of amplifier topology.
Thread on modification of chip amps to current drive:
https://www.diyaudio.com/community/threads/chip-amp-modification-to-current-drive.389985/
https://www.current-drive.info/9
It is stated that the bass with current drive would become fat and undefined but on classical fullrange drivers without crossover I cannot find this to be true.
Fullrange drivers without any crossover only profit from this type of configuration - everything sounds better with this type of amplifier topology.
Thread on modification of chip amps to current drive:
https://www.diyaudio.com/community/threads/chip-amp-modification-to-current-drive.389985/
For a resistor the current and voltage are proportional. Speakers have a dominant resistance which is almost equal to their impedance. In other words, in a speaker driven by a voltage the current tends to be proportional to the voltage. This seems to suggest voltage drive shares a lot with current drive.
@>Baffless< : In an impulse that was not well thought out, I pointed out that condenser microphones always distort with D2.
You have shown a technically very good measurement. For me it's the first time in a forum, done by a forum member, to see a good measurement that was done with an obviously very good current control amplifier. Mills and Hawksford had "trouble" in 1989. In my approach, I measure the distortions in the current to avoid my difficulties with acoustic measurements of the distortion (microphone, measuring room...). Of course I'm interested in a comparison between a current control amplifier and the 20R approach.
Is the relative reduction similar for all harmonics when comparing current control drive with voltage control drive?
Agree that they are not LTI. Non linear, but time invariant i would say.
But current is not linear to actual SPL. So measuring current doesn’t give you much
Of course not all harmonics are better with current drive, but from what i have measured usually current looks cleaner above resonance. Around 2 times above resonance and current looks better. H2 doesn't seem to benefit usually but H3 a lot better. Higher harmonics also lower with current. Around resonance voltage is better. I have some more measurements, but they are not well documented. A little mess
which driver exactly, what SPL levels, what amp and so on. Here is 2khz+5khz IMD both graphs the same measurement only diference which is on top
Nice to see these measurements.
It astonished me a lot while doing my listening tests with voltage and current driven amplifiers that the bass region profits from a more defined presentation, too. I did not expect that from the telling everywhere current drive would make the bass sound blown up.
No, with a current driven amp and fullrange loudspeakers without crossovers it sounds well defined in the bass because the reduced distortion makes you hear better the original impulse. Every impulse loses energy through distortion. Is the distortion reduced you can hear the original low tone better.
Then - most fullrange drivers without crossover network lack bass and high frequencies, so using the reduced damping in the bass just corrects for shy bass energy with voltage amps.
Nelson Pass did some measurements using Merilainens simple resistor method to verify by listening and measureing the linearization of fullrange drivers which are directly connected to the amplifier.
Due to these measurements and Merilainens measurements
https://www.current-drive.info/disto/68
there is not necessity to give a further proof of concept.
It astonished me a lot while doing my listening tests with voltage and current driven amplifiers that the bass region profits from a more defined presentation, too. I did not expect that from the telling everywhere current drive would make the bass sound blown up.
No, with a current driven amp and fullrange loudspeakers without crossovers it sounds well defined in the bass because the reduced distortion makes you hear better the original impulse. Every impulse loses energy through distortion. Is the distortion reduced you can hear the original low tone better.
Then - most fullrange drivers without crossover network lack bass and high frequencies, so using the reduced damping in the bass just corrects for shy bass energy with voltage amps.
Nelson Pass did some measurements using Merilainens simple resistor method to verify by listening and measureing the linearization of fullrange drivers which are directly connected to the amplifier.
Due to these measurements and Merilainens measurements
https://www.current-drive.info/disto/68
there is not necessity to give a further proof of concept.
Yep. Lots of improvements to be had. Pretty much my only reservation would be the possiblity of blowing up speakers, especially tweeters, if the amplifier has "too much" head room and the feedback system isn't thermally matched to the speaker.
Seems unlikely, but probably a deal-breaker for a surprisingly large cohort who like to keep the volume knob right on the edge of power compression.
Seems unlikely, but probably a deal-breaker for a surprisingly large cohort who like to keep the volume knob right on the edge of power compression.
....because the loudspeaker cone and the sound path selected with the microphone position have filter functions?
All harmonics are definitely better with current drive, exactly by the expected amount. My measurement of the Current proves that. You can't refute that with a condenser microphone. It's not your fault, we all have these things, including Esa Meriläinen.
According to my educated guess, this is most likely coming from the microphone, not from your excellent amplifier.
I find my current measurement allows the assumption that current drive also significantly reduces the distortion at resonance? My best guess is, the ripples in the graphs come from my room, which feeds back to the sound transducer via microphonic action of itself?
I have some more measurements, but they are not well documented. A little mess
same here.Good measurement. A deep bow to your flawless presentation.
I hope my translator didn't upset you.
Last edited:
This may be a matter of implementation. Over-powered class-D chips with their very high sweet spot (5-10W for a 100W chip) will likely purr with the resistor hack, as they will run at higher signal voltages with softer PWM switching ratios and lower noise.
On the other hand, a low-NFB class-A/AB would likely have distortion proportional to Vout², and a 3× change in voltage levels could have it running sub-optimally. Even a fuller implementation using a sense resistor would have to contend with the bass EQ being directly proportional to reduced negative feedback.
Hello, you know any tweak for an existing Class D amp? Like the popular TPAs for example?
Would be grateful to hear for any mod which can be done on them as I always search for amplifiers which can be hacked to be transformed to "basic current drive" or the like.
I wanted to report that the "last argument" from "voltage fans" against current drive that the bass will sound bad simply did not happen with my fullrange loudspeakers. Its all the contrary!
Maybe for existing multi-ways everything is different but usually you do not need current drive for them if they have well made classical crossovers which help the speakers already applying the "simple resistor method" for mids and highs.
Here you can see some of my fullrange loudspeakers I do DIY tweaking their paper cones to become true high end contenders.
https://www.diyaudio.com/community/threads/how-to-make-a-loudspeaker-sandwich-cone.402917/
LOTS of speakers are NOT Time Invariant. As I said, with Angelo's method, it is now easy to check on this .. and how much. Making a speaker more TI, is one of the secrets of making a small speaker sound 'big'.
BTW, there is no doubt that current drive reduces harmonic distortion, compression bla bla .. in moving coil speakers. https://www.aes.org/e-lib/browse.cfm?elib=6099
But the 'distortions' it reduces are mostly not audible https://www.aes.org/e-lib/browse.cfm?elib=2476 This paper isn't only about Intermodulation. See the appendix for the many 'distortions' that ARE audible
Last edited:
current drive eliminates the hysteresis distortion from the iron around the voice coil. This has infinite memory and is thus very much not time invariant. I presents it self as harmonic distortion for sine wave signals but more complex signals reveal the non linear memory. I find personally and subjectively this type of distortion very destructive.
These are 2 different things.
With Ohm's law you can work out that 22 ohm added to a nominal 8 ohm speaker will change a 100W class D amplifier into approximately 7.5W.
I'm not sure there's much more that could be done with most pre-packaged chips because they are too tightly integrated. What you would need is a board with open access to the H-bridge switches. Any takers??
In particular, full switching control provides more than just old-fashioned 'up' or 'down' PWM control, but also SLEEP ratios where both halves are switched off, and SHORT-CIRCUIT as a 4th option. Then 99% of the hard work would be providing a solid algorithm where high frequencies have a high sleep ratio, and low frequencies are mostly shorted. That's a bit more than a small hack.
OK, I have no idea about D amps and how they work.
First: there are no commercially available D amps with current drive. If someone would offer a kit or ready made pcb it would be a guaranteed niche for a seller (stand alone criteria) .
Second: a simple tweak like with classical chip amps seems to be unmanageable. OK. Did not know that.
First: there are no commercially available D amps with current drive. If someone would offer a kit or ready made pcb it would be a guaranteed niche for a seller (stand alone criteria) .
Second: a simple tweak like with classical chip amps seems to be unmanageable. OK. Did not know that.
I used until now only fullrange drivers with copper /Faraday rings on the pole piece. But these are rarely produced.
listening test shows that these drivers profit from current drive, too.
It's clearly audible in the bass region which gets more defined and sounds more convincing.
Simple fullrange drivers without copper rings profit much more.
They become high end transducers when current driven after the paper cone got more damped and stability by converting it to a sandwich cone.
It's an incredibly economical and elegant solution to the inferiority in distortion they had in comparison to multi way speakers.