Fortunately , the global sound is good enough, as distorsion is way lower
at lower levels..
although i m not a big fan of LR output circuits, i implement one , even
an amp doesn t need one at all...
i stay at 1R // 1uH
regards,
wahab
Bob, as I recall from your EC amp schematic you didn´t use either a Zobel filter or output inductance. Did that amp offer sufficient stability into capacitive loads? It´s a very fast design, so I assume it won´t take much to bring it into oscillation? An output coil certainly ruins the slewrate when the amp itself is capable of 300 volts/us.
I don't think high slew rate for it's own sake is important. However an output inductor will have a definite load-dependent effect on the frequency response.
Attached is the result of a sim that attempts to show the effect of a 5uH output inductor on response (measured at the loudspeaker terminals) for:
a) An ESL
b) An MC tweeter
c) An 8 ohm resistive load
While 5uH might seem high by today's standards, it wasn't always so. For example I'm pretty sure the Quad 303 had at least that much, which could help explain the "magical synergy" between Quad components that magazine reviewers and such-like folk used to harp on about back in the day.
Given that the old Quad ESL was somewhat lacking in output in the upper treble, a decibel or so of boost would certainly have been welcome.
Hmm... an amplifier that measures ruler-flat into 8 ohms but sounds a little dull on normal speakers, while managing to add sparkle to an otherwise slightly dull-sounding ESL. No magic there - It's all in the inductor.
Cheers - Godfrey
p.s. Anyone else want to get back on topic?
Attachments
Last edited:
Hi John,
As I said, some experienced designers can get away without a coil. You are definitely an experienced designer. It is likely that your amplifeirs are able to remain stable under any combination of speaker load and cables that one would ever encounter, and that is whatt matters. Your amp might even be stable if a low-ESR ceramic capacitor of any value was connected directly across the output terminals of your amplifier. Of course, those large 10-ohm base stopper resistors you use contribute to your ability to remain stable, but introduce other compromises, like reduced bias stability and a slower output stage.
The use of a coil is a judgment call, but given that even you have said that 1 uH is OK, there is no need to compromise other aspects of the amplifier, or stability, by not using one at all.
Cheers,
Bob
Hi Arthur,
Near the speaker output terminals and no closer than an inch to the chassis.
Cheers,
Bob
Bob, as I recall from your EC amp schematic you didn´t use either a Zobel filter or output inductance. Did that amp offer sufficient stability into capacitive loads? It´s a very fast design, so I assume it won´t take much to bring it into oscillation? An output coil certainly ruins the slewrate when the amp itself is capable of 300 volts/us.
Hi Rikard,
Good question. The EC amp design published in the JAES was not equipped with an output coil in the schematic because it was not intened as a construction project. If I recall correctly, I used something like 0.5 uH and one ohm, and may have mentioned it in the text. There was also a Zobel in the actual circuit.
That amplifier definitely would not be tolerant of a pure capacitive load were it not for a coil.
Cheers,
Bob
I don't think high slew rate for it's own sake is important. However an output inductor will have a definite load-dependent effect on the frequency response.
Attached is the result of a sim that attempts to show the effect of a 5uH output inductor on response (measured at the loudspeaker terminals) for:
a) An ESL
b) An MC tweeter
c) An 8 ohm resistive load
While 5uH might seem high by today's standards, it wasn't always so. For example I'm pretty sure the Quad 303 had at least that much, which could help explain the "magical synergy" between Quad components that magazine reviewers and such-like folk used to harp on about back in the day.
Given that the old Quad ESL was somewhat lacking in output in the upper treble, a decibel or so of boost would certainly have been welcome.
Hmm... an amplifier that measures ruler-flat into 8 ohms but sounds a little dull on normal speakers, while managing to add sparkle to an otherwise slightly dull-sounding ESL. No magic there - It's all in the inductor.
Cheers - Godfrey
p.s. Anyone else want to get back on topic?
Hi Godfrey,
These are really good points. In many amps it is also not just the inductor that causes load-dependent frequency response anomolies. I have always wished that reviewers, when reviewing an amp, would make a frequency response measurement of the amplifier AS SEEN AT THE SPEAKER TERMINALS of the system being reviewed. It might reveal a lot.
Too many people assert that there is no correlation between measurement and sound, when those very same people have done little more than measure the amplifier into an 8-ohm load (sometimes only at 1 kHz!).
Cheers,
Bob
Hi Bob,
As for using the speaker terminals as the measurement point, that really is the only reasonable way to do this. I was taught to do it this way. What appears at the speaker terminals is the only thing a designer can control, the rest is a big question mark. Some designs go as far as to take negative feedback off right at the terminal.
Most reviewers are not technical. They certainly do not seem to approach amplifier measurements and listening tests in a manner that proves they understand what is going on. But then, if they were never trained as a technician, why would anything else be expected? There were a couple that actually did do responsible work, but I stopped reading these reports many, many years ago.
-Chris
Yes! I completely agree with you on that. I think most good bench technicians would also agree. Thank you for making that point with the wording you used.
As for using the speaker terminals as the measurement point, that really is the only reasonable way to do this. I was taught to do it this way. What appears at the speaker terminals is the only thing a designer can control, the rest is a big question mark. Some designs go as far as to take negative feedback off right at the terminal.
Most reviewers are not technical. They certainly do not seem to approach amplifier measurements and listening tests in a manner that proves they understand what is going on. But then, if they were never trained as a technician, why would anything else be expected? There were a couple that actually did do responsible work, but I stopped reading these reports many, many years ago.
-Chris
Thanks, Chris.
Many reviewers do hear differences that make a product sound better. Sometimes a soft high end will tame an overy-bright recording, for example. Another interesting thing is that a pair of loudspeakers often do not have identical impedance curves or crossover points. When driven by an amplifier with less DF, the differential frequency response may lead to imaging anomolies, some of which may be appealing to the listener.
We need to bear in mind that some louspeakers go through pretty wild impedance variations in the vicinity of their crossover frequencies.
Cheers,
Bob
Cheers,
Bob
Hi Bob,
Yep. There are speakers out there that present one heck of a load to an amplifier. I always scratched my head trying to figure out if the designer of the speakers in question were striving for the "best sound", why would they make them difficult to drive? An amplifier in distress never sounds as good as it does normally, so it would seem they are shooting themselves (and the poor customers) in the foot. Getting more to your point concerning wild impedance variations, even the wire may become important as an added component in a circuit with high current flows (above what is normal). I wonder if a zobel network and output coil may react in ways that were not considered.
I don't like complications, so I simply stay away from speakers like this. I have seen salesmen using the fact that a speaker is difficult to drive as a selling feature!
Pairs of loudspeakers may easily have drivers with mismatched driver efficiencies - as much as 2 dB. The low end speakers are worse still! Then, add to that the crossover variations and the chances of proper imaging are diminishing fast. Should we also consider room acoustics? That is probably one reason more people don't notice this. The other big one would be volume control mis-tracking.
There are more uncertainties at work here than you can shake a stick at. What's a reviewer to do?
-Chris
Yep. There are speakers out there that present one heck of a load to an amplifier. I always scratched my head trying to figure out if the designer of the speakers in question were striving for the "best sound", why would they make them difficult to drive? An amplifier in distress never sounds as good as it does normally, so it would seem they are shooting themselves (and the poor customers) in the foot. Getting more to your point concerning wild impedance variations, even the wire may become important as an added component in a circuit with high current flows (above what is normal). I wonder if a zobel network and output coil may react in ways that were not considered.
I don't like complications, so I simply stay away from speakers like this. I have seen salesmen using the fact that a speaker is difficult to drive as a selling feature!
Pairs of loudspeakers may easily have drivers with mismatched driver efficiencies - as much as 2 dB. The low end speakers are worse still! Then, add to that the crossover variations and the chances of proper imaging are diminishing fast. Should we also consider room acoustics? That is probably one reason more people don't notice this. The other big one would be volume control mis-tracking.
There are more uncertainties at work here than you can shake a stick at. What's a reviewer to do?
-Chris
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.