DF or not to DF
RT,
The accepted way is to first lower all associated impedances like emitter resistors, Dc resistance of the output inductor and equivalent transistor impedances. This can be done with local feedback like in a collector output circuit or going to multiple devices, then add overall negative feedback. With enough NF you can get any number you want at the cost of little things like stability, complexity and sound quality. In solid state any number much over 400 has to be suspect as the compromises become too great and sound quality suffers. Even much over 200 starts on the road to compromised sound. Been there, done that!
Roger
RT,
The accepted way is to first lower all associated impedances like emitter resistors, Dc resistance of the output inductor and equivalent transistor impedances. This can be done with local feedback like in a collector output circuit or going to multiple devices, then add overall negative feedback. With enough NF you can get any number you want at the cost of little things like stability, complexity and sound quality. In solid state any number much over 400 has to be suspect as the compromises become too great and sound quality suffers. Even much over 200 starts on the road to compromised sound. Been there, done that!
Roger
Typo on my part! See question below with change made from "lowering DF" to "raising DF" (I brain farted; I wanted to ask about RAISING the DF by LOWERING the amp's output resistance):
I know a lot has been said about DF and the law of diminishing return (that is, extremely high DFs won't really get you any more bang for your buck).
==> Setting all things aside, what are some techniques for increasing DF in a solid state amp? More output transistor pairs? A higher bias current in the output pairs? Higher beta transistors?
-RT
I know a lot has been said about DF and the law of diminishing return (that is, extremely high DFs won't really get you any more bang for your buck).
==> Setting all things aside, what are some techniques for increasing DF in a solid state amp? More output transistor pairs? A higher bias current in the output pairs? Higher beta transistors?
-RT
OK RT, hope I can try to give some perspective. I am a little concerned that considerations to achieve a high DF often refer more to NFB and not enough to the output stage internal impedance (but I suspect you might know this!)
All this was admirably explained long ago in a 2-part article by Prof. Matti Otala and colleague Lammasniemi, called "Intermodulation at the amplifier-loudspeaker interface". (Wireless World, November and December 1980). I will give a brief and very incomplete overview here for the benefit of interested folks.
NFB is of course a great tool here, but with limitations. If signals come along outside the frequency band where compensating networks for stability start reacting, there will be an increasing phase angle - the feedback will arrive "late" in a manner of speaking. It is here where overload can easily occur with a high feedback factor, causing havoc. Also, a loudspeaker is reactive, especially at low frequency, and stores energy which it can pump through the feedback circuit back to the input. Otala stresses (with mathematics) the need for low impedance in the output stage itself to "short" the loudspeaker locally.
He postulated: "Interface intermodulation is a form of distortion in a feedback two-port network, caused by non-linear interaction between the input signal of the two-port and a signal externally injected to the output port propagating into the input via the feedback network." (I will come to DF.) In persuit of this he compared typical intermodulation figures for 4 topologies of a 4 transistor output stage: (a) full-complimentary, (b) double emitter-follower, (c) quasi-complimentary and (d) b-grounded emitter (i.e. without any local feedback). The load was a loudspeaker equivalent circuit.
He found relative figures under the same conditions, of for topology (a) = 0,005%, for (b) = 0,01%, for (c) = 0,1% and for (d) = 0,2%. It is thus clear that there is great advantage to first start with a low-impedance output stage before contemplating the effect of global feedback per se. (The article shows much more than this simplified summary.) It is clear from the article that the need for NFB (only) generated damping should be minimized. In other words, the output should first provide as low an open-loop output impedance as possible - then one can go from there.
It may now be said that e.g. the full-complimentary topology has in itself just a feedback-generated low impedance. That is true, but the point is that that is in order as long as the stage's internal feedback is significantly wider in frequency than the rest of the amplifier, which it usually is. To put it differently, local output stage internal phase shift should be small compared to that of the rest of the amplifier at the band extremities.
I consider this an important consideration. To this one can then add any beneficial effect of high current, higher beta transistors, etc.
This description is both long and incomplete, but will hopefully indicate the right way to approach the goal of increased DFs. I would dare to state that non-observance of the above principles is a major reason for the sometimes unacceptable sound from even high-priced amplifiers.
Regards.
All this was admirably explained long ago in a 2-part article by Prof. Matti Otala and colleague Lammasniemi, called "Intermodulation at the amplifier-loudspeaker interface". (Wireless World, November and December 1980). I will give a brief and very incomplete overview here for the benefit of interested folks.
NFB is of course a great tool here, but with limitations. If signals come along outside the frequency band where compensating networks for stability start reacting, there will be an increasing phase angle - the feedback will arrive "late" in a manner of speaking. It is here where overload can easily occur with a high feedback factor, causing havoc. Also, a loudspeaker is reactive, especially at low frequency, and stores energy which it can pump through the feedback circuit back to the input. Otala stresses (with mathematics) the need for low impedance in the output stage itself to "short" the loudspeaker locally.
He postulated: "Interface intermodulation is a form of distortion in a feedback two-port network, caused by non-linear interaction between the input signal of the two-port and a signal externally injected to the output port propagating into the input via the feedback network." (I will come to DF.) In persuit of this he compared typical intermodulation figures for 4 topologies of a 4 transistor output stage: (a) full-complimentary, (b) double emitter-follower, (c) quasi-complimentary and (d) b-grounded emitter (i.e. without any local feedback). The load was a loudspeaker equivalent circuit.
He found relative figures under the same conditions, of for topology (a) = 0,005%, for (b) = 0,01%, for (c) = 0,1% and for (d) = 0,2%. It is thus clear that there is great advantage to first start with a low-impedance output stage before contemplating the effect of global feedback per se. (The article shows much more than this simplified summary.) It is clear from the article that the need for NFB (only) generated damping should be minimized. In other words, the output should first provide as low an open-loop output impedance as possible - then one can go from there.
It may now be said that e.g. the full-complimentary topology has in itself just a feedback-generated low impedance. That is true, but the point is that that is in order as long as the stage's internal feedback is significantly wider in frequency than the rest of the amplifier, which it usually is. To put it differently, local output stage internal phase shift should be small compared to that of the rest of the amplifier at the band extremities.
I consider this an important consideration. To this one can then add any beneficial effect of high current, higher beta transistors, etc.
This description is both long and incomplete, but will hopefully indicate the right way to approach the goal of increased DFs. I would dare to state that non-observance of the above principles is a major reason for the sometimes unacceptable sound from even high-priced amplifiers.
Regards.
first excuse me for my bad english writing
and
thank'S for your good post
i have one question??
in most of datasheet of amplifiers write damping factor but in jbl's amplifier data sheet write effective DF(EDF)
for example:
i have one jbl 2200.1 and in manual of this write :
EDF=6.28 at 4OHM
My subwoofer is KICKER CVX15.it has dvc 2 ohm
i load this at 4ohm
how can i convert this EDF to DF ?
What is my amplifier output impedance??
what is my subwoofer DCR??
and
thank'S for your good post
i have one question??
in most of datasheet of amplifiers write damping factor but in jbl's amplifier data sheet write effective DF(EDF)
for example:
i have one jbl 2200.1 and in manual of this write :
EDF=6.28 at 4OHM
My subwoofer is KICKER CVX15.it has dvc 2 ohm
i load this at 4ohm
how can i convert this EDF to DF ?
What is my amplifier output impedance??
what is my subwoofer DCR??
Hi
Where did you find that data? I don't see it anywhere.
EDF could mean Effective DF, but that would be just selling thing of DF. But since you say it is 6.28 and that when you use 4ohme load, I can't say what it is, it sure is not output impedance.
DCR is DC resistance? Is it important? use DVM if you have one.
Where did you find that data? I don't see it anywhere.
EDF could mean Effective DF, but that would be just selling thing of DF. But since you say it is 6.28 and that when you use 4ohme load, I can't say what it is, it sure is not output impedance.
DCR is DC resistance? Is it important? use DVM if you have one.
in this article:luka said:
http://www.monstercable.com/mpc/stable/tech/A2412_Damping_Factor_Article.pdf
my orginal question:luka said:
do you know damping factor of jbl 2200.1??
see this data sheet:
it says that damping factor=6.28 at 4 ohm
An externally hosted image should be here but it was not working when we last tested it.
That article indicates quite clearly a matter that is mostly disregarded, leading to ridiculous claims and arguments regarding the value of high amplifier damping factor.
As said there, the damping of a loudspeaker is governed by the "braking" current that can flow after cessation of signal. That is limited by the total impedance in the circuit. This includes amplifier output impedance, cable resistance and driver voice coil resistance. Since for a 4 ohm driver the practical voice coil resistance is of the order of 2,7 ohm (varying slightly with make), it will be clear that any amplifier output impedance lower than say 0,2 ohm is of no relevance. The amplifier manufacturer will quote a damping factor of 4/0,2 = 20, but the actual electrical damping factor will be 4/(2,7 + 0,2) = 1,38. (I exclude cable resistance which should also be negligible.)
I have excluded impedance as mentioned in the given article. The article is correct in saying that the highest effective damping factor should be calculated by dividing the driver impedance at resonance by the total resistance, but the former is specific for any loudspeaker design, since in most cases the enclosure has a significant effect on this as a result of its damping. In that sense the EFD mentioned in the Jbl 2200 specs of 6,28 will only be true for a driver (in the cabinet) resonant impedance of 25 ohms. As said, we do not know that - it depends on the (subwoofer) design. At least the idea is correct.
The bottom line is that, unless the Jbl "damping factor" is lower than about 7 (and I doubt that it will be), it can be disregarded as an influence in loudspeaker damping.
As said there, the damping of a loudspeaker is governed by the "braking" current that can flow after cessation of signal. That is limited by the total impedance in the circuit. This includes amplifier output impedance, cable resistance and driver voice coil resistance. Since for a 4 ohm driver the practical voice coil resistance is of the order of 2,7 ohm (varying slightly with make), it will be clear that any amplifier output impedance lower than say 0,2 ohm is of no relevance. The amplifier manufacturer will quote a damping factor of 4/0,2 = 20, but the actual electrical damping factor will be 4/(2,7 + 0,2) = 1,38. (I exclude cable resistance which should also be negligible.)
I have excluded impedance as mentioned in the given article. The article is correct in saying that the highest effective damping factor should be calculated by dividing the driver impedance at resonance by the total resistance, but the former is specific for any loudspeaker design, since in most cases the enclosure has a significant effect on this as a result of its damping. In that sense the EFD mentioned in the Jbl 2200 specs of 6,28 will only be true for a driver (in the cabinet) resonant impedance of 25 ohms. As said, we do not know that - it depends on the (subwoofer) design. At least the idea is correct.
The bottom line is that, unless the Jbl "damping factor" is lower than about 7 (and I doubt that it will be), it can be disregarded as an influence in loudspeaker damping.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.