Hi,
I would like to know why there is a resistor at the output of inverted gainclone at http://www.decdun.fsnet.co.uk. Is it not against the belief that in orger to keep the driver in tight control, an amp should have a very low impedance presented to the driver?
Thanks in advance,
Goldy
I would like to know why there is a resistor at the output of inverted gainclone at http://www.decdun.fsnet.co.uk. Is it not against the belief that in orger to keep the driver in tight control, an amp should have a very low impedance presented to the driver?
Thanks in advance,
Goldy
As Nick explained to me, it's purpose is to increase stability of the amp. I have to admit that this is a common practice in almost every topology, sometimes combined with paralell inductor. Few amps I have built had the same circuit on exit, but I also don't quite understand purpose of it.
Can it be that amp sees more impedance flatt load with this resistor? I am sure somebody with greater knowledge would explain this.
Can it be that amp sees more impedance flatt load with this resistor? I am sure somebody with greater knowledge would explain this.
Konnichiwa,
Because I put there.
It is one way to implrove the stability margin of the Amplifier. Often a large value resistor with a parallel inductor is used, but I felt this arrangement, borrowed from Naim, is preferrable.
Yes, absolutely, it is against this belief. But luckily this belief is in the same category as Santa Claus or the Easter Bunny, so we can safely ignore it.
Sayonara
goldyrathore said:
Because I put there.
It is one way to implrove the stability margin of the Amplifier. Often a large value resistor with a parallel inductor is used, but I felt this arrangement, borrowed from Naim, is preferrable.
goldyrathore said:
Yes, absolutely, it is against this belief. But luckily this belief is in the same category as Santa Claus or the Easter Bunny, so we can safely ignore it.
Sayonara
Re: Re: Why does the inverted gainclone have a resistor at the output
Actually, OFTEN an inductor is used, SOMETIMES with a resistor across it. Without a boucherot cell on the output, the inductor and resistor will not do much good (and may do bad).
The purpose is to define a load for HF by uncoupling the actual load (hence inductor) and coupling a resistive load (hence resistor in the R-C boucherot cell).
This makes it easyer for the amp to work on reactive loads at HF because they are actually uncoupled from the output of the amp.
No we can't safely ignore it because it's not a belief but a fact. If we could just ignore it, we could just as well put 100 ohms in series and see how well our loudspeaker is controlled.
That being said, wether that fact means anything in a given installation is very debatable. Let me give 3 examples:
1) amplifier with small R in series with output driving a regular speaker system - in general, the crossover series components will have much higher resistance than the small R. So, the small R effect will be completely negligible.
2) amplifier is part of an active speaker driving a woofer - in general, you want as little influence of amplifier's small R because it alters the response of the speaker (you get a peak at the resonant frequency).
3) amplifier driving an overdamped speaker enclosure (Q factor too low). A series resistance, sometimes 10% of the nominal Z, is commonly used to increase tge Q factor and change the speaker characteristic so you get more usable bass. If the resistance is too high, you end up with case 2, above.
In other words, you need to be aware of the effect of a series resistor at the output, and evaluate it's virtues or shortcomings depending on your application. Still, if the resistor has very low resistance, it will be a negligible influence, but also it will not do much for stability, if anything. In fact, using a low R inductive resistor may well prove to be a good idea because you get the stability because of resistor inductance uncoupling the load.
Kuei Yang Wang said:
Actually, OFTEN an inductor is used, SOMETIMES with a resistor across it. Without a boucherot cell on the output, the inductor and resistor will not do much good (and may do bad).
The purpose is to define a load for HF by uncoupling the actual load (hence inductor) and coupling a resistive load (hence resistor in the R-C boucherot cell).
This makes it easyer for the amp to work on reactive loads at HF because they are actually uncoupled from the output of the amp.
Kuei Yang Wang said:
No we can't safely ignore it because it's not a belief but a fact. If we could just ignore it, we could just as well put 100 ohms in series and see how well our loudspeaker is controlled.
That being said, wether that fact means anything in a given installation is very debatable. Let me give 3 examples:
1) amplifier with small R in series with output driving a regular speaker system - in general, the crossover series components will have much higher resistance than the small R. So, the small R effect will be completely negligible.
2) amplifier is part of an active speaker driving a woofer - in general, you want as little influence of amplifier's small R because it alters the response of the speaker (you get a peak at the resonant frequency).
3) amplifier driving an overdamped speaker enclosure (Q factor too low). A series resistance, sometimes 10% of the nominal Z, is commonly used to increase tge Q factor and change the speaker characteristic so you get more usable bass. If the resistance is too high, you end up with case 2, above.
In other words, you need to be aware of the effect of a series resistor at the output, and evaluate it's virtues or shortcomings depending on your application. Still, if the resistor has very low resistance, it will be a negligible influence, but also it will not do much for stability, if anything. In fact, using a low R inductive resistor may well prove to be a good idea because you get the stability because of resistor inductance uncoupling the load.
Re: Re: Re: Why does the inverted gainclone have a resistor at the output
Konnichiwa,
Nope, i have yet to see an amplifier with a choke only output, the reson is simple, the resistor damps the HF resonance circuit otherwise formed by the choke and speaker cable capactiance.
Yup Exactly. We will find that the near complete removal of electrical damping makes the speakers LF performance somewhat unpredictable (largely dependent on the Drivers Qm), but above resonance we will find that speaker is considerably MORE controlled, as the cone movement is proportional to the force excerted on it which in turn is STRICTLY AND ONLY proportional to the current flowing in the voicecoil. Using a low impedance source establishes a condition leading to maximum current distortion do to using a coil effectively filled and surrounded with solid iron.
Further, the electrical damping of ANY driver is limited by the voice coils DCR. As long as any series resistance remains below around 1/4er of the drivers voicecoil DCR the change in electrical damping is at best notional and no worse than the production tolerance for many drivers.
Yup. And for the specific application the value is low enough to have no relevant influence. Further, the concept of driving a transducer in a manner and fashion that maximises distortion merely to be allowed to avoid having to deal with mechanical problems in a mechanical fashion and save a buck or two strikes as severely counterintuitive and monumentally stupid.
Sayonara
Konnichiwa,
ilimzn said:
Nope, i have yet to see an amplifier with a choke only output, the reson is simple, the resistor damps the HF resonance circuit otherwise formed by the choke and speaker cable capactiance.
ilimzn said:
Yup Exactly. We will find that the near complete removal of electrical damping makes the speakers LF performance somewhat unpredictable (largely dependent on the Drivers Qm), but above resonance we will find that speaker is considerably MORE controlled, as the cone movement is proportional to the force excerted on it which in turn is STRICTLY AND ONLY proportional to the current flowing in the voicecoil. Using a low impedance source establishes a condition leading to maximum current distortion do to using a coil effectively filled and surrounded with solid iron.
Further, the electrical damping of ANY driver is limited by the voice coils DCR. As long as any series resistance remains below around 1/4er of the drivers voicecoil DCR the change in electrical damping is at best notional and no worse than the production tolerance for many drivers.
ilimzn said:
Yup. And for the specific application the value is low enough to have no relevant influence. Further, the concept of driving a transducer in a manner and fashion that maximises distortion merely to be allowed to avoid having to deal with mechanical problems in a mechanical fashion and save a buck or two strikes as severely counterintuitive and monumentally stupid.
Sayonara
Re: Re: Re: Re: Why does the inverted gainclone have a resistor at the output
Yes - to be honest, I would also say that using just an inductor is bad design, but I have seen quite a few amps like that. On the other hand, sometimes the resistor is not obvious - it can also be across the speaker outputs, and again, sometimes in a form of a boucherot cell (one on both ends of the inductor). Leaving this undamped will create a virtual short circuit at the resonant frequency of the output L and cable C. For very fast amplifiers that's an invitation to oscillation, but for moderatly fast ones it's no big deal (NFB already largely reduced as wel las gain results in highish output resistance).
Great point! I wonder how much distortion you do get as you are after all limited by DCR of the voicecoil. Do you have any concrete data on this (intuitively it corresponds to my own experiences with lodspeaker design)?
Still, I would have to agree - speaker motor design is very often a showcase of truly incompetent engineering. I really can't see how a shorting ring or protruding pole piece can make a speaker THAT much more expensive...
Of course, but as you said as long as you are well away from resonance. In fact, once a passive crossover is in the picture, we are likely talking of large fractions of ohms series resistance at least, and further, thermal DCR changes tend to be larger.
However, once you start with active boxes, the difference is felt and needs to be accounted for. As I said, just one thing you cannot dismiss out of hand. After all, I've used relatively high DCR series inductors on purpose in one of my boxes to get the characteristic I wanted
I just prefer lumping the required components where they intuitively go, it's much easyer to combine amps and speakers that way on an equal a basis as possible
Kuei Yang Wang said:
Yes - to be honest, I would also say that using just an inductor is bad design, but I have seen quite a few amps like that. On the other hand, sometimes the resistor is not obvious - it can also be across the speaker outputs, and again, sometimes in a form of a boucherot cell (one on both ends of the inductor). Leaving this undamped will create a virtual short circuit at the resonant frequency of the output L and cable C. For very fast amplifiers that's an invitation to oscillation, but for moderatly fast ones it's no big deal (NFB already largely reduced as wel las gain results in highish output resistance).
Kuei Yang Wang said:
Great point! I wonder how much distortion you do get as you are after all limited by DCR of the voicecoil. Do you have any concrete data on this (intuitively it corresponds to my own experiences with lodspeaker design)?
Still, I would have to agree - speaker motor design is very often a showcase of truly incompetent engineering. I really can't see how a shorting ring or protruding pole piece can make a speaker THAT much more expensive...
Kuei Yang Wang said:
Of course, but as you said as long as you are well away from resonance. In fact, once a passive crossover is in the picture, we are likely talking of large fractions of ohms series resistance at least, and further, thermal DCR changes tend to be larger.
However, once you start with active boxes, the difference is felt and needs to be accounted for. As I said, just one thing you cannot dismiss out of hand. After all, I've used relatively high DCR series inductors on purpose in one of my boxes to get the characteristic I wanted
I just prefer lumping the required components where they intuitively go, it's much easyer to combine amps and speakers that way on an equal a basis as possible
"but I felt this arrangement, borrowed from Naim, is preferrable."
Yamaha did this too on their CR620/820/1020 series of receivers made in 1978~1980. They used 0.12 ohms.
Pre-echos of the AKSA to come, they also used a bootstraped VAS and an EF II output stage.
These were spec'd at less than 0.1% distortion from 20hz~20Khz, from phono in to speaker out, at any power level from 100mW to full rated output (and 0.05%IM).
Have we really made much progress?
Yamaha did this too on their CR620/820/1020 series of receivers made in 1978~1980. They used 0.12 ohms.
Pre-echos of the AKSA to come, they also used a bootstraped VAS and an EF II output stage.
These were spec'd at less than 0.1% distortion from 20hz~20Khz, from phono in to speaker out, at any power level from 100mW to full rated output (and 0.05%IM).
Have we really made much progress?
Hi DJK,
I believe the speed of the output stage, and methods of reducing crossover, have made modern Class AB PP amps better. There has also been progress with the speed of the voltage amplifier, and more accuracy with lag compensation. All these things have improved the sound, but significantly, by technological convergence, the topology remains pretty much the same.
Cheers,
Hugh
I believe the speed of the output stage, and methods of reducing crossover, have made modern Class AB PP amps better. There has also been progress with the speed of the voltage amplifier, and more accuracy with lag compensation. All these things have improved the sound, but significantly, by technological convergence, the topology remains pretty much the same.
Cheers,
Hugh
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.