Hi all,
I am currently designing/building my own Class-AB power amp.
Thus I came across selecting a proper output inductor. Spice simulations have shown that I need a 2µH air inductor.
So in simulation as well as in reality I recognized that the inductor followed by the nominal 4 ohms load (resistive) reduces the effective slew rate that is seen by the load. That is, measured in front of the inductor my amp is capable of producing 72V/µs into 4 ohms load. But under the same conditions measured right behind the inductor the slew rate is just about 25V/µs.
I am now wondering when reading about amplifiers that have slew rates of >100V/µs and also have quit large output inductors.
How can that be? Do manufacturers measure in front of or behind the output inductor? Or: How to achieve >100V/µs slew rates behind the inductor with a 4 ohm load present?
Hope you can help me with this because searching the internet as well as reading books from Bob Cordell or Douglas Self could not help me with this.
Thanks!
PS: Please apologizes my not perfect English. Greetings from Germany 🙂
I am currently designing/building my own Class-AB power amp.
Thus I came across selecting a proper output inductor. Spice simulations have shown that I need a 2µH air inductor.
So in simulation as well as in reality I recognized that the inductor followed by the nominal 4 ohms load (resistive) reduces the effective slew rate that is seen by the load. That is, measured in front of the inductor my amp is capable of producing 72V/µs into 4 ohms load. But under the same conditions measured right behind the inductor the slew rate is just about 25V/µs.
I am now wondering when reading about amplifiers that have slew rates of >100V/µs and also have quit large output inductors.
How can that be? Do manufacturers measure in front of or behind the output inductor? Or: How to achieve >100V/µs slew rates behind the inductor with a 4 ohm load present?
Hope you can help me with this because searching the internet as well as reading books from Bob Cordell or Douglas Self could not help me with this.
Thanks!
PS: Please apologizes my not perfect English. Greetings from Germany 🙂
Last edited:
The inductor will certainly make a lowpass filter at the output of the amp. The fastest it will slew, if everything in front is infinitely fast, is R/L volts per second per volt of output step. That slew rate is only observed at the beginning of the transition, as later in the transition, it will slow down as it completes the exponential approach.
Using your values, the fastest you'd expect it to slew is 4/2e-6 = 2V/us per volt of step amplitude. So, if the amplifier output before the inductor slewed from -15V to +15 Volts, you would see at most 60 V/us across the 4 Ohms.
I think most people would measure the slewing on the amp side of the inductor, since the lowpass action of the L-R filter would limit the observed slew rate. That shouldn't be an issue, since it's just a linear filter, with a rolloff at 318 kHz. The slewing test is a way to see that the amp is fast enough not to slope overload. Looking at the output of the L-R filter would kind of obscure that measurement.
Finally (getting too long a post here), many people would shunt the L with about 10 Ohms. That would increase the output slew rate, and also damp hi-Q resonances that might otherwise cause a very low impedance to appear across the output of the amp when driving capacitive loads.
Akitika GT-101
Update My Dynaco
Using your values, the fastest you'd expect it to slew is 4/2e-6 = 2V/us per volt of step amplitude. So, if the amplifier output before the inductor slewed from -15V to +15 Volts, you would see at most 60 V/us across the 4 Ohms.
I think most people would measure the slewing on the amp side of the inductor, since the lowpass action of the L-R filter would limit the observed slew rate. That shouldn't be an issue, since it's just a linear filter, with a rolloff at 318 kHz. The slewing test is a way to see that the amp is fast enough not to slope overload. Looking at the output of the L-R filter would kind of obscure that measurement.
Finally (getting too long a post here), many people would shunt the L with about 10 Ohms. That would increase the output slew rate, and also damp hi-Q resonances that might otherwise cause a very low impedance to appear across the output of the amp when driving capacitive loads.
Akitika GT-101
Update My Dynaco
Last edited:
IMO it's not as important as you think. First - there is a difference between "slewing" and rise time. Slewing is when one of the stages is usually when one stage of an amp is in a state where the voltage can't increase any faster and additional signal is lost (and usually other misbehavior) IMO it seems to be a form of "current clipping" in one stage. Good examples can be seen in most op amp data sheets. The limitation in HF response caused by an inductor will not result in other signal loss or other misbehavior.
Also, output inductor won't be a factor with most speakers, as tweeter inductance is usually significantly higher, making output indctance a moot point.
Also, output inductor won't be a factor with most speakers, as tweeter inductance is usually significantly higher, making output indctance a moot point.
Last edited:
Thank you djoffe and sregor. Your both answers helped me a lot. Now I know that I am on the right way with my little amp. And I know where I will measure the slew rate when it comes to writing the specifications sheet ;-)
Hi,
A properly designed amplifiers slew rate is hypothetical. The section that
limits slew rate is isolated and the designs max "slew rate" established,
is usually how it works, but you then set about designing the amplifier
so that slew rate limiting will never occur under normal operational
conditions for the amplifier and it is a gross overload condition.
5V/uS is adequate for 100W/8ohm/20KHz, (200W/4ohm/20KHz).
10V/uS is adequate for 400W/8ohm/20KHz, (800W/4ohm/20KHz).
Slew rate, power and bandwidth are inextricably linked at high frequencies.
A good amplifier might need say 5V rms input at 40KHz to induce
slew rate limiting in an intermediate stage, though this is clearly
going to cause gross overall amplifier clipping and distortion.
The output inductor doesn't really come into practical considerations.
You will never see a good amplifiers claimed slew rate at the output.
rgds, sreten.
Of course, badly designed amplifiers can have slew rate problems.
A properly designed amplifiers slew rate is hypothetical. The section that
limits slew rate is isolated and the designs max "slew rate" established,
is usually how it works, but you then set about designing the amplifier
so that slew rate limiting will never occur under normal operational
conditions for the amplifier and it is a gross overload condition.
5V/uS is adequate for 100W/8ohm/20KHz, (200W/4ohm/20KHz).
10V/uS is adequate for 400W/8ohm/20KHz, (800W/4ohm/20KHz).
Slew rate, power and bandwidth are inextricably linked at high frequencies.
A good amplifier might need say 5V rms input at 40KHz to induce
slew rate limiting in an intermediate stage, though this is clearly
going to cause gross overall amplifier clipping and distortion.
The output inductor doesn't really come into practical considerations.
You will never see a good amplifiers claimed slew rate at the output.
rgds, sreten.
Of course, badly designed amplifiers can have slew rate problems.
Last edited:
Slew rate is measured in the amps non linear state, and has very little to do with the amp playing music. As long as its considerably faster (2x) than the numbers sreten has provided (which are easy to accomplish) you have no worries. The Zobel network is not part of the equation.
- Status
- Not open for further replies.