what factors affect the slew rates? is it possible to achieve slew rates of order of 200V/us or 300V/us? Ive read the book frm Douglas slef where he states that there is no use of getting higher slew rates and doesnt make difference...
but when I see krell amps they have that super fast slam and iron fist bass. checked the slew rates its 120V/us but hows that value achieved? what factors really make that slew rate.. compensation capacitors?
reducing it to very low value might incur stabilities isnt it?
making a 10pf comp capacitor gives 64v/us and 5pf can give 110v/us
but are there any other parameters to be considered?
but when I see krell amps they have that super fast slam and iron fist bass. checked the slew rates its 120V/us but hows that value achieved? what factors really make that slew rate.. compensation capacitors?
reducing it to very low value might incur stabilities isnt it?
making a 10pf comp capacitor gives 64v/us and 5pf can give 110v/us
but are there any other parameters to be considered?
That can be achieved relatively easily, even using a relatively simple circuit and crappy components.
This buffer example exceeds 500V/µs.
http://www.diyaudio.com/forums/solid-state/223762-alternative-buffer-topologies.html#post3245432
The paramount ingredient is the topology: by simply decreasing the cap values in a conventional circuit, you hit the limit very quickly
Basically, for amplifiers with one dominant pole that is determined by a compensation capacitor:
Slew rate = (maximum output current of the stage before the compensation capacitor)*(voltage gain of all stages after the compensation capacitor)/(capacitance of the compensation capacitor)
When Miller compensation is used, the actual stage having the Miller capacitor is neither before nor after the compensation capacitor.
A well-known trick to improve slew rate is increasing the bias current of the input stage and applying local series feedback to it (assuming that the input stage is the one and only stage before the compensation capacitor). Only increasing the bias current usually means that you will need a bigger compensation capacitor, so you don't win anything. Another trick is to use a class AB-biased input stage that can deliver large peak currents when needed, like the examples that Elvee showed.
Slew rate = (maximum output current of the stage before the compensation capacitor)*(voltage gain of all stages after the compensation capacitor)/(capacitance of the compensation capacitor)
When Miller compensation is used, the actual stage having the Miller capacitor is neither before nor after the compensation capacitor.
A well-known trick to improve slew rate is increasing the bias current of the input stage and applying local series feedback to it (assuming that the input stage is the one and only stage before the compensation capacitor). Only increasing the bias current usually means that you will need a bigger compensation capacitor, so you don't win anything. Another trick is to use a class AB-biased input stage that can deliver large peak currents when needed, like the examples that Elvee showed.
PA audio don't have high slew rate. About <80V/uS
High-end amp will do it. They have very large BW and Extremely high slew rate.
Spectral DMA50: 1000V/uS BW: 0 - 1.2MHz
Magtech amp: 500V/uS
Goldmund Telos: 300V/uS
....
Marcel, the LTP current can be set independently of the ULG frequency. You set the ULG based on the closed loop amplifier gain, LTP degen resistors and Cdom. In my e-Amp design, the slew rate is 155 V/us and uses conventional Miller comp (you can set it for TMC also using jumpers).
However, it's quite possible get 300 or 400 V/us using Miller inclusive comp using this same design, although I did not pursue this option.
CFA topologies offer the possibility of even higher SR's because they do away with the input transconductance stage (i.e voltage to current LTP), with the feedback current effectively driving the compensation capacitor directly via the trans impedance stage. Opamp CFA's thus configured have beef designed that offer in excess of 1000 V/us.
So, how much SR do you really need? Some practitioners say 1 V/us per peak output voltage is a good guide, which I would think is ok.
However, it's quite possible get 300 or 400 V/us using Miller inclusive comp using this same design, although I did not pursue this option.
CFA topologies offer the possibility of even higher SR's because they do away with the input transconductance stage (i.e voltage to current LTP), with the feedback current effectively driving the compensation capacitor directly via the trans impedance stage. Opamp CFA's thus configured have beef designed that offer in excess of 1000 V/us.
So, how much SR do you really need? Some practitioners say 1 V/us per peak output voltage is a good guide, which I would think is ok.
At bass frequencies, slew rates don't matter as much as damping factor. High slew rates are important for keeping IMD under control at higher signal frequencies. But for tight, controlled bass, you want low output impedance.
You say so. Please explain why it would be the case (not the bandwidth anyway, it will be increased in the same proportions of the FB ratio, not robbed)
With amplifiers that use global feedback, a high open loop gain at low frequencies will produce a very low impedance once the loop is closed. For example, if your amp has an open loop output impedance of 1 Ohm, and an open loop gain of 100dB at 50Hz, if you close the loop with 20dB of total gain, you're left with 80dB reduction in output impedance. That turns the 1 Ohm into .0001 Ohm at 50Hz (one decimal place for each 20dB). Of course, that's theoretical, and you still have wire resistance from the amp output to the speaker that adds to the total.
So, lots of feedback and large gauge wire will improve the chances for tight, smooth, effortless bass. That and some decent speakers of course.
So, lots of feedback and large gauge wire will improve the chances for tight, smooth, effortless bass. That and some decent speakers of course.
Last edited:
Sandy is asking better questions and he has touched upon one of the big marketing hypes in hi fi.
The best (subjectively) amplifier I ever built had a slow slew rate of 1.25 v/uS. Tube amps (with their output transformers) are notoriously slow and faster speeds were easily available with ordinary parts and topologies once transistors became practical.
A slower amplifier will not produce full power bandwith. Is full power bandwidth necessary for high fidelity reproduction of musical programs? No, it is not.
As far as "slamming" bass, damping factor was mentioned and is an important consideration. Damping factor is however highly overrated as well and anything over 50 or so is more than adequate. But the power supply has a great effect on bass performance. Adequate transformers to reduce voltage droop and generous filter capacitors have a tremendous effect on percieved bass response. High slew rates have no effect on bass response.
There are different philosophies and the present mindset is high speed and very high loop gain. Many excellent amplifiers have only had 20-30 dB of loop gain though; if your circuit is linear enough this can work quite well. Tube amps were like that; it would be nigh on impossible to design a stable tube amp with higher loop gain.
The best (subjectively) amplifier I ever built had a slow slew rate of 1.25 v/uS. Tube amps (with their output transformers) are notoriously slow and faster speeds were easily available with ordinary parts and topologies once transistors became practical.
A slower amplifier will not produce full power bandwith. Is full power bandwidth necessary for high fidelity reproduction of musical programs? No, it is not.
As far as "slamming" bass, damping factor was mentioned and is an important consideration. Damping factor is however highly overrated as well and anything over 50 or so is more than adequate. But the power supply has a great effect on bass performance. Adequate transformers to reduce voltage droop and generous filter capacitors have a tremendous effect on percieved bass response. High slew rates have no effect on bass response.
There are different philosophies and the present mindset is high speed and very high loop gain. Many excellent amplifiers have only had 20-30 dB of loop gain though; if your circuit is linear enough this can work quite well. Tube amps were like that; it would be nigh on impossible to design a stable tube amp with higher loop gain.
He's not listening to how amps sound, he's saying that based on his technical understanding.
I reckon there's a correlation between higher slew rates and better sound, but I very much doubt that the higher slew rate is directly responsible. Rather its a marker for better HF linearity and its that linearity which is what gives rise to better sound.
true linearity is the key but here we do not want amp to stint at any instance no matter what kind of current demands happens...
Hi Bonsai,
The point I was trying to make is that you need some degree of local feedback (such as long tailed pair degeneration resistors) to be able to increase the bias current without (substantially) increasing the unity loop gain frequency. If I understand you correctly, that is the same point you are making, in which case we agree.
By the way, I hate the use of the term current feedback for an amplifier with series feedback at the input, shunt feedback at the output and a low open-loop input impedance at the negative input, because the term current feedback originally refers to feedback configurations with series rather than shunt feedback at the output, which has nothing to do with whatever the open-loop input impedance at the negative input may be. Anyway, that's a different discussion.
Best regards,
Marcel
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.