Your explanation makes sense to me.
It also brings to current the previous posts in another Thread, where I claimed that very Low Vdrop across the amplifier results in the amplifier gain changing. Your Vin-Vload shows the changing Vdrop.
Many said my claim is wrong. It is simply clipping distortion, not changing gain.
I wonder if your alternative simulation could shed light?
It also brings to current the previous posts in another Thread, where I claimed that very Low Vdrop across the amplifier results in the amplifier gain changing. Your Vin-Vload shows the changing Vdrop.
Many said my claim is wrong. It is simply clipping distortion, not changing gain.
I wonder if your alternative simulation could shed light?
Xer, the amplifier should have three sets of capacitors.
a.)Smoothing, b.)MF decoupling, c.)HF decoupling.
You can omit one or two sets of these. You CANNOT omit all three sets.
If you are willing to experiment you can try
a+b+c, or a+b, or a+c, or b+c, or a, or b, or c.
Peter Daniel for his original implementation developed to suit his very high efficiency 4ohms speaker ( I seem to recall 95dB/watt or 96dB/watt) adopted b+c and his claims were for improved mid+treble, by omitting a.
I am guessing that very high efficiency speakers may have a somewhat restricted low bass ability and that this may not show up the low bass shortcomings of the amplifier by omitting a, the smoothing capacitance.
I have argued that his was a quite special case and that buyers should be forewarned of his nearly unique criteria for his adopted design philosophy. My plea fell on deaf ears.
a.)Smoothing, b.)MF decoupling, c.)HF decoupling.
You can omit one or two sets of these. You CANNOT omit all three sets.
If you are willing to experiment you can try
a+b+c, or a+b, or a+c, or b+c, or a, or b, or c.
Peter Daniel for his original implementation developed to suit his very high efficiency 4ohms speaker ( I seem to recall 95dB/watt or 96dB/watt) adopted b+c and his claims were for improved mid+treble, by omitting a.
I am guessing that very high efficiency speakers may have a somewhat restricted low bass ability and that this may not show up the low bass shortcomings of the amplifier by omitting a, the smoothing capacitance.
I have argued that his was a quite special case and that buyers should be forewarned of his nearly unique criteria for his adopted design philosophy. My plea fell on deaf ears.
Thanks Andrew,
Your explanation makes it a bit clearer. I drive a few High efficient 8 ohm speakers and they play beautiful, adding a high efficient bass 8 ohm in this system did not work well. Adding the extra caps on the PS made a huge difference. (so now it plays pretty fine).
I don't know if people should be warned, but some clear explanation would be fine.
I'm sure that all the "answers" are in the different threads, but a a noob it is pretty hard to find the answers if you don't know what to look for. Building this amp brought me to this questions (now I only have to understand the answers...
)
I am planning to Bi-amp my system and I was planning to do that with another set of LM4780 from PD. But after reading your last post and some concerns of others I am thinking about a different kind of amp for the bass section.
Since I am not an expert I am thinking about a Class-D, but before I start buying I will read some more on the subject. (what is fun).
Your explanation makes it a bit clearer. I drive a few High efficient 8 ohm speakers and they play beautiful, adding a high efficient bass 8 ohm in this system did not work well. Adding the extra caps on the PS made a huge difference. (so now it plays pretty fine).
I don't know if people should be warned, but some clear explanation would be fine.
I'm sure that all the "answers" are in the different threads, but a a noob it is pretty hard to find the answers if you don't know what to look for. Building this amp brought me to this questions (now I only have to understand the answers...
)I am planning to Bi-amp my system and I was planning to do that with another set of LM4780 from PD. But after reading your last post and some concerns of others I am thinking about a different kind of amp for the bass section.
Since I am not an expert I am thinking about a Class-D, but before I start buying I will read some more on the subject. (what is fun).
Follow on from #40
I am making this up as I go along so its taking a few attempts to get to the end.
Using the earlier used example, the signal going to the input via the feedback can be calculated:
Ripple compensation at input = U ripple / gain
Ripple compensation at input = 10V / 20 = 0.5V
In the sims the Rail Ripple Voltage has a sawtooth wave shape and its spectrum contains both even and odd harmonics, a 100 Hz sawtooth wave has harmonics way up in the MHZ.
Conclusion:
A chip-amp does not only amplify the input signal but also the Rail Ripple Voltage.
This can't be new, but I have not seen it explained like this in the chip-amp forum.
I am making this up as I go along so its taking a few attempts to get to the end.
Using the earlier used example, the signal going to the input via the feedback can be calculated:
Ripple compensation at input = U ripple / gain
Ripple compensation at input = 10V / 20 = 0.5V
In the sims the Rail Ripple Voltage has a sawtooth wave shape and its spectrum contains both even and odd harmonics, a 100 Hz sawtooth wave has harmonics way up in the MHZ.
Conclusion:
A chip-amp does not only amplify the input signal but also the Rail Ripple Voltage.
This can't be new, but I have not seen it explained like this in the chip-amp forum.
Last edited:
To my opinion its seems almost impossible that the amp chip is amplifying as well the input signal and the the rail ripple voltage.
As shown in the sims you showed the ripple voltage with low capacitance is quiet extreme and this should be easily noticed in the sound/output signal as a bad distortion
As shown in the sims you showed the ripple voltage with low capacitance is quiet extreme and this should be easily noticed in the sound/output signal as a bad distortion
Yes the bass becomes pretty bad as you approach maximum output and by the time maximum output has been reached it then starts to affect the mid and treble performance.
Low current demand (= high efficiency speaker) hides this defect.
Isn't the ripple on the rail proportional with the output demand?
I this case the distortion would not only be heard with maximum output but with every volume played, right?
If only heard with maximum volume its much more likely the distortion is caused because of clipping and that I can understand because of the dropping of rail voltage with high output demand
I this case the distortion would not only be heard with maximum output but with every volume played, right?
If only heard with maximum volume its much more likely the distortion is caused because of clipping and that I can understand because of the dropping of rail voltage with high output demand
Last edited:
But if you don't hear it, why you think its causing problems with the sound quality?
I read this thread as objective as I can and I agree with lots of opinions showed. I'm not an electronic expert but can observe sound quality quiet well as I was a sound technician for some time and spend a lot of time designing speakers some time ago.
My problem is that actually I agree with what many people here explain about capacitance in theory but when I listen again to the amp I have here in front of me the end result of what I myself do not agree with sounds REALLY GREAT.
So, what is happening here?
My problem is that actually I agree with what many people here explain about capacitance in theory but when I listen again to the amp I have here in front of me the end result of what I myself do not agree with sounds REALLY GREAT.
So, what is happening here?
Do some types of distortion sound pleasant? Some like the sound of tube-amps, maybe because they have a different distortion than solid state-amps.
Have you considered (tested) if your system will not sound even better with more capacitance?
In this thread there are now three technical explanations why low capacitance is a bad thing and none explaining why it is a good thing. There are many examples of increased capacitance improving sound, only a few claiming the opposite.
Have you considered (tested) if your system will not sound even better with more capacitance?
In this thread there are now three technical explanations why low capacitance is a bad thing and none explaining why it is a good thing. There are many examples of increased capacitance improving sound, only a few claiming the opposite.
I will never understand how people think or believe that using no capacitors or small capacitance in the power supply will sound better than using a hefty power supply. We should start asking those companies names like Bryston, Krell, Adcom etc to start using 10uf in their power supply...maybe then those amp can sound even better lol. Well, I guess if people are happy "thinking" or "believing" they sound better with tiny or no caps in their power supply...let them be
You are basically correct. But the bottom of the ripple voltage can't go down to vout, without clipping. It can only go down to vout + vclip, where vclip is the minimum voltage between the amplifier output stage's power input and its output. At least one transistor and probably a resistor have to be in there, after all. Vclip is given in a graph in the LM3886 datasheet, versus rail voltage. It's usually 2 to 5 volts. It's easier to see if you sim with a discrete transistor amp model, like the Cordell one I used in the thread started by Nico Ras about power supply reservoir capacitance size (where "resevoir" was mis-spelled in the thread title). That thread has everything you ever wanted to know about this stuff.
And yes, the amplifier has to amplify something proportional to the rail voltage, but out of phase through the feedback, to get rid of it from the output, as you stated. That is related to the PSRR (power supply rejection ratio) of the amp. But note too that its success is frequency dependent. Also, since i = C dv/dt, and I (the current) is proportional to the music signal, the voltage is proportional to the integral of the music signal, not the signal itself. That's one component of it, anyway, along with the superimposed sawtooth from the draining between charging pulses.
To preclude clipping at maximum output power, we can simply pretend, when doing the math to calculate the minimum required reservoir capacitance, that the output might be a constant DC current with a level equal to the maximum peak of a sine output that would give the RATED maximum average output power. That way even the lowest bass frequencies will be covered. Also, if less capacitance is used, it simply lowers the rated maximum output power. I wrote some papers about all of that, which are also posted in Nico's thread. Sorry, I am on a different PC and don't have any of the links handy.
The spreadsheet that I posted in several threads takes all of that into account, along with the solution of the differential equation involving the capacitance, including ESR. I will try to post the latest version of it, here, possibly this evening. (I also posted a spreadsheet that has a scalable transformer model, that solves the second-order differential equation for a linear power supply with a constant-current load and plots the output. If you get serious about simulating this stuff, that transformer model could be very helpful, in spice.)
Last edited:
Please come over and listen to my setup and compare, I also would love to listen yours. You know me in the way I handle complex analog problems and I am more hands on, only interested in the end result, not bothered by conservative theoretical reasons why things could not work because the past calculations is showing it
I'm not sure if I'm right but I'm not afraid to state my findings even the whole world is against it. this to my opinion can make a difference in the world to new ideas
Last edited:
If a certain person called Ron K. did not have the same ideas you would not do what your doing now and I would not have the life experience that I have.
The moment he had his idea how to do his things the whole world was against him like you said, 10 reasons it could not work and none that confirmed what he thought was working and see how it worked out...!!!
The moment he had his idea how to do his things the whole world was against him like you said, 10 reasons it could not work and none that confirmed what he thought was working and see how it worked out...!!!
Joery, I am already convinced your amp sounds great! To think it's because of the low capacitance is far to easy, there is a lot more going on. Peter Daniel has gone to great lengths to use the best sounding components, everything is designed for better sound.
If Peter was to build a car, I bet it would be a single seater with a wooden frame, no cup holders, no power steering, it would be controversial and everyone would want one.
Joery, I have great respect for you and your way of working, but its no good asking advise if you have already made up your mind and are just looking for conformation.
My explanation goes further than just saying, it sounds good to me, so it must be good, I read this forum to learn why it sounds good and how can it sound better, that involves a lot of theory. Thinking outside of the box still needs to be backed by theory.
Edit:
Ron's great ideas are always supported by sound engineering principals even if others don't see it.
If Peter was to build a car, I bet it would be a single seater with a wooden frame, no cup holders, no power steering, it would be controversial and everyone would want one.
Joery, I have great respect for you and your way of working, but its no good asking advise if you have already made up your mind and are just looking for conformation.
My explanation goes further than just saying, it sounds good to me, so it must be good, I read this forum to learn why it sounds good and how can it sound better, that involves a lot of theory. Thinking outside of the box still needs to be backed by theory.
Edit:
Ron's great ideas are always supported by sound engineering principals even if others don't see it.
Last edited:
gootee,
I was playing around with your spreadsheet only last week. I do not know if I am using the latest version. Just an idea, why don't you add an amp and speaker to it? Is it not the load that is most important for cap selection?
Thank you for confirming my explanation. Did you see the sims at #32? the red plots show what is happening at the power transistors of the chip. It does not look good.
I am wondering if it influences HF stability.
I was playing around with your spreadsheet only last week. I do not know if I am using the latest version. Just an idea, why don't you add an amp and speaker to it? Is it not the load that is most important for cap selection?
Thank you for confirming my explanation. Did you see the sims at #32? the red plots show what is happening at the power transistors of the chip. It does not look good.
I am wondering if it influences HF stability.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.