zenmasterbrian said:
😀 😀 Do we have parrot in here or what? Stubborn or just plain stupid?
You are living(dreaming on) in seventies when mosfet technology was still taking babysteps. Its a different world out there now...
Take your eye to your hand and look at the datasheets of the examples I gave? Find 100v TO-220 bipolar with less than 0.5 volt vcesat at 50Amps? Comprende amigo?
mzzj,
Tone!
No one is requiring that you post on this thread.
I wasn't prepared to get into specific semiconductor technologies at this time. I just wanted to address the frequency issue. How low is it safe to go, before it becomes an physiological hazard?
Clearly a great deal has gone into MOS technology in the recent decades. Vertical structures, and highly parallel layouts.
I'm going to have to look at what the latest bipolar technology is. I also want to look at some of the new device structures, some of which require induced communtation.
MOS is prefered for high switching frequencies, and a few other situations. But there are still situations where BJT, as well as some of the other technologies are still used.
If you have data sheets or vendor info for MOS, BJT, or any other devices, I will be interested in looking at them.
I'm still considering the frequency issue only tentative.
As far as looking at device technologies, that has not even started yet.
I'll be getting some familiar books back from a university library soon.
Tone!
No one is requiring that you post on this thread.
I wasn't prepared to get into specific semiconductor technologies at this time. I just wanted to address the frequency issue. How low is it safe to go, before it becomes an physiological hazard?
Clearly a great deal has gone into MOS technology in the recent decades. Vertical structures, and highly parallel layouts.
I'm going to have to look at what the latest bipolar technology is. I also want to look at some of the new device structures, some of which require induced communtation.
MOS is prefered for high switching frequencies, and a few other situations. But there are still situations where BJT, as well as some of the other technologies are still used.
If you have data sheets or vendor info for MOS, BJT, or any other devices, I will be interested in looking at them.
I'm still considering the frequency issue only tentative.
As far as looking at device technologies, that has not even started yet.
I'll be getting some familiar books back from a university library soon.
"Tone!
No one is requiring that you post on this thread."
Nor you for that matter. .
"I wasn't prepared to get into specific semiconductor technologies at this time." Wasn't it you that wanted to use BJT's.... hence the requirement of the lowest possible switching frequency? You brought them both up, it's been obvious to us for awhile you weren't prepared to discuss any part of it. I guess then you just want someone to say "yaaaaaaaaah build it cooool" which I gave you 🙂
It's really very easy to find the lowest switching frequency...... you're discussing how dark it will be if the lights go out without bothering to throw the switch and find out.. but then if you're already in the dark why bother, right?
Which brings to mind the following question. Are you able to or have you ever actually switched a mosfet? Bet ya haven't/can't?
What kind of an engineer that still has a marble or two left would actually want to build such an inefficient circuit just to get an award pinned on that say's "most efficient output stage ever". Add to that the fact it wouldn't even be an efficient output stage, which you don't seem to get at all.
You won't be convinced by the efforts of others, won't make your own, won't listen to reason, we dont' know what would convince you.. you're spastic and irrational. You haven't even been able to clearly define your objectives at all, or the reasoning behind them seemingly none from an informed background to say the least, yet you'd try and have us believe you're some sort of engineer? It's fun to play online..... isn't it.
At best I grant you manager of marketing department.
Please.... delete this thread.
No one is requiring that you post on this thread."
Nor you for that matter. .
"I wasn't prepared to get into specific semiconductor technologies at this time." Wasn't it you that wanted to use BJT's.... hence the requirement of the lowest possible switching frequency? You brought them both up, it's been obvious to us for awhile you weren't prepared to discuss any part of it. I guess then you just want someone to say "yaaaaaaaaah build it cooool" which I gave you 🙂
It's really very easy to find the lowest switching frequency...... you're discussing how dark it will be if the lights go out without bothering to throw the switch and find out.. but then if you're already in the dark why bother, right?
Which brings to mind the following question. Are you able to or have you ever actually switched a mosfet? Bet ya haven't/can't?
What kind of an engineer that still has a marble or two left would actually want to build such an inefficient circuit just to get an award pinned on that say's "most efficient output stage ever". Add to that the fact it wouldn't even be an efficient output stage, which you don't seem to get at all.
You won't be convinced by the efforts of others, won't make your own, won't listen to reason, we dont' know what would convince you.. you're spastic and irrational. You haven't even been able to clearly define your objectives at all, or the reasoning behind them seemingly none from an informed background to say the least, yet you'd try and have us believe you're some sort of engineer? It's fun to play online..... isn't it.
At best I grant you manager of marketing department.
Please.... delete this thread.
"So, it could sample as low as 200hz, but of course that is audible."
the 2x figure assumes analog / infinite bit sampling. PWM is a one-bit sampling process so I would imagine it requires much higher sampling frequencies. Without any fancy processing, for example, a 8-bit sampling can be duplicated by a 1-bit sampling at 256x frequency.
that gives you some perspective.
the 2x figure assumes analog / infinite bit sampling. PWM is a one-bit sampling process so I would imagine it requires much higher sampling frequencies. Without any fancy processing, for example, a 8-bit sampling can be duplicated by a 1-bit sampling at 256x frequency.
that gives you some perspective.
"This is their opinion, and it is given regarding the design of power supplies for commercial markets. DIYaudio does not have to compete in that kind of a market. "
the devices don't know if they are in a commercial power supply or a DIY amplifier. they react exactly the same given the same conditions. so as far as the devices are concerned, what quoted about their behaviors in a power supply applies to how they will behavior in an amplifier.
Can you find conditions in which a BJT is a superior switching device? I am sure you can. For example, in an application where very large current needs to be switch, I would imagine a BJT is desirable.
Is your application such one? only you can tell.
the devices don't know if they are in a commercial power supply or a DIY amplifier. they react exactly the same given the same conditions. so as far as the devices are concerned, what quoted about their behaviors in a power supply applies to how they will behavior in an amplifier.
Can you find conditions in which a BJT is a superior switching device? I am sure you can. For example, in an application where very large current needs to be switch, I would imagine a BJT is desirable.
Is your application such one? only you can tell.
zenmasterbrian said:
The answer to this question depends not just on frequency, but SPL at that frequency. What kind of SPL will a sub have if you feed it a 20kHz signal? Think about it because I'm not going to tell you.
You're going to use an 8th order low pass filter way down around, what was it, 300Hz? Here's a simple yes or no question: Will a high frequency signal entering a 300Hz 8th order lowpass filter be attenuated enough to not be audible? Having designed lots of filters before you should know the answer.
Now, will some moderator please nuke this thread?
Hey did anyone notice that the crazy and whacko graemlins are the same thing?
Hmm, you enter : whacko : but it just displays the crazy.gif image. Interesting.
Right on. Two votes to kill this thread. Can we get a third? Third time's the charm.
How do such characters get past moderation anyway.... I kind of thought that was the whole point to it.
How do such characters get past moderation anyway.... I kind of thought that was the whole point to it.
poobah get over to that energy conservation thread and help me out! You have me unsure of my mathematics. 🙂
These kinds of posts get past the moderators because none of them frequent the class d section.
These kinds of posts get past the moderators because none of them frequent the class d section.
fokker, thanks for your comments. I'm going to wait until things settle down a bit here, before responding to them.
classd4sure, workhorse, BWRX, poobah, there is no reason any of you need to be posting on this thread. You aren't adding anything to it.
classd4sure, workhorse, BWRX, poobah, there is no reason any of you need to be posting on this thread. You aren't adding anything to it.
zenmasterbrian,
The first couple of responses to your post were somewhat helpful. I have a few thought on the PWM frequency you should yse, so here's my 2 cents worth.
First of all, in your original post you suggest that for a BW of only 100 Hz, 200 Hz sampling rate should be good. This (the Nyquist sampling theorem) does not apply here.
Basically, even if the amp -3 db bandwidth is 100 Hz, its output will have higher frequency components that will be very difficult to filter. And if you you are using feedback, such a low frequency filter will cause stability problems.
In addition this low PWM frequency will cause large current ripple in the speaker and that will be audible. (This is just another way of saying what the previous paragraph says.)
A rule of thumb is that the PWM frequency should be at least 10 times the 3 db frequency of the closed loop system (assuming unit gain). This means you need at least 10 KHz. I would recommend 25 KHz minimum.
Edit: Oops, I made a mistake. Using the times 10 rule of thumb, you would need a PWM frequency of 1 KHz (again assuming unity gain). Obviously, this would be in the audio range, so I believe a frequency above the audio range should be used; maybe 20 KHz. Otherwise a LPF will be needed.
Rick
The first couple of responses to your post were somewhat helpful. I have a few thought on the PWM frequency you should yse, so here's my 2 cents worth.
First of all, in your original post you suggest that for a BW of only 100 Hz, 200 Hz sampling rate should be good. This (the Nyquist sampling theorem) does not apply here.
Basically, even if the amp -3 db bandwidth is 100 Hz, its output will have higher frequency components that will be very difficult to filter. And if you you are using feedback, such a low frequency filter will cause stability problems.
In addition this low PWM frequency will cause large current ripple in the speaker and that will be audible. (This is just another way of saying what the previous paragraph says.)
A rule of thumb is that the PWM frequency should be at least 10 times the 3 db frequency of the closed loop system (assuming unit gain). This means you need at least 10 KHz. I would recommend 25 KHz minimum.
Edit: Oops, I made a mistake. Using the times 10 rule of thumb, you would need a PWM frequency of 1 KHz (again assuming unity gain). Obviously, this would be in the audio range, so I believe a frequency above the audio range should be used; maybe 20 KHz. Otherwise a LPF will be needed.
Rick
fokker, thanks for your posts.
A PWM signal is a two state signal, but the exact width of the pulses is continuously variable.
So it is not a one bit situation. Usually the pulse width modulation is done in analog, by integrating a square wave to a triangle, and then putting that and the analog input signal into a comparator. There are other ways.
The basic theory is the Nyquist theroem, which says you have to sample at 2x the highest frequency you want to reproduce. But in practice you sometimes need to be higher for filtering considerations. Also, some audiophile concerns have promoted higher sampling rates.
A PWM signal is a two state signal, but the exact width of the pulses is continuously variable.
So it is not a one bit situation. Usually the pulse width modulation is done in analog, by integrating a square wave to a triangle, and then putting that and the analog input signal into a comparator. There are other ways.
The basic theory is the Nyquist theroem, which says you have to sample at 2x the highest frequency you want to reproduce. But in practice you sometimes need to be higher for filtering considerations. Also, some audiophile concerns have promoted higher sampling rates.
sawreyrw, thanks for your thoughts.
I may have talked about a smaller bandwidth in the beginning.
But lets for now say that the absolute upper audio band limit would be 300hz.
I'm looking at filtering at 1000hz. So the filter would be 3dB down at 1000hz, and should not have a noticable drop at 300hz. The specifics of it, like damping ratio, will determine that.
Typically this amp would not be used over 100hz.
I do want it to be a very high power amp, so I do want to have large currents available.
I'm believing that I should not us an audible or perceptible switching frequency, so I'm wondering what the lower limit is.
Tentatively I have targetted 23.5khz. This would be minimum. It could actually to a bit higher, as much as 25.5khz, to sync up with other amps and sources.
I know this syncing is not necessary. Its just a design feature for purism.
What I am not clear on is whether the small amout of 23.5khz that would get to the driver might still be somehow perceptible.
I have condsidered a backup of having a thin foam rubber grill cloth, as an acoustic filter.
My interest in this project is driven by a commitment to multi-amping, and designing each amp specifically for what it is expected to do. So this amp would be very high power, but only for a very low band width.
I would like to be able to translate this is into a low switching frequency. The question still is how low can it be.
Once that is a little more clear, I want to start talking about device and circuit technologies. Just not yet.
I have benefited greatly from all the links people have posted.
I may have talked about a smaller bandwidth in the beginning.
But lets for now say that the absolute upper audio band limit would be 300hz.
I'm looking at filtering at 1000hz. So the filter would be 3dB down at 1000hz, and should not have a noticable drop at 300hz. The specifics of it, like damping ratio, will determine that.
Typically this amp would not be used over 100hz.
I do want it to be a very high power amp, so I do want to have large currents available.
I'm believing that I should not us an audible or perceptible switching frequency, so I'm wondering what the lower limit is.
Tentatively I have targetted 23.5khz. This would be minimum. It could actually to a bit higher, as much as 25.5khz, to sync up with other amps and sources.
I know this syncing is not necessary. Its just a design feature for purism.
What I am not clear on is whether the small amout of 23.5khz that would get to the driver might still be somehow perceptible.
I have condsidered a backup of having a thin foam rubber grill cloth, as an acoustic filter.
My interest in this project is driven by a commitment to multi-amping, and designing each amp specifically for what it is expected to do. So this amp would be very high power, but only for a very low band width.
I would like to be able to translate this is into a low switching frequency. The question still is how low can it be.
Once that is a little more clear, I want to start talking about device and circuit technologies. Just not yet.
I have benefited greatly from all the links people have posted.
Hi Brian,
I am very sad because so far you told so many time about frquency issue.
It is fine you use BJT, OK. I don't care....!
But I am really care when you decide to ignore my free of charge googling for creature hearing frequency range and decide a 23.5kHz??
This is the reason that this thread shall be closed! As you say that you want lowest frequency for class d so far by 'external' reason, your neighbor's cats, dogs whales, duck? But then you ignore the data.
I am very sad because so far you told so many time about frquency issue.
It is fine you use BJT, OK. I don't care....!
But I am really care when you decide to ignore my free of charge googling for creature hearing frequency range and decide a 23.5kHz??
This is the reason that this thread shall be closed! As you say that you want lowest frequency for class d so far by 'external' reason, your neighbor's cats, dogs whales, duck? But then you ignore the data.
zenmasterbrian said:
Count me on. I cant stand this !#%¤ anymore.
And just to save your time I remind myself that I am not requested to post here and bipolars have lover voltage drop. Plus designing 8-order filters for class-d amps is so easy.
I'll add another layer of aluminium to my foil hat to keep troll-radiation at 23khz out.
- Status
- Not open for further replies.