The switching signal itself would be about 1000W.
But the amount of 23.5khz that gets through to the woofer would be down over 100dB, worst case.
But at an audible frequency, that would still be plenty to hear.
I could consider a foam rubber grill cloth too.
A subwoofer won't propagate well at 23.5khz, but still some.
As far as dogs and cats, 23.5khz will not move far in open air. It certainly will not go through walls or glass.
So, it would be only a localized problems.
I'll look up any links or referenes posted. But what I think is really important is what precedents are there?
There must be precedents for swithing amps that run this low. People must have dealt with this issue. Power supplies used to run this low.
Lets just let this thread sit, and see if anyone has info like that. Maybe they do, maybe they don't
I'd rather not keep rejustifying this over and over.
So lets let it sit and see if anyone has dealt with this issue, low ultrasonics. How much can you allow before it becomes perceptible, or an environmental issue?
But the amount of 23.5khz that gets through to the woofer would be down over 100dB, worst case.
But at an audible frequency, that would still be plenty to hear.
I could consider a foam rubber grill cloth too.
A subwoofer won't propagate well at 23.5khz, but still some.
As far as dogs and cats, 23.5khz will not move far in open air. It certainly will not go through walls or glass.
So, it would be only a localized problems.
I'll look up any links or referenes posted. But what I think is really important is what precedents are there?
There must be precedents for swithing amps that run this low. People must have dealt with this issue. Power supplies used to run this low.
Lets just let this thread sit, and see if anyone has info like that. Maybe they do, maybe they don't
I'd rather not keep rejustifying this over and over.
So lets let it sit and see if anyone has dealt with this issue, low ultrasonics. How much can you allow before it becomes perceptible, or an environmental issue?
Actually PWM represents analog sampling, but sampling is not fully periodical, it is phase modulated wich causes nonharmonical base-band modulation product. This is the reason why does fs have to be much higher then 2*fmax, but 10*fmax is enough to eliminate these products most of the time.
Interesting, I never thought of such, or heard of such.
So, at 23.5khz, I would meet your 10x req easily.
I still want to know if there is pecident for class D running this low, and it not posing a physiological or environmental hazard.
So, at 23.5khz, I would meet your 10x req easily.
I still want to know if there is pecident for class D running this low, and it not posing a physiological or environmental hazard.
zenmasterbrian said:
"I'll look up any links or referenes posted. But what I think is really important is what precedents are there?"
Use a search engine, find your own links, do your own research, arrive at your own conclusions. Don't be such a leech. Don't be such a troll. 15 pages of this is just rediculous, come on mods, lock it up!
classD4sure,
Lets just let this thread sit, unless there are people who do have some information to contribute.
I've not found any, or I would have posted it.
Lets not fill this thread up with arguing. If there are people with information they want to contribute, fine. Otherwise just let this thread sit.
Lets just let this thread sit, unless there are people who do have some information to contribute.
I've not found any, or I would have posted it.
Lets not fill this thread up with arguing. If there are people with information they want to contribute, fine. Otherwise just let this thread sit.
Hi zenmasterbrian,
As Variac has stated, you can not stop members from posting.
These members I know as helpful people and are trying to help you out. You should look in the direction they've been pointing you too. I understand why they are frustrated, and you as well.
At this point I really have to agree that you should do some digging and reading. There are very good reasons why switching power amps do not run at 23 KHz, at least accept that and figure out why. Do not expect people to design this for you, or spoon feed you. It's DIY and you have to do some work as well.
-Chris
As Variac has stated, you can not stop members from posting.
These members I know as helpful people and are trying to help you out. You should look in the direction they've been pointing you too. I understand why they are frustrated, and you as well.
At this point I really have to agree that you should do some digging and reading. There are very good reasons why switching power amps do not run at 23 KHz, at least accept that and figure out why. Do not expect people to design this for you, or spoon feed you. It's DIY and you have to do some work as well.
-Chris
there is no reason it *couldn't* run at 23khz - some of the earlier smps ran at around that frequency.
some pluses with running that slow: no emi to worry about, you can use cheap and slow switchers - especially power bjts to handle high current, easier layout, etc.
some minuses with running that slow: around the audiable band, large magenetics, etc.
you can also come up with a list of pluses / minuses with the use of a bjt vs. mosfet.
I do agree with others that you may want to listen to folks here. the critism you are hearing came from experience and knowledge that takes time to accumulate. so maybe you want to redirect the discussion to pluses / minuses rather than a validation of your ideas.
Just my thought.
some pluses with running that slow: no emi to worry about, you can use cheap and slow switchers - especially power bjts to handle high current, easier layout, etc.
some minuses with running that slow: around the audiable band, large magenetics, etc.
you can also come up with a list of pluses / minuses with the use of a bjt vs. mosfet.
I do agree with others that you may want to listen to folks here. the critism you are hearing came from experience and knowledge that takes time to accumulate. so maybe you want to redirect the discussion to pluses / minuses rather than a validation of your ideas.
Just my thought.
I would wager the concerns about ultra-sonic radiation are a bit misplaced. I wouldn't worry about the woofer... I would worry about shutting up the magnetics. The magnetic forces involved will make noise. I have built alot of inverters (huge class d amps) that run between 8 & 30 kHz... quieting the magnetics is always a problem. Vacuum impregnation, best with epoxy, is one remedy. You must also pay close attention to magnetic loops as they will rattle any iron around (and heat it).
Eva, makes a very good point, that you can look at in reverse. Unless you are switching at high voltage, over 200-300 Volts, there is no reason to consider BJTs at this power level. Mosfets are king in low voltages and there is a reason for it: faster switching = less iron = less money = better filtering.
Don't equate less costly, or more practical, with cheap or comprimised... not always the same thing.
🙂
Eva, makes a very good point, that you can look at in reverse. Unless you are switching at high voltage, over 200-300 Volts, there is no reason to consider BJTs at this power level. Mosfets are king in low voltages and there is a reason for it: faster switching = less iron = less money = better filtering.
Don't equate less costly, or more practical, with cheap or comprimised... not always the same thing.
🙂
anatech, the reason for starting this thread was to explore the frequency issue, particularly how low it could be.
Clearly, wanting such a low frequency is not the orthodox approach.
If people have differing views they are free to state them.
But if they just want to be beligereant, I don't think they should post here.
If I had been able to come up with any materials, I would have posted them already.
If this thread dies because I don't have anything to post, and nor does anybody else, OK.
In the 1970s, switching power supplies did run in the 20khz to 25khz range.
I am pleased to see that there are now some topical posts!
Clearly, wanting such a low frequency is not the orthodox approach.
If people have differing views they are free to state them.
But if they just want to be beligereant, I don't think they should post here.
If I had been able to come up with any materials, I would have posted them already.
If this thread dies because I don't have anything to post, and nor does anybody else, OK.
In the 1970s, switching power supplies did run in the 20khz to 25khz range.
I am pleased to see that there are now some topical posts!
fokker, thanks for your well stated views.
There certainly are lots of interrelated issues to consider when designing something in an unorthodox way.
My own interest is only in multiamped systems. The amp I am envisioning would only be for a subwoofer. It would be designed specifically for that appliation, not as a broad band audio amp.
As far as BJt or MOS, I want to defer exploring that on this thread.
Now I only want to look at a more basic issue. How low can the switching frequency be?
There certainly are lots of interrelated issues to consider when designing something in an unorthodox way.
My own interest is only in multiamped systems. The amp I am envisioning would only be for a subwoofer. It would be designed specifically for that appliation, not as a broad band audio amp.
As far as BJt or MOS, I want to defer exploring that on this thread.
Now I only want to look at a more basic issue. How low can the switching frequency be?
All other factors aside then, the lower limit for your switching frequency will be determined from you filtering. A lossless filter implies 2nd order LC filtering... 40 dB per decade.
So... 200 Hz LP corner freq will give you 80 dB carrier suppression at 20 kHz... less than ideal.
So... 200 Hz LP corner freq will give you 80 dB carrier suppression at 20 kHz... less than ideal.
poobah, I had already assumed that shutting up the magnetics may be necessary.
I know that using a higher frequency would mean for smaller and lower cost magnetics. I know that switching power supplies are designed with that objective in mind.
You say you have designed some in the 8khz to 30khz range?
What I have read is that they don't normally go into the audible range for fear of audibility?
But it is precedents like you are describing that I am looking for, examples of low ultra sonics.
You say shut up the magnetics? Did some of these power supplies run 20khz to 25khz? Did those also have shut up magnetics? Was that audible, perceptible, deemed a hazard to animals?
I know that using a higher frequency would mean for smaller and lower cost magnetics. I know that switching power supplies are designed with that objective in mind.
You say you have designed some in the 8khz to 30khz range?
What I have read is that they don't normally go into the audible range for fear of audibility?
But it is precedents like you are describing that I am looking for, examples of low ultra sonics.
You say shut up the magnetics? Did some of these power supplies run 20khz to 25khz? Did those also have shut up magnetics? Was that audible, perceptible, deemed a hazard to animals?
This came up before. I'm prepared to go to a higher order filter.
I don't have any need to keep this small or compact.
Resistance added by the filter will have a negligdeable effect on output resistance because with the low bandwidth I can have higher loop gain.
Using an acoustic filter, in the form of a thin foam rubber grill cloth is not out of the question.
Even going to some variation of a class H amp, or tracking switching power supply amp is not out of the question. In this case, the switching transistors still carry the same load, but there are also linear transistors. These act kinda like a filter, but a nonlinear filter.
So there are many possibilities.
Right now I just want to look at one angle:
Assuming some small amount of the switching signal did get to the driver, how low could the frequency be before it posed a problem?
23.5? 22khz? What exactly?
I want to look at this question in isolation, for now, and on this thread.
I don't have any need to keep this small or compact.
Resistance added by the filter will have a negligdeable effect on output resistance because with the low bandwidth I can have higher loop gain.
Using an acoustic filter, in the form of a thin foam rubber grill cloth is not out of the question.
Even going to some variation of a class H amp, or tracking switching power supply amp is not out of the question. In this case, the switching transistors still carry the same load, but there are also linear transistors. These act kinda like a filter, but a nonlinear filter.
So there are many possibilities.
Right now I just want to look at one angle:
Assuming some small amount of the switching signal did get to the driver, how low could the frequency be before it posed a problem?
23.5? 22khz? What exactly?
I want to look at this question in isolation, for now, and on this thread.
These are large power inverters I'm talking about. Animals, other than myself weren't a concern.
In these app's switching losses are the concern... not so much for the power they waste, but the heat they create in the switching devices. In some odd situations, freq's had to be low to avoid eddy losses in iron.
In all cases, the desire is to push freq's up. Read my previous post. Practical realizations of low loss filters is your real limtation on frequency.
And... to beat a dead horse, you should not separate the issues of freq/voltage with device type... BJT or MOS. That is fine for Phd's making textbooks, but they get paid to make graduates... not switching power convertors.
🙂
In these app's switching losses are the concern... not so much for the power they waste, but the heat they create in the switching devices. In some odd situations, freq's had to be low to avoid eddy losses in iron.
In all cases, the desire is to push freq's up. Read my previous post. Practical realizations of low loss filters is your real limtation on frequency.
And... to beat a dead horse, you should not separate the issues of freq/voltage with device type... BJT or MOS. That is fine for Phd's making textbooks, but they get paid to make graduates... not switching power convertors.
🙂
zenmasterbrian said:
I usually do 10x of the highest frequency, for full-band amps. Don't know the math but I assume poobah's is right.
Another way to look at this issue: whether it is 20khz or 2khz isn't going to make a difference as to your choice of devices (save the magenetics): you aren't going to find a device that works at 2khz but not at 20khz, or vise versa. Both are low enough even the least capable switchers will be sufficient for your job. so discussing this further is moot, in my view.
I want to separate the issues because that is the only way to get data on them. Certainly there are a whole bunch of other relevant dimensions to this. I just want to wait.
I do understand that the switching deviced that work at 20khz should also work at 30khz. But I still want to see how low will work, just in case.
It seem that some of this low ultrasonic could get out, become sonic.
So I'm looking for precedent, where someone has written about this, or designed things like this.
In the 70's, switchers did usually run that low.
There were some early class D audio amps that did too, for bass only.
I do understand that the switching deviced that work at 20khz should also work at 30khz. But I still want to see how low will work, just in case.
It seem that some of this low ultrasonic could get out, become sonic.
So I'm looking for precedent, where someone has written about this, or designed things like this.
In the 70's, switchers did usually run that low.
There were some early class D audio amps that did too, for bass only.
I am at a loss for what you are asking...
I do know that when looking for something, it helps to what you're looking for.
😕
I do know that when looking for something, it helps to what you're looking for.
😕
I'm looking for examples of when people have designed switching power supplies, or espcially class D audio amps that run below 28khz.
I'm also looking for books and articles where people have analyzed the issues.
I have not been able to find such myself, or I would have posted them.
Maybe some people have had such experiences, or know of such material. Maybe not.
Lets just leave this thread be, and see if anyone comes up with anything.
I'm also looking for books and articles where people have analyzed the issues.
I have not been able to find such myself, or I would have posted them.
Maybe some people have had such experiences, or know of such material. Maybe not.
Lets just leave this thread be, and see if anyone comes up with anything.
- Status
- Not open for further replies.