Hi ClassD !
I plan to check that thread since one week, but I am getting afraid more and more by the horrible length....
....but I know, I really SHOULD !!!
Who can read- takes advantage! .... an already experienced topic....
Thanks
Markus
I plan to check that thread since one week, but I am getting afraid more and more by the horrible length....
....but I know, I really SHOULD !!!
Who can read- takes advantage! .... an already experienced topic....
Thanks
Markus
Hi 🙂
Don't be afraid of the length of it we went to alot of effort to work things out so life could be easier for people interested in class d..
I have looked at all these IC's before and I found it hard to nail it down to anything workable as they are all designed for such things as power supplies or pfc, I'm sure there must be one out there somewhere that could be made to work but the time required to find it is more than I can stand. Some are current feedback some voltage some both some phase some hysteresis some both...It's very confusing. Then you see Bruno's nice little UCD comparator..and you'll feel it's pull sucking you in.
I think it's likely safe to skip the first few pages at least, but there's good info throughout. Not saying you should give up and start over with a new design but I'm sure you'll get some good ideas from it, I think you'll really like the comparator from it for one thing. Dual floating outputs 180 deg. out of phase with a nice low part count. Would likely take less time making that than in finding a suitable PSU chip that could be adapted. I've looked high and low for one before but they either lacked features or speed..or something. Even the fabled Mueta IC will require 4 op amps, transformer...etc.
Best Regards,
Chris
Don't be afraid of the length of it we went to alot of effort to work things out so life could be easier for people interested in class d..
I have looked at all these IC's before and I found it hard to nail it down to anything workable as they are all designed for such things as power supplies or pfc, I'm sure there must be one out there somewhere that could be made to work but the time required to find it is more than I can stand. Some are current feedback some voltage some both some phase some hysteresis some both...It's very confusing. Then you see Bruno's nice little UCD comparator..and you'll feel it's pull sucking you in.
I think it's likely safe to skip the first few pages at least, but there's good info throughout. Not saying you should give up and start over with a new design but I'm sure you'll get some good ideas from it, I think you'll really like the comparator from it for one thing. Dual floating outputs 180 deg. out of phase with a nice low part count. Would likely take less time making that than in finding a suitable PSU chip that could be adapted. I've looked high and low for one before but they either lacked features or speed..or something. Even the fabled Mueta IC will require 4 op amps, transformer...etc.
Best Regards,
Chris
Hello Chris!
You are very lucky! For some months Fairchild doesn't send free samples to Hungary, I don't know why! 🙁 Furthermore it seems to be impossible to buy from them via distributor either.
Any idea?
You are very lucky! For some months Fairchild doesn't send free samples to Hungary, I don't know why! 🙁 Furthermore it seems to be impossible to buy from them via distributor either.
Any idea?
Pafi, rules exist, usually for good reason. Get your friend in Canada to get you samples.
For example: About 1980, I had to supply Fairchild transistors to the University of California, Berkeley for a real project, to try. They were only 50 miles away! Still, they were a university, not a real company. You are not being excluded without a general guideline being in place.
For example: About 1980, I had to supply Fairchild transistors to the University of California, Berkeley for a real project, to try. They were only 50 miles away! Still, they were a university, not a real company. You are not being excluded without a general guideline being in place.
Pafi said:
Hi, I must say I haven't actually tried obtaining any of their samples yet. Maybe I should try monday morning and see....Might not look too good if i tried to order them at 4am sunday
Have you tried emailing them to find out what the problem is?
Regards,
Chris
Thank both of you!
Chris!
The registration done by computers, so you can order at any time! They deny my requests immediately as I log on. They write: "We are currently unable to process your samples request." "Currently", but this is permanent.
Please inform me, if your order is accepted!
I haven't try to email them. I asked their local distributor, who said that this is a world-wide problem, there aren't enough SMPS semiconductor device produced to satisfy needs. But then how can they send samples to others?
Chris!
The registration done by computers, so you can order at any time! They deny my requests immediately as I log on. They write: "We are currently unable to process your samples request." "Currently", but this is permanent.
Please inform me, if your order is accepted!
I haven't try to email them. I asked their local distributor, who said that this is a world-wide problem, there aren't enough SMPS semiconductor device produced to satisfy needs. But then how can they send samples to others?
Hi, yeah sounds like something is up.
Have you tried creating a new account, new email..etc?
What is they allow? Two samples per device? Maybe they only let you try two or three.. I will let you know but I'll wait until business hours just in case it red flags me for ordering at an oddball time or something.
Have you tried creating a new account, new email..etc?
What is they allow? Two samples per device? Maybe they only let you try two or three.. I will let you know but I'll wait until business hours just in case it red flags me for ordering at an oddball time or something.
"Have you tried creating a new account, new email..etc?"
I met others who experienced the same thing.
They used to allow 10-25 samples per device, depending on part number. Now I can't reach the point where I could specify quantity, or anything. I quess you just believe that you will order sample. I'm really sorry! Their products are (would be) really good!
I met others who experienced the same thing.
They used to allow 10-25 samples per device, depending on part number. Now I can't reach the point where I could specify quantity, or anything. I quess you just believe that you will order sample. I'm really sorry! Their products are (would be) really good!
It took half a dozen tries but I got through the order process. Don't know if it will be approved though. 4 FDP3682's...nice!
OT: Sampling Rate
4 samples... jeez. Getting an amplifier circuit from the first sketches to something useful requires upward of 50pcs.
IMO, manufacturers of class-d-esque components (fets, coils, drivers etc) would do well to have a flexible sampling regime towards hobbyists and students.
The whole world is currently bracing for an onslaught of new class D products, but the developers aren't there! Schools no longer teach analogue electronics (let alone power) and companies can't or won't invest in proper r&d (quarterly result fixation).
Many audio companies are simply reassigning their "AB" designers to "D", with predictable end results. We're even seeing a backswing in the Deep East where companies are moving back to class AB because they even can't handle application of class D IC's (the local app engineers aren't helping much either).
The only chance we have at getting class D off the ground is facilitating enthousiasts to do their hobby so some can become pros in due time. For the companies to provide easy access to samples is a small investment, and the returns could be huge.
4 samples... jeez. Getting an amplifier circuit from the first sketches to something useful requires upward of 50pcs.
IMO, manufacturers of class-d-esque components (fets, coils, drivers etc) would do well to have a flexible sampling regime towards hobbyists and students.
The whole world is currently bracing for an onslaught of new class D products, but the developers aren't there! Schools no longer teach analogue electronics (let alone power) and companies can't or won't invest in proper r&d (quarterly result fixation).
Many audio companies are simply reassigning their "AB" designers to "D", with predictable end results. We're even seeing a backswing in the Deep East where companies are moving back to class AB because they even can't handle application of class D IC's (the local app engineers aren't helping much either).
The only chance we have at getting class D off the ground is facilitating enthousiasts to do their hobby so some can become pros in due time. For the companies to provide easy access to samples is a small investment, and the returns could be huge.
Actually I got 12 I just didn't want to seem greedy
What's worse being greedy or a liar? 😀
At most right now I'm concentrating on one half bridge, possibly two, with the ability to convert to two full bridges, and I still get to smoke a few of em!
I know some people feel if you get 10+ samples they consider it abusing the system, but yeah if I could simply go to a store and pick up a few of them for a few bucks I sure would have.
It isn't healthy for me to think about the sort of education I got in school so I wont' get started on it, all I will say is thank god for the internet, decent app notes, and people like yourself who are willing to help someone learn and steer them in the right direction.
Thanks.
What's worse being greedy or a liar? 😀
At most right now I'm concentrating on one half bridge, possibly two, with the ability to convert to two full bridges, and I still get to smoke a few of em!
I know some people feel if you get 10+ samples they consider it abusing the system, but yeah if I could simply go to a store and pick up a few of them for a few bucks I sure would have.
It isn't healthy for me to think about the sort of education I got in school so I wont' get started on it, all I will say is thank god for the internet, decent app notes, and people like yourself who are willing to help someone learn and steer them in the right direction.
Thanks.
I think none of us would mind paying a fair price for samples. They could simply charge $2 per fet plus transport. That would not cover the actual cost of the operation (which is an investment from their point of view) but it would prevent abuse.
Re: getting class D off the ground
(Or were you thinking that eventually those hobbyists becoming pros will take employment at the large manufacturers?)
Sorry for the off-topic post 😉
Hehe, it takes hell-of-a-lot for some startup to compete with Sony, Phillips, etc. My guesstimate is that the big manufacturers of plastic junk have 99% of the market, and that the same percentage of consumers don't care enough to even try to shop for "quality" in terms of durability and sound quality.Bruno Putzeys said:
(Or were you thinking that eventually those hobbyists becoming pros will take employment at the large manufacturers?)
Sorry for the off-topic post 😉
Re: Re: getting class D off the ground
I wouldn't mind that, of course.
It was with mild amusement that I found out that Jeff Rowland had put out a biiig shiny box with a huge transformer and two ICEPower modules in it.
Needless to say the press was raving😀
johanps said:
Should they necessarily? I think folks like NAD, Arcam etc could use a good class D hand just as well. Unless you are suggesting they all buy UcD modules 🙂johanps said:
I wouldn't mind that, of course.
It was with mild amusement that I found out that Jeff Rowland had put out a biiig shiny box with a huge transformer and two ICEPower modules in it.
Needless to say the press was raving😀
One has to admit that even the sometimes overhyped-but-maybe-not-so-good-in-the-end implementations of class-d help to increase the acceptance of this "technology".
Regards
Charles
To continue http://www.diyaudio.com/forums/showthread.php?s=&postid=424402#post424402
Yeah, seems 50x50mm PCB for two channel UcD it is too much imprudence.. for me at least..After all, the single channel is worked, but PCB now quite similar P2P (many scalpel efforts, and alternate wires, filter coil [still unclosed core] keep away from PCB). 1khz THD measurement is ready:
real .016% 588khz
PSpice .00097% 750khz
MC7 .011% 870khz
after adjusting switching to 588khz:
PSpice .004%
MC7 .009%
Tomorrow, i hope, will try same object, but with MC7 optimal values. Also, expect that result will be worse 🙂
Yeah, seems 50x50mm PCB for two channel UcD it is too much imprudence.. for me at least..After all, the single channel is worked, but PCB now quite similar P2P (many scalpel efforts, and alternate wires, filter coil [still unclosed core] keep away from PCB). 1khz THD measurement is ready:
real .016% 588khz
PSpice .00097% 750khz
MC7 .011% 870khz
after adjusting switching to 588khz:
PSpice .004%
MC7 .009%
Tomorrow, i hope, will try same object, but with MC7 optimal values. Also, expect that result will be worse 🙂
Hobbyists turning Pro
I'm one of those hobbyists who turned pro a long time ago. I built my first amp in 1968. I started making good amps around 1977. Now, I'm the senior electronics engineer at a pro sound manufacturer recognized for having about the best sounding powered pro speakers on the market. An engineering driven company without a single accountant or lawyer in management.
And we have gone to all Class-D in our own products and we have found that the ICE modules are the best sounding of the lot. No, they are not as good as the finest Class AB MOSFET amplifiers on the market but they are as good as any of the linear pro amplifiers and the best of the consumer amplifiers. Only the most "audiophile" grade amplifiers sound better than ICE modules.
As well, I have the latest ICE prototypes sitting on my desk, and they are coming out with stuff that will just blow away almost every amp designer's work with product of such low cost per watt with sound quality better than 98% of what's out there that any designer would be deficient in their duty not to look at the technology.
I have right now a module that delivers 200W into 4 ohms that includes a switching regulated power supply that operates from 90 to 26-VAC. About 4" x 6" and while I cannot reveal the price, I can barely buy the power transformer and filter capacitors for a class AB amplifier for the price.
Now that I have your attention, the only real catch is that, as far as I know, their stuff is not available for the hobby market. Sorry guys.
My point was that it is possible for hobbyists to get into the engineering departments of manufacturers and make a difference.
I'm one of those hobbyists who turned pro a long time ago. I built my first amp in 1968. I started making good amps around 1977. Now, I'm the senior electronics engineer at a pro sound manufacturer recognized for having about the best sounding powered pro speakers on the market. An engineering driven company without a single accountant or lawyer in management.
And we have gone to all Class-D in our own products and we have found that the ICE modules are the best sounding of the lot. No, they are not as good as the finest Class AB MOSFET amplifiers on the market but they are as good as any of the linear pro amplifiers and the best of the consumer amplifiers. Only the most "audiophile" grade amplifiers sound better than ICE modules.
As well, I have the latest ICE prototypes sitting on my desk, and they are coming out with stuff that will just blow away almost every amp designer's work with product of such low cost per watt with sound quality better than 98% of what's out there that any designer would be deficient in their duty not to look at the technology.
I have right now a module that delivers 200W into 4 ohms that includes a switching regulated power supply that operates from 90 to 26-VAC. About 4" x 6" and while I cannot reveal the price, I can barely buy the power transformer and filter capacitors for a class AB amplifier for the price.
Now that I have your attention, the only real catch is that, as far as I know, their stuff is not available for the hobby market. Sorry guys.
My point was that it is possible for hobbyists to get into the engineering departments of manufacturers and make a difference.
Hi Chris!
So it is true. They discriminate us by country. But why?
"FDP3682"
Good choice! It's low Qg is very nice if you drive it with discrete parts. I use bigger ones (FDP2532), since IR2113 can drive it. (And this is the only available FD.... transistor in Hungary.
)
Best regards:
Peter Sass
So it is true. They discriminate us by country. But why?
"FDP3682"
Good choice! It's low Qg is very nice if you drive it with discrete parts. I use bigger ones (FDP2532), since IR2113 can drive it. (And this is the only available FD.... transistor in Hungary.
)Best regards:
Peter Sass
Fraser,
Interesting, in this topic; http://www.diyaudio.com/forums/showthread.php?postid=432708#post432708 ,
one of mine "possible" customer did test our proto 400W Class-D (UcD400) against the ICEpower ICE500A.
After a visit in our companíes lab, where they have compared by themself on an AP System 2 the UcD400 & ICE500A, they were even more impressed.......
Indeed a SMPS is very nice, for shure when it is compatible for 90 to 250VAC. But a SMPS is FAR more expensive as a good build toroideltransformer. By a special construction our toroidel transformer is lighter and more powerful as a standard toroidel. So for the moment I don't want serve mine customers with a FAR much more complex powersupply as a SMPS definatly is.
The ICE250PS has a size of 5.9" x 6.4" and an overall eff. of 80%
Our Custom Made 200W has size of 3.15" x 3.94" + transformer size 3.94" x 1.92", and an overall eff. of 92%.
Weight of the both amps will be more of less the same.
For a manufacture our custom made amp will be highly attractive I guess.
If you are interested for a 1 tot 1 acoustically test?
Regards,
Jan-Peter
www.hypex.nl
Interesting, in this topic; http://www.diyaudio.com/forums/showthread.php?postid=432708#post432708 ,
one of mine "possible" customer did test our proto 400W Class-D (UcD400) against the ICEpower ICE500A.
After a visit in our companíes lab, where they have compared by themself on an AP System 2 the UcD400 & ICE500A, they were even more impressed.......
Indeed a SMPS is very nice, for shure when it is compatible for 90 to 250VAC. But a SMPS is FAR more expensive as a good build toroideltransformer. By a special construction our toroidel transformer is lighter and more powerful as a standard toroidel. So for the moment I don't want serve mine customers with a FAR much more complex powersupply as a SMPS definatly is.
The ICE250PS has a size of 5.9" x 6.4" and an overall eff. of 80%
Our Custom Made 200W has size of 3.15" x 3.94" + transformer size 3.94" x 1.92", and an overall eff. of 92%.
Weight of the both amps will be more of less the same.
For a manufacture our custom made amp will be highly attractive I guess.
If you are interested for a 1 tot 1 acoustically test?
Regards,
Jan-Peter
www.hypex.nl
self oscilating
Hi again!
I got inspiration from the UcD-thread. I did some simulations and... -OK, the beast was working... Also I was able to get a brief idea how it is working.... But to be honest: Theory of UcD is no fun.
Calculations ...OK, but Bruno (!!!) how do you manage to model your design in your brain??? I mean, if my brain cannot intuitively think what's going on, then debugging a real amp is getting tough!
But I like the idea of selfoscilation.
I think a simple hysteresis design is much easier to understand and
up to now I do not see much advantages of the UcD, except that the feedback is picked behind the output filter.
I am getting more and more a friend of a circuit as attached.
It is easier to understand and already gives impressive results in simulation. At 1kHz: THD=0.03 %, 3rd harmonic = 0.01%....
The attached file is showing a simplified full bridge. Well the distorsion of may not reflect perfectly the reality. But I think, if I manage a proper short dead time, then I will come close.
DC COMPONENT = -1.329206E-02
HARMONIC FREQUENCY FOURIER NORMALIZED PHASE NORMALIZED
NO (HZ) COMPONENT COMPONENT (DEG) PHASE (DEG)
1 1.000E+03 4.606E+00 1.000E+00 -6.769E+00 0.000E+00
2 2.000E+03 5.486E-04 1.191E-04 8.646E+01 1.000E+02
3 3.000E+03 4.604E-04 9.994E-05 1.127E+02 1.330E+02
4 4.000E+03 5.125E-04 1.113E-04 8.953E+01 1.166E+02
5 5.000E+03 7.015E-04 1.523E-04 5.169E+01 8.554E+01
6 6.000E+03 2.762E-04 5.996E-05 1.271E+02 1.677E+02
7 7.000E+03 6.567E-04 1.426E-04 2.739E+01 7.477E+01
8 8.000E+03 5.398E-04 1.172E-04 1.341E+02 1.883E+02
9 9.000E+03 8.001E-05 1.737E-05 1.038E+02 1.648E+02
10 1.000E+04 3.606E-04 7.830E-05 1.236E+02 1.913E+02
TOTAL HARMONIC DISTORTION = 3.222336E-02 PERCENT
In order to improve further, it might be possible to add a differential amp (giving gain and the possibility to pick differential feedback behind the filter) and settle an outer loop. For a subwoofer, the frequency compensation would be no bif deal. For full audio range the outer loop may require some brain work.
Charles already said "PID".... 🙄
Back to my basic question:
What is the main advantage of the UcD, except that already the simple structure picks the feedback from behind the filter?
Looking forward to your comments.
Bye
Markus
Hi again!
I got inspiration from the UcD-thread. I did some simulations and... -OK, the beast was working... Also I was able to get a brief idea how it is working.... But to be honest: Theory of UcD is no fun.
Calculations ...OK, but Bruno (!!!) how do you manage to model your design in your brain??? I mean, if my brain cannot intuitively think what's going on, then debugging a real amp is getting tough!
But I like the idea of selfoscilation.
I think a simple hysteresis design is much easier to understand and
up to now I do not see much advantages of the UcD, except that the feedback is picked behind the output filter.
I am getting more and more a friend of a circuit as attached.
It is easier to understand and already gives impressive results in simulation. At 1kHz: THD=0.03 %, 3rd harmonic = 0.01%....
The attached file is showing a simplified full bridge. Well the distorsion of may not reflect perfectly the reality. But I think, if I manage a proper short dead time, then I will come close.
DC COMPONENT = -1.329206E-02
HARMONIC FREQUENCY FOURIER NORMALIZED PHASE NORMALIZED
NO (HZ) COMPONENT COMPONENT (DEG) PHASE (DEG)
1 1.000E+03 4.606E+00 1.000E+00 -6.769E+00 0.000E+00
2 2.000E+03 5.486E-04 1.191E-04 8.646E+01 1.000E+02
3 3.000E+03 4.604E-04 9.994E-05 1.127E+02 1.330E+02
4 4.000E+03 5.125E-04 1.113E-04 8.953E+01 1.166E+02
5 5.000E+03 7.015E-04 1.523E-04 5.169E+01 8.554E+01
6 6.000E+03 2.762E-04 5.996E-05 1.271E+02 1.677E+02
7 7.000E+03 6.567E-04 1.426E-04 2.739E+01 7.477E+01
8 8.000E+03 5.398E-04 1.172E-04 1.341E+02 1.883E+02
9 9.000E+03 8.001E-05 1.737E-05 1.038E+02 1.648E+02
10 1.000E+04 3.606E-04 7.830E-05 1.236E+02 1.913E+02
TOTAL HARMONIC DISTORTION = 3.222336E-02 PERCENT
In order to improve further, it might be possible to add a differential amp (giving gain and the possibility to pick differential feedback behind the filter) and settle an outer loop. For a subwoofer, the frequency compensation would be no bif deal. For full audio range the outer loop may require some brain work.
Charles already said "PID".... 🙄
Back to my basic question:
What is the main advantage of the UcD, except that already the simple structure picks the feedback from behind the filter?
Looking forward to your comments.
Bye
Markus
Attachments
- Status
- Not open for further replies.