Hello,
I am a bass player, in the process of building a bass amplifier with class d modules (Zappulse 2.3SE). I have been following some of the class d threads here for some 3 months and I enjoy very much reading the various opinions and ideas of all those different people. I also, in my modest way, learn a lot of them, although I am far from a natural born technician. So far I have constructed a rather untidy (to say the least) kind of test version. I am rather ashamed to show it, considering some of the really beautiful designs of other people in other threads; the picture serves only to indicate in which stage and at what level I am presently. Sadly, it also demonstrates that I need a better camera
The sound of it, however, is beautiful; Lows and mid are deep and warm but fast responding and accurate; the highs and top-highs are crunchy, with only very simple and minimal equalization. It also has a nice, substantial and compact thump in the gut and a pleasant attack.
I feel rather presumptious to start a new topic just like that. My knowledge of electronics is kind of limited; I understand some of the behaviour of some common components in some simple circuits and I am aware of some of the topologies in class D and smps. So go learn, you would say. Well, I decided to accept my limitations and also accept the fact that I don’t have enough time (ok, maybe also energy…) to go as deep into it myself as would be neccessary to find real solutions. Still, I want to build that nice amplifier and there are a couple of aspects on building with the populare modules (like from Hypex, LCAudio, B&O…) that I cannot find specific information on that matches my level and/or purposes. My hope is that other people out there may be in the same line, or at least are willing to contribute with their knowledge on the issue(s) below:
SMPS. (Oh, no, not again…. yes, again, hear me out).
As I have understood, switched mode power supplies are much too noisy to be acceptable in a hifi configuration. Hiss, hum or crakle from the speakers can destroy any good music that needs some dynamics, and let’s forget about recording altogether. However, I am not primarily interested in extremely low noise-levels. Any bass-stack, whether it be an Ampeg, SWR, Eden or even Aguilar or whatever, will produce some significant hiss and hum when you turn up the volume and listen close. Maybe a few, more or less dedicated, studio amplifiers will be very quiet, but then, in many cases those are not very good in live situations. So I am completely willing to accept some steady noise, as long as it does not follow the dynamics in my playing. One should think that this would loosen some of the restrictions on an smps design. I mean, I sense a tendency that certain specifications are to be matched by default, while I (and I hope more people) would happily give them up, to a certain extend, in return for suffucient power (talking in the range of 1000-2000W) and “resistance” to the phenomenon of power supply pumping. So that is my aim: A switched mode power supply that is powerfull and suitable for steady bass amplification and that not neccessarily has to be high-end. It would make the difference between a 3 or 4 unit high, 50cm (20”) deep, 25 kg (50 pounds) 19” housing containing large toroids and a 2 unit high, 12” deep 15 to 20 pounds housing. This is very important when you have to carry around a 19” rack, possibly also containing a preamp, sound/effectsprocessor, BassPod or the likes.)
There have been some threads dealing with smps for class d: in the thread “SMPS design for power amps (Class D in particular)” ( http://www.diyaudio.com/forums/showthread.php?s=&threadid=41211 ), some very creative and skilled technicians (a.o. Bruno Putzeys, analogspiceman, ….) have been discussing SMPS design on a high theoretical level. The real content of their postings was way byond my comprehension, but very interesting at the same time. The thread “died” at the end of november 2004.
In the thread about Icpower modules for PA applications, the issue emerges as well on a somewhat more human level. It seems a very complicated matter, e.g.: because of the limit that the smps design seems to impose on the capacitance of the output caps, the supply is in danger in case power supply pumping occurs. There are some smps’s available, though, as many who have been around here may already know. One is the A and T labs smps ( http://www.a-and-t-labs.com/K6_Sw_Amp/index.htm ) of 1000W, designed to power some 2 x 300W or so class A/B amplifier. However, it’s design stems from 1995 and one may assume that new insights on the issue have evolved since then. Furthermore, it has small output caps and it seems to control only the output voltage rate and not the symmetry, which would make it defenseless against a pumping class D module. (thanks for the explanation, Ghemink, I hope I have understood it well). Other smps’s for poweramplifiers are available at http://www.cadaudio.com . They are designed to work well with their own cadaudio modules and with the Icpower ones, which are fullbridged and do not suffer from power supply pumping. Sadly, mainly for this reason they also are not sold seperately.
There MUST be some design possible that would match my (and, again, hopefully other’s) requirements. This is illustrated, for example, by the fact that Hevos, a Dutch manufacturer of high quality bassgear, will be introducing a bass top this spring that has a tubedriven preamp and a (or more) UCD module(s) from Hypex (not coincidentally the company that Bruno Putzeys works for….), all together in a 2 unit high lightweight 19” housing. I mailed them in the hope to get some information about the setup, but understandably they would not go into details, as they told me in their friendly reply.
So, concluding this too long story: that is what it’s about. A “practical” SMPS. I would be happy to see some postings containing info that could help me and possibly others. If not, ok, the thread will sink in the list and dissapear out of sight, big deal. But i think it’s woth a try. And if, let’s try to stick to the topic. I must admit that I myself posted some messages here and there in the past that were more or less off-topic, or not helpful or relevant at all. I’m still a little uncertain in my forum-manners (maybe demonstrated by the novel-length of this message..? Sorry…)
Besides the SMPS, any ideas on preamping the class D amp would also be highly appreciated. But the main issue will stay smps.
Though it’s high time to end this introduction, still a few words about myself:
I have been playing bass on (semi-) professional level from the late 70’s till the early 90’s. I played jazz, fusion, rock, blues and aside from that basically anything that would get me some mony, on fretted/unfretted electric bass and upright bass. Sometimes it was good for a living, sometimes not nearly close. I was a “studier”, practiced a lot and yet did not reach the level I had in mind. On the contrary, eventually I did not know anymore how to play a normal, functional bassline and in 1990 I decided to try and find a decent job, in which I succeeded. Presently, I am a database consultant (like many flipped musicians, philosophers, sociologists, teachers and the likes) and I like what I am doing. Recently, I started playing again, motivated by a small gig I had been asked for and I noticed how much I missed the feeling and how much a musician I am in my heart. I am also less blocked by ambition and I like searching for good basslines now, played with a good sound.
I am, and have always been, a great admiror of the late Jaco Pastorius.
Electronics: I am pretty comfortable as long as Ohm’s law and some other basic formulas are applicable in simple equations. Byond that, I tend to get heavily lost. My most creative idea so far was a power supply without transformer, i.e. using rectified and buffered mains AC without any galvanic isolation. In Holland, we use 230V rms, which would render some 320V DC on the rail. The Zappulse in my testamp uses +/- 80V (so, 160V). In some kind of symmetrical setup my supply could power 2 modules. After some rethinking I began to suspect that it would probably mean the surest one way ticked to heaven, where I could serve as a human 320V capacitor for Jaco’s bass rig…
Best regards, Joep
I am a bass player, in the process of building a bass amplifier with class d modules (Zappulse 2.3SE). I have been following some of the class d threads here for some 3 months and I enjoy very much reading the various opinions and ideas of all those different people. I also, in my modest way, learn a lot of them, although I am far from a natural born technician. So far I have constructed a rather untidy (to say the least) kind of test version. I am rather ashamed to show it, considering some of the really beautiful designs of other people in other threads; the picture serves only to indicate in which stage and at what level I am presently. Sadly, it also demonstrates that I need a better camera
The sound of it, however, is beautiful; Lows and mid are deep and warm but fast responding and accurate; the highs and top-highs are crunchy, with only very simple and minimal equalization. It also has a nice, substantial and compact thump in the gut and a pleasant attack.
I feel rather presumptious to start a new topic just like that. My knowledge of electronics is kind of limited; I understand some of the behaviour of some common components in some simple circuits and I am aware of some of the topologies in class D and smps. So go learn, you would say. Well, I decided to accept my limitations and also accept the fact that I don’t have enough time (ok, maybe also energy…) to go as deep into it myself as would be neccessary to find real solutions. Still, I want to build that nice amplifier and there are a couple of aspects on building with the populare modules (like from Hypex, LCAudio, B&O…) that I cannot find specific information on that matches my level and/or purposes. My hope is that other people out there may be in the same line, or at least are willing to contribute with their knowledge on the issue(s) below:
SMPS. (Oh, no, not again…. yes, again, hear me out).
As I have understood, switched mode power supplies are much too noisy to be acceptable in a hifi configuration. Hiss, hum or crakle from the speakers can destroy any good music that needs some dynamics, and let’s forget about recording altogether. However, I am not primarily interested in extremely low noise-levels. Any bass-stack, whether it be an Ampeg, SWR, Eden or even Aguilar or whatever, will produce some significant hiss and hum when you turn up the volume and listen close. Maybe a few, more or less dedicated, studio amplifiers will be very quiet, but then, in many cases those are not very good in live situations. So I am completely willing to accept some steady noise, as long as it does not follow the dynamics in my playing. One should think that this would loosen some of the restrictions on an smps design. I mean, I sense a tendency that certain specifications are to be matched by default, while I (and I hope more people) would happily give them up, to a certain extend, in return for suffucient power (talking in the range of 1000-2000W) and “resistance” to the phenomenon of power supply pumping. So that is my aim: A switched mode power supply that is powerfull and suitable for steady bass amplification and that not neccessarily has to be high-end. It would make the difference between a 3 or 4 unit high, 50cm (20”) deep, 25 kg (50 pounds) 19” housing containing large toroids and a 2 unit high, 12” deep 15 to 20 pounds housing. This is very important when you have to carry around a 19” rack, possibly also containing a preamp, sound/effectsprocessor, BassPod or the likes.)
There have been some threads dealing with smps for class d: in the thread “SMPS design for power amps (Class D in particular)” ( http://www.diyaudio.com/forums/showthread.php?s=&threadid=41211 ), some very creative and skilled technicians (a.o. Bruno Putzeys, analogspiceman, ….) have been discussing SMPS design on a high theoretical level. The real content of their postings was way byond my comprehension, but very interesting at the same time. The thread “died” at the end of november 2004.
In the thread about Icpower modules for PA applications, the issue emerges as well on a somewhat more human level. It seems a very complicated matter, e.g.: because of the limit that the smps design seems to impose on the capacitance of the output caps, the supply is in danger in case power supply pumping occurs. There are some smps’s available, though, as many who have been around here may already know. One is the A and T labs smps ( http://www.a-and-t-labs.com/K6_Sw_Amp/index.htm ) of 1000W, designed to power some 2 x 300W or so class A/B amplifier. However, it’s design stems from 1995 and one may assume that new insights on the issue have evolved since then. Furthermore, it has small output caps and it seems to control only the output voltage rate and not the symmetry, which would make it defenseless against a pumping class D module. (thanks for the explanation, Ghemink, I hope I have understood it well). Other smps’s for poweramplifiers are available at http://www.cadaudio.com . They are designed to work well with their own cadaudio modules and with the Icpower ones, which are fullbridged and do not suffer from power supply pumping. Sadly, mainly for this reason they also are not sold seperately.
There MUST be some design possible that would match my (and, again, hopefully other’s) requirements. This is illustrated, for example, by the fact that Hevos, a Dutch manufacturer of high quality bassgear, will be introducing a bass top this spring that has a tubedriven preamp and a (or more) UCD module(s) from Hypex (not coincidentally the company that Bruno Putzeys works for….), all together in a 2 unit high lightweight 19” housing. I mailed them in the hope to get some information about the setup, but understandably they would not go into details, as they told me in their friendly reply.
So, concluding this too long story: that is what it’s about. A “practical” SMPS. I would be happy to see some postings containing info that could help me and possibly others. If not, ok, the thread will sink in the list and dissapear out of sight, big deal. But i think it’s woth a try. And if, let’s try to stick to the topic. I must admit that I myself posted some messages here and there in the past that were more or less off-topic, or not helpful or relevant at all. I’m still a little uncertain in my forum-manners (maybe demonstrated by the novel-length of this message..? Sorry…)
Besides the SMPS, any ideas on preamping the class D amp would also be highly appreciated. But the main issue will stay smps.
Though it’s high time to end this introduction, still a few words about myself:
I have been playing bass on (semi-) professional level from the late 70’s till the early 90’s. I played jazz, fusion, rock, blues and aside from that basically anything that would get me some mony, on fretted/unfretted electric bass and upright bass. Sometimes it was good for a living, sometimes not nearly close. I was a “studier”, practiced a lot and yet did not reach the level I had in mind. On the contrary, eventually I did not know anymore how to play a normal, functional bassline and in 1990 I decided to try and find a decent job, in which I succeeded. Presently, I am a database consultant (like many flipped musicians, philosophers, sociologists, teachers and the likes) and I like what I am doing. Recently, I started playing again, motivated by a small gig I had been asked for and I noticed how much I missed the feeling and how much a musician I am in my heart. I am also less blocked by ambition and I like searching for good basslines now, played with a good sound.
I am, and have always been, a great admiror of the late Jaco Pastorius.
Electronics: I am pretty comfortable as long as Ohm’s law and some other basic formulas are applicable in simple equations. Byond that, I tend to get heavily lost. My most creative idea so far was a power supply without transformer, i.e. using rectified and buffered mains AC without any galvanic isolation. In Holland, we use 230V rms, which would render some 320V DC on the rail. The Zappulse in my testamp uses +/- 80V (so, 160V). In some kind of symmetrical setup my supply could power 2 modules. After some rethinking I began to suspect that it would probably mean the surest one way ticked to heaven, where I could serve as a human 320V capacitor for Jaco’s bass rig…
Best regards, Joep
Attachments
Joep Zonnebloem said:Hello,
I am a bass player, in the process of building a bass amplifier with class d modules (Zappulse 2.3SE). I have been following some of the class d threads here for some 3 months and I enjoy very much reading the various opinions and ideas of all those different people. I also, in my modest way, learn a lot of them, although I am far from a natural born technician. So far I have constructed a rather untidy (to say the least) kind of test version. I am rather ashamed to show it, considering some of the really beautiful designs of other people in other threads; the picture serves only to indicate in which stage and at what level I am presently. Sadly, it also demonstrates that I need a better camera
The sound of it, however, is beautiful; Lows and mid are deep and warm but fast responding and accurate; the highs and top-highs are crunchy, with only very simple and minimal equalization. It also has a nice, substantial and compact thump in the gut and a pleasant attack.
I feel rather presumptious to start a new topic just like that. My knowledge of electronics is kind of limited; I understand some of the behaviour of some common components in some simple circuits and I am aware of some of the topologies in class D and smps. So go learn, you would say. Well, I decided to accept my limitations and also accept the fact that I don’t have enough time (ok, maybe also energy�E to go as deep into it myself as would be neccessary to find real solutions. Still, I want to build that nice amplifier and there are a couple of aspects on building with the populare modules (like from Hypex, LCAudio, B&O�E that I cannot find specific information on that matches my level and/or purposes. My hope is that other people out there may be in the same line, or at least are willing to contribute with their knowledge on the issue(s) below:
SMPS. (Oh, no, not again�E yes, again, hear me out).
As I have understood, switched mode power supplies are much too noisy to be acceptable in a hifi configuration. Hiss, hum or crakle from the speakers can destroy any good music that needs some dynamics, and let’s forget about recording altogether. However, I am not primarily interested in extremely low noise-levels. Any bass-stack, whether it be an Ampeg, SWR, Eden or even Aguilar or whatever, will produce some significant hiss and hum when you turn up the volume and listen close. Maybe a few, more or less dedicated, studio amplifiers will be very quiet, but then, in many cases those are not very good in live situations. So I am completely willing to accept some steady noise, as long as it does not follow the dynamics in my playing. One should think that this would loosen some of the restrictions on an smps design. I mean, I sense a tendency that certain specifications are to be matched by default, while I (and I hope more people) would happily give them up, to a certain extend, in return for suffucient power (talking in the range of 1000-2000W) and “resistance�Eto the phenomenon of power supply pumping. So that is my aim: A switched mode power supply that is powerfull and suitable for steady bass amplification and that not neccessarily has to be high-end. It would make the difference between a 3 or 4 unit high, 50cm (20�E deep, 25 kg (50 pounds) 19�Ehousing containing large toroids and a 2 unit high, 12�Edeep 15 to 20 pounds housing. This is very important when you have to carry around a 19�Erack, possibly also containing a preamp, sound/effectsprocessor, BassPod or the likes.)
There have been some threads dealing with smps for class d: in the thread “SMPS design for power amps (Class D in particular)�E( http://www.diyaudio.com/forums/showthread.php?s=&threadid=41211 ), some very creative and skilled technicians (a.o. Bruno Putzeys, analogspiceman, �E) have been discussing SMPS design on a high theoretical level. The real content of their postings was way byond my comprehension, but very interesting at the same time. The thread “died�Eat the end of november 2004.
In the thread about Icpower modules for PA applications, the issue emerges as well on a somewhat more human level. It seems a very complicated matter, e.g.: because of the limit that the smps design seems to impose on the capacitance of the output caps, the supply is in danger in case power supply pumping occurs. There are some smps’s available, though, as many who have been around here may already know. One is the A and T labs smps ( http://www.a-and-t-labs.com/K6_Sw_Amp/index.htm ) of 1000W, designed to power some 2 x 300W or so class A/B amplifier. However, it’s design stems from 1995 and one may assume that new insights on the issue have evolved since then. Furthermore, it has small output caps and it seems to control only the output voltage rate and not the symmetry, which would make it defenseless against a pumping class D module. (thanks for the explanation, Ghemink, I hope I have understood it well). Other smps’s for poweramplifiers are available at http://www.cadaudio.com . They are designed to work well with their own cadaudio modules and with the Icpower ones, which are fullbridged and do not suffer from power supply pumping. Sadly, mainly for this reason they also are not sold seperately.
There MUST be some design possible that would match my (and, again, hopefully other’s) requirements. This is illustrated, for example, by the fact that Hevos, a Dutch manufacturer of high quality bassgear, will be introducing a bass top this spring that has a tubedriven preamp and a (or more) UCD module(s) from Hypex (not coincidentally the company that Bruno Putzeys works for�E), all together in a 2 unit high lightweight 19�Ehousing. I mailed them in the hope to get some information about the setup, but understandably they would not go into details, as they told me in their friendly reply.
So, concluding this too long story: that is what it’s about. A “practical�ESMPS. I would be happy to see some postings containing info that could help me and possibly others. If not, ok, the thread will sink in the list and dissapear out of sight, big deal. But i think it’s woth a try. And if, let’s try to stick to the topic. I must admit that I myself posted some messages here and there in the past that were more or less off-topic, or not helpful or relevant at all. I’m still a little uncertain in my forum-manners (maybe demonstrated by the novel-length of this message..? Sorry�E
Besides the SMPS, any ideas on preamping the class D amp would also be highly appreciated. But the main issue will stay smps.
Though it’s high time to end this introduction, still a few words about myself:
I have been playing bass on (semi-) professional level from the late 70’s till the early 90’s. I played jazz, fusion, rock, blues and aside from that basically anything that would get me some mony, on fretted/unfretted electric bass and upright bass. Sometimes it was good for a living, sometimes not nearly close. I was a “studier�E practiced a lot and yet did not reach the level I had in mind. On the contrary, eventually I did not know anymore how to play a normal, functional bassline and in 1990 I decided to try and find a decent job, in which I succeeded. Presently, I am a database consultant (like many flipped musicians, philosophers, sociologists, teachers and the likes) and I like what I am doing. Recently, I started playing again, motivated by a small gig I had been asked for and I noticed how much I missed the feeling and how much a musician I am in my heart. I am also less blocked by ambition and I like searching for good basslines now, played with a good sound.
I am, and have always been, a great admiror of the late Jaco Pastorius.
Electronics: I am pretty comfortable as long as Ohm’s law and some other basic formulas are applicable in simple equations. Byond that, I tend to get heavily lost. My most creative idea so far was a power supply without transformer, i.e. using rectified and buffered mains AC without any galvanic isolation. In Holland, we use 230V rms, which would render some 320V DC on the rail. The Zappulse in my testamp uses +/- 80V (so, 160V). In some kind of symmetrical setup my supply could power 2 modules. After some rethinking I began to suspect that it would probably mean the surest one way ticked to heaven, where I could serve as a human 320V capacitor for Jaco’s bass rig�E
Best regards, Joep
Hello Joep,
That was a long post, but interesting and quite well summarized. However, never, never try to power your amp directly from the AC mains. You will really electrocute yourself or others. Always use transformers between the mains and the power for the amps. Fortunately, you came to that conclusion yourself, this is really killing.
I think the best way to avoid problems with an SMPS is still to go for a bridged amp. I know, that increases the cost of the project, but you could buy almost two UcD400 modules for that 1 ZAPpulse that you have there. So hardly increased cost and guaranteed no power supply pumping.
I still don`t have my SMPS, it is delayed a bit because of not all components being in stock. After that I still have to build it in the small number of evening hours I have (my day-job takes a lot of my time
, no I`m not a database consultant, I develop Flash memories)Anyway, have fun with your project, have fun making music and stay alive and away from the mains without any transformers.
Best regards
Gertjan
Hi Joep,
A power factor corrected switched mode power supply is almost just like a low performance class d amplifier running in reverse. Kind of like a bad sub amp really - it only needs to poorly produce one low frequency (that of the net at a few percent distortion). That, plus it needs to provide galvanic isolation. If a class d amp can do this quietly while handling the music power peaks, a power supply that can do the same should be only a little bit larger and more expensive the class d amp (because it needs nearly an equivalent amount of stuff plus a high frequency isolation transformer).
There simply has not been a large enough market to support the slightly different, but not more difficult design-wise needs of a SMPS for class d amps. The recent flood of high performance, low cost class d amps will soon change this situation, IMO.
What's different about SMPS for class d?
Like the d amps themselves, SMPS's for class d need a high peak to average capability. IMO, they also should have power factor correction with a lot of energy storage (needed to suppress audio line flicker). And they should be designed for very low common mode output noise. This last performance feature would not be especially difficult to do, but it usually not required by the typical SMPS customer so it is not provided.
Regards -- analogspiceman
PS: The unisolated Euro mains can easily be lethal and is much too dangerous for the average DIYer to be fooling around with.
Re: Re: Re: Small BIG power for not so high-end class D diy’s
Hi Gertjan,
I already have experienced some interesting phenomena with Mains AC the last few weeks, like 4 resistors expelling very rapidly 4 very neat and identical little feathers of smoke, in a split second, completely synchronously. That was a fascinating sight. On another occasion I found, also in a split second, one pen of my multimeter welded to a copperwire after a very loud “pang” that made my ears hiss for some time. You can imagine I have gotten a little (lot) more cautious now.. Thanks for the concern, anyway
As for the power supply pumping: how bad will it be? Who has some experience with this? I don’t read a lot on it, apart from the theoretical possibility. I have not been able yet to test my amp at its full power. I could try it in my living room, but, depending on the mood my neighbour is in, the effect on my health could eventually be as disastrous as fooling around with AC mains carelessly . So I bought 5 100W lightbulbs with fittings and I am going to connect them in parallel and then the lot in series with a little 8 Ohm speaker (to give me some acoustic feedback) to my output and then try to apply maximum power (in steps, using a regulator) I realize it’s not a perfect dummyload, but it may serve my purpose, i.e. investigating the tendency to power supply pumping with a +/- 80V supply at full power in 4 Ohms. I will report here on this little test.
Success in assembling the PS! I’m very curious for the eventual result.
Best regards, Joep
Hi analogspiceman,
I feel honoured by your posting. As I said earlier, I have been reading your discussions with a.o. Bruno and admire the ease with which you people are able to understand and analyse these complex matters. It’s encouraging to read your comment about the expected availability of suitable SMPS in the future.
Do you mean with “high peak to average capability” that the supply should be able to deliver high peaks, above its normal operational current? And with “a lot of energy storage (needed to suppress audio line flicker)” that it should have large caps on the output to flatten the switching noise? Then I’m starting to get the picture. Sorry for my ignorance, I’m still learning…
As for the AC mains adventures, see my reply to Ghemink above. In the mean time I have learned to always check and double check whether the apparatus is disconnected from the mains and will always discharge the caps with a lightbulb before touching anything. My proposed powersupply was only mentioned to illustrate my even greater ignorance some time ago…
Best regards, Joep
PS, of course I read your correction on the peak to average capability...
ghemink said:
Hello Joep,
That was a long post, but interesting and quite well summarized. However, never, never try to power your amp directly from the AC mains. You will really electrocute yourself or others. Always use transformers between the mains and the power for the amps. Fortunately, you came to that conclusion yourself, this is really killing.
I think the best way to avoid problems with an SMPS is still to go for a bridged amp. I know, that increases the cost of the project, but you could buy almost two UcD400 modules for that 1 ZAPpulse that you have there. So hardly increased cost and guaranteed no power supply pumping.
I still don`t have my SMPS, it is delayed a bit because of not all components being in stock. After that I still have to build it in the small number of evening hours I have (my day-job takes a lot of my time
Anyway, have fun with your project, have fun making music and stay alive and away from the mains without any transformers.
Best regards
Gertjan
Hi Gertjan,
I already have experienced some interesting phenomena with Mains AC the last few weeks, like 4 resistors expelling very rapidly 4 very neat and identical little feathers of smoke, in a split second, completely synchronously. That was a fascinating sight. On another occasion I found, also in a split second, one pen of my multimeter welded to a copperwire after a very loud “pang” that made my ears hiss for some time. You can imagine I have gotten a little (lot) more cautious now.. Thanks for the concern, anyway
As for the power supply pumping: how bad will it be? Who has some experience with this? I don’t read a lot on it, apart from the theoretical possibility. I have not been able yet to test my amp at its full power. I could try it in my living room, but, depending on the mood my neighbour is in, the effect on my health could eventually be as disastrous as fooling around with AC mains carelessly
. So I bought 5 100W lightbulbs with fittings and I am going to connect them in parallel and then the lot in series with a little 8 Ohm speaker (to give me some acoustic feedback) to my output and then try to apply maximum power (in steps, using a regulator) I realize it’s not a perfect dummyload, but it may serve my purpose, i.e. investigating the tendency to power supply pumping with a +/- 80V supply at full power in 4 Ohms. I will report here on this little test. Success in assembling the PS! I’m very curious for the eventual result.
Best regards, Joep
analogspiceman said:
Hi analogspiceman,
I feel honoured by your posting. As I said earlier, I have been reading your discussions with a.o. Bruno and admire the ease with which you people are able to understand and analyse these complex matters. It’s encouraging to read your comment about the expected availability of suitable SMPS in the future.
Do you mean with “high peak to average capability” that the supply should be able to deliver high peaks, above its normal operational current? And with “a lot of energy storage (needed to suppress audio line flicker)” that it should have large caps on the output to flatten the switching noise? Then I’m starting to get the picture. Sorry for my ignorance, I’m still learning…
As for the AC mains adventures, see my reply to Ghemink above. In the mean time I have learned to always check and double check whether the apparatus is disconnected from the mains and will always discharge the caps with a lightbulb before touching anything. My proposed powersupply was only mentioned to illustrate my even greater ignorance some time ago…
Best regards, Joep
PS, of course I read your correction on the peak to average capability...
Re: Re: Re: Re: Small BIG power for not so high-end class D diy’s
Hm... On second thought I'd better give this concept some more thought.... It may kill my little testspeaker
@Analogspiceman: Are you still developping (theoretically or practically) any SMPS for audio, like you did with Bruno at the time? In that case, do you already have some sort of design that could be tested? I would be very happy to do that! (being VERY carefull on the mains, of course )
Best regards, Joep
Joep Zonnebloem said:
Hm... On second thought I'd better give this concept some more thought.... It may kill my little testspeaker
@Analogspiceman: Are you still developping (theoretically or practically) any SMPS for audio, like you did with Bruno at the time? In that case, do you already have some sort of design that could be tested? I would be very happy to do that! (being VERY carefull on the mains, of course
)Best regards, Joep
Joep,
You think your test-setup is ugly?
Behold my test setup... My girlfriend is lovingly reffering to it as "Het Gedrocht"
Essentially it's a hacked old integrated amp. Just need to put the speaker terminals on the back plate so I can close it up. All electronics in the amp are disconnected, and can be repaired when I move to the real project (budget, budget...)
Yes, the top lid of the icecream box is on the part list...
You think your test-setup is ugly?
Behold my test setup... My girlfriend is lovingly reffering to it as "Het Gedrocht"
Essentially it's a hacked old integrated amp. Just need to put the speaker terminals on the back plate so I can close it up. All electronics in the amp are disconnected, and can be repaired when I move to the real project (budget, budget...)
Yes, the top lid of the icecream box is on the part list...
Attachments
Re: Re: Re: Re: Re: Small BIG power for not so high-end class D diy’s
Hello Joep,
You should put the 8Ohm speaker in parallel with the lightbulbs and connect a resistance in series with the little speaker if you want to protect it.
Joep Zonnebloem said:
Hm... On second thought I'd better give this concept some more thought.... It may kill my little testspeaker
@Analogspiceman: Are you still developping (theoretically or practically) any SMPS for audio, like you did with Bruno at the time? In that case, do you already have some sort of design that could be tested? I would be very happy to do that! (being VERY carefull on the mains, of course
Best regards, Joep
Hello Joep,
You should put the 8Ohm speaker in parallel with the lightbulbs and connect a resistance in series with the little speaker if you want to protect it.
Re: Re: Re: Re: Small BIG power for not so high-end class D diy’s
So powersupply pumping would be worst at low frequencies and impedances. So a 20Hz signal at high power maybe the worst test signal.
I have done some simulations to see what happens, forgot about the results, but you can see it clearly happen at 20Hz depending on capacitor size.
Gertjan
Joep Zonnebloem said:
Hi Gertjan,
I already have experienced some interesting phenomena with Mains AC the last few weeks, like 4 resistors expelling very rapidly 4 very neat and identical little feathers of smoke, in a split second, completely synchronously. That was a fascinating sight. On another occasion I found, also in a split second, one pen of my multimeter welded to a copperwire after a very loud “pang�Ethat made my ears hiss for some time. You can imagine I have gotten a little (lot) more cautious now.. Thanks for the concern, anyway
As for the power supply pumping: how bad will it be? Who has some experience with this? I don’t read a lot on it, apart from the theoretical possibility. I have not been able yet to test my amp at its full power. I could try it in my living room, but, depending on the mood my neighbour is in, the effect on my health could eventually be as disastrous as fooling around with AC mains carelessly
Success in assembling the PS! I’m very curious for the eventual result.
Best regards, Joep
Hi analogspiceman,
I feel honoured by your posting. As I said earlier, I have been reading your discussions with a.o. Bruno and admire the ease with which you people are able to understand and analyse these complex matters. It’s encouraging to read your comment about the expected availability of suitable SMPS in the future.
Do you mean with “high peak to average capability�Ethat the supply should be able to deliver high peaks, above its normal operational current? And with “a lot of energy storage (needed to suppress audio line flicker)�Ethat it should have large caps on the output to flatten the switching noise? Then I’m starting to get the picture. Sorry for my ignorance, I’m still learning�E
As for the AC mains adventures, see my reply to Ghemink above. In the mean time I have learned to always check and double check whether the apparatus is disconnected from the mains and will always discharge the caps with a lightbulb before touching anything. My proposed powersupply was only mentioned to illustrate my even greater ignorance some time ago�E
Best regards, Joep
PS, of course I read your correction on the peak to average capability...
So powersupply pumping would be worst at low frequencies and impedances. So a 20Hz signal at high power maybe the worst test signal.
I have done some simulations to see what happens, forgot about the results, but you can see it clearly happen at 20Hz depending on capacitor size.
Gertjan
Re: Re: Re: Re: Re: Re: Small BIG power for not so high-end class D diy’s
Hi GertJan,
That's what I had figured out in the meantime... Only, to simulate a load of say 600W, I need a lot of bulbs. One 100W bulb has a resistance of around 50 Ohm, so I would need some 12 bulbs in parallel to reach around 4 ohms. And their resistance might vary with their temperature. I am now looking for an alternative. There is a circuit on the Dutch site circuitsonline.net ( http://www.circuitsonline.net/circuits/view/90 ) for a dummyload up to 100W that would simulate speaker behaviour. Maybe you understand some Dutch (your name points in that direction ), so maybe you could take a look. But for 600W. IMO it would need modifying the value of R1 (with multiple resistors in parallel, I guess) which would make it not particularly cheap... Aside from that, I cannot tell whether this modification would influence the speaker-like behaviour of the circuit. If so, it would be a waste to go that way, because I am not capable (yet) of redesigning the whole thing with inductances etc.
I am now wandering around my house with my multimeter trying to find some domestic gear ( irons, percolators, heaters ...) whith the right impedance and sufficient power consumption. Any suggestions maybe?
The lowest note (low E) on a regular bass is around 43 HZ. A fivestring bass (which I own but practically never use) has a low B of 31 HZ. Maybe, when you play massive chords (which I rarely do), "difference-ferquencies" (I don't know if this is the correct English word for it) could amount to significantly lower values, but then I doubt if they would be very prominent. ( EDIT Correction! They can be, if all strings are strummed together in some cases) I would think 20HZ is a typically (home-)theater surround sound frequency. So anyway, i think it's worth to try testing it.
Thanks for the input! Regards, Joep
ghemink said:
Hi GertJan,
That's what I had figured out in the meantime... Only, to simulate a load of say 600W, I need a lot of bulbs. One 100W bulb has a resistance of around 50 Ohm, so I would need some 12 bulbs in parallel to reach around 4 ohms. And their resistance might vary with their temperature. I am now looking for an alternative. There is a circuit on the Dutch site circuitsonline.net ( http://www.circuitsonline.net/circuits/view/90 ) for a dummyload up to 100W that would simulate speaker behaviour. Maybe you understand some Dutch (your name points in that direction
), so maybe you could take a look. But for 600W. IMO it would need modifying the value of R1 (with multiple resistors in parallel, I guess) which would make it not particularly cheap... Aside from that, I cannot tell whether this modification would influence the speaker-like behaviour of the circuit. If so, it would be a waste to go that way, because I am not capable (yet) of redesigning the whole thing with inductances etc.I am now wandering around my house with my multimeter trying to find some domestic gear ( irons, percolators, heaters ...) whith the right impedance and sufficient power consumption. Any suggestions maybe?
The lowest note (low E) on a regular bass is around 43 HZ. A fivestring bass (which I own but practically never use) has a low B of 31 HZ. Maybe, when you play massive chords (which I rarely do), "difference-ferquencies" (I don't know if this is the correct English word for it) could amount to significantly lower values, but then I doubt if they would be very prominent. ( EDIT Correction! They can be, if all strings are strummed together in some cases) I would think 20HZ is a typically (home-)theater surround sound frequency. So anyway, i think it's worth to try testing it.
Thanks for the input! Regards, Joep
Re: Re: Re: Re: Re: Re: Re: Small BIG power for not so high-end class D diy’s
Yes, I'm from Holland, just happen to live and work in Japan. Dus we kunnen gewoon in het NLs mailen.
Those light bulbs seem to be quite interesting loads to me. They are cheap, give visual feedback and you can put any number in parallel to get any load that you want. Plan to use them for power supply testing as well. I have made some dummy loads with resistors, but you need many resistors to share the power, I will try lamps. I would not bother too much to try to simulate a load that resembles a speaker, every speaker is different anyway.
Groeten
Gertjan
Joep Zonnebloem said:
Hi GertJan,
That's what I had figured out in the meantime... Only, to simulate a load of say 600W, I need a lot of bulbs. One 100W bulb has a resistance of around 50 Ohm, so I would need some 12 bulbs in parallel to reach around 4 ohms. And their resistance might vary with their temperature. I am now looking for an alternative. There is a circuit on the Dutch site circuitsonline.net ( http://www.circuitsonline.net/circuits/view/90 ) for a dummyload up to 100W that would simulate speaker behaviour. Maybe you understand some Dutch (your name points in that direction
I am now wandering around my house with my multimeter trying to find some domestic gear ( irons, percolators, heaters ...) whith the right impedance and sufficient power consumption. Any suggestions maybe?
The lowest note (low E) on a regular bass is around 43 HZ. A fivestring bass (which I own but practically never use) has a low B of 31 HZ. Maybe, when you play massive chords (which I rarely do), "difference-ferquencies" (I don't know if this is the correct English word for it) could amount to significantly lower values, but then I doubt if they would be very prominent. ( EDIT Correction! They can be, if all strings are strummed together in some cases) I would think 20HZ is a typically (home-)theater surround sound frequency. So anyway, i think it's worth to try testing it.
Thanks for the input! Regards, Joep
Yes, I'm from Holland, just happen to live and work in Japan. Dus we kunnen gewoon in het NLs mailen.
Those light bulbs seem to be quite interesting loads to me. They are cheap, give visual feedback and you can put any number in parallel to get any load that you want. Plan to use them for power supply testing as well. I have made some dummy loads with resistors, but you need many resistors to share the power, I will try lamps. I would not bother too much to try to simulate a load that resembles a speaker, every speaker is different anyway.
Groeten
Gertjan
It may be worth Googling up the Sony TA-N88(B) power amp.
This is perhaps the grand-daddy of all decent class D amps, at 170WPC and 0.1% distortion.
It uses a SMPS for power.
And, it was first demoed in 1977, production started in 1978.
It uses VFETs in the output stage (2 P-N pairs per channel). These are NOT MOSFETs but power JFETs (therefore depletion devices), 200V/10A, VERY fast with only about 180pF gate capacitance, but they need about -50V to turn them off.
It is also a fully discrete design except for a high gain chip amp speciffically made for it wired as a 'compare with zero' comparator (today this could easily be replaced by a half-decent current feedback opamp or even one of the new ultrafast voltage feedback amps, or a fast comparator).
Interestingly, it uses a 4th order output filter, with air cored inductors.
I have the amp, and it's schematic - learned lots from it.
This is perhaps the grand-daddy of all decent class D amps, at 170WPC and 0.1% distortion.
It uses a SMPS for power.
And, it was first demoed in 1977, production started in 1978.
It uses VFETs in the output stage (2 P-N pairs per channel). These are NOT MOSFETs but power JFETs (therefore depletion devices), 200V/10A, VERY fast with only about 180pF gate capacitance, but they need about -50V to turn them off.
It is also a fully discrete design except for a high gain chip amp speciffically made for it wired as a 'compare with zero' comparator (today this could easily be replaced by a half-decent current feedback opamp or even one of the new ultrafast voltage feedback amps, or a fast comparator).
Interestingly, it uses a 4th order output filter, with air cored inductors.
I have the amp, and it's schematic - learned lots from it.
The real great-grandaddy of switching power amps was the Infinity power amp circa 1974. It was supposedly rated at 250W per channel, and used Solitron milspec bipolar output devices. It also used a switching power supply, which was its downfall. There was not enough capacitance on the output, so the amp had serious problems with pumping for sustained high power bass tones. This is the beast that got me interested in switching power supplies and amplifiers.
It would be a good idea to use simple iron in the power supply at first to get all the amplifier issues ironed out. Trying to develop both a switching amp and power supply at the same go is rather a large plate full. Also, disasters involving mains operated switching supplies tend to be pretty destructive, especially when there is a PFC involved, not to mention the safety issues. You will need to pay a lot of attention to EMI control, unless you want to be carrying around a broadband RF noise generator masquerading as a power amplifer.
The big problem with using a switching power supply to power an audio amp is the large dynamic range of the load current. Switching power supplies like to see a fairly steady load, and can lose regulation at light load unless you know the tricks to get around those problems. Also, switching power supplies will generally have a fairly limited bandwidth and can have trouble slewing fast enough to follow sudden load changes. Unless you have special equipment, it will be difficult to even measure the loop bandwidth and phase margin for a switcher.
The best approach may be an open loop type power supply that programs the duty cycle as a function of the input voltage. This will get you approximately the right output voltage and also enable ripple cancellation to get rid of the mains frequency ripple introduced by the PFC. Loop stability will not be an issue. Since you will be using large output capacitance to get around the bass pumping problem, you will have a fairly large energy reservoir on the output to take care of load transients. Use of swinging chokes in the output filters will allow one to regulate at low output current. What you would essentially have is a high frequency equivalent to a big chunk of iron. One more thing - if you are serious about developing a switching power supply, get your hands on a current probe. Otherwise, you'll be completely blind to a lot of the problems that can crop up.
There are successul professional amplifier designs out in the field that have solved the problems involved with using switching power supplies. QSC uses a switching power supply for their new professional power amplifiers.
It would be a good idea to use simple iron in the power supply at first to get all the amplifier issues ironed out. Trying to develop both a switching amp and power supply at the same go is rather a large plate full. Also, disasters involving mains operated switching supplies tend to be pretty destructive, especially when there is a PFC involved, not to mention the safety issues. You will need to pay a lot of attention to EMI control, unless you want to be carrying around a broadband RF noise generator masquerading as a power amplifer.
The big problem with using a switching power supply to power an audio amp is the large dynamic range of the load current. Switching power supplies like to see a fairly steady load, and can lose regulation at light load unless you know the tricks to get around those problems. Also, switching power supplies will generally have a fairly limited bandwidth and can have trouble slewing fast enough to follow sudden load changes. Unless you have special equipment, it will be difficult to even measure the loop bandwidth and phase margin for a switcher.
The best approach may be an open loop type power supply that programs the duty cycle as a function of the input voltage. This will get you approximately the right output voltage and also enable ripple cancellation to get rid of the mains frequency ripple introduced by the PFC. Loop stability will not be an issue. Since you will be using large output capacitance to get around the bass pumping problem, you will have a fairly large energy reservoir on the output to take care of load transients. Use of swinging chokes in the output filters will allow one to regulate at low output current. What you would essentially have is a high frequency equivalent to a big chunk of iron. One more thing - if you are serious about developing a switching power supply, get your hands on a current probe. Otherwise, you'll be completely blind to a lot of the problems that can crop up.
There are successul professional amplifier designs out in the field that have solved the problems involved with using switching power supplies. QSC uses a switching power supply for their new professional power amplifiers.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.