I've been toying with the idea of building an amplifier and from what I can tell Class D appears to be a cost effective way. (if I'm wrong correct me) The problem is that I know little if any of the theory or what is and isn't possible and at what expense and effort. Where do I start? Are there books you could recommend or proven kits? I want an attractive simple to use amplifier. I need a starting point and much of what I have read here reuires a little more than I know, not much more, but enough to confuse me. Oh yeah I thought that building an amp should be cost effective. Is this realistic?
Bill-of-material wise, class D and linear amplifiers cost pretty much the same over a wide range of power levels. What you save in heatsink and maybe power supply goes into the output inductors and the construction technique.
The bill of material is not going to determine how much money is going into your project though. DIY is not a good way to save money. For one, even if you had a perfectly flawless design to start, buying all the parts from hobby stores would be more expensive than the equivalent thing bought off the shelf (like you don't build a car out of spare parts and hope to save money). You can save some money over high-end gear by buying a finished module (Hypex and LC are the only ones that sell to DIY'ers if I'm not mistaken) and add a power supply.
From your question however I gather you really want to start from scratch. I must warn you that class D is about the worst thing to try (bar a lunar lander possibly) if you want to have a shot at a first-time-right project.
So why do DIY? For the fun of building, experimenting and learning. Not to save money. Certainly not to save time.
There are no books about class D amplifier design. Some information can be found in IEEE and AES papers but no recipes. When someone makes a good design, this is usually with the intention of making a product. As much as I'd love to, I can't see how I could publish a design for DIY'ers without disadvantaging myself (financially).
There's also much more detail information floating on the class D section of diyaudio than you think, but you have to sift through a lot of more practical discussions to find the interesting bits.
(I resist the temptation of naming the people who write the most interesting posts for fear of forgetting someone!)
By all means, join the club
The bill of material is not going to determine how much money is going into your project though. DIY is not a good way to save money. For one, even if you had a perfectly flawless design to start, buying all the parts from hobby stores would be more expensive than the equivalent thing bought off the shelf (like you don't build a car out of spare parts and hope to save money). You can save some money over high-end gear by buying a finished module (Hypex and LC are the only ones that sell to DIY'ers if I'm not mistaken) and add a power supply.
From your question however I gather you really want to start from scratch. I must warn you that class D is about the worst thing to try (bar a lunar lander possibly) if you want to have a shot at a first-time-right project.
So why do DIY? For the fun of building, experimenting and learning. Not to save money. Certainly not to save time.
There are no books about class D amplifier design. Some information can be found in IEEE and AES papers but no recipes. When someone makes a good design, this is usually with the intention of making a product. As much as I'd love to, I can't see how I could publish a design for DIY'ers without disadvantaging myself (financially).
There's also much more detail information floating on the class D section of diyaudio than you think, but you have to sift through a lot of more practical discussions to find the interesting bits.
(I resist the temptation of naming the people who write the most interesting posts for fear of forgetting someone!)
By all means, join the club
I've seen the modules from hypex, but what else do you need are there preamp modules out there as well? The obvious pluses of the building experience are clear I guess I kinda figured that by eliminating a lot of the gizmos that they pack in receivers these days there maybe a sound quality increase for equal money and since most of the amps I find are a little on the steep side I hoped to have found the middle ground. Complete from scratch would far exceed my capabilities. Thanks for the feedback.
If you don't want to use a separate preamp, the Hypex modules can be driven straight from a potentiometer. The input impedance is high enough to do that. This means that if you want a working solution with the least effort, you can order two 180 or 400 modules and a power supply from Hypex, buy a good 2x20k pot and wire the whole thing together in a box. This must be about the quickest diy integrated amp possible.flaevor said:
Dear Flaevor:
What Bruno may not be allowed to say is that building a power amp from Hypex' modules (even with "regular" versions) will be the most enjoyable an cheap way to end with an effortlest, balanced, natural and detailed sounding power amp.
You can base on Hypex' power supply diagram and pick the parts from RS-export of Farnell, I think. Main UCD threads are very long but almost every topic is touched there.
Personally, I'm also a beginner (ethernal, I must say) that built a four channel UCD180 amp that is driven directly from a Behringer active crossover (without pre) to form an "active" speaker system. The only negative point is that I have only -20dB attenuation: good for classical; party level on pop music
This basic untweaked DIY system has dynamics to die for.
Enjoy.
Mauricio
What Bruno may not be allowed to say is that building a power amp from Hypex' modules (even with "regular" versions) will be the most enjoyable an cheap way to end with an effortlest, balanced, natural and detailed sounding power amp.
You can base on Hypex' power supply diagram and pick the parts from RS-export of Farnell, I think. Main UCD threads are very long but almost every topic is touched there.
Personally, I'm also a beginner (ethernal, I must say) that built a four channel UCD180 amp that is driven directly from a Behringer active crossover (without pre) to form an "active" speaker system. The only negative point is that I have only -20dB attenuation: good for classical; party level on pop music
This basic untweaked DIY system has dynamics to die for.
Enjoy.
Mauricio
speaker sensitivity / efficiency
I have been following and learning from the forum threads dealing with modding and upgrading the SI. There has been much advice about the SI needing to be coupled with sensitive / efficient loudspeakers for best results.
My questions relate to how loudspeaker sensitivity / efficiency is measured. If I couple an amplifier to an un-enclosed loudspeaker chassis, the sound will be "thin" and not very "full". If I then mount that same speaker on an open baffle, or in a box for examples, the sound will immediately become "more extended and fuller". (Subjective terms I know). I might also say that I percieve the sound to be louder, - increased efficiency?- but this may be just that, perceived, and would not be bourne out by measurement.
Is the sensitivity / efficiency of a loudspeaker soley to do with the size of the magnet on the loudspeaker chasis and its construction materials, or is more to do with the type of enclosure in which it is housed? Maybe there is some synergy with loudspeaker chassis and enclosure?
My understanding is that a horn enclosure is likely to be more efficient than say an infinite baffle when using the same type of drive unit. Therefore it would seem that the type of enclosure does - can? - play a key part in efficiency. (Whether ones preference is for the sound of a horn or for the sound of an infinite baffle, is perhaps another question for another day!)
Would anyone like to advise please - or kindly refer me to the Forum where loudspeaker sensitivity / efficiency may already have been discussed?
CambshireGordon
I have been following and learning from the forum threads dealing with modding and upgrading the SI. There has been much advice about the SI needing to be coupled with sensitive / efficient loudspeakers for best results.
My questions relate to how loudspeaker sensitivity / efficiency is measured. If I couple an amplifier to an un-enclosed loudspeaker chassis, the sound will be "thin" and not very "full". If I then mount that same speaker on an open baffle, or in a box for examples, the sound will immediately become "more extended and fuller". (Subjective terms I know). I might also say that I percieve the sound to be louder, - increased efficiency?- but this may be just that, perceived, and would not be bourne out by measurement.
Is the sensitivity / efficiency of a loudspeaker soley to do with the size of the magnet on the loudspeaker chasis and its construction materials, or is more to do with the type of enclosure in which it is housed? Maybe there is some synergy with loudspeaker chassis and enclosure?
My understanding is that a horn enclosure is likely to be more efficient than say an infinite baffle when using the same type of drive unit. Therefore it would seem that the type of enclosure does - can? - play a key part in efficiency. (Whether ones preference is for the sound of a horn or for the sound of an infinite baffle, is perhaps another question for another day!)
Would anyone like to advise please - or kindly refer me to the Forum where loudspeaker sensitivity / efficiency may already have been discussed?
CambshireGordon
Although we're a little off-topic, there is logic in describing the sound from a single driver (without enclosure) as thin.
This is because the low frequencies are "shor circuited" accoustically, as the air from the front of the cone travels around to the back of the cone (in stead of being sent to your ears) when the frequency gets low enough.
I have verified the phenomenon with a 213 mm woofer unit (without enclosure):
The speed of sound is about 334 m/s.
The sound level was measured with a microphone to drop off below approx. 770 Hz.
The wavelength of this frequency is 334 / 770 = 0.434 m.
Half wavelength is 0.216 m or 216 mm, which fits nicely with the drivers diameter. (Fortunately, theory supports these measurements )
This is why your driver (without enclosure) won't play deep frequencies (audibly). The cone will move, but the sound presure doesn't get to your ears.
Putting the driver in a baffle increases the effective "diameter" in the example above, which would lower the "dropp off" frequency.
In a sealed enclosure, the sound will never make it to the back of the cone, and thus the speaker will (audibly) reproduce lower frequencies. At this point, your frequency range is limited by other (driver/enclosure) parameters.
Hope this helps you.
Jennice
This is because the low frequencies are "shor circuited" accoustically, as the air from the front of the cone travels around to the back of the cone (in stead of being sent to your ears) when the frequency gets low enough.
I have verified the phenomenon with a 213 mm woofer unit (without enclosure):
The speed of sound is about 334 m/s.
The sound level was measured with a microphone to drop off below approx. 770 Hz.
The wavelength of this frequency is 334 / 770 = 0.434 m.
Half wavelength is 0.216 m or 216 mm, which fits nicely with the drivers diameter. (Fortunately, theory supports these measurements
)This is why your driver (without enclosure) won't play deep frequencies (audibly). The cone will move, but the sound presure doesn't get to your ears.
Putting the driver in a baffle increases the effective "diameter" in the example above, which would lower the "dropp off" frequency.
In a sealed enclosure, the sound will never make it to the back of the cone, and thus the speaker will (audibly) reproduce lower frequencies. At this point, your frequency range is limited by other (driver/enclosure) parameters.
Hope this helps you.
Jennice
Rod Elliot has added a new article to his site about Class D amps.
http://sound.westhost.com/articles/pwm.htm
http://sound.westhost.com/articles/pwm.htm
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.