In keeping with the whole low power Tripath amp craze, I decided to design some modules that can compete with the Tripath amps but have a few key differences that make them more DIY friendly.
This is not a discrete amp. It uses a chip, however the chip comes in an SOIC8 package which is much easier to solder than the 36 pin surface mount Tripath jobbies. Since this is a single ended amp it requires +/- rails. This particular chip uses up to 12V rails. Being single ended with split supply rails also gives it the advantage of not needing any input or output coupling caps. The amp is self oscillating, which is why the chip comes in such a small package, but it does not use quite the same scheme as UCD.
The number of parts needed to build a working amp of this type is quite low. I've added some extra parts to the basic design so the modules you see in the pictures could probably be reduced to half the size if I did away with the stuff I added.
The on-board supply section is similar to what was implemented on the Ref-T. There are two 680uF Panasonic FM caps per rail, with an optional ferrite bead in between for filtering. It might sound better without it but I have yet to try that.
There are indicator LEDs for each power rail to show that each rail is powered up. They're also there to supply a constant load for the rails and to drain the supply caps at a constant current when the power is turned off. Right now I have them configured to draw about 5mA and the LEDs are still really bright, so I may decrease that a mA or two!
I used the same toroid inductor that comes with the 41Hz amp3 and amp6 kits because it's what I have available and I can wind them in my sleep by now (I still hate doing it though...).
There is no turn on thump with this circuit but there is a small turn off pop that's nothing to worry about.
This is not a discrete amp. It uses a chip, however the chip comes in an SOIC8 package which is much easier to solder than the 36 pin surface mount Tripath jobbies. Since this is a single ended amp it requires +/- rails. This particular chip uses up to 12V rails. Being single ended with split supply rails also gives it the advantage of not needing any input or output coupling caps. The amp is self oscillating, which is why the chip comes in such a small package, but it does not use quite the same scheme as UCD.
The number of parts needed to build a working amp of this type is quite low. I've added some extra parts to the basic design so the modules you see in the pictures could probably be reduced to half the size if I did away with the stuff I added.
The on-board supply section is similar to what was implemented on the Ref-T. There are two 680uF Panasonic FM caps per rail, with an optional ferrite bead in between for filtering. It might sound better without it but I have yet to try that.
There are indicator LEDs for each power rail to show that each rail is powered up. They're also there to supply a constant load for the rails and to drain the supply caps at a constant current when the power is turned off. Right now I have them configured to draw about 5mA and the LEDs are still really bright, so I may decrease that a mA or two!
I used the same toroid inductor that comes with the 41Hz amp3 and amp6 kits because it's what I have available and I can wind them in my sleep by now (I still hate doing it though...).
There is no turn on thump with this circuit but there is a small turn off pop that's nothing to worry about.
Attachments
The parts values I chose for the circuit make the amp oscillate at approximately 660kHz. Using such a small chip has its advantages when it comes to optimizing the PCB layout. The switching node distance between the chip's output pin and the inductor is only 8mm long, and the high current switching paths were kept as short as possible. The power supply and output ground points were kept as close as possible to each other. As such, the film output cap is hard to see because it is nestled in between the supply caps 
The board itself is about 54mm x 35mm and is a two layer design with 1oz copper.
The board itself is about 54mm x 35mm and is a two layer design with 1oz copper.
Attachments
And how does it sound? So far I'm liking it just as much as my Ref-T. The channel separation is very good because of the dual mono implementation even though both modules in my setup share the same supply. The detail of string instruments sounds excellent much like with the Tripath amps.
Because of the small size and mono output I will eventually upgrade to an active line level crossover and have one amp module per speaker I just need to find some good 9V AC wallwarts so I can make some separate supplies for these guys.
Because of the small size and mono output I will eventually upgrade to an active line level crossover and have one amp module per speaker
I just need to find some good 9V AC wallwarts so I can make some separate supplies for these guys.Attachments
Very nice.
Back in the late 1980s I soldered PCB (ICs and the lot) for computers used on oil tankers. I didn't put much thought in it back then. But later I thought that was a big deal. Imagine a single cold joint! It must cost a fortune to have a tanker stuck in some port.
The point to this is that I still am afraid I'm not up to soldering those tiny SMTs. I will have to give it a go eventually. But I'm not rushing it.
Back in the late 1980s I soldered PCB (ICs and the lot) for computers used on oil tankers. I didn't put much thought in it back then. But later I thought that was a big deal. Imagine a single cold joint! It must cost a fortune to have a tanker stuck in some port.
The point to this is that I still am afraid I'm not up to soldering those tiny SMTs. I will have to give it a go eventually. But I'm not rushing it.
Hey Brian,
I see you've been up to no good.
Cool board! I don't know that chip. How do you keep it cool?
As for toroids, you might try the 2100 series from Jameco. Price is good, so is the sound. Too big to fit the Sonic boards, but should work well for you.
They don't kill the high RF harmonics like some cores, but they aren't bad. Only problem is wire thickness. You'll want bigger holes on the board. But that big wire and short windings will give very low DCR. And those inductors are NOT going to saturate!
Good work on the new board. Keep us informed.
I see you've been up to no good.
Cool board! I don't know that chip. How do you keep it cool?
As for toroids, you might try the 2100 series from Jameco. Price is good, so is the sound. Too big to fit the Sonic boards, but should work well for you.
They don't kill the high RF harmonics like some cores, but they aren't bad. Only problem is wire thickness. You'll want bigger holes on the board. But that big wire and short windings will give very low DCR. And those inductors are NOT going to saturate!
Good work on the new board. Keep us informed.
How do you keep the chip cool? So far I just let it sit there. It hasn't gotten hot enough to require additional heatsinking 
Thanks for the tip about the Jameco toroids. I'll look into them. I was also contemplating the use of a surface mount inductor and was trying to get some samples without any luck so far.
Here's a link to the chip's datasheet for those interested:
http://www.monolithicpower.com/files/MP7720_r1.9.pdf
Thanks for the tip about the Jameco toroids. I'll look into them. I was also contemplating the use of a surface mount inductor and was trying to get some samples without any luck so far.
Here's a link to the chip's datasheet for those interested:
http://www.monolithicpower.com/files/MP7720_r1.9.pdf
Mp7720
Brian,
Do you use split rails for this circuit? They show how to do this in one of their app notes. This allows complete DC coupling, a desirable thing especially at the output. I would do all the critical caps in through hole so they can be upgraded to poly film types.
Aside from this how does it compare to the sonic impact or other amps we might be familiar with?
Good job on the layout looks very professional.
Roger
Brian,
Do you use split rails for this circuit? They show how to do this in one of their app notes. This allows complete DC coupling, a desirable thing especially at the output. I would do all the critical caps in through hole so they can be upgraded to poly film types.
Aside from this how does it compare to the sonic impact or other amps we might be familiar with?
Good job on the layout looks very professional.
Roger
Hi Roger. Thanks! I designed the board for a split rail configuration as opposed to their single rail design that requires input and output coupling caps for the reason you mentioned and for some others as well! One being that you can easily add a preamp stage and use the split +/-12V rails to power it.
The critical caps could be changed to through hole but I don't see anything wrong staying with surface mount types. The lowest quality ceramic caps used on the board are X7R, with C0G or NP0 ceramics used wherever possible - especially for the feedback cap and the cap that is a factor in setting the oscillation frequency. I also used 2% or 5% tolerance varieties in those spots.
I've really only compared it to my amp3 so far, and I'd have to say that I like it better than the amp3. And the way I have them powered and wired up right now can be improved quite a bit! You can see my setup in the photo from the first post of this topic. I'll be comparing the D10.1s to my Ref-T as well, which I also like better than the amp3.
I'll probe around with my scope and see what's going on. I'll also try to do the AM radio test to see about EMI. I don't think they emit much because I have them sitting within inches of each other and can't hear any audible interaction like beat frequencies and such. I could see what the switching frequency of each module is as well.
The critical caps could be changed to through hole but I don't see anything wrong staying with surface mount types. The lowest quality ceramic caps used on the board are X7R, with C0G or NP0 ceramics used wherever possible - especially for the feedback cap and the cap that is a factor in setting the oscillation frequency. I also used 2% or 5% tolerance varieties in those spots.
I've really only compared it to my amp3 so far, and I'd have to say that I like it better than the amp3. And the way I have them powered and wired up right now can be improved quite a bit! You can see my setup in the photo from the first post of this topic. I'll be comparing the D10.1s to my Ref-T as well, which I also like better than the amp3.
I'll probe around with my scope and see what's going on. I'll also try to do the AM radio test to see about EMI. I don't think they emit much because I have them sitting within inches of each other and can't hear any audible interaction like beat frequencies and such. I could see what the switching frequency of each module is as well.
Here you go.
http://www.monolithicpower.com/files/DN0001_r1.0.pdf
I just tweaked the design a bit and added my own circuitry to drive the LEDs and get 5V referenced to the - rail.
http://www.monolithicpower.com/files/DN0001_r1.0.pdf
I just tweaked the design a bit and added my own circuitry to drive the LEDs and get 5V referenced to the - rail.
I've been comparing the D10.1s to my Ref-T (with Sonicap input coupling caps) and have enjoyed every minute so far! So far I've concluded that I prefer the sound of the D10.1 modules with my 2 way speakers. The Ref-T sounds very detailed with a great sound stage and has a tad more power with more bass authority, but the dual mono modules have the same great sound stage but with a slightly warmer sound (warm meaning less clinical and not less detailed). I also like the simple self oscillating design and no need for input coupling caps.
As such, I'm now looking into making D10.1 kits instead of a production run for the Ref-T. This would be an excellent way to get into the hobby without having to worry about soldering a fine lead pitch pin like the Tripath TA2021B. In a few days I will be able to put kits together for my 4 remaining boards. I haven't come up with a price yet, but would anyone be interested in buying a couple kits to try out? If so, please email me.
As such, I'm now looking into making D10.1 kits instead of a production run for the Ref-T. This would be an excellent way to get into the hobby without having to worry about soldering a fine lead pitch pin like the Tripath TA2021B. In a few days I will be able to put kits together for my 4 remaining boards. I haven't come up with a price yet, but would anyone be interested in buying a couple kits to try out? If so, please email me.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.