Not all compromises are the same, however I don't really want to discuss which compromises you or I or others would rather live with. You obviously have your preference and I have mine. Let's leave it at that and stay on topic.
Kitchen radios are for casual off-axis listening. Lobing errors is wrong place to focus when designing excellent speakers for stereo listening in the spot.
I've had Kef Reference 205/2. My listening and measurement show that I would still select lobing errors instead of coaxial driver.
Its not about off axis listening, its about neutral direct and reverberant fields - unless your listening room is like an anechoic chambers the off-axis response matters.
Correction - what was meant that a speaker could be considered a minimum phase device and using a minimum phase eq will correct anomalies. Forget about linear phase eq, useless in the above situation.
Many people really enjoy the sound of the TPA3116D2 amp. Some have compared it with Hypex and like the 3116 better, and some have said it sounds like a good class A amp. 4300 posts and 430,000 views of this thread in 14 months is testament to the interest it produces.
http://www.diyaudio.com/forums/class-d/237086-tpa3116d2-amp.html
A stock $30 TPA3116D2 with about $20 worth of mods in caps and inductors can sound quite amazing. Or you can buy a fully-built 3118 variant with high quality parts in a box with knobs for $50. If you haven't tried it, not much to lose and a lot to gain if you like it.
Thanks for correcting that. I wasn't sure what you meant, but I thought you might have said it wrong.
Like everything else, one can buy on Amazon... 🙂
Champagne box with 24v supply:
Amazon.com: SMSL SA-36A Pro 20WPC TPA3118D2 Digital Amplifier AMP + 24V Power Supply Golden: Electronics
Black box with 12v supply:
http://www.amazon.com/SMSL-SA-36A-TPA3118D2-Digital-Amplifier/dp/B00L5Y5GOK/ref=sr_1_5?s=electronics&ie=UTF8&qid=1408036621&sr=1-5&keywords=tpa3118
Note that one can use a standard 19v laptop supply and that works great. Even better if you have a linear regulated supply.
Based on examination of circuit board in this model and from user's feedback, this amp sounds great stock out of the box as it has many of the "mods" one would apply if one started with a cheap $20 board (upgraded input caps, bootstrap caps, bootstrap snubber, upgraded power supply caps, upgraded inductors). The 3118 is the non-external heat sink version that is cooled by conduction to PCB so max output is 25 watts/ch at 4ohms.
Yes, it was in post #687, and repeated here:
http://www.diyaudio.com/forums/class-d/237086-tpa3116d2-amp.html
Techically, the amp sold is a 3118, but the 3118 is same as 3116 but not cooled with an external heatsink but with vias to the PCB. there is technically a 3118 thread too here:
http://www.diyaudio.com/forums/class-d/219730-tpa3118d2.html
I built the first diy 3118 before they made PCB's and amps on Amazon using point-to-point dead bug with SMT parts:
Sorry for the OT... 😀
http://www.diyaudio.com/forums/class-d/237086-tpa3116d2-amp.html
Techically, the amp sold is a 3118, but the 3118 is same as 3116 but not cooled with an external heatsink but with vias to the PCB. there is technically a 3118 thread too here:
http://www.diyaudio.com/forums/class-d/219730-tpa3118d2.html
I built the first diy 3118 before they made PCB's and amps on Amazon using point-to-point dead bug with SMT parts:
Sorry for the OT... 😀
Last edited:
I'd like to see a low level analysis of those just to be sure 🙂 Yes they do measure very well, but the distortion vs output power graph shows pretty high distortion at lower output powers, this could just be due to the scaling of the measurement with the necessary headroom for measuring up to high powers.
In other words it'd be nice if the maximum level was changed to measure up to say 2 watts re 8 ohms so the noise floor was capable of showing us whether or not the amplifier can do 0.0005% distortion at 1 watt or less.
Still the nCore are ridiculously expensive.
I have a pair of their other UCD amps on my subs and there they work great.
Myself included. I have built several devices using the 3118. 4 channels run the surrounds in my surround set up. 2 channels run the near field set up in the area I work in when soldering etc and I've built two sets of powered loudspeakers that use variants on the same design.
From a brief measurement I made they did appear to throw up reasonable levels of higher order distortion products, but I haven't made a proper detailed measurement to confirm this.
Nevertheless they do sound very nice.
Well, I ordered a TPA3118D2 from here (free shipping):
SMSL SA-36A Pro 20WPC TPA3118D2 Digital Amplifier + Power Adapter - Professional Audio Store - Shenzhenaudio.com
So cheap that I might as well try it. Thanks for the suggestion. I have Ncore amps which are the cleanest and most transparent amps I've ever tried. I will do a comparison. 20W should work fine with my Klipsch K402 horn. Might be enough for my Abbeys as well.
SMSL SA-36A Pro 20WPC TPA3118D2 Digital Amplifier + Power Adapter - Professional Audio Store - Shenzhenaudio.com
So cheap that I might as well try it. Thanks for the suggestion. I have Ncore amps which are the cleanest and most transparent amps I've ever tried. I will do a comparison. 20W should work fine with my Klipsch K402 horn. Might be enough for my Abbeys as well.
You are welcome. You won't be disappointed. You may wonder why you spent so much on more expensive amps in the past. 🙂
Do you have measurements? I ask because published measurements of the 201/2 and 207/2 show simple and absolute excellence in performance: flat and smooth axial frequency response, and smoothly declining polars. I'd be really surprised if a 205/2 was very different from those two.
I'm sure that amp is fine for people who don't like high-fidelity reproduction. I quickly looked over the spec sheet and didn't any variation of the phrase that really matters in class D amps: load invariant frequency response. That means the FR will stay constant regardless of the driven load. There aren't very many load invariant class D amps. The only ones of which I know are the Hypex (UcD and Ncore), Icepower ASX2 (not earlier versions), and Anaview ones. Hopefully there are others unknown to me as well, but most class D amps offered today are strictly low-fidelity or closed system* devices. No different from tube amps, really.
*amps that aren't load invariant are fine for closed systems, i.e. speakers with active crossovers designed using measurements of specific drivers powered by a specific amp channel. But for any broader application...why waste so much time getting the crossover right if the amp is going to screw up the frequency response anyway?
Last edited:
Well, look again and compare...
http://www.cirrus.com/en/pubs/proDatasheet/CS4272_F1.pdf
What you see in fig 9.2 is the residual noise of the amp, and it is ridiculously low: 23uV, 125dB S/N at max power, 102dB at 1W, all this unweighted and with a 26dB gain ...
The straight downward line is the sign that the residual noise dominates, meaning that the distortion stays under that ultra low residual noise up to at least 10W, at any frequency.
- Status
- Not open for further replies.