TI TPA3100D2, Amazing Class D Amp

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Hmm you guys almost inspired me to get these masterkit amps to test. I have UcD 180 for years and quite happy with them. While THD charts of TI chips are unimpressing at least in comparison with Hypex, you are describing them in such colors so it's hard to resist to try these and do A/B comparing with UcD180.
 
TPA3100D2

Hi I saw your board in the previous post. looks very complex.
can your board be adapted for a simple amp as in the evaluation board ?


Hi kp93300,

For sure it can, now that I have it drawn in a schematic, parts all placed and routed.

Need to Copy design, blow away everything not needed on schematic, go back to layout & away you go again.
I could make it exactly as TI did in the app note, I could not find the TOKO inductors so I subed the Murata parts. TOKO, you can never get their parts.
Same as on the newer TPA3118 eval board, same TOKO parts used again.

Regards
Rick
 
TPA3100D2 current/power consumption

You would be surprised how little power these parts use with music as a source.
Using the Dynaudio Gemini book shelf speakers, rated 4ohm!!
I measured the portable stereo, with the lcd backlight on
1) Muted, 216mA @ 14.4V
2) Regular listening level, 250mA@14.4V
I want to do some more measurements, driving the 15" speaker cabinets, to really push it and crank it in the living room. Need to check the bass response in the lowest octave.

Rick
 
With such low power consumption, this accounts for my observation that the amp will run for a few seconds when i switched off the mains.
Maybe i should look for a high value low ESR caps and place it at the chip input !
rsavas,
Can you design a board with through hole parts and wider spacing for experimentation?

thanks

kp93300
 
Hmm you guys almost inspired me to get these masterkit amps to test. I have UcD 180 for years and quite happy with them. While THD charts of TI chips are unimpressing at least in comparison with Hypex, you are describing them in such colors so it's hard to resist to try these and do A/B comparing with UcD180.

Hi Mcdark
I do not have any hypex amp but the description of the NCORE sound in the Audiocircle has some semblance to what I am hearing.
kp93300
 
TPA3100 design factors

1) With such low power consumption, this accounts for my observation that the amp will run for a few seconds when i switched off the mains.
2)Maybe i should look for a high value low ESR caps and place it at the chip input !
rsavas,
3)Can you design a board with through hole parts and wider spacing for experimentation?

thanks
kp93300


1)Yes that makes sense, the current drawn is low, so the bulk capacitor reserve capacity takes a while to bleed off. I do not get a turn-off thump, because the MCU measures the supply voltage and mutes the TPA3100, if the supply crosses the detection threshold.
2) Not sure why you need to change any caps other than beef up the main reserve cap, which I think you said you have done already? My choice of a 4700uF/35V is I believe sufficient for the application.
3) Sure I can design with PTH parts, but the IC itself is SMT, so the logical thing to do is use SMT parts, that is what TI did on their eval board. I did things a bit different than the TI eval design, as I did not use 0603 & 0402 parts, I used 0805/1206/1210 parts in my design. Makes it slightly larger, but not by much and it is easier to solder 0805 than 0603/0402 parts. Using leaded film caps is a waste of space/cost in my mind. As local IC supply filters, leaded film caps are worse than SMT parts, because of the longer trace lengths = more lead inductance. The signal coupling caps I used are 1210/50V/X7R parts. Sufficient for this application.

Rick
 
Hi Mcdark
I do not have any hypex amp but the description of the NCORE sound in the Audiocircle has some semblance to what I am hearing.
kp93300

Which source/speakers are you using?

Most Class D amps have similiar sound signature. I've got Yamaha MCR-040 for my kitchen and it has Yamaha YDA147 Class D chip, it sounds pretty much the same way as UcD180 with same speakers (in comparison to say NAD receiver I had before), it has same excellent timing and dynamics, clarity and richness, but UcD180 just has more of that, mostly in clarity/details department. This is well confirmed by THD+N charts of them, they are similiar in shape but different in numbers. So looking at TPA3100 datasheet charts makes me expecting something different as its THD+N chart has different shape and is obviously much worse but which sound signature it shows sonically I can't predict. Why not to try, it is fun.

Jean-Paul, I'm going to order one, try it myself :) and then we'll decide on further details. Sending from Russia is costly... I'll drop you PM then, but it's not going to be quite soon, one-two weeks maybe.
 
kp93300

I see. My speakers approach is quite different, source is tweaked uDAC 2 (Sabre inside so it is close enough to your Buffalo2 source) but speakers are modified AudioPro Bravo FX-04 with selected Vifa/ScanSpeak XT25TG30-04 tweeters installed in place of stock OEM Vifa unknown but similiar units. My small Yamaha uses full-range speakers and they sound amazingly good (escpecially for their size), smooth and juicy with its own Class D, but they did not expected to reveal the differences between it and main UcD-based amp just due to resolution limitations of coaxial/full range design itself in comparison to Ring Radiator type tweeters (or say Magnepans). I do know limitations of two-way speakers in comparison with full-range but thats my sound signature choice and probably this amp is just not for me while it could fit your setup with excellent results. For the money worth trying just in curiosity though, at least to confirm my understanding of Class D properties. Gotta start saving for nCore I guess...:rolleyes:
 
Hi
This are the motor run cap that use. Rather cheap parts.
I tuned it by ear. The 2 biggest caps are the motor run caps with values 0f 1 uf and 5 uf.

The second picture is a more effective upgrade in my view.
I parallel a 1000uf 25 v cap on the underside of the board. The 1000uf is an improvement over the 470uf.
I tried 0.1uf electro cap, 100nf polyester and 101 ceramic caps. The ceramic sound better.
The type of the bypass caps has a big impact in SQ.

You should experiment and post your findings here.
cheers

kp93300
 

Attachments

  • motor run cap.jpg
    motor run cap.jpg
    311.7 KB · Views: 775
  • 1000uf .jpg
    1000uf .jpg
    333.3 KB · Views: 747
Last edited:
I want to share a little more experience with this chip .
I notice that i am playing music at lower volumes these days. Mostly around 85db at my listening position.
My wife comment that she experience good clarity of the vocal even in another room.
The sound is rather neutral and rather sensitive to upstream cable change.

cheers
kp93300
 
Oh dear looks like I'll be putting that order in then. Didn't see a pair of metal scissors under the small board of the tbi and it shorted something with a snap sound, even though it wasn't plugged in. I don't own a multi-meter but presumably if I got one I could go round testing the bigger components on that board and possibly replace the one I just blew?
 
Oh dear looks like I'll be putting that order in then. Didn't see a pair of metal scissors under the small board of the tbi and it shorted something with a snap sound, even though it wasn't plugged in. I don't own a multi-meter but presumably if I got one I could go round testing the bigger components on that board and possibly replace the one I just blew?
Hi gilesw,
Do no despair.
The board survived a short circuit in the power supply input in my board !
A multimeter is very helpful in diy work. try to get one and have some fun .
cheers
kp93300
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.