Cheap TPA3118D2 boards, modding them and everything that comes with it

I have the blue board monos and am very impressed with the sound quality. I went ahead and got a pair of the "other" monos for the heck of it. The inductors look beefier and I like the fact that it has the terminals. Seems to have much less components in the circuit. We'll see how it turns out.
 

Attachments

  • s-l1600.jpg
    s-l1600.jpg
    85.5 KB · Views: 1,003
That's not what I asked. That I can do but do not wish to.

I wanted to know if anyone knew of any compatible enclosures for this project.

Hi Carlmart

Yes you are right. :confused:....i checked at aliexpress 14 pages with the "TPA3116 cases"

i found nothing fitting to your pcb. your pcb is special. you have pot and input jack on one side and at the opposite side the power jack.

i just found other configurations which i found very often:
Nobsound HiFi TPA3116 Digital Power Amplifier Mini Stereo Audio Class D Amp 2.0 Channel 50W*2-in Amplifier from Consumer Electronics on Aliexpress.com | Alibaba Group

mn36 Mini TPA3116 Bluetooth 4.2 Digital Amplifier 50W+50W Stereo Audio Amp 2018 New Listing-in Amplifier from Consumer Electronics on Aliexpress.com | Alibaba Group


Aiyima 100W TPA3116 Subwoofer Audio Amplifier TPA3116D2 Mono Digital Power Amplifiers NE5532 OP AMP With Case-in Amplifier from Consumer Electronics on Aliexpress.com | Alibaba Group


only this board is nearly the "same" just the power + speakers terminals is instead of the power jack at the back side
New Version High power Amplifiers Dual Chip TPA3116 50W x 2 Digital Amplifier Board 12 24V Dual Channel Stereo-in Amplifier from Consumer Electronics on Aliexpress.com | Alibaba Group


chris
 
Thanks for your search, Chris.

Problem is they do not sell ready made drilled boxes, just the case, for you to put your board into.

The boards can also be single, mono types, that would be fine.

Yesterday I went ahead and bought this amp complete, board and case:

TPA3116 HIFI Class 2.0 Stereo Audio Digital Amplifier Advanced 2X50W Breeze Amp 656112245688 | eBay

Also the external supply:

Douk Audio 19V 3.42A 65W AC/DC Power Adapter Power Supply For Digital Amplifiers | eBay

It's interesting that all you seem to find are mostly using the TPA3116, instead of the TPA3118.

In any case, I am thinking of assembling an active crossover system with two of above amps, one foe each channel, feeding a 2-way box.

But my idea is not to use an active equalizer before the amps, but to use a passive filter at each input, one low pass, the other high pass, feeding tweeter and bass-mid drivers respectively.

Has anyone used these amps, 3118 or 3116, for an active system?
 
Interesting measurement/observation

I recently got around to finally wiring up some of the tpa3118 pbtl mono boards ( and a sanwu blue tda7498) ........
they seem to be running ok sound and dc offset ( the 3118 0vdc and the 7498 ~50mv) additionally if I measure ac volts with music playing the output at each of the inductors are similar ( as much as I can measure fluctuating ac with a voltmeter)
What is interesting/ confusing to me is on both board types there is one of the four ps cap and a inductor on the 7498 board that are running essentially at room temp (+-1-2c)while the other3 are 10-15c above ....same with the 3118 but just one of the ps cap.
Is this normal? Or is it sign of nonfunctional/counterfeit caps and a inductor that the actual value is so low there is little resistance?
 
I'm fairly sure I read in this thread of noise issues occasioned when running two or more 3118 modules off the same power supply. The search function is not really helping, I'd be grateful if someone can point me in the right direction.

Edit: I see mention of this in post #1226, but I don't see where a solution is posted.
 
Last edited:
It may very well be worth trying syncing the two (or more) modules together. That's why there are two sets of gain settings.

Although as another user pointed out, the resistor & capacitor values mentioned in the datasheet for connecting the "master" chip to the "slave" one(s) are incorrect.

I'm fairly sure I read in this thread of noise issues occasioned when running two or more 3118 modules off the same power supply. The search function is not really helping, I'd be grateful if someone can point me in the right direction.

Edit: I see mention of this in post #1226, but I don't see where a solution is posted.
 
I got up to reading through page 100 or so and want to take the plunge. Given the quality difference we see on ebay direct amps is there a current "winning" seller/link for these amps? How about the 3116's and the TDA8932's?

I will be powering 2 martello's with FR89EX's installed with two notch filters. 8ohm versions.
Martello enclosure for FR88EX


It looks like I'll want to mod this to 26db gain and put in nicer inductors and caps.

I'm trying to decide the source for this build as well but I'll save that guidance for another thread. Most likely a selector with bluetooth and a pi zero w with a DAC->sure 1701DSP->2.1 martello's and voxel mini sub. Mini sub will be powered by 3116 plate amp. Cases for speakers will be old IPE decking material S4'd at 1/2".
 
Member
Joined 2008
Paid Member
I've made a little comparison of a modified and a stock amplifier board, with regard to idle power usage. As you can see, there's quite something to be gained when using lower-DCR coils, and even tossing out all the ceramics on the output side yields some milliwatts. Might be of interest for battery powered applications.

attachment.php


The first column shows the stock board as delivered, with a gain setting of 32dB and neither input nor output connected to anything (didn't make a difference here). The idle power at 24V is almost 1.5W, which I had expected to be somewhat less - the datasheet states a typical idle current of 32mA (50mA max though) at 24V, which would have been half of that.

The second column shows the first mods. I removed the power LED with its resistor and the reverse-voltage protection diode. The inductors were replaced with a 10uH Sagami 2-in-1 with a typical DCR of 15mOhms. This reduced the 24V idle power draw by ~33%.

The last column takes the mod a bit further, in that now I removed the ceramic output caps together with the "EMI C-RC snubber" and replaced them with 1uF WIMA film capacitors. This saves almost another 100mW, which are only converted to heat.
 

Attachments

  • 20181105_085921.jpg
    20181105_085921.jpg
    468.3 KB · Views: 692
  • tpa3118_idle_power.png
    tpa3118_idle_power.png
    12.1 KB · Views: 1,571
For the heat maybe, but certainly not for saving some power... ;)


Thorough and well structured work. Interesting to investigate the potential of these versatile 3116/3118 chips. Depending on the price level and power focus such measures can be implemented. For low cost applications, Sagami chokes and Wima foil caps are likely to be seen as excessive.
 
It all adds up in battery-powered applications, though :rolleyes:

"A penny saved is a penny earned" - i'd argue it kinda works the same way with mAh :D

Thorough and well structured work. Interesting to investigate the potential of these versatile 3116/3118 chips. Depending on the price level and power focus such measures can be implemented. For low cost applications, Sagami chokes and Wima foil caps are likely to be seen as excessive.
 
The second column shows the first mods. I removed the power LED with its resistor and the reverse-voltage protection diode. The inductors were replaced with a 10uH Sagami 2-in-1 with a typical DCR of 15mOhms. This reduced the 24V idle power draw by ~33%.

The last column takes the mod a bit further, in that now I removed the ceramic output caps together with the "EMI C-RC snubber" and replaced them with 1uF WIMA film capacitors. This saves almost another 100mW, which are only converted to heat.


Wouldn't it have been better to measure after each step to see the potential?


LED + Diode + Inductor is one part
EMI C-RC + MLCC-Foil change is another


I suspect the Diode has the least impact on idle-power, given a 0.3V voltage-drop, this renders to 9mW for 30mA idle current. The Led is having an influence of about 6.3mW at 9V, as the resistor is 10k at Vf of the LED ~2V. LED current is then 7V/10k = 0.7mA.


At 24V this renders to 24V-2V/10k= 2.2mA -> 53mW


The main power savers are the inductors here. Removing the CRC-Snubbers helps a bit more, but not that much.


For more saving check Pin1 at the TPA3118. If its on GND, the Chip is running BD-Modulation. When connecting this pin (lift up to do so) with AVCC, the IC is running 1SPW modulation, which uses way less power.


Care must be taken when using 1SPW, as ringing might occur, see page 19 datasheet:
http://www.ti.com/lit/ds/symlink/tpa3118d2.pdf


I did quite alot of measurements regarding BD/1SPW/Hybrid in the past:
[Resolved] TPA3128D2: THD Performance in Hybrid/1SPW vs. BD modulation when using single-ended input (compared to fully differential drive) - Audio forum - Audio - TI E2E Community


Regards.
 
Hi all.

I get two of these boards.
I must say, at first (stock, I just removed protection diodes and add some 470uf 25v cap on supply rail) I was really disappointed with the sound I get from, huge lack of bass, harsh and blurry highs... Terrific result on 8Ohms load with 19v 5a laptop brick.

I replaced input cap (c4) with 2.2uf x7r, original one were somewhere between 560nf and 640nf... Really to low, sorted out r27 to reduce gain and increase input Zin and bass were starting to came out but highs where sooooo ugly, still un listenable...
I added some snubber 10Ohms + 300pf between choke input and ground, not enough, I'm not sure it helps a lot.

Then I downloaded the ti filter calculator spreadsheet and according to some simulations, swapped 10uf smd inductors with 22uf toroid I removed from my tda7498e board.
With the stock 1uf at stock filter it's just a bit under-damped at 8Ohms load and over-damped at 6.

Now it sounds really excellent. Nothing to say, I'm impressed !!!
For 1.5€ + 2€ more on parts and some solder job per channel you can get a pretty decent listening experience.
Better to my taste to the tpa3116 boards I get before
It just lacks a bit of the juice I like a lot with tda7498e.

So, just an advice if you have to drive some 8R speakers, just go to 22uh beefier inductors, you'll never regret.

@Preamp, where did you get you inductors and what are the values? 22uh is the perfect match with 680nf for 8R speakers.
 

Attachments

  • IMG_20181119_190852.jpg
    IMG_20181119_190852.jpg
    567.2 KB · Views: 509
Last edited:
Member
Joined 2008
Paid Member
attachment.php


These are the measured responses of the speakers that I have around.
The magenta trace is for my main speakers, Spendor S6. As you can see, the impedance rises pretty steep above 10kHz, and probably continues to do so above my measurement limit. Unlike the green trace, which shows some heavy Canton speakers, that simulate pretty much like an ideal 4 Ohm load!

Here's the LTspice sim file again.
 

Attachments

  • speaker_impedance.png
    speaker_impedance.png
    18.2 KB · Views: 1,626
  • Class-D_Output_Filter_02_all.asc
    35.8 KB · Views: 51