TPA3116 D2 mono amplifier as a stereo system?

Most small d-chip amps got a polarity protection in the input. The anti idiot device. This loweres the voltage the amp can use. So if you drive your amp hard, consider removing or bridgeing the diode. Only if you are sure the polarity can never be reversed by mistake, like in a fixed installation with PSU, of course!
This should give you about 0.5 volt extra at the output, with bridged amps even 1 volt more p-p.
 
PS nice idea is to use a diode from positive to ground, so if polarity reverses, as with some SMP´s while shuting down, the voltage gets neutrelized. Combined with a fuse it can even protect agains wrong polarity, the fuse blows , cause the diode is a short, but the cirquit does not get more than 0.7 volt the wrong way.
 
Any DIYS amp can use an input transformer. Makes it near to impossible to run ground loops. A pair of nice HIFI usable Lundahls is about 150 Euros, but before you order:
Why not buy a good T-amp or Behringer KM750 for the money, have no ground loop problems and a case, power supply and amp for free? If you have tried these in a HIFI system and compared them to some of the D-amps this is all about, I´m pretty sure you will change direction. Transformers do not make mediocre amps sound better, they CAN solve impedance problems. Just like a 2$ pre amp PCB with a NE5532 op amp can do, with much less distortion. If you know how to prevent ground loop´s, of course.

PS just "fitting a transformer" is not the case with these amps, as often the input is running some potential, making capacitors neccesary. This complicates the input cirquit, because the transformer needs a constant impedance. No problem if you know about all this stuff, but needs some thought if you want a good result.
 
Last edited:
@FauxFrench - I'm referring to your post #54 in this thread.

TPA3116 D2 mono amplifier as a stereo system?

I have the identical board and am trying to reduce hum/hiss. If I understand your instructions correctly, I can connect a 4k7 resistor directly across the input terminals, from positive input terminal to ground input terminal, then connect signal inputs as usual?

Also, I'm using a stereo TPA 3116 board in the same setup, so can I do the same with both signal inputs on that board?
 
Hi Erik,
This board should be without a volume potentiometer. Then, you just do as you say - connect 4K7 across the input terminal and ground. For the stereo board, between each of the two inputs and ground. Signal input is as usual.
If you have a board with a 50K volume potentiometer, then from the potentiometer output and ground.
 
Thank you FF, both boards have volume pots, so I'll have to see if I can identify which is the potentiometer output on each board.


You can connect the 4K7 directly on the volume potentiometer:
With the volume potentiometer axis (the one that is turned) pointing towards you and the three potentiometer terminals pointing downwards, you connect the 4K7 between the left-most terminal and the middle terminal. The middle terminal is actually the potentiometer output.
 
Hi graham

lets go from the speakers "backwards"...

if your volume knob request more power the amps should be deliver the power.
The bass burst need the most power- for this the caps should deliver inbetween a very short time the needed current if the PSU is not able to hold the voltage and deliver the needed current.
so for some power brick the requestet short time is a problem. the concept of them is to deliver constatn a current and react very bad in dynamic requests by an amp. so a big cap could help if the PSU is too weak or your party should roll:D
so if you need much power the "correct" way is to choose i very strong PSU.e.g. 24V 10Amps = about 250-300W and not a brick with 24V 2,5A;)

chris

As Chris says, the PSU has to be sufficient to supply the power needed for the amplifier.

The supply line decoupling capacitors act as local buffers for the amplifier such that the PSU only needs to supply the average energy used by the amplifier. The capacitors will collect the energy from the PSU and forward the energy to the amplifier with the current needed by the amplifier. That current may heavily exceed the current the PSU can handle but only for a short moment. Thus you are right, a poorly performing PSU may be improved with good decoupling capacitors. The bigger the capacitors, the longer they can supply the amplifier with current without loosing too much voltage.

With a very good "capacitor bank" for decoupling, you can use a PSU of half the power that is maximum demanded by the amplifier. That is due to the "crest-factor" of music and the ability of the capacitor bank to smoothen out the amplifier energy consumption to near the average consumption.

Thanks guys, so...

If I have no option but to use a 20v 6.7a PSU to power two boards (could be either TPA3118 or TDA7492) would I have enough reserves with a 10,000uf cap per board? I am trying to wangle a set up for as cheap as possible for a friend.
 
Hi Graham

The problem with big cap banks with a smps (brick) is that it could be that your PSU run into the over current protection mode because of the high current flow for the cap bank -during switch on. additionally i measure that a big cap bank can have a overvoltage = overswing because of its regulation.

regulation:
the smps try to have the recommenden voltage up to its max power - so the current is regulated by the smps. the cap bank is a big storage of energy. if the amp (music) request a big current then the caps at the amp board must deliver this. the cap bank load the caps at the amp board + the smps load the cap bank. this current demand force the smps to regulate as fast as designed the voltag/current/ and this you can see at the scope. eg. about 3-6V overshoot.

i learnt that the design of the smps is mandatory for your needed power and for the used impedance of your speakers. e.g. 8R more voltage - but for TPA3116 its max is 24V. tpa3118 26V , TDA7492 26V
so your 20V 6,7A psu is fits to your amps.

my proposal is to solder good caps at the amp board = low ESR/ESL as big as possible = 2200µ -3300µF 35V and try during loud music if the bass is stable enough. then you can try as "cap bank" additionally a 2200 - 4700µF /35V.

chris
 
Hi Graham

The problem with big cap banks with a smps (brick) is that it could be that your PSU run into the over current protection mode because of the high current flow for the cap bank -during switch on. additionally i measure that a big cap bank can have a overvoltage = overswing because of its regulation.

regulation:
the smps try to have the recommenden voltage up to its max power - so the current is regulated by the smps. the cap bank is a big storage of energy. if the amp (music) request a big current then the caps at the amp board must deliver this. the cap bank load the caps at the amp board + the smps load the cap bank. this current demand force the smps to regulate as fast as designed the voltag/current/ and this you can see at the scope. eg. about 3-6V overshoot.

i learnt that the design of the smps is mandatory for your needed power and for the used impedance of your speakers. e.g. 8R more voltage - but for TPA3116 its max is 24V. tpa3118 26V , TDA7492 26V
so your 20V 6,7A psu is fits to your amps.

my proposal is to solder good caps at the amp board = low ESR/ESL as big as possible = 2200µ -3300µF 35V and try during loud music if the bass is stable enough. then you can try as "cap bank" additionally a 2200 - 4700µF /35V.

chris

Cheers man, that is what I will do then! :)

As far as running any 2 of this type of amp board from 1 PSU goes, its osc. sync essential? I can't find the thread where I guy used a diode.

Can a TPA and a TDA be synced?
 
Hi Graham,

Good advice from Chris.

To your question, can a TPA and a TDA be synchronized to one another?, the reply is that it is not foreseen but perhaps possible.
The heterodyne effect typically relates to frequencies that close to one another. A TPA3118 operates at 400KHz and upwards. A TDA7492 running on internal oscillator operates at 310KHz. The difference in operational frequency is so large that the heterodyne frequency should be no bother. Therefore, the possibility of synchronization of the two is more theoretical.
A TDA7492 can operate up to 400KHz with an external clock. The TPA can generate a clock for 400KHz operation. So, in theory you can let the TPA be master and the TDA be slave at 400KHz. But, it requires that the synchronization signal, often a clock signal at double the frequency of the operational frequency, is suitable for the TDA7492 and the clock signal level is right. You can try but you will hardly notice an improvement even if you succeed.
 
You can connect the 4K7 directly on the volume potentiometer:
With the volume potentiometer axis (the one that is turned) pointing towards you and the three potentiometer terminals pointing downwards, you connect the 4K7 between the left-most terminal and the middle terminal. The middle terminal is actually the potentiometer output.

I'd need to dismantle pretty much the whole setup to get the boards out in order to connect resistors to the volume potentiometers. Instead, will the same effect be achieved if I connect the resistors at the signal inputs, since I can access these without any fuss. If this solves the issue I have, I can then remove the boards and solder the resistors in at the potentiometers.
 
Hi Erik,
I understand that the volume potentiometer(s) are mounted on the amplifier board(s). If you connect the resistors before the potentiometer, you lower the impedance for the source but for the TPA3116 input it changes very little.
Could you please make me a couple of photos showing access to the potentiometers?
 
Hi Erik,
I understand that the volume potentiometer(s) are mounted on the amplifier board(s). If you connect the resistors before the potentiometer, you lower the impedance for the source but for the TPA3116 input it changes very little.
Could you please make me a couple of photos showing access to the potentiometers?

Sure, I'll take the boards out and post some photos.
 

Attachments

  • IMG_1662.jpg
    IMG_1662.jpg
    595.8 KB · Views: 382
  • IMG_1667.jpg
    IMG_1667.jpg
    524.4 KB · Views: 387
  • IMG_1663.jpg
    IMG_1663.jpg
    606.6 KB · Views: 382
  • IMG_1664.jpg
    IMG_1664.jpg
    731.5 KB · Views: 372


Many thanks Erik,

As I see the rear of the boards, it leaves the possibility to mount the 4K7 resistors on the rear side. Am I right?

You said it was quite a job to get access to the boards, are they normally inside active speaker cabinets (the white background)? I went a bit back in the thread history but could not find evidence of your construction.
 
I think I gave you the wrong impression - it's not a big problem to access the boards, they are mounted in a simple enclosure, on stand-offs. To add resistors to the potentiometers will entail me removing the boards entirely from the enclosure, so my alternative suggestion was to connect resistors at the easily accessible signal inputs, as a test to check whether this would cure the noise, and if so, I would then proceed to remove the boards and solder the resistors in place on the potentiometers. This was my attempt at a 'quick fix' to test what may work.

You have since explained to me that connecting resistors at the signal inputs will not have the same effect as connecting at the potentiometer outputs, so I am happy to remove the boards and solder the resistors to the potentiometer outputs on the back of the boards, having been guided by yourself as to connection points.
 
I think I gave you the wrong impression - it's not a big problem to access the boards, they are mounted in a simple enclosure, on stand-offs. To add resistors to the potentiometers will entail me removing the boards entirely from the enclosure, so my alternative suggestion was to connect resistors at the easily accessible signal inputs, as a test to check whether this would cure the noise, and if so, I would then proceed to remove the boards and solder the resistors in place on the potentiometers. This was my attempt at a 'quick fix' to test what may work.

You have since explained to me that connecting resistors at the signal inputs will not have the same effect as connecting at the potentiometer outputs, so I am happy to remove the boards and solder the resistors to the potentiometer outputs on the back of the boards, having been guided by yourself as to connection points.

Hi Erik,

I checked on my own TPA3116 mono-board. The small trim-potentiometer used as volume control on that board also increase the volume when you turn clock-wise. For that board and looking at the rear side, the 4K7 resistor has to connect the lower pad/pot.terminal (furthest away from the edge) to the pad/pot.terminal to the right (direction of the power supply terminals).

For the stereo board and as you show the rear side, the pad/pot.terminals to the left have to be connected with two 4K7 resistors to the pad/pot.terminals in the middle. One resistor pointing upwards and one downwards.

Good luck!