I have the following 15 watt amp kit and have a question about it.









Datasheet for the TDA-2005

https://www.st.com/resource/en/datasheet/tda2005.pdf

With a speaker connected upon turn on it puts out -4.5Vdc for about 1-2 seconds then settles to under 100mV and draws about 2A briefly.
No load it is about -11Vdc.

If the cap connected between pin 5 and ground is lower than the recommended value of 2.2uF it makes the amp have a high turn on pop. The stock value of the amp kit is .1uF.

I tried 2.2uF and it didn't reduce the turn on voltage all that much if any.

Looking at the datasheet example circuit I saw that those two caps were the same value.

So I tried a 10uF cap and now the issue is greatly reduced with maybe a 1/2 second dc voltage that is maybe -2Vdc. I can take a voltage measurement tomorrow.

One thing I did to eliminate a phase shift at 20Hz was to increase the cap in series with the 470 ohm resistor to 1,000uF from its stock value of 100uF, but this issue was happening before I made that change.

Also I eliminated the 680pF cap as it caused a phase shift at 20KHz and a slight reduction in output starting maybe about 13KHz. I saw no instability at all measuring the amp with my scope (had to do just one speaker wire and ground), but I know that isn't a true indicator of 100% stability


What I don't get is the datasheet has a ready made example circuit so why did the kit designer use a different circuit unless the different circuit was better in some way shape or form?

Any idea how to fix that issue?

I may try another 10uF cap in parallel with the existing one to see if that further reduces the issue. If not I'll need some sort of fix if one is available.

Also are there any improvements that can be made?

If all else fails I can use a relay, cap and resistor to make a time delay circuit to where the speaker isn't connected until after 2 seconds.

The same company who makes this kit also makes a speaker protection circuit, but one side of the circuit is referenced to ground and unless I use two supplies that are not connected to each other I could not make that protection circuit work.

The amp is flat from at least 20Hz to 20kHz with minimal phase shift at 20Hz and 20kHz. The amp can go out to maybe 50kHz (can take a measurement tomorrow) without the 680pF capacitor. I don't need that high of a response, although the amp seems fully stable.

EDIT:

The heatsink does seem seem get warm in normal operation, but not too warm. I've thought of adding a small fan but that requires a hole in the case and a 12 volt regulator or resistor to power the fan.

Also I didn't post a picture of the bottom of the case, but I put a hole in it so that I could adjust the pot on the board without having to remove the case.

The pot was the original way to control the volume, but I added a separate volume control and that trimpot now sets the maximum level so that the main volume control isn't overly sensitive.

Also I'll do a load test with a 4 ohm load to see how much wattage it actually produces.

I do have another question though.

How can I measure the output of the amp across both speaker terminals when the scope is grounded and the audio generator is also grounded without using an isolation transformer or transformer between the speaker terminals and scope?

EDIT:

I tried a 10uF cap in parallel with the 10uF cap connected to pin 5 and the issue got worse so apparently the cap on pin 1 and the cap on pin 5 must be the same exact value to minimize the turn on DC voltage across the speaker terminals.

I am thinking of adding a relay cap and resistor to form a delayed connection of the speaker. That way no DC will pass to the speaker.

I don't think your schematic is complete. There should be some feedback resisors to return some of the output signal of pin 10 to the inverting inputs of pin 2 and pin 4.

The speaker is connected between the two amp outputs, not from an amp output to ground. Thus the term "bridge" output. With a single polarity supply, both amp outputs operate with a DC bias well above 0v, but both outputs should be at the same value above zero so the speaker does not "see" any of it.


This is a common issue with single polarity supplies since most of the circuit, and the amp's output, operate at a voltage point above 0v. At turn on, it takes a bit of time for the circuit points to settle at their design voltages. This time is altered by capacitors connected to the circuit. They behave as RC filters using circuit resistances as the "R" part of "RC". Altering the "C" simply changes the time of charge/discharge. In other words, the charge cycle that causes and audible "pop" or "thump" at turn on can be spread out over time, hopefully to make it sound like a gentle "whump". The energy is mostly dissipated as heat by the speaker voice coil so it must be considered during design. If the turn-on time of the two bridged amps is not the same, the size and shape of the "pop" will be increased.

The most effective way to reduce or eliminate this is to use a bipolar supply, 6v pos and 6v neg supplies. Assuming they both turn on with near-equal ramp time, most of the capacitors can be eliminated and the "pop" can be relatively small.

Not needed unless you're worried about excessive rf noise present on the audio input.


Good question

Datasheet says the TDA chips include overheat protection so maybe you're okay as it is.

Both outputs should have equal audio amplitudes but of inverted polarity. So, using a grounded 'scope you can put a probe on one or the other speaker output and see a decent representation of what is being applied to the speaker. If your 'scope is dual channel, you can use both inputs and view the entire output using the 'scope's "subtract" or "differential" view mode.

For sinetones, you could simply use your volt meter across the speaker as long as you confine the audio generator's frequency to be within the meter's range. Some meters don't handle frequencies above a few hundred hertz but most modern dmms are accurate with AC up to and above the audio band. Since all meters can read the power line frequency, a 60Hz generator frequency will always be reliable.

I did wind up using a digital scope with two probes and used subtract mode to view the whole signal.

Sure the chip has overheat protection, but I find the heatsinks the kit manufacturer uses for their audio amps to be undersized for the application.

When I test an amp I like to see what is the max it can do into its rated load impedance while maintaining a non-distorted sinewave. Given that is a worst case scenario it causes the most heat to be generated by the output devices. If the heatsink is inadequate it will get hot enough to burn you. Personally I don't like that so if it happens I will either use a larger heatsink or a fan on the existing heatsink to keep the temperature lower.

While testing this amp I thought my load resistors were 8 ohms so I put them in parallel.

Turns out they were 4 ohms each.

However the amp managed to put out 5.95Vrms at nearly 3A on a regulated 15 volt supply across a load of 2.2 ohms for 16.09 watts, but the heatsink got quite hot.

Now using just one 4 ohm load it put out 8.1 Vrms across 4.33 ohms for 15.15 watts.

So I need a toroidal transformer rated at 3 amps, although I am not quite sure the secondary voltage needed. 12 Vrms gives me 16.968Vdc unloaded, but I don't know the voltage it will be loaded.

I'm thinking this transformer will work.

https://www.antekinc.com/as-1212-100va-12v-transformer/

All I'll need is a bridge rectifier and 10,000uF of capacitance.

When I build an amp I have no problem using multiples of 10,000uF capacitors to ensure a very stout power supply.

I'll see if I can find a several amp relay that can be used with a resistor and cap to create about a 2 second turn on delay.

I see why removing the 680pF cap didn't cause any instability. The example circuit in the datasheet shows no cap between the input and ground so it must not have been needed.

The schematic indeed is drawn right as I traced the connections and it is as drawn as connected, however I do know adjusting the value of the 470 ohm resistor alters the gain of the amplifier.

The individual outputs are 11.56 Vpp and 11.09 Vpp, however I see no actual way to adjust the signals to be equal unless it can be adjusted by adding resistance in series with pin 5. I may try that and who knows it might further reduce the DC across the speaker upon turn on.

I've learned the Futurekit audio kits are decent enough, but often benefit from a few improvements.

EDIT:

I will not be rocking out with the amp at 15 watts all the time so I feel a fan is not needed at this time.

Hete's a link to the ST Microelectronics datasheet: https://www.st.com/resource/en/datasheet/tda2005.pdf
Here is their circuit for a 20W bridge amp:

They even give a parts layout of the PCB for it:

As you can see their schematic is quite different from the one you posted.
John

Indeed it is different.

I suppose maybe Futurekit didn't want to pay royalties or whatever has to be paid to use the datasheet design so they did their own design instead.

No royalties involved when using schematic suggestions from a datasheet, indeed the chip manufacturer WANTS you to have a good result with their products.

I don't understand why Futurekit left out feedback components or how the resulting amp works properly. There must be some internal feedback built into the IC , based on its fixed gain of (iirc) 40db, or else the kit would be wildly unstable. I also don't understand how the lower half of the amplifier in Futurekit's design can receive input signal from the upper half's inverting input. It doesn't make sense to me, and the datasheet doesn't explain it to me. STMicro's dircuit looks perfectly normal and I can see the signal path necessary to the lower amp's inverting input. Not so in the circuit you originally posted.

I see how its done.

Here's a couple photos.





Pins 1 and 5 each go to the base of a transistor and pins 2 and 4 go to the same transistor emitters.

Yup, there it is, obvious as all get out. I just didn't see it before. Thanks!

You're welcome.

I didn't see it either until you mentioned it and i looked more in depth at it.

Here's a photo with the fan installed.

A 100 ohm series resistor will be used with the fan.



There's three existing holes where other things were mounted when it was used for a different project so the air will be drawn in through them. If the airflow isn't enough I'll drill more holes.

The plan for the power supply is to install a bridge rectifier and at least two 10,000uF caps in the amp case and put the power transformer in a computer power supply case as it has an IEC power jack so that I can easily replace the cord if need be. I'll also add a power switch to the case.

I'm tempted to install the caps and rectifier in the computer power supply case, but I don't want to do that because it would be possible to accidentally connect the wiring in reverse and screw up the amp, whereas with AC being fed to the amp that won't be possible as it won't matter how the wiring is connected.

I'll also experiment with a series resistor added between the cap and pin 5 to see if that will make both outputs equal in amplitude as that's the only way I see of doing it. The caps I used for replacements are axial as that's all I had so I can cut the exposed lead, wire in a resistance substitution box and see what resistance is needed. The substitution box I have at work is in a metal case which has its own ground terminal so that the box is effectively shielded.

I used that with the amp putting it in parallel with the 470 ohm resistor seeing if that would alter the output amplitude to allow me to perfectly balance both sections, but all it did was alter the gain. The amp went haywire until I connected the substitution box ground terminal to ground and then the amp was perfectly normal.

Now I'm tempted to put a 1K variable in series with the 470 ohm resistor or perhaps some lower value than the 470 ohm resistor as a way to adjust the gain instead of having the trimpot after the volume control, but I don't want to risk causing instability by doing so plus with the 10K trimpot after the main volume control I can adjust it so that the amp could be driven from pretty much any signal level greater than line level such as the speaker level output of an amp without the volume control being overly sensitive.

I am quite pleased with the amp, however if I ever need to buy another of these amp kits I'll find a larger heatsink to use with it.

Have you considered a protection diode in series with the positive supply lead? You'd install it inside the amp case. So if the plus/minus connections to the case are reversed the diode won't conduct and the amp is protected. Only downside is 0.6v drop across the diode. Worth it in my view. A Schottky diode would drop only half that much. Pay attention to current draw through the diode, IN400x could probably handle normal use and should handle the high current draw at turn on.

I have thought of that and might do that if there's enough room to put the caps and bridge rectifier inside the power supply case I'm gonna use for the power transformer.

EDIT:

The fan seems to be keeping the heatsink a good bit cooler under full output wattage.

Also adding resistance in series with the cap on pin 5 did nothing.

I see why the datasheet circuit may have been the better circuit.

It allows for the outputs to be trimmed by the feedback resistor values so that they are equal when measured to ground, whereas the kit circuit is at the mercy of the chip's tolerances as to how equal the outputs are.

However, the kit circuit's outputs are close enough to being equal that I'm not too worried about it. Plus it does do the claimed 15 watts into 4 ohms so I'm satisfied.

Plus the gain wouldn't be as high, which you probably don't need and are using a 2 volume pots to deall with. I suppose the kit was designed for those who avoid feedback under all circustances because they say it sounds better. Glad you've got it working to your liking!

The gain is proper for a line level input far as I know. The 470 ohm resistor sets the gain so altering that resistor value increases or decreases the gain.

I just chose to use the two pots for dealing with higher than line level signals should I ever feed a higher than line level signal into the amp plus it's a good way to set the maximum output level of the amp for a given input signal voltage.

I'm thinking the trimpot is almost set to full output anyways.

The reason for the trimpot and not a trimpot in place of the 470 ohm resistor is because if for instance I needed this amp to be fed from a speaker level signal I could adjust the trimpot lower to where the main volume control isn't overly sensitive, whereas merely lowering the gain wouldn't necessarily work as the signal may be too high and cause the amp to distort.

Regarding your toroidal transformer, when calculating the power supply's no-load voltage you need to also factor in the forward voltage drop of each rectifier diode. As your amplifier apparently uses a single voltage (vs an amp using bipolar power, e.g + and - supply rails) you will no doubt be using a full wave bridge rectifier. Therefore the forward voltage drop of the type diode you use will be the drop of one diode X4. This would be a factor if you need to extract as much output power from the amp as possible.
Most transformers intended to be used for audio amplifier D.C. power and also some intended for use in power supplies have a secondary output voltage above that of the desired rail voltage(s) produced after rectification, and after integration by the filter capacitor(s).
I think the basic calculation would be:
The transformer output voltage for a given DC no-load rail voltage would equal:
The desired DC rail voltage X .707 + the rectifiers' voltage drops.
If your amp's voltage need is non-critical, the toss all of that out the window and find something close.

For now I'm going to use a 12 volt transformer unless I can use a 15 volt transformer and remove some of the winding until I get the desired voltage under load.

I could also use the 12 volt transformer and add a winding to develop 15Vdc at full output.

However I'll try the 12 volt transformer first just to see how well it does.

I know that I will not be running the amp at a continuous 15 watts and average power may be about 5-7 watts.