12V DC gainclone?

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Alright, I have done a horrible job of explaining myself. Perhaps if I draw what I am thinking somebody can correct my errors.

I think it's safe to say I am 100% lost on the TDA2040.

both schematics are right, it will work...
just one thing with the tda2040 used in single supply mode,
you must add two 1N4004 diodes connected in reverse mode
from the output pins of the IC to positive rail and ground...
here a schematic that will help....
 

Attachments

  • TDA2040 protection diodes.zip
    2.2 KB · Views: 77
Well, I bought this "significant wall wart plug" for $20.
Toshiba AC Power Adapter Cord for PA2521U-3ACA 15V 6A 9 - eBay (item 170421661028 end time Feb-16-10 02:34:36 PST)

The 15 x 6 = 90va, and that seems workable. Except, that is a switch mode, which is a frail technology, so a good size tank capacitor will be needed as a helper. Its a slightly dodgy prospect, but it was quite inexpensive.
Perhaps it will also be nicely clean power? We'll find out.

The voltage is too high for TDA1558Q; but I couldn't find 14.4v available. Almost all car amplifiers can run on 14.8 volts DC (a car actively trying to charge its battery), so I'm hoping at 15v regulated isn't too much for it.
I'm wondering if I can use diode drop or if that would make additional noise?

For durability purposes, I have 2 chips, so they can run 1 chip per speaker (parallel mode). I'd planned it this way because the thermal interface on that chip appears to be insufficient, due to its small size. This trouble is because I have 4 ohm speakers and have also selected too much voltage (for that size thermal interface).

don t worry daniel, i use a 16.4V switch mode power supply with the 1558,
and it work very well..
anyway, the max power supply is 18v, although i wouldn t go further than the
voltage i m using...
the circuit is made to work with 3.2 ohm loads in bridge mode...
the thermal surface is largely enough, provided you have a decent
heatspreader, i used one that was not very big, and it made no problems.
 
don t worry daniel, i use a 16.4V switch mode power supply with the 1558,
and it work very well..
anyway, the max power supply is 18v, although i wouldn t go further than the
voltage i m using...
the circuit is made to work with 3.2 ohm loads in bridge mode...
the thermal surface is largely enough, provided you have a decent
heatspreader, i used one that was not very big, and it made no problems.

Did you test yours with 4 ohm speakers?

The amplifier described is either incredibly efficient or its hooked to 8 ohm speakers. ;)
 
i tested with 3.2 ohm loads at full power,
actually it s used with 4 ohm speakers...
the TDP depend not on the surface of the ic METAL
contact, but firstly fom the thermal resistance between
the chip and the metal surface of the ic.....
just don t forget some thermal past for better heatspriding...
anyway, try, you won t be deceived..
 
Bad design

The TDA1558Q is really misbehaving for me.

I build the bridge design from the datasheet.

It seems like something is missing. Input ground seems missing. It has an input ground marked but that doesn't seem to connect to anything whatsoever, according to the multimeter.

With one input channel plugged in there's no output.

With the other input channel plugged in, there's a tiny output plus a great deal of noise.

With both input channels plugged in (stereo) it plays decently on both speakers, but unfortunately. . .

In any condition, the datasheet design draws more than 12 ampers, and my diodes start smoking.

Help!
 
The TDA1558Q is really misbehaving for me.

I build the bridge design from the datasheet.

It seems like something is missing. Input ground seems missing. It has an input ground marked but that doesn't seem to connect to anything whatsoever, according to the multimeter.

With one input channel plugged in there's no output.

With the other input channel plugged in, there's a tiny output plus a great deal of noise.

With both input channels plugged in (stereo) it plays decently on both speakers, but unfortunately. . .

In any condition, the datasheet design draws more than 12 ampers, and my diodes start smoking.

Help!

the input ground must be connected to the other grounds,
not only to the source ground...
 
Okay, have mute to v+ via a 10k resistor. The mute switch didn't need to pull full current--it only needs voltage. Stupid datasheet. lol!

It sure does run cool now!

Next problem:
There's a little buzz
Now trying to find out how to impose an input load without also increasing noise. Any attempt to put a load between ground and input makes a noise. darnit! Can't use a potentiometer?

However, pin15 with about 2k to input does make an input load (amp goes silent and I had to increase the input cap size too). Pin 15 doesn't work for input ground though. OH, this is one weird little chip!
 
pin15 is to be let unused...
you can add a 100uF cap from pin4 to ground to improve
the supply ripple rejection...at first , i didn t add it, and it
also worked...
the little buzz is surely due to a ground loop from the
earth connection of the switch mode supply to the earth
connecction of your source if this latter has also a earth
connection...
i use a potentiometer at the input,
as the amp has high gain, it s mandatory..
 
both schematics are right, it will work...
just one thing with the tda2040 used in single supply mode,
you must add two 1N4004 diodes connected in reverse mode
from the output pins of the IC to positive rail and ground...
here a schematic that will help....

I am not sure I understand. There's a bunch of stuff from the data sheet missing, like C5, C3, R1, and R5. I assume your C1 and R1 replace the data sheet's C6 and R6. Also not sure what the TLO72 is mentioned for.

Am I eliminating the R5 connection between pins 4 and 2; as well as eliminating the C3 and R1 connection between pins 5 and 1; and eliminating C5? Because that's what I got out of it.
 
the input ground must be connected to the other grounds,
not only to the source ground...

So the signal grounds connect to pin 3... and pin 3 also connects to the other grounds?

Here a picture of what I understanding.

By the way, did I actually get polarity or phase of speaker correct?
 

Attachments

  • TDA1558Q2.JPG
    TDA1558Q2.JPG
    46.5 KB · Views: 431
So I need a 10k resistor between pins 14 and 5? In addition to a switch?

You don't need to add a switch.

Well the amplifier draws less current when you connect pins 14 (mute) and 13 (v+) together with a 10k resistor.

Typically, you'd also cause the 10k resistor to charge up a capacitor (voltage goes through the resistor in order to charge a cap), size from 10uF to 100uF (larger value is longer time delay), and the amplifier won't switch fully on until the cap is charged up. Its a time delay, typically called "soft start".
That's also a luxury item that serves as a sort of speaker protection--the delay stops the pops and clunks when you power-up the amp.

Since pin 14 was meant for this sort of circuit and/or to operate from a digital controller unit, this fact is what gave me the idea to use the 10k resistor (to decrease the current draw of the chip's inbuilt switch).

I have not yet connected a capacitor between pin 14 and ground for time delay start, have not yet determined that 10k is the most appropriate value for mute powering, and have not yet gotten the amp up to high fidelity performance. Currently, my own TDA1558Q sounds exactly like a car radio and this is not impressive. So, bear that in mind--I'm just experimenting at this point.

.

Here's the redo build:
(the main idea here is to get the amp into "expected operation" to create a working baseline)

I'm thinking to bend pins 5 and 13 out just far enough to create a "3rd row" and solder a sold wire across them--makes V+ rail.

Now that they're out of the way, that also allows to solder a sold wire across pins 3, 7, 9, 11--makes ground rail.

These two "rails" can be held apart by soldering a 100nF capacitor between them (as shown in the datasheet).

Next step, connect the speaker wires--the spacing of 20ga zip cord is remarkably convenient as it is the same as the pin spacing (thus the pins don't accidentally touch).

Then I have to figure out the most appropriate value for the resistor between V+ and mute (range probably 1k to 22k, minimum and maximum values).
After that its probably good to connect a 100uF cap from pin 4 to ground, and an additional cap (range is 10uF to 100uF) from pin 14 to ground (when in combination with the resistor, it creates the expected "soft start" function).

After these steps to get the amplifier into "expected operation" the next thing is to figure out the mystery of Pin15 (which can be used to remove noise, completely remove DC offset and several other uses), and fill in the input circuit (such as how to make a load without also making a noise), and fill in the rest of the power circuit.

Wish me luck on the redo!
If anyone has any comments on how to do this better, please let me know!
 
Last edited:
Pin 15, the noisebuster

So far, I found only one thing to do with pin 15, and it has been useful.

I wanted a stout 10k load internal to the amplifier to bust the noise. Well, 15 is the centerpoint voltage internal to the amplifier. Pin 15 = every input that isn't provided on pins 1,2, 16, 17

So, I put two 10k resistors on pin 15--intending one load for left channel and one load for right channel.
-and-
I joined pins 1+2, and also joined pins 16+17 as seen in the datasheet for bridged amplifier.
-and-
As for the 10k resistors that are hooked to pin 15, they are connected, one of the resistors to pin 1+2, and the other resistor to pin 16+17.

This did increase the input load.
It was not completely effective in removing the "guitar amp type noise" but it did reduce noise dramatically.
It also changed the DC offset (as expected) from the previous 52.1mv to 00.6mv. My meter cannot read reliably at 00.6mv, so it may be zero DC offset.

I had to change input filter capacitors (located in-between the potentiometer and the amplifier) to correspond to 10k. I used 47uF//1uF electrolytic (per each right and left). This capacitor size is overlarge, but I'm planning to experiment with 5k input load. 1k is too much (bad tone) and 10k isn't quite enough (not completely silent during no input).

The amplifier is performing much more clearly and the efficiency (45 watts barely warming a 3" heatsink) and power output are prodigious indeed! Still working on sound quality (s/n ratio and stereo seperation) but made major progress with pin 15.

Curiosity problem: This isn't a class AB amplifier (impossible efficiency and output per voltage). So, what is it?
 
Sorry to hear that. Really sorry to hear that.

I hope your experimentation solves the problem. As I will likely follow in your footsteps in such a case.

Ah, the pin 15 mod brought the entertainment level up considerably. Its clear as can be, no longer sounding quite so much like a car amp. It is as Wahab reported, clear, detailed and quite lively.

All extreme gain amplifiers are lively and they all suffer a bit of a challenge for noise control. TDA1558Q is not an exception to that. In trade for the bother (extra design challenge), you get a sort of "uncompressor" feature.

Currently, I'm playing with:
1). Pin 15's centerpoint voltage employed (plus 2 resistors, left and right) as an all-bandwidth load upon the inputs. This removes noise at Inside the amplifier and reduces DC offset to zero. Need to find an optimized resistor value.
2). A potentiometer ( also called "pot" "L-pad" "voltage divider") located in-between the input filter caps and the RCA jack. The potentiometer doesn't place any load on dc but it does decrease the level of input input (ac audio), and that also reduces noise (and everything).
Wahab is indeed correct that a potentiometer volume control is necessary because the amplifier's gain is considerably excessive. Need to find an optimized potentiometer load value.
3). Soft start function at the mute pin for about 1/4 second timing. This is mute to v+ with 10k -and- also mute to ground with 22uF.

After this, I think we can have some sketches and photography. :) Perhaps Wahab can post some construction photos as well? That would be nice to see.
 
So far, I found only one thing to do with pin 15, and it has been useful.

I wanted a stout 10k load internal to the amplifier to bust the noise. Well, 15 is the centerpoint voltage internal to the amplifier. Pin 15 = every input that isn't provided on pins 1,2, 16, 17

So, I put two 10k resistors on pin 15--intending one load for left channel and one load for right channel.
-and-
I joined pins 1+2, and also joined pins 16+17 as seen in the datasheet for bridged amplifier.
-and-
As for the 10k resistors that are hooked to pin 15, they are connected, one of the resistors to pin 1+2, and the other resistor to pin 16+17.

This did increase the input load.
It was not completely effective in removing the "guitar amp type noise" but it did reduce noise dramatically.
It also changed the DC offset (as expected) from the previous 52.1mv to 00.6mv. My meter cannot read reliably at 00.6mv, so it may be zero DC offset.

I had to change input filter capacitors (located in-between the potentiometer and the amplifier) to correspond to 10k. I used 47uF//1uF electrolytic (per each right and left). This capacitor size is overlarge, but I'm planning to experiment with 5k input load. 1k is too much (bad tone) and 10k isn't quite enough (not completely silent during no input).

The amplifier is performing much more clearly and the efficiency (45 watts barely warming a 3" heatsink) and power output are prodigious indeed! Still working on sound quality (s/n ratio and stereo seperation) but made major progress with pin 15.

Curiosity problem: This isn't a class AB amplifier (impossible efficiency and output per voltage). So, what is it?


great work, daniel, you modified the thing considerably...
it s a classic class AB amp, with each amp draining about
25mA at iddle...
it s true that the gain is too large, at 46 db for the bridged version,
which make use of an input potentiometer mandatory...
this wasn t missed by phillips that did reduce the gain to
26 db in the successor of this circuit, the TDA8561Q...
i used this one because i had some left, and the output
level of a laptop is not very high, so i retained a high gain circuit
after having exausted a ten different circuits, this one seemed the
most practical...
anyway, good tries, i ll publish some photo of the proto soon...
 
Looking at the specs of that one.. I sort of wish I grabbed that one. Not a huge difference between the two. But I would trade THD for a watt or two any day of the week. Maybe I should pick that one up.

For example a 90db @ 1 watt efficiency rating on a speaker:
90db@1w
93db@2w
96db@4w
99db@8w
102db@16w
105db@32w
108db@64w
111db@128w
114db@256w
117db@512w
120db@1024w
You get 3 more decibels every time you double the amplifier power.

Compare the 16w amplifier with the sound you want to crank up -versus- the 64w amplifier with the sound you'd never want to hear again. Only one of these options is useful, so lets make clean amplifiers. :)
 
Last edited:
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.