In that schematic, inputs are looped through on the master board from K1 to K2, then a cable with SGND only, from K2 to K1 on the slave board. If you don't use that second cable then the master - slave SGNDs connects via the power GND and the two loop breaker resistors. Probably not causing the DC but it would cause hum.
Do you have a link to the original german article?
Do you have a link to the original german article?
Last edited:
http://schematic.narod.ru/Audio/semi/Amp/a100-daten-301.pdf
this is revision 1, revison 2, then downrated to 80W you won't find anymore.
I checked, it makes its way back to signal ground. So in slave input is shortened to SGND and inverting iput gets output of master through a feedback resistor.
Seems something weird happens after they slave boards have been switched on a few times. The other boards have 0 V on their outputs, regardless if the PSU caps have discharged, mute is on or off and power is on or off.
But the slave board is at -4V this morning. The PSU caps are at +1 and -14V.
this is revision 1, revison 2, then downrated to 80W you won't find anymore.
I checked, it makes its way back to signal ground. So in slave input is shortened to SGND and inverting iput gets output of master through a feedback resistor.
Seems something weird happens after they slave boards have been switched on a few times. The other boards have 0 V on their outputs, regardless if the PSU caps have discharged, mute is on or off and power is on or off.
But the slave board is at -4V this morning. The PSU caps are at +1 and -14V.
On your board, you connect slave K2(2) to master K4(2), set jumper JP1 (1-2) on both boards and jumper on K6-IN (1-2) of the slave board.
You still need to connect the input ground to K6-IN of the slave board to get it to work correctly.
You still need to connect the input ground to K6-IN of the slave board to get it to work correctly.
JP1 will disappear, its always same setting.
When shortening K6-IN, IN+ pin 3 is connected to SGND pin 4. At the moment, ok, there are 2.2 ohm between them. But it did not work before, when SGND did not have the 2.2 ohm to PGND. Something is not right. Shortened out the resistor between SGND and {GND, same result. Maybe it blows up a diode or the mute internals 1sttime I use the slave as slave. 1st time I switch it on, there is no plop, but after a few times there is one. Now, when 240V power is disconnected, after a few minutes, the output of the slave is -4V, all others is at 0V. Switching mains on,but mute of, after a minute or so its at 0V, then slowly goes up to 200mV. All the others stay at 0V. Cannot see any blown up part in that mute circuit.
When shortening K6-IN, IN+ pin 3 is connected to SGND pin 4. At the moment, ok, there are 2.2 ohm between them. But it did not work before, when SGND did not have the 2.2 ohm to PGND. Something is not right. Shortened out the resistor between SGND and {GND, same result. Maybe it blows up a diode or the mute internals 1sttime I use the slave as slave. 1st time I switch it on, there is no plop, but after a few times there is one. Now, when 240V power is disconnected, after a few minutes, the output of the slave is -4V, all others is at 0V. Switching mains on,but mute of, after a minute or so its at 0V, then slowly goes up to 200mV. All the others stay at 0V. Cannot see any blown up part in that mute circuit.
Does that mean that there is -4V between the SGNDs of the two amps
? Are you measuring this relative to Power GND?
? Are you measuring this relative to Power GND?
relative to power ground, essentially its -4 V all the way from output to inverting input, its all at the same level
I measured it. If both inputs have the same signal at the input, it would cancel out and the result is 0. So the bridge amp works by making an mirror image of the same waveform, then the speaker sees twice the voltage across.
At present the amplification is about 40 times, so to have -4V out there should be -40mV at the input.
When I switch power on, input and output voltage slowly creep up, the input has 10mV less. I.e. input -100mV, output -110mV
Should make 2 pins on the PCB layout so its easier to check what the TDA iputs see.
At present the amplification is about 40 times, so to have -4V out there should be -40mV at the input.
When I switch power on, input and output voltage slowly creep up, the input has 10mV less. I.e. input -100mV, output -110mV
Should make 2 pins on the PCB layout so its easier to check what the TDA iputs see.
Last edited:
So this is like the feedback is open, it would amplify until it nearly reaches the supply voltage, that is what it does when I swithch off mute, it reaches -20V for a split second, the driver membrane nearly falls out.Then it goes to 10mV or so and its ok. So if the PSU caps are +0.6V and -15V, the TDA sits at -4V, so whatever supply voltage it gets, it can reach that -10 or so V.
I guess there should be some capacitor at the negative input, otherwise it gets whatever the master out put does when mute is switched off.
Spec sheet says signal inputs can handle 90V, mute/standby inputs 120V. I doubt it.
I guess there should be some capacitor at the negative input, otherwise it gets whatever the master out put does when mute is switched off.
Spec sheet says signal inputs can handle 90V, mute/standby inputs 120V. I doubt it.
Last edited:
the bit where it tries to keep same voltage at both inputs? Only works if it gets symmetrical supply? If it gets +0.6 and -15V it would do something else.Something is not right in there. The "standard" boards are at 0V at the output even when they get +0.6 and -15V on their supplies. Mute does not switch it off completely, just 60dB less.
Measured again, Pin 2 and pin 14 at -4V, as inverting amp it should be +40mV in and -4V out. The PSU is at 30Vin less than 1 s, pin 14 jumps from -4 to -140mV and then slowly creeps up. Pin 2 is at the same voltage as pin 2 + 10mV, so it would be at -130mV then.
The standard boards stay at 0V all the time.
edit 1 Put pin 2 onto ground, pin 14 slowly went to 0V.
edit 2 took jumper off, pin 2 open again, pin 14 stays at 0V. So its C3 I guess. When its discharged, its not putting a negative voltage into pin 2.
Rev 1 put the 2 elcos between pin 14 and pin 2, Rev 2 put the elcos between pin 2 and GND, so one elco would discharge into pin 2.
Measured again, Pin 2 and pin 14 at -4V, as inverting amp it should be +40mV in and -4V out. The PSU is at 30Vin less than 1 s, pin 14 jumps from -4 to -140mV and then slowly creeps up. Pin 2 is at the same voltage as pin 2 + 10mV, so it would be at -130mV then.
The standard boards stay at 0V all the time.
edit 1 Put pin 2 onto ground, pin 14 slowly went to 0V.
edit 2 took jumper off, pin 2 open again, pin 14 stays at 0V. So its C3 I guess. When its discharged, its not putting a negative voltage into pin 2.
Rev 1 put the 2 elcos between pin 14 and pin 2, Rev 2 put the elcos between pin 2 and GND, so one elco would discharge into pin 2.
Last edited:
The ST datasheet just has a small onewith positive side facing pin 2. There is something, the bigger these caps, the better the low frequency performance.
in the video: at 35min there is something how to reject common mode currents on the cheap.
in the video: at 35min there is something how to reject common mode currents on the cheap.
No, it would still try to set the feedback input to the same voltage as the other input. The problem is that the output has to stay within the rail voltages for the feedback to work.
Master pin2 = pin3 = input signal.
Master pin14 = input signal x gain.
Master sgnd = pin2 = pin3 = 0V.
slave sgnd = 0V.
Slave pin14 = -input signal x gain.
On both boards Pins 1+4 = 0V.
Here they do something similar. They have set a jumper over the caps.
diy-spot: Power Amplifier bridge 180W or stereo 2 x 80 Watts
Last edited:
took one cap out, the one with the negative facing pin 2. Now its to -2.5V. Ok, will take the other one out as well.
took all caps out, now there is just a small amount of noise, like air escaping. Will try tomorrow, if it does music. Still think there should atleas be a capacitor somewhere to block DC from entering the inverting input. I know it should not be there, because its been filtered by the master.
Case look at the top pic in post36.
The two input filters are set to
Low Pass: 390r & 180nF giving an F-3dB ~2.3kHz
High Pass: 2u2F & 22k giving an F-3dB ~3.3Hz
Now look at your gain plot
You have a marker @ ~3.3Hz where the gain is down by ~3dB and you have another marker @ ~2.3kHz where the gain is down ~3dB
The two input filters are setting the passband of the amplifier.
Look at the NFB,
the gain is set by the ratio of R3:R2
gain = {22k/680r} +1 = 33.35times = +30.46dB
This agrees with the four middle range markers sitting roughly on the +30dB level.
Now consider an input at any frequency on that flat line of +30.46dB
input at Pin +IN = 100mV
The output must be at 33.35times 100mV = 3.335V
Look at the resistor divider consisting of 22k, two capacitors with an impedance of zero and 680r.
The voltage at pin -IN = 680r/{22k+capimpedance+680r}*Vout = 680/{22680}*3.335V = 0.0999918 the voltage difference at the two input pins is ~ {0.1V - 0.0999918V} = 0.0000088V = ~9uV for an input of 100mV the difference between the two inputs is set by the output voltage divided by the opamp's open loop gain and ends up being ultra low (almost zero).
The open loop gain is Vout/Vin diff = 3.335V/0.0000088V = 378977times = +111.6dB
This is very approximate but gives rough numbers that you may be able to measure, (clearly you can't measure 0.0000088V).
Now look at the middle pic.
You have changed the DC blocking cap from 470uF to 22uF.
The roll-off created by this new value has an F-3dB ~ 10.6Hz
The marker in the third plot @ ~10Hz is indeed sitting ~3dB down.
You have screwed the performance of the Power Amplifier by choosing a DC blocking capacitor that is far too small. It now has very significant AC voltage across it, whereas the assumption I made earlier for pic1 was "two capacitors with an impedance of zero"
The minimum value for the DC blocking capacitor is:
C2 >= sqrt(2) * C1 *R1 / R2 >= 1.414*2u2F*22k/680r >= 100.7uF
use 120uF or 150uF or 220uF or 330uF or 470uF
Don't use 22uF !
And put back in your two input filters to define the pass band you require.
The two input filters are set to
Low Pass: 390r & 180nF giving an F-3dB ~2.3kHz
High Pass: 2u2F & 22k giving an F-3dB ~3.3Hz
Now look at your gain plot
You have a marker @ ~3.3Hz where the gain is down by ~3dB and you have another marker @ ~2.3kHz where the gain is down ~3dB
The two input filters are setting the passband of the amplifier.
Look at the NFB,
the gain is set by the ratio of R3:R2
gain = {22k/680r} +1 = 33.35times = +30.46dB
This agrees with the four middle range markers sitting roughly on the +30dB level.
Now consider an input at any frequency on that flat line of +30.46dB
input at Pin +IN = 100mV
The output must be at 33.35times 100mV = 3.335V
Look at the resistor divider consisting of 22k, two capacitors with an impedance of zero and 680r.
The voltage at pin -IN = 680r/{22k+capimpedance+680r}*Vout = 680/{22680}*3.335V = 0.0999918 the voltage difference at the two input pins is ~ {0.1V - 0.0999918V} = 0.0000088V = ~9uV for an input of 100mV the difference between the two inputs is set by the output voltage divided by the opamp's open loop gain and ends up being ultra low (almost zero).
The open loop gain is Vout/Vin diff = 3.335V/0.0000088V = 378977times = +111.6dB
This is very approximate but gives rough numbers that you may be able to measure, (clearly you can't measure 0.0000088V).
Now look at the middle pic.
You have changed the DC blocking cap from 470uF to 22uF.
The roll-off created by this new value has an F-3dB ~ 10.6Hz
The marker in the third plot @ ~10Hz is indeed sitting ~3dB down.
You have screwed the performance of the Power Amplifier by choosing a DC blocking capacitor that is far too small. It now has very significant AC voltage across it, whereas the assumption I made earlier for pic1 was "two capacitors with an impedance of zero"
The minimum value for the DC blocking capacitor is:
C2 >= sqrt(2) * C1 *R1 / R2 >= 1.414*2u2F*22k/680r >= 100.7uF
use 120uF or 150uF or 220uF or 330uF or 470uF
Don't use 22uF !
And put back in your two input filters to define the pass band you require.
Last edited:
- Status
- Not open for further replies.