LM1875 with NE5532 low frequency problem

[gekkehenk]

I have the following amplifier. Its a Lm1875 amp with a NE5532 opamp. I've changed the opamp for a ST4558C

The problem is that it is holding back at low frequenties. The power is there but it doesn't put it into the speakers. The LM1875 frequency response is somewere from 10Hz -3dB so I asume that's not the problem.

Is there a problem in the schematic? Can I change anything to make it better?



Thnx!

[jaycee]

I'd suggest some resistors inline with your DC blocking caps, 1K should do.

If you are using a volume control, where is it connected?

I suspect your problem is that you need more than 4700uF in the power supply for 2 channels. If you put another pair of 4700uF it should improve. Also, what VA rating is your PSU transformer ? I'd go for at least 160VA.

Oh, why did you swap NE5532 for a 4558? The 4558 is a dreadful opamp. Put the 5532 back.

[gekkehenk]

Originally Posted by jaycee

I'd suggest some resistors inline with your DC blocking caps, 1K should do.

If you are using a volume control, where is it connected?

I suspect your problem is that you need more than 4700uF in the power supply for 2 channels. If you put another pair of 4700uF it should improve. Also, what VA rating is your PSU transformer ? I'd go for at least 160VA.

Oh, why did you swap NE5532 for a 4558? The 4558 is a dreadful opamp. Put the 5532 back.

No volume control, I use a Mitsubishi DA-P10 preamp.

The amplifier uses 4x 4700uF capacitors for both amps and a 120VA transformer 2x18volts

Allright, tomorrow I'll change the opamp and listen to the difference. Otherwise I'll buy and try a Burr Brown OPA2604.

I'd suggest some resistors inline with your DC blocking caps, 1K should do. What do you mean by that? My english isn't that good, sorry. You free to use the schematic and change it over there

[infinia]

Nothing wrong with the schematic, BW of U1 should be 1.5 Hz from 1/ (2*PI*R17*C23). Maybe try using bigger/better 220uF caps & using low ESR for C25-28, would be my only subjective suggestion.
Maybe it's not living up to a discrete design, or compared to what you normally use?
Image of the actual circut board would be helpful.

[jaycee]

OK, if you have a preamp already, then i would remove the op amp circuit completely. You have enough gain in your preamp that there is no point in having more in the amp itself, it will only add noise.

Connect your preamp to the +ve terminals of C21/22, and remove the opamp. You might then want to put the 200pF capacitors across R7 and R8 for RF suppression.

[AndrewT]

Hi,
you have a gain of 20/1+1 for each power amp.
Balanced doubles this gain to 42.
You have a further gain of 20/10+1 in the opamps giving a total gain of 126 (+42dB).
Why?

The High pass filter at the input is ~1000ms.
after the opamps it is ~1030ms.
in the NFB it is ~100ms.
in the PSU it is ~4700uF*2*4ohm~38ms.
This needs a lot more thought.
In fact it is just about back to front.

Could this "Mitsubishi DA-P10 preamp" be making the loss worse?

How many chipamps are powered by the +-9400uF? two or four?

[infinia]

Voltage gain end to end (total) Av = 20LOG(3*21) = 36 dB
Why is this balanced? this requires one channel to be inverted, which is clearly not the case here, and because the schematic is also marked L and R.

Gain is a bit high but not unreasonable Av = Voutput/Vin
solve for Vin = 40/(3*21) = 0.64 Vpk-pk

[djoffe]

Assume every capacitor has the potential for some blame...are there some caps we could remove?

C21 and C22 serve very little purpose. The input stage is already DC blocked by C1 and C2, and the DC gain of the input stage is unity, based upon C3 and C4, so the DC voltage at the output of the input stage will be a few millivolts.

Thus, you replace C21 and C22 by short circuits, with very little effect on the DC operating point, and in the process, do an experiment to eliminate a potential source of your problem. Note that if you do replace C21 and C22 by shorts, that you could also remove R7 and R8.

Please let us all know how this experiment turns out...Even if there is a null result, perhaps there are some other capacitor changes that might help...that is, we might have picked out a problem, but it might not be the worst of the problems hurting your sound.

[pacificblue]

Originally Posted by gekkehenk

The power is there but it doesn't put it into the speakers.

What exactly does that mean?

Could the speakers or their set-up be the culprits? Is AndrewT's assumption correct that you use a pair of 8 Ohm speakers?

Originally Posted by gekkehenk

The LM1875 frequency response is somewere from 10Hz -3dB

1,8 Hz.

Originally Posted by AndrewT

Hi,
you have a gain of 20/1+1 for each power amp.
Balanced doubles this gain to 42.
You have a further gain of 20/10+1 in the opamps giving a total gain of 126 (+42dB).

Originally Posted by infinia

Voltage gain end to end (total) Av = 20LOG(3*21) = 36 dB

Interesting. I see a gain of 3 around U1 and 21 around U4 and U5, which should add up to 24 times or 27,6 dB. Do we all see the same schematic?

Originally Posted by AndrewT

The High pass filter at the input is ~1000ms.
after the opamps it is ~1030ms.
in the NFB it is ~100ms.
in the PSU it is ~4700uF*2*4ohm~38ms.
This needs a lot more thought.

Definitely. And the NFB around the op amps is 220 ms. The good news is that all filters, except the PSU have low corner frequenies, which should push most of the bad effects below the audio range.

First priority should be to check the speakers and their positioning. Second to try with more power supply capacitance. And to test the influence of the pre-amplifier, you could replace it with a potentiometer for a while.

[AndrewT]

Originally Posted by pacificblue

I see a gain of 3 around U1 and 21 around U4 and U5, which should add up to 24 times or 27,6 dB. Do we all see the same schematic?

gain of 3times (U1 = +9.5dB) and gain of 21times (U4 or U5 = +26.4dB) and gain of 2times (U4+U5 balanced doubles the drive voltage seen at the load = +6dB).
Total gain = 3*21*2 = 126times (9.5dB + 26.4dB + 6dB = +42dB) this will clip @ 28Vpk when the signal is 157mVac.
This gain is completely inappropriate for line level signal into a power amplifier.