Check this schematic please

[CJ900RR]

So i have made this schematic for a LM3875. Well most of it is the original National schematic but... I have some questions regarding R3, R4 and C6.

I understand every other component but how do I calculate this three? Yes I have read the datasheet and I know the formula to use, but how mutch GAIN do I want for example.

I feel a little stupid because i know that this is basic knowledge. But anyhow, if someone would like to explain this for me as you would explain it for a child, then I would be happy

Thank's.

[Mick_F]

R3 and R4 set the gain of the amplifier. The gain is calculated by G=1+R3/R4. If you want to use the amp as an integrated , i.e. without a pre, you should choose a gain of approx. 33. If you use it as a power amp with a pre, you can keep the gain around 20.

C6 is there to lower the DC offset on the output. DC coming into the amp see an infinite resistance to ground due to the presence of C6, and as a result the gain for DC is 1. I don't think you need to calculate a particular value here, just use something in the range of what's recommended in the datasheet.

Mick

[CJ900RR]

Danke!

That cleared it up

So if I change R4 on the schematic to 680 ohm, it would give me a GAIN of somewhere around 29,41? Is that correct?

[AndrewT]

Hi Cj,
well almost, remember to add 1.
20k/0k68 + 1 = 30.4 = +29.7db.

PA equipment tends to run with power amp gains in the range +20db to +28db. Helps keep out noise when LONG cable runs are used and when equipment is used in a hostile environment.

Domestic equipment is generally more suited to a higher power amp gain, ranging from +30db to +36db.
This better suits some of the low level sources available and allows opamp chips running on +-12V to develop sufficent output voltage (without risking getting into the distortion areas that can ensue) to drive the higher gain equipment.

The RC time constant of the Negative Feedback Loop (NFB) determines the bass response of your amplifier.

You generally want a bass response down to 20Hz, but to allow this to play properly I recommend a high pass roll-off a full decade below this i.e. 2Hz. This requires a time constant on the input filter of 80mS and of the NFB of 110mS (0.11Seconds) or both could be a little longer.

To achieve 110mS for the NFB using 680r for R4 then 0.11/680=0.00016F=0.16mF=160uF, 150uF is close enough.
I use 90mS and 120mS for these two and 180mS for the PSU RC.

BTW you input filter is set at 47mS.
You would benefit by adding a low pass filter. 0.5uS to 2uS usually does.
1nF to ground after R1 will give 1uS.

Shout if you want/need more.

[juma]

Originally posted by Mick_F

C6 is there to lower the DC offset on the output. DC coming into the amp see an infinite resistance to ground due to the presence of C6, and as a result the gain for DC is 1. I don't think you need to calculate a particular value here, just use something in the range of what's recommended in the datasheet.
Mick

Also, R4 and C6 together form a high-pass filter in feedback network which influences the frequency response of the amplifier. For R4=680 and C6=47uF lower -3dB cutoff is at 8 Hz. If you really want very low DC offset at output (less than 10mV) choose carefully the brand and make of C6 because it makes a significant difference to the quality of the sound. Peter Daniel and others wrote much about this, search the forum about this issue. The other path is not to use C6 at all and as Carlosfm (NIGC ver. 4) proposed, make R3=100 and R4=3K. I tried this and offset was about 20 mV which is negligible.

Good luck !

[AndrewT]

Hi Juma,
I have never tried omitting the NFB DC blocking cap.

With your experience, does it improve the sound?

Would you say that a DC blocking cap must be fitted to the input if the NFB one is omitted?

less than 10mV) choose carefully the brand and make of C6

is this a leakage problem?

[cpemma]

Originally posted by Mick_F
R3 and R4 set the gain of the amplifier. The gain is calculated by G=1+R3/R4. If you want to use the amp as an integrated , i.e. without a pre, you should choose a gain of approx. 33. If you use it as a power amp with a pre, you can keep the gain around 20.

Or in child-speak, you're selecting a gain to suit the ingoing signal level and the supply voltage (or what's left after transistor losses), so you can make full use of the volume control without excessive clipping at the top end. So for say a +/-24V supply a voltage gain of around 23 might be appropriate.

[juma]

Originally posted by AndrewT
Hi Juma,
I have never tried omitting the NFB DC blocking cap.

With your experience, does it improve the sound?

Would you say that a DC blocking cap must be fitted to the input if the NFB one is omitted? is this a leakage problem?

I was not able to buy Blackgate or other expencive and much praised capacitors but I made a comparison with and without ordinary caps (Jamicon, Philips, Panasonic) in NFB. Without this cap LM3886 amplifier had much better sound, more detail in mids, no grains in upper mids and highs. Also, bass sounds more realistic and controled (tight, as the word is).

I generally don't use blocking caps, because I allways put a buffering/preamplifing stage before the GC and I make that stage without DC Offset on its output so that it can't transfer any to GC.
Surely there are good coupling caps, but they are a bit on the expencive side and not easy to obtain where I live. Other people (Peter Daniel especially) researched this subject into considerable depth.

No it's not a leakage problem. Simply, GC chip has some DC offset on it's input (in most cases 1-5 mV - I tried about 20 chips, all of them LM3886, neither had input offset bigger than 2 mV). How big it wil it be on the speaker depends on DC gain of the amp. 100 mV or less of DC voltage on the speaker is generally considered harmless. DC offset on the chip's output becomes important in paralleled topology where it (if existant) is the cause of excessive heating of the amp (chips' outputs are connected through 0R1 resistors and for two chips having 50 mV difference in output offset the unwanted current flow between chips will be 250 mA).

Cheers!

[AndrewT]

Hi Juma,
sounds like I need to try removing the NFB cap.

Sure cheaper than buying an expensive, big, high quality, bi-polar replacement.

Why is the quality of C6 important to DC offset if it's not leakage?
I understand the rest of your argument, but that bit beat me.

[juma]

Originally posted by AndrewT
Why is the quality of C6 important to DC offset if it's not leakage?
I understand the rest of your argument, but that bit beat me.

a) Quality of C6 is important only for the quality of the sound.

b) For the value of DC offset it is the sheer presence of C6 in the circuit that is important.

c) And, it is the values of C6 and R4 that are important for the low frequency roll off point of the amp.

Cheers !