Many years ago I built 2 LM3875 monp amps that have been going ever since. I have started using them with my Lampizator P-17 open baffles and they sound really good. However the input is getting saturated and I am getting some odd static type noise. I have 10dB attenuators and these have removed it,
The circuit is shown below. If I wanted to alter the input sensitivity would I change the value of the input resistir (56k) or is it not that simple.
http://www.decdun.me.uk/gainclone.html
The circuit is shown below. If I wanted to alter the input sensitivity would I change the value of the input resistir (56k) or is it not that simple.
http://www.decdun.me.uk/gainclone.html
The gain is feedback R / input R = (432k / 2) / 10k = 21.6 (I assume the 432k resistors are in parallel). The values are that high (vs. the usual 1k / 20k in non-inverting config) to keep a reasonable input impedance. The 56k is there to discharge the input cap and prevent pops when connecting an input to the amp. It doesn't affect the sensitivity, only the input impedance (which is 56k !! 10k = 8k5).
To reduce sensitivity, depending on which resistors you have at hand, increase the input one, decrease the feedback one or both, but note that per the datasheet the minimum gain for stability is 10, so don't go below that value.
As for the "static" noise, can you be more specific? Is it hiss, hum, both, constant, intermittent, etc.?
To reduce sensitivity, depending on which resistors you have at hand, increase the input one, decrease the feedback one or both, but note that per the datasheet the minimum gain for stability is 10, so don't go below that value.
As for the "static" noise, can you be more specific? Is it hiss, hum, both, constant, intermittent, etc.?
Last edited:
It is nearly that simple. You add a resistor in series with the input (to the left of the 56k) and that forms a divider. The overall input impedance of the circuit (defined by the 10k in parallel with the 56k) is also modified by the actual values in your attenuator... so trial and error to get the values that work best for you.
A 9k1 or 10k series resistor would halve the speaker voltage output for a given input level. If its a bit quiet then lower the series value to say 6k8 or 4k7 and try again.
A 9k1 or 10k series resistor would halve the speaker voltage output for a given input level. If its a bit quiet then lower the series value to say 6k8 or 4k7 and try again.
I have used these amps many times over the years with various speakers and have not experienced what I am now. It only happens on certain tracks streamed from Spotify and not during the whole time the track is playing. Volume level seems to be connected, but it is so random. My speakers are 97dB so very sensitive maybe there is a mis-match with the amps and speakers I can't say as some track play clear and loud without this momentary static/glitch type noise. in the Mid/HF area.
I have used Class A Pass Amps, several Class D and Mosfet amps without this happening.
I have used Class A Pass Amps, several Class D and Mosfet amps without this happening.
Thanks Mooly. That looks like quite a variation to my elementary knowledge of circuits. Would this also be applicable to the split rail supply I am using? Would the fact that I added the reccomended regulated PS (reccomended by Nick Whetstone on the Decibel Dungeon site) have any bearing on the glitches I hear? Other gainclones I built without regulation would have more voltage avaialble?
Yes, those additions are pretty standard to any power amp whether discrete or chip.
Its possible but unlikely if they have been built with good layout and wiring practices.
Stabilised supplies for power amps can limit peak transient current peaks (as the regulator goes into a current limit situation) but given these are 5 amp devices its unlikely.
Ideally you should use a scope to look at the output signal of the amp and see if you can see anything happening when these noises occur.
To properly test with a scope, it should have a minimum 20 MHz bandwidth (more is better) to catch high frequency oscillation in the circuit. The "Gain Clone" topology is incomplete, it leaves out components that should be considered necessary for reliable operation...that's not just my opinion, it is the LM3886's manufacture's guide lines.
Mike
Mike
Do you have a PCB layout for your design?
I've seen some strange noise caused by demodulated RF from the feedback loop. The virtual ground formed by your PS using the unsymmetrical regs can cause that under certain conditions, depending on where your chassis and input is grounded. Thats why I like to use simpler RC filtering without vregs.
Perhaps the lower noise LT versions of 3 pin regs are better in terms of specifically the negative reg itself. You may also want to try a "softer" chassis ground terminated with a small series resistance to the PS ground. In your case thats after the regs and not at the rectifier, close to the main filter caps. Sometimes it can also be more effective to ground through a 100nF film cap instead of a direct wired ground.
I've seen some strange noise caused by demodulated RF from the feedback loop. The virtual ground formed by your PS using the unsymmetrical regs can cause that under certain conditions, depending on where your chassis and input is grounded. Thats why I like to use simpler RC filtering without vregs.
Perhaps the lower noise LT versions of 3 pin regs are better in terms of specifically the negative reg itself. You may also want to try a "softer" chassis ground terminated with a small series resistance to the PS ground. In your case thats after the regs and not at the rectifier, close to the main filter caps. Sometimes it can also be more effective to ground through a 100nF film cap instead of a direct wired ground.
Attachments
Last edited:
For my LM3886 and 4780 implementation, the Cc and Cf Rf as shown in mooly's post eliminated static noise issues. Oddly these very crucial components are absent in many "well known" implementations
Read more -
https://www.circuitbasics.com/design-hi-fi-audio-amplifier-lm3886/
Read more -
https://www.circuitbasics.com/design-hi-fi-audio-amplifier-lm3886/
Last edited:
In relation to the above I have added a 10k resistor at the input but this has not cured the problem. It is frustrating as it does not happen with every track played and is limited to the HF range and is very momentary when it does happen.
The circuit I used as shown above has a feedback resistor value of around 216k. I have not seen any other schematic with this value used there. I changed this to 20k as this is shown in several schematics and in post 9, but it did not like that. The schematic shown at post 9 shows 20k but with a cap and resistors added to the circuit.
Would I do better to start again and build them as per the schematic in post 9?
The circuit I used as shown above has a feedback resistor value of around 216k. I have not seen any other schematic with this value used there. I changed this to 20k as this is shown in several schematics and in post 9, but it did not like that. The schematic shown at post 9 shows 20k but with a cap and resistors added to the circuit.
Would I do better to start again and build them as per the schematic in post 9?