I built a LM3875 chipamp from Audiosector years ago and added a discrete buffer circuit from Decibel Dungeon. See link below.
http://www.decdun.me.uk/gainclone2.html
I just realized after all these years that I'm supposed to add a DC blocking capacitor to the input and output of the discrete buffer circuit. Which value capacitor do I use? Would a 1uF polypropylene capacitor be ok?
http://www.decdun.me.uk/gainclone2.html
I just realized after all these years that I'm supposed to add a DC blocking capacitor to the input and output of the discrete buffer circuit. Which value capacitor do I use? Would a 1uF polypropylene capacitor be ok?
They say 2uF and 6uF, and that's about right. But the op amp is unity gain, has fet inputs,
and bipolar supply, so you may be able to do without them after all.
Really all you need is a coupling capacitor for the source output (before the volume control),
which it likely has already.
and bipolar supply, so you may be able to do without them after all.
Really all you need is a coupling capacitor for the source output (before the volume control),
which it likely has already.
Last edited:
Thanks you both for your help. I don't think I have any DC in the signal. I've been using it without the DC blocking capacitors ever since the amp was built but it always made a pop sound through the speakers when I turn it on. After realizing I didn't use DC blocking capacitors I thought this might be why I hear the pop sound. My idea to get around this issue was to add a relay switch circuit that will delay the connection between the output of the buffer and the amp during power up.
Well more accurately either the sound or your wallet! Large PP caps aren't cheap. Or you can go for even larger electrolytic values, cheaper but will age faster. As a coupling cap's value increases the percentage distortion it can impart to the signal drops since the signal voltage across it drops. Very physically large caps can end up picking up EMI too.
Don't think/assume, just measure. The 2 discrete buffer circuits in the link definitely need coupling caps!!!! If you use either one without coupling caps then it is a miracle your speakers are still alive as those circuits will have DC offset and the amplifier will be happy to amplify that. You then also never measured/tested/checked temperature of the setup which is gambling by assumption.
The pop sound may be a double barreled issue as the DC coupled amplifier outputs need some settling time after being powered on and that DC is transferred to the woofers. That can be solved with a speaker protection relay + circuit. Delayed contacts at power on, instant open contacts at power off.
Hint: µPC1237
The pop sound may be a double barreled issue as the DC coupled amplifier outputs need some settling time after being powered on and that DC is transferred to the woofers. That can be solved with a speaker protection relay + circuit. Delayed contacts at power on, instant open contacts at power off.
Hint: µPC1237
Last edited:
As an added negative you might be throwing away amplifier output power if it is forced to output the DC offset. If the amp has a cap at its input and the offset from the buffers isn’t huge you are ok. If the buffer has a big offset it could be causing a distortion due to asymmetry of the signal. The pop you hear is just the electronics coming up to working voltages. As was said, turn the amp on last.
If there is DC offset at the output at the amp then it should be measurable with a multimeter right. With the amp on I measured 56mV and 27mV at each speaker terminal. I think this is still ok because it’s been this way ever since I built it and my speakers still work. I also read the manual again for the audio sector kit and found out that I can put a 4.7uF in the R1 position of the PCBA to protect the amp from DC.
To fix the thump on power up I installed a small circuit that I built with a timer and relay switch at the output of the buffer. I have it set so that it waits 3 sec before it connect the buffer to the input of the amp.
To fix the thump on power up I installed a small circuit that I built with a timer and relay switch at the output of the buffer. I have it set so that it waits 3 sec before it connect the buffer to the input of the amp.
Either you are lucky or things have been done as they should be as both discrete circuits have considerable DC offset and definitely need coupling caps as the text also points out. We don't see the amplifiers schematic and the exact buffer circuit (there are 2 in the link) nor a picture of the built device so we rely on words. The words indicate a high risk situation.
Just 1 cap in the buffer to amplifier chain is enough to prevent amplification of DC offset and in the case of the discrete bipolar buffer the source should also be protected against the buffers input DC offset. That is all I can tell from the given information. Power on thumps are normally mostly caused by outputs of amplifiers settling at power on and not by buffers settling as there is a cap to prevent that DC. If the relay prevents a severe power on thump you have proven that coupling caps are needed. You could have measured what exactly happens at the amplifiers outputs at power on and I think it would be shocking. Despite the relay the amplifier will still amplify any DC occurring because of a bad connection or the like and you will still have the amplifiers outputs generating a thump. The coupling caps are mandatory in the described situation. In a DC coupled chain that single cap is the "fuse" protecting the loudspeakers.
BTW are you sure you have measured the offset in the multimeters DC mode?
Just 1 cap in the buffer to amplifier chain is enough to prevent amplification of DC offset and in the case of the discrete bipolar buffer the source should also be protected against the buffers input DC offset. That is all I can tell from the given information. Power on thumps are normally mostly caused by outputs of amplifiers settling at power on and not by buffers settling as there is a cap to prevent that DC. If the relay prevents a severe power on thump you have proven that coupling caps are needed. You could have measured what exactly happens at the amplifiers outputs at power on and I think it would be shocking. Despite the relay the amplifier will still amplify any DC occurring because of a bad connection or the like and you will still have the amplifiers outputs generating a thump. The coupling caps are mandatory in the described situation. In a DC coupled chain that single cap is the "fuse" protecting the loudspeakers.
BTW are you sure you have measured the offset in the multimeters DC mode?
Attachments
Last edited:
I guess I have been lucky for all these years that I've been using this amp in my main system. I am using the discrete circuit in the photo you attached but it wasn't until recently that I went back to check the schematic that it had a note saying it needed DC blocking caps at the input and output. It didn't say what kind of capacitor to use which is why I came on here for advice.
I did try measuring the output of the amp when it powered on but my multimeter doesn't record peak voltage. It does appear to spike upwards very briefly before settling down. I'm going to purchase some capacitors from digikey once I figure out which ones to get.
And yes I'm 100% sure I had the meter dial in DC mode and not AC.
I did try measuring the output of the amp when it powered on but my multimeter doesn't record peak voltage. It does appear to spike upwards very briefly before settling down. I'm going to purchase some capacitors from digikey once I figure out which ones to get.
And yes I'm 100% sure I had the meter dial in DC mode and not AC.
Maybe the luck was in more than 1 area so congrats as the times I made such mistakes always costed money, time and effort. Essential parts for correct operation were omitted, it are not superfluous caps. You could inspect the woofers voice coils if possible as they absolutely had to endure severe DC at power on.
You will need film caps and the standard and minimum quality depending on manufacturer is polyester/MKS/PET/mylar. I would use 4.7 µF 50V in 5 mm pitch so Wima MKS2 for C2 and a 1 µf 50V Wima MKS2 as input cap. Now cap wars may start because of this remark as many are believers of various cap brands and dieelelectric differences. If you want a better type then polypropylene or MKP caps are a choice but they are larger. Please note that using "better" but ridiculously large film caps may introduce new phenomena like that caps and long lead wires functioning as antennas. You might gain a little and loose a lot. Compact building with short connections is always better. You can not go wrong with the small Wima MKS2 types.
Please rethink the relay. Having a suitable 8A rated relay with silver based contacts for delayed switching the LM3875s outputs solves 2 power on issues and will result in total silence powering on the amplifier. If you use a circuit with µPC1237 you will even protect the loudspeakers against amplifier failure as an extra.
You will need film caps and the standard and minimum quality depending on manufacturer is polyester/MKS/PET/mylar. I would use 4.7 µF 50V in 5 mm pitch so Wima MKS2 for C2 and a 1 µf 50V Wima MKS2 as input cap. Now cap wars may start because of this remark as many are believers of various cap brands and dieelelectric differences. If you want a better type then polypropylene or MKP caps are a choice but they are larger. Please note that using "better" but ridiculously large film caps may introduce new phenomena like that caps and long lead wires functioning as antennas. You might gain a little and loose a lot. Compact building with short connections is always better. You can not go wrong with the small Wima MKS2 types.
Please rethink the relay. Having a suitable 8A rated relay with silver based contacts for delayed switching the LM3875s outputs solves 2 power on issues and will result in total silence powering on the amplifier. If you use a circuit with µPC1237 you will even protect the loudspeakers against amplifier failure as an extra.
Last edited:
Thanks, I have ordered the capacitors you suggested from digikey. I will feel more comfortable putting this amp back into my system once they're installed. I don't know how it worked all this time without those dc blocking capacitors. But it doesn't appear to have damaged the speakers. I already sold them so we'll never know if there was ever any damage. I will look into the speaker protection circuit but that will be a project for later.
That voice coils were either black or brown but they are sold which is kind of a solution. We wait for the new owner opening a thread on why his loudspeakers have an issue.
It is surprising to see that errors like these are not even noticed as this power on thump (that was not simply a standard power on thump) must have been quite loud. Same goes for missing channels, crackling, channel differences, distortion etc. If one makes a remark it can even be taken as an insult
