Hi Guys,
I've made a double layer PCB for stereo Lm3875 recently. It works fine but the ground buzz annoys me. Hope you can help me to find a solution to this.
The grounding scheme is as shown in the pictures below. Amplifier ground is not connected to Mains Earth.
Troubleshoot Scenarios :
1) Amplifier connected to CD player (without earth connection also). Ground buzz get loudest when volume pot is at middle position. Weak buzz when volume knob is at lowest and highest position. When I use my finger to touch the RCA body, buzz noise becomes weaker. When amplifier is connected to main Earth, no more buzz noise. (Loud buzz means it is audible at 1.5 metres away. Weak buzz means it is audible only when place ear close to speaker cone)
2) Amplifier connected to CD player (with earth connection). No buzz noise for all volume pot position.
3) Amplifier not connected to CD player. medium buzz level with all volume pot position.
I know that the easiest solution is to connect amplifier ground to main earth but would like to find out more why/how this happen.
1) Why the ground buzz level is influenced by the volume pot position? Would it be that the amplifier is unstable?
2) Why the buzz goes off when amplifier is connected to main earth. My previous LM3875 does not produce buzz noise even though it is not connected to main earth.
3) Is my grounding scheme done correctly? If not how to improve it?
Thanks in advance for your help.
I've made a double layer PCB for stereo Lm3875 recently. It works fine but the ground buzz annoys me. Hope you can help me to find a solution to this.
The grounding scheme is as shown in the pictures below. Amplifier ground is not connected to Mains Earth.
Troubleshoot Scenarios :
1) Amplifier connected to CD player (without earth connection also). Ground buzz get loudest when volume pot is at middle position. Weak buzz when volume knob is at lowest and highest position. When I use my finger to touch the RCA body, buzz noise becomes weaker. When amplifier is connected to main Earth, no more buzz noise. (Loud buzz means it is audible at 1.5 metres away. Weak buzz means it is audible only when place ear close to speaker cone)
2) Amplifier connected to CD player (with earth connection). No buzz noise for all volume pot position.
3) Amplifier not connected to CD player. medium buzz level with all volume pot position.
I know that the easiest solution is to connect amplifier ground to main earth but would like to find out more why/how this happen.
1) Why the ground buzz level is influenced by the volume pot position? Would it be that the amplifier is unstable?
2) Why the buzz goes off when amplifier is connected to main earth. My previous LM3875 does not produce buzz noise even though it is not connected to main earth.
3) Is my grounding scheme done correctly? If not how to improve it?
Thanks in advance for your help.
Attachments
You are on the right track. Make sure all input cables are shielded. Connect the star ground to the safety earth by a loop breaker circuit. (10R, 100nF, 2 diodes in parallel). Connect all exposed metal parts to the safety earth. This includes the heatsink, standoffs, potentiometer case, chassis, metal plate on diode bridge. Also check for loose connections and solder joints. 50-60hz hum is caused by a ground loop. The star ground and loop breaker should take care of that. Variable noises are usually caused by improper shielding and failure to ground metal parts. If you have a 100-120hz buzz, try beefing up the ground traces around your main filter caps.
Where is the center tap of your tranformer attached? With your setup, it should be attached to the center of the main filter caps.
Check out the grounding article at http:sound.westhost.com
Where is the center tap of your tranformer attached? With your setup, it should be attached to the center of the main filter caps.
Check out the grounding article at http:sound.westhost.com
Hi,
are the input RCA grounds insulated from the Chassis?
Bill is correct to point your centre tap to the power supply common (Main Ground IN).
As an alternative to connecting the chassis (safety) earth, using the disconnecting network, to the audio ground (Star Ground) try taking the safety earth via the disconnecting network to the Input RCA ground.
Another alternative worth trying. Take the Input RCA grounds and the pot common off the GND (local clean ground) and take them to the audio ground. It might help here to link the two RCA grounds together and take one wire to the audio ground.
Unrelated to the buzz, I would recommend changing the relative ratio of pot to input impedance setting resistor.
pot=10k and resistor = 50k. This unfortunately will also require the NFB upper and lower legs to be raised to 50k and 1k41 (use either1k3 or 1k5). You could just change the pot and try leaving the in Z and NFB "as is".
Q. When a chip amp is wired as a DC opamp, the inverting input sees the NFB resistors in parallel, but with a pot and DC blocking cap on the non-inverting input, does the different loading (600r & 22k) on the differential inputs lead to an output offset and does this offset vary more widely than when the inputs are balanced.
are the input RCA grounds insulated from the Chassis?
Bill is correct to point your centre tap to the power supply common (Main Ground IN).
As an alternative to connecting the chassis (safety) earth, using the disconnecting network, to the audio ground (Star Ground) try taking the safety earth via the disconnecting network to the Input RCA ground.
Another alternative worth trying. Take the Input RCA grounds and the pot common off the GND (local clean ground) and take them to the audio ground. It might help here to link the two RCA grounds together and take one wire to the audio ground.
Unrelated to the buzz, I would recommend changing the relative ratio of pot to input impedance setting resistor.
pot=10k and resistor = 50k. This unfortunately will also require the NFB upper and lower legs to be raised to 50k and 1k41 (use either1k3 or 1k5). You could just change the pot and try leaving the in Z and NFB "as is".
Q. When a chip amp is wired as a DC opamp, the inverting input sees the NFB resistors in parallel, but with a pot and DC blocking cap on the non-inverting input, does the different loading (600r & 22k) on the differential inputs lead to an output offset and does this offset vary more widely than when the inputs are balanced.
Thank you for replying.
Hi Bill,
I'll try measure the hum frequency with scope. My input cable is not shielded. I'm use WireWrapwire for all signal wiring. My Transformer and rectifier is located in a seperate box. In this case does the input cable shielding mandatory? If yes, where can I get those cable? Center tap is connected to main ground in.
Hi Rob,
Yes, the speaker GND is connected to star ground.
Hi Andrew,
RCA body is insulated from chassis. The chassis is not connected to any ground. Why does it buzz when amp is not connected safety earth?
Thanks.
Hi Bill,
I'll try measure the hum frequency with scope. My input cable is not shielded. I'm use WireWrapwire for all signal wiring. My Transformer and rectifier is located in a seperate box. In this case does the input cable shielding mandatory? If yes, where can I get those cable? Center tap is connected to main ground in.
Hi Rob,
Yes, the speaker GND is connected to star ground.
Hi Andrew,
RCA body is insulated from chassis. The chassis is not connected to any ground. Why does it buzz when amp is not connected safety earth?
Thanks.
The center tap needs to be connected to center of the main power caps and the star ground and then to the safter ground by a loop breaker.
For shielding you can use aluminium foil and connect it to the chassis. RadioShack sells "shield audio cable". It has two insulated conductors surround by foil, copper braid, and outer insulation. The shielding is to protect against ambient noise such as flouresent lights, nearby computers, smps, rf, etc.
For shielding you can use aluminium foil and connect it to the chassis. RadioShack sells "shield audio cable". It has two insulated conductors surround by foil, copper braid, and outer insulation. The shielding is to protect against ambient noise such as flouresent lights, nearby computers, smps, rf, etc.
Hi Ipa,
shielded two core comes with or without twist.
Twisted two core and shield will reject very slightly better than parallel two core.
That double shield that Bill mentions should be pretty good.
Bill
The one I am concerned about is the pot case, I would have thought it could go to audio ground, but maybe not.
Is pot case a safety issue?
shielded two core comes with or without twist.
Twisted two core and shield will reject very slightly better than parallel two core.
That double shield that Bill mentions should be pretty good.
Bill
How important is it to use chassis shielding for all the components you have listed?
The one I am concerned about is the pot case, I would have thought it could go to audio ground, but maybe not.
Is pot case a safety issue?
Hi Bill and Andrew,
Thanks for the tips. The ground buzz goes away when I connect the chassis to RCA body, even though the chassis is still not connected to safety earth. Does the ground need more metal area to act as a real ground? FYI, my PCB ground plane is already quite huge.
The output offset for both channel are 90 and 110mV respectively, which is rather high. Shall I change the non-inverting input resistor to the parallel combination of 680//22K to get minimum offset? I'm afraid that this would load the input stage. What is the best solution to lower the output offset?
I measured the ground buzz with scope yesterday with volume turn to middle position. This is measure before I connect chassis to RCA ground. The buzz frequency is 70Hz. Does this noise come from power supply?
Thanks,
Ipanema
Thanks for the tips.
The ground buzz goes away when I connect the chassis to RCA body, even though the chassis is still not connected to safety earth. Does the ground need more metal area to act as a real ground? FYI, my PCB ground plane is already quite huge. The output offset for both channel are 90 and 110mV respectively, which is rather high. Shall I change the non-inverting input resistor to the parallel combination of 680//22K to get minimum offset? I'm afraid that this would load the input stage. What is the best solution to lower the output offset?
I measured the ground buzz with scope yesterday with volume turn to middle position. This is measure before I connect chassis to RCA ground. The buzz frequency is 70Hz. Does this noise come from power supply?
Thanks,
Ipanema
Hi Ipanema,
the chassis connection to safety ground is not an option.
You MUST safety earth all EXPOSED conductive parts.
Do this before you ever connect a project to the mains.
Then experiment with the audio, power, clean and star grounds.
Never remove your safety ground when you are maintaining a piece of equipment.
If you can weld the safety to the chassis, then that's best.
Next best is a bolted safety earth dedicated to chassis only.
Third is a bolted safety next to chassis with a nut on top. Then your disconnecting network with a second nut on top.
the chassis connection to safety ground is not an option.
You MUST safety earth all EXPOSED conductive parts.
Do this before you ever connect a project to the mains.
Then experiment with the audio, power, clean and star grounds.
Never remove your safety ground when you are maintaining a piece of equipment.
If you can weld the safety to the chassis, then that's best.
Next best is a bolted safety earth dedicated to chassis only.
Third is a bolted safety next to chassis with a nut on top. Then your disconnecting network with a second nut on top.
Hi,
they are probably double insulated, ClassII products.
If they do not carry the appropriate symbol, do not touch them.
Can you develop and test for all operational and fault circumstances to meet the ClassII standard? Then go ahead and build a ClassII product.
If you cannot guarantee to meet ClassII standard then apply a permanent safety earth to all exposed conductive parts.
they are probably double insulated, ClassII products.
If they do not carry the appropriate symbol, do not touch them.
Can you develop and test for all operational and fault circumstances to meet the ClassII standard? Then go ahead and build a ClassII product.
If you cannot guarantee to meet ClassII standard then apply a permanent safety earth to all exposed conductive parts.
Strange Behaviour on LM3886
Sorry for the thread hijack.
I've got BrianGT's LM3886 boards and after assembling them there's this problem:
The music comes from both channels but in opposite phases. It seems that the left speaker is playing the positive tones for ~1s, then the right speaker will be playing the negative tones for ~1s, all alternating between each other.
To begin with:
I have 300VA transformers wired up to two discrete diode bridges and connected to one capacitor bank, consisting of 4 x 4700uF per voltage rail, 0.2R resistors between each capacitor.
Output voltage at the supply is ~+/-36V referenced to ground.
With the ACV meter, there's about 0.050ACV ripple.
All grounds are connected via Star Grounding at the 0V point of the capacitor bank, which is earthed directly. With a CD-player connected, there is no hum, tizz or buzz, so I guess my grounding is correctly done.
The heatsinks do get suspiciously hot too. I'm not able to give an indication of the temperature though. I have check that the chips are securely connected to the heatsink, and they're the LM3886TF's, so a short to the heatsink is out.
Thanks if anyone can help.
Sorry for the thread hijack.
I've got BrianGT's LM3886 boards and after assembling them there's this problem:
The music comes from both channels but in opposite phases. It seems that the left speaker is playing the positive tones for ~1s, then the right speaker will be playing the negative tones for ~1s, all alternating between each other.
To begin with:
I have 300VA transformers wired up to two discrete diode bridges and connected to one capacitor bank, consisting of 4 x 4700uF per voltage rail, 0.2R resistors between each capacitor.
Output voltage at the supply is ~+/-36V referenced to ground.
With the ACV meter, there's about 0.050ACV ripple.
All grounds are connected via Star Grounding at the 0V point of the capacitor bank, which is earthed directly. With a CD-player connected, there is no hum, tizz or buzz, so I guess my grounding is correctly done.
The heatsinks do get suspiciously hot too. I'm not able to give an indication of the temperature though. I have check that the chips are securely connected to the heatsink, and they're the LM3886TF's, so a short to the heatsink is out.
Thanks if anyone can help.
problems too
Ipanema,
Have you solved your problem? If so, what was the specific problem? I have the same issue. Having built many SS discrete amps and preamps with textbook star grounding, it really bothers me that I am having this issue here.
The only thing that I have not done is take the center tap from the rectifier board directly to the star first, then the ground from the chip board to the star. I wired the boards as labeled by audiosector. I guess this is the net thing to try.
My dc offset is a little over 100mV. Did you have any luck with that? I am looking into servos but think that a simple low pass filter might be easier. Not sure how either would sound though.
Thanks.
Ipanema,
Have you solved your problem? If so, what was the specific problem? I have the same issue. Having built many SS discrete amps and preamps with textbook star grounding, it really bothers me that I am having this issue here.
The only thing that I have not done is take the center tap from the rectifier board directly to the star first, then the ground from the chip board to the star. I wired the boards as labeled by audiosector. I guess this is the net thing to try.
My dc offset is a little over 100mV. Did you have any luck with that? I am looking into servos but think that a simple low pass filter might be easier. Not sure how either would sound though.
Thanks.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.