Have just completed my first amp project, an inverted two channel LM3875 using this schematic. The only changes I have made are that I used 0.1uF caps across 1000uF, used an additional 4000uF for each power supply rail and used a 56k resistor to ground at the input instead of vaiable resistor (using as power amp). I am getting some noise at the speakers. When the amp is connected to pre-amp and volume of pre-amp is turned past 1/2 way I am getting a low level 100hz hum from the woofers that increases in volume when volume knob on pre is increased. This noise is inaudible with volume less than 1/2. When pre amp is on there is also a hiss coming from the tweeters. This hiss does not change volume with movement of the pre-amp volume control, but stops completely when pre-amp is switched off. There is NO noise at all from speakers when there is no input to amp (pre-amp turned off or RCA cables disconnected). Any help on where to start looking to solve these problems would be appreciated.
Thanks,
Chris
Thanks,
Chris
Attachments
Try removing the 22k, it injects bias current into the signal ground. This resistor will set the DC gain to 10 which in my opinion is unnecessary. The 10k + 4.7u + source impedance will provide a gentle rolloff to DC unity gain without it.
I myself have 330k/15k + 2.2u and a 100k pot and a 1k from + to ground on my 3886's and they are totally quiet. I do have a zobel and a 10 ohm + coil on the output though, no resistor. (I still don't understand why people add a resistor to the output).
If you still want that 22k, try moving it to the power ground instead.
Have you checked your preamp if it can drive this low impedance (<10k)? Hum and noise could come from it.
Is the input connector insulated from the chassis? If not, insulate it and wire its ground to the signal ground point. I have pots, so I have wired the connector ground to the CCW lug of the pot and from there to the signal ground.
Are both the preamp and GC signal grounds connected to protective earth? If so, you have a nice ground loop. You could try letting the GC signal ground float above the chassis but
do not break the protective ground
I myself have 330k/15k + 2.2u and a 100k pot and a 1k from + to ground on my 3886's and they are totally quiet. I do have a zobel and a 10 ohm + coil on the output though, no resistor. (I still don't understand why people add a resistor to the output).
If you still want that 22k, try moving it to the power ground instead.
Have you checked your preamp if it can drive this low impedance (<10k)? Hum and noise could come from it.
Is the input connector insulated from the chassis? If not, insulate it and wire its ground to the signal ground point. I have pots, so I have wired the connector ground to the CCW lug of the pot and from there to the signal ground.
Are both the preamp and GC signal grounds connected to protective earth? If so, you have a nice ground loop. You could try letting the GC signal ground float above the chassis but
do not break the protective ground
My knowledge of electronics is rather basic, so I have stuck to the schematic that I downloaded from one of the chip amp sites. I am happy to experiment, but would really appreciate some direction ratther than just randomly changing parts... No, I have not made any changes yet. You suggest try removing 22k to ground and trying 10k at input?
Thanks guys,
Chris
Thanks guys,
Chris
You can get much infomation just to read the datasheet and availabe application notes which in this case is very good. If you don't understand so much you can still compare good truths and what you have found on the net.chrish said:
If you want an inverting amp, start with a "clean" connectionm described in datasheets and application notes. Then if it works start changing.
Much of the LM3886 stuff is valid for LM3875.
http://www.national.com/ds.cgi/LM/LM3875.pdf
http://www.national.com/ds.cgi/LM/LM3886.pdf
http://www.national.com/an/AN/AN-1192.pdf
Thanks for the help guys. I tried connecting the amp directly to an old Luxman cd player that I have that has a variable output. There is virtually no noise when cd on pause and volume full. Maybe it is a problem interracting with the pre amp? For info the pre amp is an old Adcom GFP555 II.
GregGC asked me to describe grounding. I have used star grounding, all signal to point on chassis next to input and power ground next to the 1000uF filter caps (behind caps in this view).
GregGC asked me to describe grounding. I have used star grounding, all signal to point on chassis next to input and power ground next to the 1000uF filter caps (behind caps in this view).
Attachments
At what volume does the output start to clip (distort) when connected to the pre-amp? If it is at less than about "2:00" on the volume dial, then decrease the gain of your amp by reducing the value of the feedback resistor. Try 100K instead of 220K. This will reduce noise, especially the omni-present background hiss that your pre-amp is creating.
Thanks guys, really appreciate the help. When I first built the amp I checked for output with a throwaway $2 speaker, there was sound and no smoke (first test passed...). I then hooked up my old Bose 901s (I was young and did not know any better when I bought them) with nothing hooked up to the input. There was absolutely no noise. Hooked up old Luxman CD player with variable output directly to amp - no problem. It was only after hooking up through pre-amp and connecting to new speakers (Lynn Olson Ariels) that I noticed some noise...
I have gone back to old system, Adcom 555II pre-amp and Adcom 555 power amp on old speakers and the noise is there! So in short when I noticed the noise I assumed it was the new GC amp, but in fact it had been there in my pre-amp all along...
Have just done a little more fault finding and noticed that the hum is present only when inputs are connected to the pre-amp and only when connected with reasonable quality sheilded interconnects. There is no hum with 50 cent unsheilded interconnect (curious).
Sorry for the red herring, will try the reduced gain suggested for the hiss. I am not sure where the amp is clipping as it is connected the Ariel speakers I have just spent months building and don't want to damage them (it does go to 12 o'clock with no problem). Will have to connect to old speakers again if I want to check that. Once again thanks guys...
I have gone back to old system, Adcom 555II pre-amp and Adcom 555 power amp on old speakers and the noise is there! So in short when I noticed the noise I assumed it was the new GC amp, but in fact it had been there in my pre-amp all along...
Have just done a little more fault finding and noticed that the hum is present only when inputs are connected to the pre-amp and only when connected with reasonable quality sheilded interconnects. There is no hum with 50 cent unsheilded interconnect (curious).
Sorry for the red herring, will try the reduced gain suggested for the hiss. I am not sure where the amp is clipping as it is connected the Ariel speakers I have just spent months building and don't want to damage them (it does go to 12 o'clock with no problem). Will have to connect to old speakers again if I want to check that. Once again thanks guys...
I know that this has probably now become a little off-topic... The pre-amp was purchased about 13 years ago in USA, I live in Australia and it was just very recently that I converted the pre-amp to 240 volts (by just changing a bridge on power supply pcb). Could it be that I need to look at some additional filtering for the pre-amp power supply?
Thanks,
Chris
Thanks,
Chris
Check if the hum is there if you connect only one channel to the preamp and short the other input.
If the hum disappears I would guess the cause is the ground loop formed from the preamp signal ground -> left interconnect screen -> left input ground star -> chassis/protective earth wires -> right input signal ground -> right interconnect screen -> preamp signal ground.
If this is the case, I would try this to see if this helps:
Insulate all the star grounds from the enclosure and only connect protective earth to the enclosure. Also break the connection between left and right ground. Then add ground loop breakers (4.7 ohm 5W in parallel with a 0.1uF capacitor) from the signal ground of each channel to the point where the protective earth is connected to the enclosure.
I myself use class II transformers (approved for equipment without protective earth) and leave the amp signal grounds floating and no ground loop breakers, works fine for me.
Just my 2 coins.
If the hum disappears I would guess the cause is the ground loop formed from the preamp signal ground -> left interconnect screen -> left input ground star -> chassis/protective earth wires -> right input signal ground -> right interconnect screen -> preamp signal ground.
If this is the case, I would try this to see if this helps:
Insulate all the star grounds from the enclosure and only connect protective earth to the enclosure. Also break the connection between left and right ground. Then add ground loop breakers (4.7 ohm 5W in parallel with a 0.1uF capacitor) from the signal ground of each channel to the point where the protective earth is connected to the enclosure.
I myself use class II transformers (approved for equipment without protective earth) and leave the amp signal grounds floating and no ground loop breakers, works fine for me.
Just my 2 coins.
There are 2 interpretations.
One is that the preamp sources the hiss and hum.
The other is that your amp oscillates when connected to the preamp.
If the hissing channel becomes hot then it could be oscillations, if not or just a tad warmer it's the preamp, especially if the hiss volume follows the volume control.
Other than adding a "better safe than sorry" capacitor between v+ and v- as close to the cip as possible (0.1uF >100V) I have no other ideas.
One is that the preamp sources the hiss and hum.
The other is that your amp oscillates when connected to the preamp.
If the hissing channel becomes hot then it could be oscillations, if not or just a tad warmer it's the preamp, especially if the hiss volume follows the volume control.
Other than adding a "better safe than sorry" capacitor between v+ and v- as close to the cip as possible (0.1uF >100V) I have no other ideas.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.