Hi All,
I am using a pair of LM3886s in parallel mode per channel, and am getting DC along with AC on output pins. The chips share the power and the ground, and if I plug in any of the chips' output, I get a good amplified signal. If I measure the potential difference between the 2 output pins, I see DC in the range of 40-80V.
As I said, surprisingly either of the pins connected to the output terminal work alone in the same circuit, but not together.
I have spent over 10 hrs trying to troubleshoot, but to no avail. I do not have any fancy test equipment for testing, and the circuit is connected using a perfboard.
I have checked the ground wiring thoroughly too, and redid the circuit...dont want to unsolder the wires from the pins because it is such a pain to put them all together, but have tested multiple times and there is no shorting of the pins.
Any help is greatly appreciated.
Thanks,
Mohit
I am using a pair of LM3886s in parallel mode per channel, and am getting DC along with AC on output pins. The chips share the power and the ground, and if I plug in any of the chips' output, I get a good amplified signal. If I measure the potential difference between the 2 output pins, I see DC in the range of 40-80V.
As I said, surprisingly either of the pins connected to the output terminal work alone in the same circuit, but not together.
I have spent over 10 hrs trying to troubleshoot, but to no avail. I do not have any fancy test equipment for testing, and the circuit is connected using a perfboard.
I have checked the ground wiring thoroughly too, and redid the circuit...dont want to unsolder the wires from the pins because it is such a pain to put them all together, but have tested multiple times and there is no shorting of the pins.
Any help is greatly appreciated.
Thanks,
Mohit
.
Random thoughts. When several high power amps interact ground loops can become positive feedback, and then things go all to pieces. Are you using a star ground?
Do you really have the problem? Check with the input(s) grounded, so output should be zero. Audio signals present at the output can give strange meter readings.
By "parallel mode" is this what you mean? http://www.ti.com/litv/pdf/snaa021b
.
Random thoughts. When several high power amps interact ground loops can become positive feedback, and then things go all to pieces. Are you using a star ground?
Do you really have the problem? Check with the input(s) grounded, so output should be zero. Audio signals present at the output can give strange meter readings.
By "parallel mode" is this what you mean? http://www.ti.com/litv/pdf/snaa021b
.
Hi All,
I will post pictures tomorrow. I am using the standard PA100 layout (http://www.ti.com/lit/an/snaa021b/snaa021b.pdf)
There is nothing more to the amp than that. Also, I took out all the wires from the pins of the chips and redid them, still the same result.
The chips are genuine for sure, I bought them from a gentleman who stocks thousands of electronic parts, and he does not sell any counterfeit stuff.
I am a hobbyist, and have built a couple of single, parallel and bridged gainclones before as well. I am not using a star ground, but will try it. Before this, I have just connected all the grounds together (I have ground from the power supply caps, ground or negative from the audio input, and the pin7 from LM3886s joined together), but have a question on star ground. Does the ground from pin 7 of LM3886 goto your audio ground or power ground?
I will also try with the input grounded to see what the output is.
Thanks all for your help, and I will let you know if I am able to resolve the problem.
Thanks,
Mohit
I will post pictures tomorrow. I am using the standard PA100 layout (http://www.ti.com/lit/an/snaa021b/snaa021b.pdf)
There is nothing more to the amp than that. Also, I took out all the wires from the pins of the chips and redid them, still the same result.
The chips are genuine for sure, I bought them from a gentleman who stocks thousands of electronic parts, and he does not sell any counterfeit stuff.
I am a hobbyist, and have built a couple of single, parallel and bridged gainclones before as well. I am not using a star ground, but will try it. Before this, I have just connected all the grounds together (I have ground from the power supply caps, ground or negative from the audio input, and the pin7 from LM3886s joined together), but have a question on star ground. Does the ground from pin 7 of LM3886 goto your audio ground or power ground?
I will also try with the input grounded to see what the output is.
Thanks all for your help, and I will let you know if I am able to resolve the problem.
Thanks,
Mohit
Last edited:
.
Layout refers to the physical arrangement of parts, schematics are something else. Are you using the schematic on page 13, figure 6, "6.2.1 Parallel Amplifier Schematic"? I ask mainly because...well, it's not like there were a lot of parts.
<< if I plug in any of the chips' output, I get a good amplified signal >>
So you're saying the amp works?
<< I am using the standard PA100 layout...There is nothing more to the amp than that. >>
Actually, yes there is. That schematic doesn't reference the mute circuit(s), you'll have to state what you're doing with those. Or much better, post your full circuit.
If you didn't draw up your full circuit, including mute(s), before you plugged in your soldering iron, then I'm pretty sure I found your problem. Moving along from there...
The DC voltage you speak of is measured exactly where? Across the speaker terminals? Pin 3 (output) to ground? Pin 3 to pin 3? Somewhere else? Note that the op amps' output pins are not the speaker terminals.
<< have a question on star ground >>
With a star ground there is one ground point, and only one, and all ground connections go to that one point, and never go anywhere else.
<< Does the ground from pin 7 of LM3886 go to your audio ground or power ground? >>
What pin 7 might be is anybody's guess unless somebody emails the manufacturer and asks. Interestingly, the manufacturer will probably reply to an email--though the reply might be in hurried engineerese (senior engineers don't pull email duty). Anyway, you seem to imply that you have several ground points, which is always iffy, and with high power amps more so.
Suggestion: have your circuit board and your [full, complete, correct] schematic lying on a table side by side. Also your magnifying glass.
Start at pin 1 of op amp 1 and go outbound from the op amp. Check what is connected against what the schematic says should be connected.
Now go to pin 2, and same thing. And pin 3...but you get the idea.
This seems like it would take forever, but it really doesn't. And even if it did, it would still be a lot faster than standing across the room and guessing.
.
Layout refers to the physical arrangement of parts, schematics are something else. Are you using the schematic on page 13, figure 6, "6.2.1 Parallel Amplifier Schematic"? I ask mainly because...well, it's not like there were a lot of parts.
<< if I plug in any of the chips' output, I get a good amplified signal >>
So you're saying the amp works?
<< I am using the standard PA100 layout...There is nothing more to the amp than that. >>
Actually, yes there is. That schematic doesn't reference the mute circuit(s), you'll have to state what you're doing with those. Or much better, post your full circuit.
If you didn't draw up your full circuit, including mute(s), before you plugged in your soldering iron, then I'm pretty sure I found your problem. Moving along from there...
The DC voltage you speak of is measured exactly where? Across the speaker terminals? Pin 3 (output) to ground? Pin 3 to pin 3? Somewhere else? Note that the op amps' output pins are not the speaker terminals.
<< have a question on star ground >>
With a star ground there is one ground point, and only one, and all ground connections go to that one point, and never go anywhere else.
<< Does the ground from pin 7 of LM3886 go to your audio ground or power ground? >>
What pin 7 might be is anybody's guess unless somebody emails the manufacturer and asks. Interestingly, the manufacturer will probably reply to an email--though the reply might be in hurried engineerese (senior engineers don't pull email duty). Anyway, you seem to imply that you have several ground points, which is always iffy, and with high power amps more so.
Suggestion: have your circuit board and your [full, complete, correct] schematic lying on a table side by side. Also your magnifying glass.
Start at pin 1 of op amp 1 and go outbound from the op amp. Check what is connected against what the schematic says should be connected.
Now go to pin 2, and same thing. And pin 3...but you get the idea.
This seems like it would take forever, but it really doesn't. And even if it did, it would still be a lot faster than standing across the room and guessing.
.
Last edited:
AFAIK Pin 7 is ground reference for the diffamps of the input and VAS stage, and should not connect to signal ground.
Such an odd voltage points to an error in layout or implementationeg a missing component, and without a few pictures or a drawing of the wiring, it will be impossible to suggest causes.
Hope you can get a handle on the problem.
Such an odd voltage points to an error in layout or implementationeg a missing component, and without a few pictures or a drawing of the wiring, it will be impossible to suggest causes.
Hope you can get a handle on the problem.
Hi Everybody,
Thanks a ton for the detailed comments and suggestions, especially @bentsnake! I really appreciate everybody's time and help! The amp is still giving the same problem. I redid all the pin wiring again, but still the same issue
I am attaching the layout and the schematic for your reference with points called out for easy identification.
For mute, I have soldered a 10K ohms resistor between pins 8 and 4.
I was seeing some sort of short when I connected the two pins 3 via 0.3 ballast resistors, and got to admit that I was wrong in measuring the voltage between pin 3 of one chip and the pin 3 of the other connected to the circuit.
In the attached pic, point 2 is the ground. Caps are connected to it.
8 and 9 are the + (pin 10 inputs). 3 and 4 are the -ve inputs. 6 and 7 are the output pin inputs (pin 3). 10 is the speaker positive output and the ground is the negative for the speaker. I have not done the star ground yet, but ground is not an issue with either of the chips not in action (with the other chip's pin 3 open)
I am really baffled at this, but maybe it is something really basic I am missing. I redid the entire power supply again yesterday. All the pins from the chips are soldered using wired to the board. Also, none of the pins are shorting, and both the chips work totally fine when working independently.
Also, the gains are the same. Rf is 30K in my case, and R1=1K. Input voltages are a little high (+-42V rectified), but the heatsink is really beefy, and I have separate heatsinks for each channel.
Also, I have a 1uF audio grade capacitor soldered to point1. All resistance and capacitor values are the same as on the schematic (except the gain ones being 30K instead of 20K).
Please let me know your thoughts.
Thanks,
Mohit
Thanks a ton for the detailed comments and suggestions, especially @bentsnake! I really appreciate everybody's time and help! The amp is still giving the same problem. I redid all the pin wiring again, but still the same issue
I am attaching the layout and the schematic for your reference with points called out for easy identification.
For mute, I have soldered a 10K ohms resistor between pins 8 and 4.
I was seeing some sort of short when I connected the two pins 3 via 0.3 ballast resistors, and got to admit that I was wrong in measuring the voltage between pin 3 of one chip and the pin 3 of the other connected to the circuit.
In the attached pic, point 2 is the ground. Caps are connected to it.
8 and 9 are the + (pin 10 inputs). 3 and 4 are the -ve inputs. 6 and 7 are the output pin inputs (pin 3). 10 is the speaker positive output and the ground is the negative for the speaker. I have not done the star ground yet, but ground is not an issue with either of the chips not in action (with the other chip's pin 3 open)
I am really baffled at this, but maybe it is something really basic I am missing. I redid the entire power supply again yesterday. All the pins from the chips are soldered using wired to the board. Also, none of the pins are shorting, and both the chips work totally fine when working independently.
Also, the gains are the same. Rf is 30K in my case, and R1=1K. Input voltages are a little high (+-42V rectified), but the heatsink is really beefy, and I have separate heatsinks for each channel.
Also, I have a 1uF audio grade capacitor soldered to point1. All resistance and capacitor values are the same as on the schematic (except the gain ones being 30K instead of 20K).
Please let me know your thoughts.
Thanks,
Mohit
Attachments
Last edited:
Mohit, here's a few suggestions.
1. Verify that the circuit wired this way can work as a stereo amplifier. Which means that each chip connected to a speaker, works correctly and both operate at the same time.
2. There should be no DC offset either individually or between the chips this way, and channel balance must be correct. As in, gain from both channels must be equal.
3. If you do not use matched resistors to 0.1%, the ballasts need to be increased or the chips will not share current properly. You can go up to 0.47 ohms if required, though it's better to use close tolerance resistors or match them well.
4. Separate the audio and power ground, do not use a single point under the board for all of them.
5. I feel the 1k input resistor is too large and causes output imbalance as the offset voltage increases with a larger resistor there. You may go lower if needed. I've used down to 100 ohms and prefer lower values for these applications. I've gone down to 15 ohms in 4 parallel design. Obviously, using 0.1% resistors is a good alternative to reducing the size. Large input stoppers always help prevent oscillation better than small ones.
Good luck!
1. Verify that the circuit wired this way can work as a stereo amplifier. Which means that each chip connected to a speaker, works correctly and both operate at the same time.
2. There should be no DC offset either individually or between the chips this way, and channel balance must be correct. As in, gain from both channels must be equal.
3. If you do not use matched resistors to 0.1%, the ballasts need to be increased or the chips will not share current properly. You can go up to 0.47 ohms if required, though it's better to use close tolerance resistors or match them well.
4. Separate the audio and power ground, do not use a single point under the board for all of them.
5. I feel the 1k input resistor is too large and causes output imbalance as the offset voltage increases with a larger resistor there. You may go lower if needed. I've used down to 100 ohms and prefer lower values for these applications. I've gone down to 15 ohms in 4 parallel design. Obviously, using 0.1% resistors is a good alternative to reducing the size. Large input stoppers always help prevent oscillation better than small ones.
Good luck!
Thanks Sangram, I will try what you suggested. Where does one get 0.1% resistors? I am only able to get 1% ones from eBay.
Also, I am an electrical engineer myself, and don't really understand how separating these grounds and later connecting them is different from connecting them all together? Are we talking resistance from some extra length of wire? Different length power ground wires and/or different length audio ground wires would introduce some potential difference between the connecting point anyway, wouldn't they?
Thanks,
Mohit
Also, I am an electrical engineer myself, and don't really understand how separating these grounds and later connecting them is different from connecting them all together? Are we talking resistance from some extra length of wire? Different length power ground wires and/or different length audio ground wires would introduce some potential difference between the connecting point anyway, wouldn't they?
Thanks,
Mohit
Engineering treats ground as on point, in fact each wire or track leading to ground has finite resistance and will have small potential differences across its length. This drop induces noise in the system. Treat the ground as truly a single point, the only way to do this is to ensure the return for each signal is being designed to handle the current in that return. Microamperes for input, many amperes for output.
AD has some excellent papers on the topic, relating specifically to opamps, which the 3886 also is. You can refer those for more detailed analysis on the topic.
You can get precision resistors on eBay too. Here's a link for your 30k value, for example - http://www.ebay.com/itm/1x-30K000-X...662?pt=LH_DefaultDomain_0&hash=item1c346ac87e
AD has some excellent papers on the topic, relating specifically to opamps, which the 3886 also is. You can refer those for more detailed analysis on the topic.
You can get precision resistors on eBay too. Here's a link for your 30k value, for example - http://www.ebay.com/itm/1x-30K000-X...662?pt=LH_DefaultDomain_0&hash=item1c346ac87e
Last edited:
Your layout is not good for a chip amp. You should reed the LM3886 data sheet on layout.
Resistance is not the problem, its inductance. For example Rf should be connected directly to the chip pins. The same for the bypass caps.
http://www.diyaudio.com/forums/chip-amps/252436-lm3886-pcb-vs-point-point-data.html#post3845470
There are plenty more P2P examples to be found on internet.
Resistance is not the problem, its inductance. For example Rf should be connected directly to the chip pins. The same for the bypass caps.
http://www.diyaudio.com/forums/chip-amps/252436-lm3886-pcb-vs-point-point-data.html#post3845470
There are plenty more P2P examples to be found on internet.
Hmm, thanks Sangram and mark for your latest comments.
I don't know if I can change the layout right away because I have to build everything from scratch, but I think that layout will make things easier in terms of connectivity as well.
I did make a start ground connection, and did what Sangram suggested. Both chips are working simultaneously as a stereo amp. There is no voltage between their outputs (my voltmeter just reads AC 200V and has a 10mv precision.
Each output to ground is as expected, and there is no DC offset.
But when I connect the 2 outputs using the 0.3 ohm ballasts, I get a non-working configuration. Is it just a possibly mismatch in output votage or something else? My multimeter cannot seem to catch it.
Also, how I am testing the circuit is by using a light bulb in series with the amp power supply. When the circuit is run as stereo, the light bulb is dim as expected, but when outputs are in parallel, I see a bright consistent glow.
Thanks,
Mohit
I don't know if I can change the layout right away because I have to build everything from scratch, but I think that layout will make things easier in terms of connectivity as well.
I did make a start ground connection, and did what Sangram suggested. Both chips are working simultaneously as a stereo amp. There is no voltage between their outputs (my voltmeter just reads AC 200V and has a 10mv precision.
Each output to ground is as expected, and there is no DC offset.
But when I connect the 2 outputs using the 0.3 ohm ballasts, I get a non-working configuration. Is it just a possibly mismatch in output votage or something else? My multimeter cannot seem to catch it.
Also, how I am testing the circuit is by using a light bulb in series with the amp power supply. When the circuit is run as stereo, the light bulb is dim as expected, but when outputs are in parallel, I see a bright consistent glow.
Thanks,
Mohit
***Please ignore my previous post (had some typos)***
Hmm, thanks Sangram and mark for your latest comments.
I don't know if I can change the layout right away because I have to build everything from scratch, but I think that that layout will make things easier in terms of connectivity as well.
I did make a star ground connection, and did what Sangram suggested. For star, I did as was suggested. Power, power supply caps, and pins 7 go to power ground. All audio grounds are together. They are connected using a short wire.
Both chips are working simultaneously as a stereo amp. There is no voltage between their outputs (my voltmeter just reads AC 200V and has a 10mv precision.
Each output to ground is as expected, and there is no DC offset.
But when I connect the 2 outputs using the 0.3 ohm ballasts, I get a non-working configuration. Is it just a possibly mismatch in output votage or something else? My multimeter cannot seem to catch it.
Also, how I am testing the circuit is by using a light bulb in series with the amp power supply. When the circuit is run as stereo, the light bulb is dim as expected, but when outputs are in parallel, I see a bright consistent glow.
Thanks,
Mohit
Hmm, thanks Sangram and mark for your latest comments.
I don't know if I can change the layout right away because I have to build everything from scratch, but I think that that layout will make things easier in terms of connectivity as well.
I did make a star ground connection, and did what Sangram suggested. For star, I did as was suggested. Power, power supply caps, and pins 7 go to power ground. All audio grounds are together. They are connected using a short wire.
Both chips are working simultaneously as a stereo amp. There is no voltage between their outputs (my voltmeter just reads AC 200V and has a 10mv precision.
Each output to ground is as expected, and there is no DC offset.
But when I connect the 2 outputs using the 0.3 ohm ballasts, I get a non-working configuration. Is it just a possibly mismatch in output votage or something else? My multimeter cannot seem to catch it.
Also, how I am testing the circuit is by using a light bulb in series with the amp power supply. When the circuit is run as stereo, the light bulb is dim as expected, but when outputs are in parallel, I see a bright consistent glow.
Thanks,
Mohit
.
<< Both chips are working simultaneously as a stereo amp >>
I can't get over the idea that all this is somehow phase-related. Do you have your input connected exactly (that's exactly) as the schematic shows? Or are you using inputs that are somehow different at different times? Or different at the same time? Is there more than one input?
I'm thinking inputs that are 180 degrees out of phase, hence outputs out of phase, hence much voltage and impressive current flow.
You don't, perhaps, have one chip connected as a non-inverting amp, and the other connected as an inverting amp?
.
<< Both chips are working simultaneously as a stereo amp >>
I can't get over the idea that all this is somehow phase-related. Do you have your input connected exactly (that's exactly) as the schematic shows? Or are you using inputs that are somehow different at different times? Or different at the same time? Is there more than one input?
I'm thinking inputs that are 180 degrees out of phase, hence outputs out of phase, hence much voltage and impressive current flow.
You don't, perhaps, have one chip connected as a non-inverting amp, and the other connected as an inverting amp?
.
Last edited:
@bentsnake, I was thinking of that too. The input is the same (one channel), and I am pretty sure that the pins 9 and 10 of each chip are connected as per the intended connections.
Should I swap the wires between pins 9 and 10 (+ and -) of one and try?
I will let you know what I find.
Thanks,
Mohit
Should I swap the wires between pins 9 and 10 (+ and -) of one and try?
I will let you know what I find.
Thanks,
Mohit
.
<< Should I swap the wires between pins 9 and 10 (+ and -) of one and try? >>
My thinking is no, mainly because I think there's been more than enough fooling around with your circuit. In my mind the key fact is that if you built the circuit given in TI's SNAA021B, it will work, I consider this to be a given. If it doesn't work, then you built something else.
I'm now suspicious about the pins on the LM3886s. The pins are irritatingly close together, and easy to short either to each other, or to some other connection or part.
.
<< Should I swap the wires between pins 9 and 10 (+ and -) of one and try? >>
My thinking is no, mainly because I think there's been more than enough fooling around with your circuit. In my mind the key fact is that if you built the circuit given in TI's SNAA021B, it will work, I consider this to be a given. If it doesn't work, then you built something else.
I'm now suspicious about the pins on the LM3886s. The pins are irritatingly close together, and easy to short either to each other, or to some other connection or part.
.
I don't know what is going on with the circuit! I was pretty sure that it may be that the channels are putting out inverted signals, and swapping pins 9 and 10 leads will solve it. Swapping them cause huge DC in output, hence that was not it.
I took apart almost all the pins of chips, but still the same thing. There is some sort of inversion going on for sure. There are no shorts between the pins, I checked each pair using a multimeter, and do not see anything.
Also, interestingly, when I plug in one of the chips that is working, and also power the other channel (which is okay), it leads to a short.
Does anybody have any idea what might be happening?
I have rebuilt the ground and don't have the slightest clue about what is happening
Thanks all,
Mohit
I took apart almost all the pins of chips, but still the same thing. There is some sort of inversion going on for sure. There are no shorts between the pins, I checked each pair using a multimeter, and do not see anything.
Also, interestingly, when I plug in one of the chips that is working, and also power the other channel (which is okay), it leads to a short.
Does anybody have any idea what might be happening?
I have rebuilt the ground and don't have the slightest clue about what is happening
Thanks all,
Mohit
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.