I was following the guide on circuit-projects site to build my first Hi-Fi Amplifier. The Circuit there is based on the Typical Application circuit from the datasheet.
The Circuit is minimal and easy.
But Later I saw there was a PCB for LM3886 on sale on ebay. It would save me a lot of time but I am unsure of the circuit and if it's better or worse than the minimal one! It has a dc blocking capacitor.. Does it make it inferior to the first circuit? or should i avoid it and use a preamplifier? Here's the circuit..
An externally hosted image should be here but it was not working when we last tested it.
The Circuit is minimal and easy.
But Later I saw there was a PCB for LM3886 on sale on ebay. It would save me a lot of time but I am unsure of the circuit and if it's better or worse than the minimal one! It has a dc blocking capacitor.. Does it make it inferior to the first circuit? or should i avoid it and use a preamplifier? Here's the circuit..
An externally hosted image should be here but it was not working when we last tested it.
.
You've opened a can or worms. On the question of "which circuit to use?" feelings run high.
Here's one vote for: a) Use the factory circuit, it was developed by the same team of engineers who created the chip in the first place. b) Most of the worry about using capacitors is urban legend. The real gurus (Pass, Self, etc) don't hesitate to put in a capacitor when they need one, so how bad can they be?
Just to mention it, Radio Shack's perfboard is something of a best kept secret. If you're not wiring up circuits for a jet fighter it will do the job for you: Pre Punched IC Spacing Perfboard 2266451 | eBay
.
You've opened a can or worms. On the question of "which circuit to use?" feelings run high.
Here's one vote for: a) Use the factory circuit, it was developed by the same team of engineers who created the chip in the first place. b) Most of the worry about using capacitors is urban legend. The real gurus (Pass, Self, etc) don't hesitate to put in a capacitor when they need one, so how bad can they be?
Just to mention it, Radio Shack's perfboard is something of a best kept secret. If you're not wiring up circuits for a jet fighter it will do the job for you: Pre Punched IC Spacing Perfboard 2266451 | eBay
.
The typical application circuit will work fine, provided that you decouple the supply, use the Zobel (R+C to ground) and Thiele (L||R in series with the output) networks on the output, provide the LM3886 with a good sized heat sink, and pay attention when wiring it up. Basically: Read the data sheet (in particular the application section in the back) and follow the advice given there.
I have tons of information available here: Taming the LM3886 Chip Amplifier
You may also find these DIY Audio threads interesting:
http://www.diyaudio.com/forums/chip-amps/252436-lm3886-pcb-vs-point-point-data.html
http://www.diyaudio.com/forums/chip-amps/265771-lm3886-thermal-experiment-data.html
~Tom
I have tons of information available here: Taming the LM3886 Chip Amplifier
You may also find these DIY Audio threads interesting:
http://www.diyaudio.com/forums/chip-amps/252436-lm3886-pcb-vs-point-point-data.html
http://www.diyaudio.com/forums/chip-amps/265771-lm3886-thermal-experiment-data.html
~Tom
Last edited:
.b) Most of the worry about using capacitors is urban legend. The real gurus (Pass, Self, etc) don't hesitate to put in a capacitor when they need one, so how bad can they be?
They can be quite bad if you choose the wrong type. Douglas Self touches on this in his Power Amplifiers book and even more in his Small Signal Audio book. He backs up his claims with data (which I like).
The bottom line is: Dielectrics matter. Parasitics matter. Brands and snake oil do not.
If you use electrolytic capacitors in the signal path, ensure that the pole or zero caused by them is way outside the audio band. For a DC blocking cap, this means the pole is placed in the single-digit mHz range in a high-performance amp. Also ensure that no significant DC voltage can develop across the electrolytic cap if used this way.
Polypropylene capacitors are transparent for audio. They don't cause measurable distortion when used as coupling caps.
C0G/NP0 dielectric ceramic capacitors can be used for compensation caps (and coupling caps in low-voltage circuits) as they don't cause measurable distortion either.
If you're curious about the performance of components such as resistors and capacitors, I recommend reading the relevant chapter in Self's "Small Signal Audio" book. He just released a new edition of it.
~Tom
.
Did anybody see the part about "build my first Hi-Fi Amplifier"? Sounds like a fun weekend project, huh? But no, instead let's send some poor guy off into a 6-month study course. Because that really is helping a guy out, isn't it?
ayushextreme, do as you think best, of course. But I can tell you that if you build an amp with Mylar film capacitors for small values, and electrolytics for large values, you'll be pleased as punch with the result. Your wife will be impressed, your kids will be proud of dad, your friends will envy you.
Some links that might help. Disclosure: I have no association with these people, except I send them money sometimes.
Mylar Film Capacitors: items in store on eBay!
Electrolytic Capacitors: items in store on eBay!
As for what anybody else tells you, well, that's what they tell you.
.
Did anybody see the part about "build my first Hi-Fi Amplifier"? Sounds like a fun weekend project, huh? But no, instead let's send some poor guy off into a 6-month study course. Because that really is helping a guy out, isn't it?
ayushextreme, do as you think best, of course. But I can tell you that if you build an amp with Mylar film capacitors for small values, and electrolytics for large values, you'll be pleased as punch with the result. Your wife will be impressed, your kids will be proud of dad, your friends will envy you.
Some links that might help. Disclosure: I have no association with these people, except I send them money sometimes.
Mylar Film Capacitors: items in store on eBay!
Electrolytic Capacitors: items in store on eBay!
As for what anybody else tells you, well, that's what they tell you.
.
You've opened a can or worms. On the question of "which circuit to use?" feelings run high.
I guess so.
~Tom
The schematic found in this manual stays pretty close to the data sheet recommendations, with some enhancements:
http://www.akitika.com/documents/AssemblyManualGT101rev1p38.pdf
It has a few years of successful use by satisfied customers.
http://www.akitika.com/documents/AssemblyManualGT101rev1p38.pdf
It has a few years of successful use by satisfied customers.
They can be quite bad if you choose the wrong type. Douglas Self touches on this in his Power Amplifiers book and even more in his Small Signal Audio book. He backs up his claims with data (which I like).
The bottom line is: Dielectrics matter. Parasitics matter. Brands and snake oil do not.
If you use electrolytic capacitors in the signal path, ensure that the pole or zero caused by them is way outside the audio band. For a DC blocking cap, this means the pole is placed in the single-digit mHz range in a high-performance amp. Also ensure that no significant DC voltage can develop across the electrolytic cap if used this way.
Polypropylene capacitors are transparent for audio. They don't cause measurable distortion when used as coupling caps.
C0G/NP0 dielectric ceramic capacitors can be used for compensation caps (and coupling caps in low-voltage circuits) as they don't cause measurable distortion either.
If you're curious about the performance of components such as resistors and capacitors, I recommend reading the relevant chapter in Self's "Small Signal Audio" book. He just released a new edition of it.
~Tom
I think the circuit diagram was not visible before..
It has a Zobel and theile network but i am unsure of the mute circuit.
I see a pic !
What happened to the pics in post1?
It is better to attach pics. That way we don't lose them at some later time.
This sch is quite good, in that it shows many of the components that National describe as "optional", when in reality they are NOT optional. They should be considered compulsory for all beginners and many others.
C1 is OK.
C2 should not be connected to -IN pin, move it to Signal Return.
C7 & C8 should be connected to each other, not to a remote power ground.
C9 and C10 should be connected together, not to a remote Power Ground.
Then connect these two decoupling ground together.
C6 should not be connected to a remote power ground. Connect it to the local decoupling ground.
Now connect the local decoupling ground to the remote power ground.
Locate L1 and R7 Off the PCB, preferably on the cable route going to the amplifier's output terminals.
R7 is too big. It needs to be much lower to provide effective damping to the output inductor. Try 1r0 to 5r.
I use 5r here and add second Zobel across the output terminals, this second Zobel can be 100n+5r. The total load seen by the amplifier when the speaker is not connected is 10r+100n || 10r+100n = 5r+200nF If R6=10r and C6=100nF. R6 can be lower. try 5r6 with the second Zobel reduced to 1r0.
The ratio of C1:C3 is wrong.
C3 >sqrt(2)*C1*{R2+R1}/R3 use 75uF or 100uF or 150uF or 220uF, going too big will increase the time and the size of the output "thump" during switch on and maybe off, if the unmute is too quick.
Add provision for an optional "break" between Signal Return and Power Ground.
This "break" can be a wire link, or a parallel combination of R||Power Diode||reverse Power Diode.
Add Power Diodes from both supply rails to output pin 3
A nice to have but not compulsory: add a pair of Power Diodes from the supply rails to the Power Ground on the PCB. This helps protect the amplifier from damage if you connect the PSU incorrectly and is especially good when you use the Mains Bulb Tester for first amplifier power ON.
Keep C4 but do not populate until after testing the amplifier. You can find that C4 causes oscillation, or makes the stability margins too low.
What happened to the pics in post1?
It is better to attach pics. That way we don't lose them at some later time.
This sch is quite good, in that it shows many of the components that National describe as "optional", when in reality they are NOT optional. They should be considered compulsory for all beginners and many others.
C1 is OK.
C2 should not be connected to -IN pin, move it to Signal Return.
C7 & C8 should be connected to each other, not to a remote power ground.
C9 and C10 should be connected together, not to a remote Power Ground.
Then connect these two decoupling ground together.
C6 should not be connected to a remote power ground. Connect it to the local decoupling ground.
Now connect the local decoupling ground to the remote power ground.
Locate L1 and R7 Off the PCB, preferably on the cable route going to the amplifier's output terminals.
R7 is too big. It needs to be much lower to provide effective damping to the output inductor. Try 1r0 to 5r.
I use 5r here and add second Zobel across the output terminals, this second Zobel can be 100n+5r. The total load seen by the amplifier when the speaker is not connected is 10r+100n || 10r+100n = 5r+200nF If R6=10r and C6=100nF. R6 can be lower. try 5r6 with the second Zobel reduced to 1r0.
The ratio of C1:C3 is wrong.
C3 >sqrt(2)*C1*{R2+R1}/R3 use 75uF or 100uF or 150uF or 220uF, going too big will increase the time and the size of the output "thump" during switch on and maybe off, if the unmute is too quick.
Add provision for an optional "break" between Signal Return and Power Ground.
This "break" can be a wire link, or a parallel combination of R||Power Diode||reverse Power Diode.
Add Power Diodes from both supply rails to output pin 3
A nice to have but not compulsory: add a pair of Power Diodes from the supply rails to the Power Ground on the PCB. This helps protect the amplifier from damage if you connect the PSU incorrectly and is especially good when you use the Mains Bulb Tester for first amplifier power ON.
Keep C4 but do not populate until after testing the amplifier. You can find that C4 causes oscillation, or makes the stability margins too low.
Last edited:
And it's in my comment again.That 220pF capacitor should be between ground not negative feedback. Its like that on TI's datasheet but It would be better to put it like that... AndrewT commented about that once I think...
not so fast
if the intent with C2 is "noise gain" compensation it actually is taking advantage of the input series R to give the desired frequency shaping of the "noise gain" for stability
my quick estimates don't lead me to think that the feedback stability/compensation is really well thought out, safe for worst case chips
it is a "problem" that the LM3866 datasheet uses the decades old numbers when we suspect modern processes must be able to do better, have lower variance, hold tighter to the design typical spec - possibly all new production 3886 are above average
if the intent with C2 is "noise gain" compensation it actually is taking advantage of the input series R to give the desired frequency shaping of the "noise gain" for stability
my quick estimates don't lead me to think that the feedback stability/compensation is really well thought out, safe for worst case chips
it is a "problem" that the LM3866 datasheet uses the decades old numbers when we suspect modern processes must be able to do better, have lower variance, hold tighter to the design typical spec - possibly all new production 3886 are above average
Last edited:
And it's in my comment again.
It will only reduce RF rejection. Doesn't really matter anyway.
And it's in my comment again.
The schematic is available as a PCB. I can't make changes to it! Should I buy it then?
You can look at your PCB and find the easiest way to re-locate the RF filter capacitor.
For an amplifier that omits it completely, it is usually easiest to add it in parallel the the Rin resistor, 100k || 1nF is typical.
Could you show me a good LM3886 PCB to make then? A single sided one for toner transfer
I see a pic !
What happened to the pics in post1?
It is better to attach pics. That way we don't lose them at some later time.
This sch is quite good, in that it shows many of the components that National describe as "optional", when in reality they are NOT optional. They should be considered compulsory for all beginners and many others.
C1 is OK.
C2 should not be connected to -IN pin, move it to Signal Return.
C7 & C8 should be connected to each other, not to a remote power ground.
C9 and C10 should be connected together, not to a remote Power Ground.
Then connect these two decoupling ground together.
C6 should not be connected to a remote power ground. Connect it to the local decoupling ground.
Now connect the local decoupling ground to the remote power ground.
Locate L1 and R7 Off the PCB, preferably on the cable route going to the amplifier's output terminals.
R7 is too big. It needs to be much lower to provide effective damping to the output inductor. Try 1r0 to 5r.
I use 5r here and add second Zobel across the output terminals, this second Zobel can be 100n+5r. The total load seen by the amplifier when the speaker is not connected is 10r+100n || 10r+100n = 5r+200nF If R6=10r and C6=100nF. R6 can be lower. try 5r6 with the second Zobel reduced to 1r0.
The ratio of C1:C3 is wrong.
C3 >sqrt(2)*C1*{R2+R1}/R3 use 75uF or 100uF or 150uF or 220uF, going too big will increase the time and the size of the output "thump" during switch on and maybe off, if the unmute is too quick.
Add provision for an optional "break" between Signal Return and Power Ground.
This "break" can be a wire link, or a parallel combination of R||Power Diode||reverse Power Diode.
Add Power Diodes from both supply rails to output pin 3
A nice to have but not compulsory: add a pair of Power Diodes from the supply rails to the Power Ground on the PCB. This helps protect the amplifier from damage if you connect the PSU incorrectly and is especially good when you use the Mains Bulb Tester for first amplifier power ON.
Keep C4 but do not populate until after testing the amplifier. You can find that C4 causes oscillation, or makes the stability margins too low.
"Add provision for an optional "break" between Signal Return and Power Ground.
This "break" can be a wire link, or a parallel combination of R||Power Diode||reverse Power Diode.
Add Power Diodes from both supply rails to output pin 3"
Could you explain why? I am going to route a PCB for this, I am planning to add Baxandall active volume controls with an RF filter on the input. I have a ganged pot but it has an offset of 10k. Will it cause an imbalance of the channels?
The first is to allow a fault current from a faulty source to escape via the diodes to the PE inside this amplifier. That current then blows the fuse of the faulty ClassII equipment that has no PE connection.
The second is to prevent the output rail being driven outside the supply rails due to back emf of the speaker.
The second is to prevent the output rail being driven outside the supply rails due to back emf of the speaker.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
- Home
- Amplifiers
- Chip Amps
- Which LM3886 circuit is better?