My first non-inverted gainclone circuit

[eranrund]

Hello people.

Below is my first non-inverted gainclone circuit. It is powered from a toroid-based power supply which will supply around 25-30v to it (still need to decide - any help will be appritiated). The power supply has 10000uf of total capacitence on each rail (shared by two copies of this circuit...)
I plan to power 6-8ohm speakers with it.
Volume controll will be established using an external, active, pre-amplifier yet to be built

Please - try be realistic. This is my first project - dont kill me

Thanks a lot

[Nuuk]

Not sure what to say. It looks OK but I would use a smaller film capacitor for the DC blocker.

If you are using 10K caps in the PSU, I suggest that you have a look at Carlos' snubberized PSU.

[eranrund]

Heya.

I did put snubbers on the PSU
Ok, I will use a smaller capacitor. How about 4.7uF?
Except that. will it be ok?
Any other things worth adding/changing/removing?

Thanks.

[Nuuk]

How about 4.7uF?

That's fine and do make sure it is a good film type as this is the most important cap as regards sound quality.

Anything else need changing? No, I don't think so but you can always read my thoughts on GC's at Decibel Dungeon.

[eranrund]

Thanks dude
Ofcourse I've read your site. An excellent one, I must say!

[eranrund]

Hey guys.
I've made some more changes and designed a PCB:

[ScottRHinson]

Originally posted by eranrund
Hello people.

Below is my first non-inverted gainclone circuit. It is powered from a toroid-based power supply which will supply around 25-30v to it (still need to decide - any help will be appritiated). The power supply has 10000uf of total capacitence on each rail (shared by two copies of this circuit...)
I plan to power 6-8ohm speakers with it.
Volume controll will be established using an external, active, pre-amplifier yet to be built

Please - try be realistic. This is my first project - dont kill me

Thanks a lot

Eranrund,

If you don't mind...a couple of suggestions:

1. I don't see a path for DC bias currents for the input. I would move C2 directly accross pin 7 and 8. Also move R2 to the other side of C1.

2. R3 and R1 are somewhat high in value, try 22K for R3, 1K for R1 and 100uF for C3.

3. The National application note for the 3886 mentions the potential need for a resistor accross the input terminals, at the input terminals to prevent oscillations when there is nothing connected to those input terminals. Leave yourself a spot on the PCB for a 10K or a 5K resistor and a small cap (to filter RF) right at the input terminals.

4. Try to get your feedback and input components closer to the 3875 chip on your circuit layout.

5. Provide separate ground returns to the star point of your chassis for signal ground (R2 and C3 return) than power ground (R4, C7, C8, C5, C6). Study the suggested layout for the 4870.

6. The power supply bypass capacitors need to be closer to the chip.

I have just finished putting together a 3886 board. When I first started it up I was using some linearly regulated supplies with test leads that were longer than I would have liked. In the process I discovered how sensitive the high gain, wide bandwidth chip is. I had oscillation issues with that test supply even though I had all of my bypass capacitors as close as physically possible to the chip, my feedback and input circuitry was in a tight loop around the chip pins, and I had separated my signal and power grounds all the way back to the star point.

When I switched to the real supply with shorter, heavier gauge wiring all of these issues went away, but it did give me a better appreciation for the warnings posted in the data sheet.


Scott

[Russ White]

Quite a good start!

[eranrund]

Heya.

New versions:
- still need to figure out how to do star grounds - no idea how :-(
i think i'm gonna solder directly to the chip extra capacitors for +ve and -ve. i might solder the feedbacks on it as well.

Please give some more comments

[ScottRHinson]

Originally posted by eranrund
Heya.

New versions:
- still need to figure out how to do star grounds - no idea how :-(
i think i'm gonna solder directly to the chip extra capacitors for +ve and -ve. i might solder the feedbacks on it as well.

Please give some more comments

For an excellent example of star grounding look at the datasheet for the LM4870 (4780?). When doing your layout I suggest the following steps.

1. Place amplifier chip.
2. Place feedback parts next to the pins they are connected to.
3. Place input parts next to the pins they are connected to.
4. Place bypass parts next to the pins they are connected to.
5. Place everything else.
6. Route bypass capacitors with wide traces.
7. Route feedback nets keeping all traces as short as possible.
8. Route input nets keeping all traces as short as possible.
9. Route everything else.

Scott