IC audio amplifiers

This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
I am specifically referring to the overture series from national semiconductor.

what all needs to be considered when using these? Is it as simple as providing the proper rail voltages and connecting an input & output?

I would like to build some nice diy computer speakers.

has anybody played with these and would care to comment?


Hi jt,

I have recently build an Amp for my Computer with that nice LM 1876 from National Semi. I used the schematic as they described in the data sheet and have to say, that this great IC sound very good for this purpose. No hum , no noise just sound.
Try it !

Well that is certainly promising.

I am thinking about two of the LM3886 or one LM4766 coupled with the class D LM4751.

I am going to use some inexpensive 1" dome tweeters & ~5" drivers for the desk, and 2 6.5"'s or 1 8" for the subwoofer.

~40 watts per each satellite and ~100-120 for the sub.

I am a little concerned about the power supply. I have only built one in my life and it was very specific about exactly what I needed to build it.

For the power I am looking for the rail voltages on all 3 IC's needs to be right around +/- 22-25V. Can I use the same supply for all 3 IC's? I would think so, especially since this isn't uber hi-fi or anything. Secondly, where is a good place I might be able to find a transformer that would give me ~26-30V secondary windings for not a lot of money. The one I had built for my leach amp was +/-58V? and cost me 60 bucks. A lil too much for this project.

Last Q, Would there be any problems using a protoboard(other than the mess) to do this rather than making a layout/PCB?

Thanks a lot!

LM3886 are very good (for their price) but there is some tradeoff:

They can handle only 4 ohms at +-25V, which means that you cannot make a 4/8ohms amplifier with a 35v PSU.

The power protection is very good (hum, not really, I have one lm1876 with a crater on the package), but introduce very nasty spikes when close from maximum modulation.

There is also some stability problem, so stick with the reference design. Snubber network (Rsn and Csn) can cause some oscillation when used.

Don't forget also to bypass power supply with caps really close from the supply pins (value is not critical, about 470uF/1000uF + 100nF ceramic). For +-25V you will need a 2*18V transformer with about 4700uF per amp and per rail min.

When using one supply for many amp, you will probably have problem with hum from improper grounding (I always have). Useful tips: separate IC ground from analog ground (because of bypass current flow in ground), use star topology (all ground tied to one point).
I made an Amp for My Computor using an LM-3886 operating off (2) 160 watt transformers providing +24 &-24 Volts DC after Rectification. and 25,000uF X(4) rail capacitence. This sounds very good for a 30 watt Amp. The Circuit i used is showen below although it is Heavaly Modified but this will give you an Idea of the PC layout since no place is provided on this site to attach docs. Sorry cant show then
Use a PCB

I've built a few amps using National's amps and suggest you use a PCB if possible. Besides tight wiring problems, it is easy to get feedback that causes oscillation, and a real problem trying to make it stop.

Rod Elliott has PCBs for sale at reasonable prices from his site:
Project 19 is for the LM3876 or 3886. Rod says he is quite happy with the sound if it is not driven into clipping.
Elliott PCBs

I should have mentioned that I bought 2 of the Project 19 PCBs (also used in Project 73) and the quality of the boards is good. This will be an easy amp to build, not much bigger than the size of the heatsinks. My current thinking is that the power supply will go in a separate box.
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.