Gainclone Embedded

[rjm]

I'd like to run an idea past you, to get some feedback before I go ahead.

I bought a pair of Elecom MS-76 computer speakers the other day. Cost was only Y3,000 including shipping, and for that you get two reasonably well-built MDF cabinets, a pair of cheapo drivers and about 4.5wpc of amplification. The little speaker grills are in fact removable, just like on real speakers. They are pretty cute.

The back of the cabinet is held in place by four screws, so my thoughts naturally start turning to replacing the electronics with a pair of gainclone amplifiers.

Let me make it clear that this is not to improve the sound, in asmuch as it would not be "worth" the effort expended. I just think it would be a fun challenge.

I searched the forum but didn't find any references to similar projects.

Anyone know of something similar being done before?

[AndrewT]

worth doing.
Fit a monoblock amp in each cabinet.
Try 18+18Vac, 80VA transformer, +-10mF smoothing, 3886 with mute delayed, gain ~28dB, external heatsink, IEC socket, RCA input.
Add all the National optional components to limit bandwidth and ensure stability.

[boholm]

You are looking at it the wrong way. Most probably true, there are no projects that *only* targets such speakers. But Gainclones ca be used for *any* speaker (well OK, almost ;-). Therefore search for "gainclone" and you will get so many hits that you will get confused again. Most of them will be about the circuit that is showed in the data sheets, but there are variations.

Find the data sheet for your chip, do a search on your chip and have Andrews suggestion ready beside you, while looking at your search hits.

[rjm]

What I'm really after is not a standard gainclone - which I'm reasonably familiar with after all - but how to tweak the circuit for this specific application.

The two main points are;

1. its a tiny, cheap speaker

- the driver is going to distort like crazy if it ever sees more than 1W, and size, heat, and safety constraints suggest limiting the amplifier output power to the 2-5 wpc range usually seen driving these types of speakers. It's a 4 ohm load, so going with the datasheet minimum voltage rails (V+ - V- = 20V) and a fairly minute transformer (25 VA 2x7VAC) seems like a reasonable idea.

2. the computer is the preamplifier

- this is probably the more interesting aspect. Normally the computer volume slider is set about 80%. That's actually quite high, you don't need very much gain from an external amplifier at that setting. So I can trade some of the normal 30db gain of a gainclone away for something else. My cunning plan is to trade it for some extra input impedance so I can run the circuit inverted while keeping a low impedance seen from the op amp inputs.

The input attenuator also cuts any DC offset from the computer (the outputs are usually capacitively coupled anyway) so I can get away with not using an input coupling cap. Since the amp is right next to the load, the Zobel is probably ok to leave off too, but I can add it at the driver terminals if need be.

I would do this point to point, but as an example here's the Eagle schematic and 1.5" square layout for the circuit I'm describing.

/R



schematic

[pacificblue]

2x7 V AC is a bit tight, if you want to achieve ±10 V DC. Even if regulation will help you at those power levels with such a small transformer, mains voltage tolerance has to be considered.

The best value for R3 = R2||R4 = ~319 Ohm.

Is R0 the right size? That looks like a lot of attentuation.

[AndrewT]

the 1875 operates from a "wide supply range".
16Vdc to 60Vdc.

The 7Vac transformer may allow the supply drop below +-8Vdc (=16V supply) and intermittently shut down.

This will be worse if you adopt the Low Smoothing Capacitance that the GainCard is designed around.
The supply line ripple may drop the minimum voltage below +-8Vdc on every mains cycle.

[rjm]

It will actually shut down on too low a supply voltage? I didn't know that.

2x9VAC would be a safer bet then, unfortunately I have a 2x7VAC trafo sitting unused in my desk drawer...

The overall gain should be -6 dB. Judging from my current volume settings that should be about right but its a bit of a guess until I have a chance to actually try it out.

[pacificblue]

Originally posted by rjm
It will actually shut down on too low a supply voltage? I didn't know that.

AN-898 deals with the function of the SpiKe protection system.

Originally posted by rjm
The overall gain should be -6 dB. Judging from my current volume settings that should be about right but its a bit of a guess until I have a chance to actually try it out.

If you want -6 dB, you don't need an amplifier. You need an attenuator.

[rjm]

Originally posted by pacificblue

AN-898 deals with the function of the SpiKe protection system.


If you want -6 dB, you don't need an amplifier. You need an attenuator.

According to AN-898 the LM3875 undervoltage protection kicks in when V+-V- < 14 V, so 2x7VAC should be OK. The LM3886 and other "6" series chips have an additional trip point at V- under -9V, so 2x7VAC probably wouldn't work with those ICs.

An attenuator? Surely not. You could call it a transimpedance amplifier if you insist, or a buffer stage. I prefer to think of it as a voltage amplifier with a impedance multiplier on the input...

[digi01]

i think you can try TDA1521,LM series is a bit to HUGE

if you like solid stage&point to point,check this:simple otl amp

have fun,
Zang