The ChipSET - A Concept Chip Amplifier

[Steve Eddy]

For the past couple of years I've been developing an amplifier concept which uses step-up transformers to achieve totally passive voltage gain with the transformers driving an active output current buffer for driving the loudspeaker.

What I've been using so far have been various discrete single-ended buffers which sound great but are horribly inefficient. Anyway, the other day I got thinking that a power opamp could be used for the output buffer which would make for not only a more efficient (and subsequently much smaller) solution but also one which would be very simple to build.

Here's the basic circuit:



About the only transformer I'd recommend for this is the CineMag CMMI-10C. Which means that R1 would ideally be 150k ohms. However unless you'd care to make a custom 150k stepped attenuator, you're not going to have much luck finding 150k ohm pots. A 100k pot can be used instead.

Most opamps aren't unity gain stable and power opamps such as the ubiquitous LM-3875 require a minimum closed loop gain of 10.

R2 and C1 are used to keep the opamp's noise gain at a minimum of 10 while the circuit has a signal gain of 1. The values of R2 and C1 will ultimately depend on the opamp used. A nice tutorial on how to determine the appropriate values can be found in Apex Microtechnologies' Application Note 1 in section 10.2.1.

R3 should be about 25k ohms.

The intended use for this amp is driving high efficiency speakers where you're usually cruising in the sub-watt range. To that end, R4 is included, through which the opamp's output is tied to the negative supply rail, biasing its output stage higher into class A.

Unlike a typical complimentary (or quasi-complimentary) push-pull amplifier where cranking up its internal bias brings both halves of the output stage into class A, with this method, within the class A range, the output stage is actually operating single-ended, rather than push-pull, hence the "SET" in "ChipSET."

For reasonably sensitive speakers (say 96dB and above), 500mA should be sufficient that the amp will be running single-ended except on transient peaks. So if you're running say a +/- 24 volt supply, use a 48 ohm, 25 watt resistor for R4.

While this makes the amp less efficient, it's not nearly as inefficient as a pure single-ended solution for the same total output power.

If you're running less efficient speakers, I'd recommend leaving R4 out.

I call this a concept amplifier because I haven't built this actual circuit yet and I don't know when I'll get the time to do it. I just thought I'd throw this out in case someone with more time wanted to take the basic concept and run with it.

Oh, the one potential "drawback" is that this amp will have an input impedance of between 1k-1.5k ohms and should ideally be driven by a source impedance of 150 ohms or less. This won't tend to pose any problems for sources with solid state outputs but most tube output stages won't get along with it too well.

se

[carlosfm]

Steve!
You can use the OPA549, OPA541 or even the OPA548 at unity gain.
At low gains they sound very good.

[Steve Eddy]

Originally posted by carlosfm
You can use the OPA549, OPA541 or even the OPA548 at unity gain.

Thanks, Carlos. Yeah, there are some that are unity gain stable, but it seems the more popular ones aren't.

At low gains they sound very good.

Low gain, no pain, eh?

se

[analog_sa]

SE

Could you please explain the idea behind the amp. So, voltage gain is handled by a transformer at the expense of the preamp. Why is this beneficial? Would the opamp feel that much better if it is relieved of any voltage amplification duties?

[Steve Eddy]

Originally posted by analog_sa
Could you please explain the idea behind the amp. So, voltage gain is handled by a transformer at the expense of the preamp. Why is this beneficial?

Um, the expense of the preamp perhaps?

Would the opamp feel that much better if it is relieved of any voltage amplification duties?

You'd have to ask the opamp.

se

[TB]

Hello Steve, similar idea has been already discussed on the forum of VALVE web-magazine by Bottlehead. I do not think VALVE is still being published, but I found the link to members forum with amp's schematic:
http://db.audioasylum.com/cgi/m.mpl?...en&r=&session=
For those who are interested, if the link does not work go to bottlehead forum and search it for Love Kitten.
The project was a volume control ( in shunt mode ) feeding Magnaquest transformer. So the idea is similar to yours but, reversed; volume control -->X-former --> output.

[Steve Eddy]

Thanks, TB.

Yeah, I'm familiar with Jeff's Love Kitten. But it's opposite to mine in another respect in that the transformer used is a step-down transformer, which is required in order to get sufficiently low output impedance. The consequence is that you need an amp with enough excess gain to overcome the loss of gain due to the transformer's stepping down the voltage.

Jack Elliano of ElectraPrint uses a scheme that's more similar to mine in that it uses a step-up transformer for voltage gain followed by a pot for volume control but doesn't use any active circuit on the output. So it's transformer -> volume control -> output.

However in order to keep the output impedance sufficiently low in order to drive a reasonable length of cable with it (it's intended to be a stand-alone preamp), the transformer's loaded with a 10k pot. The consequence is that it has an even lower input impedance than the already fairly low 1.5k ohms of mine. In fact it's about 10 times lower at just 150 ohms.

se

[analog_sa]

SE

I am not trolling your idea at all, sorry if that's the way it appeared. If you say that you built it and it sounds incredible that's fine with me. If it's just a concept i would really like to know what's the point of getting an opamp not to do what opamps do best and a transformer to do what transformers do worst.

[Stocker]

The National chips are unity-gain stable in inverting configuration.

[carlosfm]

Where have you seen that?