London Power ad

Search the Forum

(Advanced Search)

SE power amps that are not
Hi guys

A lot of guys build what they think are single-ended (SE, not Swedish) that are not. These are solid-state designs and it is is generally hifi builds that are in question.

BJT = bipolar junction transistor, the usual kind generically called "transistors"

MOSFET = metal-oxide semiconductor field-effect transistor

Both semiconductors come in two sexes as P or N; for BJTs this is PNP and NPN.

A single-ended amp gets its name from tube circuits and the connection between the output transformer (OT) and the power tube. One end of the OT primary winding  is tied to B+ while the other end is driven by the tube plate. The tube idles at a current that is a little bit higher than the peak signal current. During the positive half of the signal, the tube conducts more current, pulling the end of the winding towards ground. On the negative half-cycle, the tube conducts less current and the energy stored in the OT causes the plate voltage to swing well above B+. The current through the tube swings from approximately Ipk, to (2 x Ipk) to almost zero, averaging at Ipk.

In the tube circuit, the OT is driven at one end, thus "single-ended". A push-pull circuit uses a centre-tapped primary OT with a tube driving each end.

Where the OT is the usual means to match the high-voltage environment of the tube to the low-voltage requirement of the speaker, the tube can be choke-loaded and capacitively-coupled to the speaker provided the power output is low AND the speaker impedance is high. This is only done with headphones. A variation on this theme is to use a choke load capacitively-coupled to an OT that carries no idle current for the tube - the choke does that. In these latter cases, the choke acts just as the SE OT does except it does not have a secondary to match the speaker load. The variation uses a separate OT for that. because this OT does not have to be gapped to withstand the DC idle current, its performance can be greatly improved and its size reduced.

We can configure solid-state SE exactly like the tube circuits above, with a conventional but lower-impedance OT, with choke loading, and with the choke - cap - OT. But...

Part of the advantage of solid-state is that semiconductors can pass the heavy speaker currents directly. So, the first attempt is to replace the load with a current source. The current source is an active circuit that does as its name suggests, it supplies a constant current regardless of the voltage across itself. Technically, the circuit can be a current "source" or a current "sink" depending on which side of the signal-driving element it is on.

An SE BJT amp output stage looks pretty much like any push-pull BJT output stage EXCEPT that the signal is only applied to half of it. Suppose we have an NPN BJT for the upper element, C tied to V+, E tied to the load and B tied to the drive circuit which we do not care about here. A second NPN has its E tied to V- and C tied to the load, which is also the emitter of the signal transistor. The lower BJT is configured one of many ways to provide a steady current. Feedback MUST be taken from the junction of the two BJTs to control the DC voltage at the very least even if we use a coupling cap to the speaker.

Suppose the idle current is 1A and the load is 8R. Suppose the supply rails are 10V each. With no signal, each BJT has 10V across it and 1A through it, and thus each device dissipates 10W at idle.

For the positive signal half, the upper BJT is turned 'on' more until it conducts 2A peak. At that point, the upper transistor dissipates 2A x 2V = 4W.  Meanwhile, the lower BJT still has 1A but now has 18V across it and 18W of heat. The speaker is conducting the difference of current, at 1A and 8V with 8W peak power.

For the negative half-cycle, the upper transistor is driven 'off' until it is at zero current with 18V across it. The current source still has 1A through itself but now at 2V, so 2W. The 1A from the current source flows through the speaker producing the 1A 8W negative peak of the signal.

It is clear that the signal-driven BJT only contributes to one half of the signal while the current source makes up the other half. This is really an aesthetic SE that is actually push-pull. It should be clear that the current source is much more relevant to performance than most people believe and that the current source should be a "good" one rather than a simple buffered voltage reference. In audio, current sources should always be the feedback type, comprised of two same-sex devices, as two BJTs or as a MOSFET pass element and BJT control element.

Have fun

Forum Jump:

Come in where it's warm!
A warm welcome to tube amp modding fans and those interested in hi-fi audio! Readers of Kevin O'Connor's The Ultimate Tone (TUT) book series form a part of our population. Kevin O'Connor is the creator of the popular Power Scaling methodology for amplifiers.
Please remember these three principles: respect, sharing, community.
Not familiar with The Ultimate Tone book series? See discussion topics, or click here to visit London Power/Power Press Publishing.

Tube Amp Forum Hosted by London Power
London Power logo