What should I see at the collector of an Emitter Follower power stage?

[christarak]

Good morning my knowledgeable mates. I am working on a Chinese bass amp and I'm having a brain freeze that requires thawing.

If my 1Khz sine wave comes into the Base of Q207/208 and the current amplified output (with a diode drop in voltage) appears at the Emitter, what signal should I see at the Collector?

The left half of my brain says I should ONLY see a clean DC level at +/-VCC as there is no resistor between the supply rail and the Collector. The filter caps would bypass the voltage amplified AC signal to ground. And yet that’s not what I see on my scope. The VCC rails have the sine wave superimposed on the DC. Wouldn’t this then contaminate the supply to Q205/206? “Aha! Faulty filter caps”, me thinks. I piggy backed sufficient capacitors to make a change…but no.

The right half of my brain says “it’s just a dumb transistor. It doesn’t know if its common collector or common emitter…and the signal will appear at both the emitter AND collector. Besides, if you shunt the collector AC signal to ground via the filter caps, you effectively kill the signal going through to the emitter”.

By the way, the amp works with that AC signal on the supply rails. I look forward to having the light of your wisdom shine upon the darkness of my ignorance.

[Enzo]

No signal, the collectors are connected directly to the power supply.. in conventional amp. But look again. Your amp is a "flying rail" design. Note the emitter resistors are grounded. The load connects between ground and the power supply common. So the power supply rails move around with the signal. The power supplies themselves remain steady, but the common end has the speaker signal with respect to ground.

[christarak]

No signal, the collectors are connected directly to the power supply.. in conventional amp. But look again. Your amp is a "flying rail" design. Note the emitter resistors are grounded. The load connects between ground and the power supply common. So the power supply rails move around with the signal. The power supplies themselves remain steady, but the common end has the speaker signal with respect to ground.

Thank you Enzo. Your knowledge blows me over. To understand, I will need to read your answer several times and do some woodshedding on "flying rail'. Thank you, Sensei.

[Enzo]

Flying rail, also called grounded emitter. Google I am sure has tons.

Look at your schematic closely, see the speaker connector? On the left is the ground connection and also the emitters of the "output" circuit. On the right is the connection to the power supply common.

Now think about it. Usually you see the power supply grounded and the emitters as the output. So if we left the emitters connected to the left pin and the power common connected to the right pin, all we need to do to convert this to a conventional circuit is to move the ground to the right side. In other words the two kinds of amp are pretty much the same, differing on which end of the speaker we ground.

In fact if you breadboarded this circuit on a wooden plank, so there was no chassis to ground to, you'd have this circuit with no ground.

Does that help?

[Mike Sulzer]

Inquiring minds want to know....why?

[nickb]

Inquiring minds want to know....why?

...because it has voltage gain meaning you can drive it from a cheap op amp.

[Mike Sulzer]

...because it has voltage gain meaning you can drive it from a cheap op amp.

But why do that? All discrete designs cost little more, and there are some good ones that have such low distortion is is hard to count the number of zeros after the decimal point. (OK, I exaggerate a bit.)

[Enzo]

Decimal points don't matter in a stage-use instrument amp. When you are cranking an SVT are you really listening for the 0.0002% distortion levels? I would have added some more zeroes, but I only have so many fingers... Cheap op amps will be around a long time, people are always scrambling to find transistors that have gone obsolete.

[J M Fahey]

It also simplifies designs **a lot** and reduces stress on all parts, except on power transistors what is unavoidable.

To put some numbers into it, suppose you have +/- 50V rails.
* power transistors must stand 100Vce, when speaker signals swing from rail to rail, "opposite" non-conducting transistor must stand that at every positive or negative signal peak.
* drivers must do the same.
* Vas (main gain stage) , and (if used) its constant current/active load must not only stand the same, but *swing* to full rail voltage (hint: none does) and every volt or two you lose there, are also lost from peak voltage applied to speakers,so is needlessly lost power output.
Not a BIG deal but, hey, it scratches me the wrong way.
* input differential pair must at least stand "1" rail voltage.

Now to flying rail amps:
only the first condition must be met, and none of the others.

All others must swing a only a few Volts, 5 or 6V peak at most, and in any case comfortably within +/-15V rail limits
Even the power transistor drivers
Like was said above, Power Transistors show voltage gain.
Why?
Because they are not emitter followers any more (Gain=1 by definition because load is at the emitters) but grounded emitter (that´s exactly what was done) and now load is effectively at the Collectors , although it´s not easy to visualize that at first sight.

I wonder why MORE of these are not used.
Possible explanations are that it´s not widely known, and the only real "problem": you can´t feed stereo amps from the same supply.

Have a look at QSC 400 amps: VERY simple and elegant design, only problem is that the main power transformer has two separate sets of secondaries, two rectifier bridges and separate main caps ... no big deal.

And for a single poweramp Instrument amplifier? (as in 99.99% of them) .... a very good and practical solution.

[Enzo]

QSC also took the added step of reversing the PNP and NPN outputs so the collectors were grounded. That let them bolt the outputs directly to the heat sink, no insulators required. So none of the TO3 bodies were electrically hot, unlike conventional amps where the rails are on the cases.

[g1]

For all those benefits, there must be some downside? (other than the separate supplies issue)

[darkfenriz]

for one, it's transconductance amp in open loop, made voltage amp by the virtue of feedback - load stability may be trickier to achieve and the specs such as damping factoror consistency of thd over loads may suffer, in case anyone cares (say 80s hi-fi mentality)
the other is it looks strange.
3rd the supplies are harder to use elsewhere, such as reg them down for the pre-amp

[Helmholtz]

Because they are not emitter followers any more (Gain=1 by definition because load is at the emitters) but grounded emitter (that´s exactly what was done) and no load is effectively at the Collectors , although it´s not easy to visualize that at first sight.

The "no" is confusing, must be a typo as the load appears in series with the collectors.
Otherwise excellent and comprehense explanation, thanks.

[J M Fahey]

The "no" is confusing, must be a typo as the load appears in series with the collectors.
Otherwise excellent and comprehense explanation, thanks.

Sorry, it means "now".
I lost both "a" and "w" keys, thanks to sandwich crumbs getting under keys (we eat French bread so crumbs are *hard*), so I had to blow them away with compressed air.
Corrected typos but clearly missed that one.
Hope I can still edit the post, not sure 24 hours later but will try.

ERRATA:
and no load is effectively at the Collectors should read and now load is effectively at the Collectors

[Mike Sulzer]

I think this over-simplified diagram makes what Juan said more clear. If you like, you can think of those voltage sources representing the power supply as batteries.