15-09-2024, 12:54 PM
It's not just badly drawn - par for the course for Grundig, IMO - but it is an unusual topology.
The radio is positive earth, and everything apart from the output stage runs from -12V. That is generated by the middle of the two windings, and stabilised by the zener diode. That much is easy to see.
What is different about this is the winding for the output stage, which yes, is smoothed by C67. This is around 12V.
I had to sketch it out to be sure, but this is a variation of the Transova circuit. It has been used by several manufacturers over the years, including Haffler and QSC. UK monitor speaker manufacturer ATC have used it (probably still do?) in their active speakers - they call it "grounded source" in their marketing. I understand the concept was engineered for them by Tim Issacs. But I think this radio might be the earliest commercial example - at least, it's the earliest I've seen so far...
With such an amplifier, the output stage is powered by a separate supply. This supply has to be isolated from all others, which means a 3 way active speaker needs 3 separate centre-tapped windings in addition to the winding for all the small-signal electronics. This makes it inconvenient for DIYers who don't want the expense of a custom transformer or multiple transformers in the same box.
Why does the supply for the power stage need to be separate? Because the whole thing moves up and down with the signal. Contrast that with a regular amplifier where the PSU is effectively grounded and the signal moves within the confines of the rails. Essentially, instead of grounding the negative side of the speaker in the usual way, the output side of the speaker is grounded (hence "grounded source").
What's the point? So far, it's extra transformer windings, and the need to wrap your head around a new topology. Standard topologies have been around forever, and work well.
Naturally people like to experiment, and innovation is useful, perhaps if only indirectly. But here, there are some advantages.
Think about the circuitry that drives a conventional unity gain output stage (usually emitter followers, as in this example). The voltage driving them is the same voltage that ends up across the loudspeaker, so 80V peak to peak for a 100W/8R amplifier. The small-signal stages prior to the output stage must work at this voltage, obviously.
But with the ground moved (and the power supply to the output stage allowed to bob up and down), the signal driving the output stage is now driving the base-emitter junctions of the same output transistors. By moving the earth to the other side of the loudspeaker, the output stage has gone from emitter follower to common-emitter. This means that, with respect to earth, the preceding stages only need to generate a voltage swing of a couple of volts, in round numbers, rather than 80V peak to peak. The output stage provides the voltage gain, rather than the usual VAS ahead of it.
So what? Small signal transistors are cheap and widely available, right? Not so much today. Which is why I've always been quite interested in this topology - it potentially makes for a cheap and simple amplifier that might just use op-amps to drive the output transistors. If you want more power, just scale up the supply to the output transistors - everything else still runs on the low value rails, which might be +/-15V or less. For that reason, it is often seen in PA amplifiers (though most have gone to class D now). Great for an active speaker where you might want 200W for the woofer, 75W for the midrange and 25W for the tweeter, say. Yes, it means a custom transformer, but for a manufacturer mass-producing these, that's not expensive.
But no real advantage for a lower power amplifier, where suitable small-signal transistors are still plentiful and cheap.
So why on earth have Grundig used it here? I've honestly no idea. As I say, it's great to experiment, but quite how it left the R&D lab is beyond me, as I'm not seeing any of the advantages in a ~2 watt table radio, and the added cost of an extra transformer winding and extra rectifiers and caps. The fact they dropped it for the RF111 says quite a lot, really...
There's a similar discussion here: https://www.vintage-radio.net/forum/show...p?t=186598
In the last post I've attached a sketch of the output stage which might make the concept slightly clearer.
The radio is positive earth, and everything apart from the output stage runs from -12V. That is generated by the middle of the two windings, and stabilised by the zener diode. That much is easy to see.
What is different about this is the winding for the output stage, which yes, is smoothed by C67. This is around 12V.
I had to sketch it out to be sure, but this is a variation of the Transova circuit. It has been used by several manufacturers over the years, including Haffler and QSC. UK monitor speaker manufacturer ATC have used it (probably still do?) in their active speakers - they call it "grounded source" in their marketing. I understand the concept was engineered for them by Tim Issacs. But I think this radio might be the earliest commercial example - at least, it's the earliest I've seen so far...
With such an amplifier, the output stage is powered by a separate supply. This supply has to be isolated from all others, which means a 3 way active speaker needs 3 separate centre-tapped windings in addition to the winding for all the small-signal electronics. This makes it inconvenient for DIYers who don't want the expense of a custom transformer or multiple transformers in the same box.
Why does the supply for the power stage need to be separate? Because the whole thing moves up and down with the signal. Contrast that with a regular amplifier where the PSU is effectively grounded and the signal moves within the confines of the rails. Essentially, instead of grounding the negative side of the speaker in the usual way, the output side of the speaker is grounded (hence "grounded source").
What's the point? So far, it's extra transformer windings, and the need to wrap your head around a new topology. Standard topologies have been around forever, and work well.
Naturally people like to experiment, and innovation is useful, perhaps if only indirectly. But here, there are some advantages.
Think about the circuitry that drives a conventional unity gain output stage (usually emitter followers, as in this example). The voltage driving them is the same voltage that ends up across the loudspeaker, so 80V peak to peak for a 100W/8R amplifier. The small-signal stages prior to the output stage must work at this voltage, obviously.
But with the ground moved (and the power supply to the output stage allowed to bob up and down), the signal driving the output stage is now driving the base-emitter junctions of the same output transistors. By moving the earth to the other side of the loudspeaker, the output stage has gone from emitter follower to common-emitter. This means that, with respect to earth, the preceding stages only need to generate a voltage swing of a couple of volts, in round numbers, rather than 80V peak to peak. The output stage provides the voltage gain, rather than the usual VAS ahead of it.
So what? Small signal transistors are cheap and widely available, right? Not so much today. Which is why I've always been quite interested in this topology - it potentially makes for a cheap and simple amplifier that might just use op-amps to drive the output transistors. If you want more power, just scale up the supply to the output transistors - everything else still runs on the low value rails, which might be +/-15V or less. For that reason, it is often seen in PA amplifiers (though most have gone to class D now). Great for an active speaker where you might want 200W for the woofer, 75W for the midrange and 25W for the tweeter, say. Yes, it means a custom transformer, but for a manufacturer mass-producing these, that's not expensive.
But no real advantage for a lower power amplifier, where suitable small-signal transistors are still plentiful and cheap.
So why on earth have Grundig used it here? I've honestly no idea. As I say, it's great to experiment, but quite how it left the R&D lab is beyond me, as I'm not seeing any of the advantages in a ~2 watt table radio, and the added cost of an extra transformer winding and extra rectifiers and caps. The fact they dropped it for the RF111 says quite a lot, really...
There's a similar discussion here: https://www.vintage-radio.net/forum/show...p?t=186598
In the last post I've attached a sketch of the output stage which might make the concept slightly clearer.
