After replacing R19 with 24 Ohm, Q15 emitter is -42.25V now.
Whole OS starts to oscillate when I try to measure voltages at some other spots, but most of them show the same voltage as before.
When this happens, R17 (1k) becomes very hot.
100mA current limit on the PSU...
Whole OS starts to oscillate when I try to measure voltages at some other spots, but most of them show the same voltage as before.
When this happens, R17 (1k) becomes very hot.
100mA current limit on the PSU...
Oscillations in those conditions are not particularly surprising or worrying: in voltage mode, multimeters often present a non-negligible input capacitance, and this can cause instability in some circuits.
A simple fix is to attach a 1K resistor to the "hot" tip
A simple fix is to attach a 1K resistor to the "hot" tip
R17 should be about 15K. That will limit Ic Q6 to about 5.5mA, more than enough to completely shut down Q14, ie > 0.65/150 = 4.3mA. Even so, 82*82/15k = 0.45 Watts, so it needs to be a 15K, 1W resistor.
For stability, you need a second Q10 (+R13) dedicated to C25 only. This provides HF drive from the input without positive HF feedback from Q15.
For stability, you need a second Q10 (+R13) dedicated to C25 only. This provides HF drive from the input without positive HF feedback from Q15.
It would be a good idea to also check Q9 & Q15: even with R19 at 600K, the voltages weren't what you would expect
Tandem in this context is definitely a novel term; two-quadrant appears from time to time. Sometimes, it is physically implemented but not explicitly mentioned.
Philips has made line of "reversible" lab supplies: they can be used as regular supplies or dummy loads, and the transition is completely seamless.
A one-quadrant supply fitted with a bleeder can already be seen as a SE two-quadrant
Philips has made line of "reversible" lab supplies: they can be used as regular supplies or dummy loads, and the transition is completely seamless.
A one-quadrant supply fitted with a bleeder can already be seen as a SE two-quadrant
Why is this power supply idea so "novel"? Isnt it just an amplifier output? If it needs to source current, the high side element provides it. If it needs to sink current, the low side element provides the path. With the requisite feedback implemented, so you can move the output voltage about.
I'd say any DC coupled amplifier design could be a power supply rail for another amplifier and give the fancy schmancy "two quadrant" or "Tandem" performance. That is, if any current for whatever reason needs to be sunk, as well as sourced, on that rail.
I'd say any DC coupled amplifier design could be a power supply rail for another amplifier and give the fancy schmancy "two quadrant" or "Tandem" performance. That is, if any current for whatever reason needs to be sunk, as well as sourced, on that rail.
This power supply idea is not novel; not the two quadrant aspect anyway: in order to improve dynamic properties, some supplies are built like a class AB amplifier.
The novelty of the tandem is that both halves interact and collaborate in a synergetic manner to achieve exceptional performances: if you look at the linearity, output impedance etc. of the tandem configuration without GNFB, they outperform many topologies based on GNFB.
Also, because of the interaction, things like quiescent are inherently under control
The novelty of the tandem is that both halves interact and collaborate in a synergetic manner to achieve exceptional performances: if you look at the linearity, output impedance etc. of the tandem configuration without GNFB, they outperform many topologies based on GNFB.
Also, because of the interaction, things like quiescent are inherently under control