A simple PushPull balance LED indicator?
I saw the thread on this here.
But I was wondering if there couldn't be something even simpler, with less parts, and perhaps more reliable. Also It should work, not by voltage per se but by detecting a voltage difference, or rather none, between sides of the push-pull output tubes.
That is, a simple circuit that measures if there is a voltage difference (indicating a current difference, because of equal resistances in the cathode resistors), and sets a red light if its either negative or positive, that you could watch while adjusting the balance (not of the drivers, but of the output currents through each half of the output tranny).
Here's that ST-70 circuit for starting off:
Gingertube's explanation is fine:
The idea is that there would also be a BIAS supply, or adjustment resistor in the case of a self-bias circuit,
with a trim-pot wired up with suitable safety bypass resistors in case of an open-circuit pot (that is a separate task, but doable).
You would simply look at two or three LEDs, or maybe a two-in-one, which showed red when there was some imbalance say more than a 1/10 of a volt or something to indicate more than say 10 mA of current difference per tube-pair.
So, it would go RED for too much current left-side, GREEN for a close balance, and RED again for too much current on the other side.
For now, I would adjust the BIAS / balance while the circuit idles without a signal.
Later, I'd like to use the same basic circuit for an auto-adjustment (something like a voltage regulator) to keep the current through the cathode resistor of each tube constant.
I know you're saying, "But the current ISN'T constant!" Its a push-pull for crying out loud!"
But in my circuits, I'm aiming for a class-A constant current output stage. Also, I'd like the circuit to perhaps either self-adjust slowly (lower than 20 Hz perhaps) or take extended samples of the instantaneous current, to then compensate by attempting a slow tilt of the current back to center on each side.
But the main point for now, is there ought to be an elegant and VERY simple circuit with only a few resistors and LEDs (maybe a zener or two) that can do this basic indicating job. (detect a voltage to ground difference between two points).
Lets take the 2nd Last picture:
Now compare that to the 2nd example (which gives a value for the measure-resistor):
Somehow we want to combine both ideas (the best from each).
Here is our basic Cathode-biased output stage:
For this to work, the separate legs of the Push-Pull stage have to operate independently. So any connection between the two cathodes has to be high resistance/impedance.
R2 will be the sensing resistor, used to catch any difference in voltage (to ground) between the cathodes. This will be a good indicator of current difference, because the resistances to ground for each side should be the same. Same voltage then means same current.
The original circuit used a fuse. When blown, power remained on the red pole of the dual color LED, giving RED. Presumably with the load on, the Zener dropped the voltage to the Red LED enough to either turn it off or leave it dim.
In our circuit, R2 must be high (to isolate the two sides), Z1 must be adjusted for a similar effect/ purpose as the original circuit,and R1 must be chosen to properly bias both LEDs for the given BIAS voltage.
The BIAS / Balance would presumably be set without a signal, and when the Red LED shuts off or dims and the Green LED lights up, we would suppose the current is balanced.
(1) Since the circuit is cathode-biased, all hell breaks loose when the music is playing, and the LEDS won't mean anything then.
(2) One obvious problem is that both LEDs might be on at the same time, making the reading confusing and also perhaps overheating the dual LED package.
(3) Also, the circuit seems cumbersome and already has a lot of components. To make a full version (indicating both left and right BIAS offsets, you'd have to again double the circuit (one for each side) and double the adjust-pots (not shown) to independantly both lower and raise current/bias on either leg.
(4) I would also like to see the LEDs powered by independent circuits instead of drawing on the cathode current, which may affect the sound, especially if they are shutting on and off.
Thats why we want to look at the other circuit, which has some advantages:
It uses an additional transistor, which presumably detects more accurately a voltage DIFFERENCE between the two legs, and perhaps can be modified to minimize noise or interference with the amplifier.
Turning to the second reference circuit,
at first it looks more complicated, with the added transistor,
but the overall parts count is low, and the circuit looks simple.
It seems to only detect current in one direction, so we need again to double up:
Also, since we still want isolation between the two cathodes (and little current flowing directly across), we have added R3 (split on either side to force the current draw of the diodes to draw equally from either leg).
Some questions remain:
(1) Voltage source to power LEDs is assumed to be from elsewhere, and must be significantly higher than the BIAS reference. We have also added R4 to allow a significantly higher voltage source, so that LEDs operate properly regardless of the actual BIAS voltage chosen. Remember, we only want to detect a DIFFERENCE, not set the overall BIAS with this balance indicator and controls.
(2) The 1 ohm detector-resistor (R1) may have to be adjusted to work with the isolating resistors (R3) properly. We want to limit overall current across while maintaining pretty fine accuracy and sensitivity to imbalance, while still being able to independently select a wide range of BIAS for the pair.
(3) We would prefer to connect to ground rather than a +ve supply, so we don't have to add another power supply with its added cost and complexity. This may require different transistors or layout.
(4) This only turns on LEDs (both RED) when there is an imbalance. We'd like to add a green resistor to indicate the circuit is BALANCED properly, but which will not turn on if the BIAS is wrong or if a tube has failed (short or open).
Still more work to do.
Solving the power supply problem:
In this next version, we shunt the current from the cathodes through the LEDs to ground.
The power essentially comes from the cathode voltage,
eliminating the need for a power supply.
To pull this off, I substituted PNP transistors for the original NPNs.
R4 now should be chosen to handle the expected BIAS voltage,
namely from 40 to 120 volts D.C., keeping the LED currents to a sensible minimum (probably about 10-20 mA).
Presuming that fluctuations during music play would average out, the current drawn by the two LEDs would remain fairly constant on average.
However, the turning on and off of the transistors might introduce unwanted noise. Some kind of bypass/shunt caps, or a high impedance device might prevent any noise from sneaking into the cathode circuit.
Both LEDs here would be red,
indicating the voltages (and currents) were out of balance,
when no music is playing, and the user is adjusting the balance.
If the new circuit introduced excessive noise,
they could be switched off or out of circuit after the balance was set.
Still missing is a green LED indicating that current balance has been achieved,
through the output transformer.
This would light up when there was no cross-current through,
or voltage across the measuring-resistors, but would shut off,
if there was no voltage at all (HV off or Amp off),
or if one tube was shorted out, open, or missing.
Too much on left : Left Red LED lights. Others off.
Too much on right: Right Red LED lights. Others off.
Perfect balance: Green LED lights. Others off.
No power: All LEDs off.
Fault Condition: (?)
Possible options for a fault condition might be both Red LEDs on (impossible normally),
or a third indicator along with safety features turns on.
Damn. keeps happening.
I try to edit a post to improve it,
and instead it gets reposted.
The Edit button keeps vanishing.
An option is to use a "tri-fferential" with suitable degeneration components.
Here is an example that feeds directly from the HV with a single resistor, and provides ~20mA to 1 of 3 LEDs: the green LED is active when the difference is smaller than +/-300mV, and the corresponding red LED lights when either threshold is exceeded.
Simple, inexpensive and flexible.
In the sim, the ground is placed on the right output, that's simply to be able to display the difference voltage directly (gray trace).
A few questions -
(1) The connections for v1 and v2 are those between the plate and cathode? This would need to be modifed to accurately measure current differences. I would not expect the tubes to ever have equivalent resistances except during instantaineous crossovers in transient variations.
I would rather measure the voltage/current through the fixed cathode resistor to ground for accuracy.
(2) Is the 300v tap necessary? On the one hand, I'll be using a 520 volt supply, and on the other, the HV seems unnecessary for a low voltage circuit connected near ground with an average voltage range of 50-120 volts (bias) and transients at the cathode hopefully not much higher than 200 v (?)
What is the QTL690C? just a part number for an ordinary red LED?
(4) Have you eliminated the sampling resistor across the cathodes in my circuit?
will any small PNP signal transistors do here?
What is the meaning of the two different types of transistor?
|All times are GMT. The time now is 11:24 AM.|
vBulletin Optimisation provided by vB Optimise (Pro) - vBulletin Mods & Addons Copyright © 2016 DragonByte Technologies Ltd.
Copyright ©1999-2016 diyAudio