Special LM317 Constant Current Power Supply

Question

An unusual problem came up that requires a substantial dip in current when the load is increased. The solenoids hold the bass audio contacts closed on a Hammond H-100 Series. The feature is called "String Bass".

Here's how it supposed to work: One pedal is depressed--then released, its corresponding solenoid is held in by a holding contact. When a second pedal is depressed, the current is divided and the first solenoid drops out due to insufficient holding current. The second pedal doesn't care since it is manually held with the foot, and consumes the full 130mA once the preceding solenoid drops.

Problem: 1/2 of the constant current is enough to hold some solenoids in due to manufacturing tolerances. The need is to have the current momentarily (~30ms) dip substantially below the 65mA to insure dropout of the pedal that had been released.

Hammond accomplished this with many transistors (some Ge), two zeners, one diode, two adjustment pots, and several resistors. Rebuilding the original scheme (that burned out) is not insurmountable, but undesirable.

Holding contacts and diode clamps not shown.

Any ideas or alternative approaches appreciated.

Answer 1

^{simulate this circuit – Schematic created using CircuitLab}

Figure 1. Modified pedalboard circuit.

My initial thoughts were that if a current sensing resistor, R2, can be added into the switch ground return as shown then you may be able to trigger a comparitor, CMP1. D4 provides a 0.7 V reference and R2 is chosen so that the voltage across it will exceed 0.7 V when two or more solenoids are on.

Since we are now monitoring the current we can change the LM317 configuration to constant voltage rather than constant current. The open collector of the comparitor can then shunt the LM317 voltage divider to drop the voltage to a couple of volts and unlatch the additional solenoids.

A pulse extender or monostable is required on the output of CMP1 (but not shown) to hold the LM317 low until the unpressed solenoids all release. Without this the voltage will probably oscillate.

Figure 2. If low-side current sensing is a problem the LT2940 power and current monitor might provide a solution.

The configuration shown in Figure 2 is monitoring power and multiplies the current and voltage. You don't require that and I guess that the V+ and V- inputs should be held at a constant voltage rather than monitor the load voltage.

Figure 3. Block schematic.

The LT2940 has two complimentary comparitor outputs which may save an inverter stage in your application.

Answer 2

Add a capacitance such that the parallel LC creates a resonance. If there is resonance, current will flow in that direction for a good amount of time.

That blue line is a movable wire. I'm not sure if this will work but it might..

Answer 3

Try adding a resistor in parallel with each coil to bleed off some of the current that is holding the solenoid on, about 20-30ma should suffice (470 - 750R). This would reduce the half current to 25mA and leave the normal (full current) at 100mA, more than enough to hold.

Answer 4

I would suggest a simple current limiting circuit and a capacitor, separate for each solenoid. Capacitor provides initial current. Once it is discharged, the current limiting circuit becomes dominant. And to release you should use logic. This is what is intended for such functions.

Answer 5

So if I understand correctly you want to shunt some current from the supply, briefly when any key closes?

Build a switchable current sink around an NPN transistor, two resistors and two diodes in the usual way, control with a non retriggerable monostable and a comparator monitoring the voltage of the constant current line.

Or implement essentially the same logic with a little 8 pin micro.

Or take a honking great electrolytic with maybe 100 ohms in series and place across each solenoid, crude, dumb and large.