Goosing a solenoid??

Proofread, proofread, proofread....

--
Cheers!
Rich
 ------
"OPTIMIST: A man who makes a motel reservation before a blind date."

Move D1's cathode to the supply end of the coil.

John

duty

Of course it's possible and desirable. This is especially true when there are 24 solenoids to be driven. Using PWM and 48 volt drivers, the total dissipation can be kept low and the parts count reasonable and the power supply simple. A couple of inexpensive microprocessors can easily do this and it should be given serious consideration. It would be the simplest, cleanest and most versatile method if it is ever to become a commercial product. See my above post. I have designed drivers like this. If interested e-mail me directly for more on the subject. Please remove the ns from the front of the email address. Bob

I'm not sure if pwm is desirable or even needed. Here are three examples of circuits that could be used depending on the performance needed:

1. Capacitor Discharge

The schematic is shown in

The capacitor C1 is charged to 48 Volts through R1. Q3 connects the bottom end of the solenoid coil L1 to ground. Capacitor C1 then discharges through the solenoid to increase the initial current rise. When the voltage across the capacitor falls below 12V, diode D1 turns on to provide sustaining current to keep the solenoid closed. Diode D2 and R6 limit the back emf voltage when the driver transistor Q3 turns off.

The waveforms are shown in

The top graph shows the capacitor voltage in red and the coil voltage in blue. The bottom graph shows the solenoid current in black.

This method offers good performance for low duty cycle applications. The rep rate is limited by the time required to charge C1. Considerable power is dissipated in R1 when the solenoid is activated.

1. Switched Capacitor

The schematic is shown in

When the solenoid is not activated, transistors Q1 and Q2 keep the capacitor C1 charged to 48 Volts.

The waveforms are shown in

As in the previous example, the top graph shows the capacitor voltage in red and the coil voltage in blue. The bottom graph shows the solenoid current in black.

This circuit provides good performance for high rep rate applications. A small amount of power is wasted keeping transistor Q1 turned on when the solenoid is not activated.

1. Pulsed Voltage

The schematic is shown in

The R3, C2 network in the emitter of Q2 form a one-shot that briefly applies 48 Volts to the solenoid when Q3 turns on. The D2, R4, D3 network provide protection against the back emf generated when Q3 turns off.

The waveforms are shown in

As in the previous examples, the top graph shows the capacitor voltage in red and the coil voltage in blue. The bottom graph shows the solenoid current in black.

This circuit provides the fastest current rise into the solenoid, since the voltage is held at a constant 48 Volts instead of decaying as in the previous examples. The current fall time could be increased by allowing a higher back emf when transistor Q3 turns off. The quiescent current is essentially zero.

These examples are intended to show the basic operation. Additional effort may be required to meet actual system requirements such as worst-case supply voltages, temperature extremes, component variation, etc.

However, it is difficult to see how pwm could provide higher performance and lower component count.

Mike Monett

what ever happen to the simple 3 component network to drive a power transistor with a peek and hold function ?

2 resistors in series to properly calculate the holding current of the relay, one of the resistors will have a Cap coupled with it that will be calculated to create a current lag long enough to give the desired peek current, then as it charges the current will drop back to the holding current when the cap gets charged. having the cap coupled with a resistor allows it to discharge faster for the next time. ect. old age analog electronics long forgotten i guess! :)

[...]

It would be nice if you provided a schematic. However, if I guess correctly, your circuit was one of the first discussed in this thread. The high power dissipation and long recovery time are the most serious limitations.

"Peek" is probably spelled "peak" in this contxt.

Mike Monett

[...]

Yes, Bob, thanks for the reply. I see how you do it now. The PIC is definitely the way to go.

I used obsolete parts in the SPICE analysis since I have them in the library. A proper design would obviously use modern parts.

Mike Monett

People can do whatever they want, of course. But keep in mind that there are

24 channels so your solutions are multiplied by 24 in parts count, money and complexity. Secondly these methods are not versitile. Lets say you need a bit more drive, what do you do? Change the 330uf caps to hold more energy? Get a new power transformer? What you have described is the solution we used back in the 1970's or before. The PWM solution I have used for a piano playing mechanism uses a microprocessor to drive 12 channels, 12 solenoids. For 12 channels, the parts count is one PIC microprocessor 12 gate resistors. 12, IRF FETS and 12 fly back diodes. There is one powersupply, 70 volts, several small bypass caps, a 5 volt voltage reg chip and that's it. Keep in mind that the micro provides the drive signal as well, a component you have not even shown in your scheme. Also, a 3055 transistor has poor current gain so if you need to drive several amps, the base drive is high as well requiring more parts. I'm sorry, but your method is a klutsy, old school solution. Of course you don't have write any code so that's an advantage. If you want more information, please e-mail me. Remove the ns from the e-mail address to respond, nsmontassocatyahoodotcom. Thanks for the argument. Bob

You guys should check out some pinball machine operator's manuals, for example, 1990's Williams pinballs used End of Stroke switches / Capacitors to help with the main drive/"hold" circuit.

There are schematics in some of the scanned operators manuals.

-cyBORG