Emitter Follower Lockup on Electropneumatic Transducer

[Daniel Bingamon]

I have a situation that occurs sometimes on a Fairchild Model TT6000-906 Electro-pneumatic Transducer.  This is a proportional device and is voice coil drive that typically puts out 0 to 90 PSIG based on 1 to 9 VDC.   Typically the reference pressure is just somewhere over 90 PSIG.
The voltage generation is a DGH 4181 RS-232 based Module with 24VDC power supply.  A command is given and it puts out 0-10 VDC.  Now, this device does not offer enough current to drive the TT6000-906 so a 2N2222 transistor is employed as an emitter follower.  The collector gets 24 VDC, base is connected to the Module output, the emitter goes to the positive terminal on the transducer (also called E/P or pronounced E2P for short).   The ground leg of the E/P goes to circuit ground.
Normally these work fine.   We have put a flyback diode in reverse across the E/P terminal to prevent coil breakdown voltage from hurting the transistor.
Once in a while these devices when put at the high end of its operation range, say 8.5 or 9 VDC that when the voltage is removed the circuit continues to put out a high pressure of say 85 to 90 or so PSIG.   It stays this way until you grab the voltage and approach the terminals of the device.   Sometimes all you have to do is get the lead close to the positive terminal and it will let loose (that is it no longer stays in this locked state).   It seems like some high impedance or capacitance phenomena is going on.
This E/P is an Electro-mechanical device.  We've used hundreds of these and this happens on a small number of units.
Any ideas what is causing this now and then and how can it be prevented?

[SeanB]

First thing would be to put a resistor from emitter to ground, probably a value around 1k will do, to make sure leakage in the transistor is not an issue. As well a resistor in the base emitter side will also help to ensure the transistor is going to stay off, as the power dissipation is likely to be high, you want to use the 2N2219A instead, with a clip on heatsink. Base emitter resistor of around 10k will not harm, and will ensure the transistor does turn off, even with leakage when hot. A resistor of around 150R 2W in the collector leg will help reduce power dissipation in the transistor, as you will need to dissipate up to 1W in the transistor, which is not going to be easy in the plastic package device with a max dissipation of 600mW.

[Benta]

I have a different suspicion (the emitter resistor is needed,yes), but emitter followers can be problematic and sometimes tend to oscillate at several MHz at high amplitude.
The way to kill this is adding a base resistor, in your case 100...470 ohms. Worth a try.

[coppercone2]

those extra parts are IMO essential if this is on some thing connected by a long wire (remote actuator). Deep in a PCB in a enclosure actuating a relay is one thing, but some 'module' connected to a industrial pipe by typical industrial wiring methods (which I assume this is), you want those parts.