Problems with transistor silent switch

[tss]

The original idea was to insert a small circuit at the output of a switcher to mute the output while mechanical relays are switching to avoid pop noises. I use a 2N5089 with a grounded emitter and the collector hooked up to the signal output. The control voltage is from a micro-controller. According to the datasheet, saturation happens at 1mA I've put a 4k resistor in series to limit the current to 1.25mA. To open the transistor slowly (to avoid pops) I inserted a cap. from the base to ground. I started with 33uF just so I could see the something on the scope. With a DC 1V voltage I check how the circuit acts and it is a bit strange. The fall time (muting) is about 1.6msec and the raise time is about 11msec. How come there is such a huge difference? I would like to slow down both the falling time. But how?

[dwmorrin]

Treat the base-emitter junction as a diode in your analysis.
Try putting a resistor in series with the base of the transistor, so that the base's voltage changes based on the RC of that resistor.
With the cap directly connected to the base, the base-emitter diode quickly discharges it.  Your 4k isn't slowing down the discharge through the diode.  Just the charging time.

[tss]

So since it is a diode the voltage does not increase after .6 or .7 only current increases? So the cap does nothing but create a delay in switching the transistor... so perhaps PWM through the resistor would work?

[dwmorrin]

So since it is a diode the voltage does not increase after .6 or .7 only current increases?

Yes.
If you have a working silicon BJT transistor, then the BE voltage must stay at 1 diode drop while it is forward active.
How accurate do you need this to be?  PWM seems like overkill.  I think 2 resistors and a cap will be fine.  Make the cap as big as it needs to be to set the main time constant, then fine tune charge and discharge times with the two resistors.  The total resistance will set the current, so divide your 4k as needed between the two.

[tss]

I had a feeling that is going to be the circuit

However I am in a pickle since I have this on a PCB and can't really add components (no room).

[amptramp]

You may want to add a high resistance from the base to ground to take care of leakage currents.  Figure out your worst-case temperature and derive the leakage current then add a resistance that guarantees the transistor will turn off at the Vbe at the temperature you used for the leakage current.

[R.G.]

However I am in a pickle since I have this on a PCB and can't really add components (no room).

Ooops. One of the fundamental rules of PCBs is to never put a new design on a PCB before you've thoroughly tested it other ways.

Going direct to PCB means you're powerfully sure of yourself - or as yet unscarred by interactions with Mother Nature.

But you now know this in Mother's most forceful version of lessons.

[Quackzed]

However I am in a pickle since I have this on a PCB and can't really add components (no room).

there's always room! 

[tss]

However I am in a pickle since I have this on a PCB and can't really add components (no room).

Ooops. One of the fundamental rules of PCBs is to never put a new design on a PCB before you've thoroughly tested it other ways.

Going direct to PCB means you're powerfully sure of yourself - or as yet unscarred by interactions with Mother Nature.

But you now know this in Mother's most forceful version of lessons.

That's true. I 'kinda' overlooked this one by doing all my test using simulation but it turns out transient analysis can't be the only tool for checking a design...