Switching phantom with NPN transistor?

[Ethan]

I've only ever seen this done with relays or switches.  Would it be possible to switch phantom with an NPN?
Could someone please enlighten me with how this might be done?

 

[clintrubber]

Whether it's the most elegant or economical is another issue, but it sure can be done.

If you have for instance a discrete regulator with NPN (so not an integrated regulator), you might even be able
to combine it, if the job is to have that NPN & make it a bit more compact.
Should be verified w.r.t. switch-on & -off behaviour & other eventual nasty side-effects.

Regards,

  Peter

 

[clintrubber]

Mmmm, but what switches this NPN ? Something like a switch again...
(or you should perhaps want to do it by means of jack-inserting, which would be 'no explicit' switch)

 

[volker]

Sure you can. But a PNP transistor would be more appropriate, because you want to switch high side.
Emitter goes to 48V, collector goes to the 6,8k resistors, base gets a resistor and the control voltage.
Depending on which voltage you want to use to switch phantom on and off, you might need an additional
high side driver, which is just another NPN.


Volker

 

[Svart]

http://www.groupdiy.com/index.php?topic=33794.0

I would use a PFET.

I have a circuit drawn up if anybody wanted to see it.

 

[JohnRoberts]

Sure you can. But a PNP transistor would be more appropriate, because you want to switch high side.
Emitter goes to 48V, collector goes to the 6,8k resistors, base gets a resistor and the control voltage.
Depending on which voltage you want to use to switch phantom on and off, you might need an additional
high side driver, which is just another NPN.


Volker

You can use a pnp for the high side switch, but you will generally need a npn to level shift from 5v logic up to 48v. Since it is also common to use a npn or darlington npn follower to buffer the 48v,  using the npn logic level translation device to just pull down the high impedance +48v base drive also works...

lots of ways to skin this cat...

JR

 

[Ethan]

Thanks for all the responses.  I'm still a little confused (dense).  Could I use just 1 NPN where the collector is connected to +48V and the emitter connected to the 6.8k resistors and put 5V on the base resisor to turn it on?  I actually thought I could do this but found that what happens is about 38V appears at pins 2&3 whether the base is at 5V or ground. 
Where did I go wrong?

 

[tv]

maybe.

 

[Svart]

I would use a Pchannel Mosfet.  You will have much less insertion loss and it's easier to control.  FETs were designed to be switches and newer fets have such low RDS that it's almost impractical to use anything else as a solid state switch.

I made a PDF of my schematic and inserted some information on usage:

http://www.theopiumdenproductions.com/phantom logic_1.pdf

 

[Ethan]

I'm using an existing supply with a 48V rail, not the discrete psu that clintrubber posted.  I have a bunch of NPN BJTs, so if it's possible p do it with what I have that would be ideal.  Thanks in advance.  

 

[Svart]

Radio Shack sells Pchannels for like 1$..

What NPN are they? can they handle the load?

I'll sim something up if you tell me what the part # is. 

 

[Ethan]

Thanks Svart,
BC536 yeah according to the datasheet they can handle 60v and max collector current is 100mA

 

[JohnRoberts]

Yup.. the schematic posted by TV is along the lines of what I was thinking for an all NPN solution.
--------
I am doing some low voltage work with PMOS high side switches and you can go much higher impedance for the control resistors.  I'm not sure if you want the cap to ground or from the plus unregulated, but you probably don't need 10 uF.

Depending on the gate breakdown voltage you might not need a divider.

A hybrid, combination of the two, could replace the switch to ground with a NPN transistor... to drive the PFET.. In fact I am doing something like that in my application to drive my high side switch from a 3.3V micro...

JR

 

[Svart]

John,

I just threw some values for the resistors in there.  They aren't really important as long as the ratio stays the same and they aren't so low as to waste power.  As for the cap, I just threw that in there too for simulation purposes.. 

 

[jdbakker]

As for the cap, I just threw that in there too for simulation purposes.. 

Be careful if you do include the cap in a real circuit. If V1 turn-on is much faster than the time constant set by R2*C1 you could still exceed the PFETs VGS(MAX).

JD 'caveat SPICEtor' B.

 

[guavatone]

Why not use a PNP with 48V going to emitter, collector to 6k8's, 100K between base and emitter, then base to 220k to a switch between the 220K and ground.

No need for another Voltage rail.  This circuit appears in MP-2NV and Fabio's API schem but are used for 24V relay switching. the resistor values will probably need to be tweeked.

 

[Svart]

JDB,  I built the thing with a 1uf poly which works pretty well.  I've not had it fry yet at least in this configuration.  The FET i used has a max VGS of +/-20v so it'll survive some reversal of G-S voltages.  Since it was for my own use, I went cheap and left out some parts I would have normally used.  To make the circuit production worthy, which I didn't add to the schematic but have used in the past on motor drives, is a reverse biased schottky between the gate and source so that if the rail falls and the gate is still charged, that charge will drain through the schottky, much like the extra protection diode on the FET itself.

 

[ppa]

maybe.

yes, it's quasi-correct, I council to use a BC639 and a resistor of 6.8k from its base to emitter if the driving voltage is from +5 to +24V and applied with a simple swich. Moreover, a diode with the catode on C6 cap and the anode to
q3 collector must be applied to avoid excess of inverting voltage to BE junction of same q3. This diode discarge c6 to avoid to broke q3 when the BC639 is on.  
 
However, I don't like this regulator so much for several reasons.

 

[Ethan]

OK I'm starting over...

Using a BC546, I'm using a 1k base resistor.  The emitter is connected to the 6.81K phantom resistors and the collector is connected to the 48V rail.
WHen I apply 5V to the base resistor, I only see 3.9V instead of close to 48V at the emitter.
It turns off properly when I ground the base resistor but why am I not seeing anything close to 48V at the emitter when 5V is at the base resistor?
Thanks in advance.

 

[JohnRoberts]

In an emitter follower the emitter follows the base, less a base voltage drop.  So 5V on base will be 5V minus .6 or so.. if darlington minus -1.2v or so.

Base needs pullup resistor to 48v then it will output 48V - base drop... so 46-47v,,

Resistor needs to be small enough to feed transistor base adequately but large enough to not over heat when pulled to ground.

JR