Using Transistor as a shorting switch

Thread Starter

dandy1

Joined Sep 30, 2017
178
First of all, I know its a bad Idea but Ive tried it and it works, at least on the face of it.

I have a negative going signal about -25v for approx 2mS. The peak voltage is up for debate but lets just say -25.

I am using NPN 3904 transistor to dead short the signal and seems to be holding up. However I'm not sure what effect negative voltage has on a transistor and the implications of driving it from MCU.

Is it even feasible to be using a transistor?

I know the proper option is to switch it with a relay but I am trying to keep the footprint small.
 

MrChips

Joined Oct 2, 2009
30,456
A P-N junction can be used as a signal clamp. In other words, the signal will not exceed the threshold voltage of the P-N junction device, which is typically about 0.35V for a Schottky diode and 0.65V for an ordinary silicon diode or transistor.

If you posted a circuit diagram we would have a better idea of what you are trying to do.
 

Thread Starter

dandy1

Joined Sep 30, 2017
178
A P-N junction can be used as a signal clamp. In other words, the signal will not exceed the threshold voltage of the P-N junction device, which is typically about 0.35V for a Schottky diode and 0.65V for an ordinary silicon diode or transistor.

If you posted a circuit diagram we would have a better idea of what you are trying to do.
1.JPG
 

BobTPH

Joined Jun 5, 2013
8,659
Just place a Schottky diode cathode to the signal, and anode to ground. This will clamp the voltage to no more -0.3V.

The circuit you have will possibly damage your micro because negative pulses will go right through the CB junction.
 

MrChips

Joined Oct 2, 2009
30,456
This is a standard method of clamping input signals.

1692221188962.png

The signal out is clamped to -0.3V and Vcc+0.3V when using Schottky diodes.
 

Thread Starter

dandy1

Joined Sep 30, 2017
178
My application is clamping the signal yes, but i also want to short it to ground when required. But as pointed out by
BobTPH
the negative signal will go straight thru the CB junction.

Could I then use a blocking diode on the MCU output?
 

Thread Starter

dandy1

Joined Sep 30, 2017
178
It sounds like you want to turn the clamping on and off. Why?
No I very much need the clamping to stay in place, I am simply trying to ground the signal as to act for a kill switch for an ignition (it's part of the project you have replied to in my other threads). I have checked to see that indeed as is shown above the negative voltage goes straight through to the transistor base however the trans will still switch will very high base resistance (1M) which may allow the internal clamps to do the work although this is a very last resort. I also had put a rectifier diode across the base to ground but that had no effect.
 

Ya’akov

Joined Jan 27, 2019
8,973
An active component in a kill switch that must be energized to kill the circuit is not failsafe. If this has any safety application, it is just wrong to do. A kill switch must kill when there is no power to it in order to fail safe.
 

wraujr

Joined Jun 28, 2022
160
Really need to know more about the 50Hz drive circuitry on the left to know how your transistor switched to ground will perform.
Also, does the signal on the right go somewhere? Are we missing the connection to MCU input?
 

Ian0

Joined Aug 7, 2020
9,467
AS this is for your generator project, there isn't a safety issue.
But you are not going to switch a 1kV ignition system with a 40V transistor. You will kill a 40V transistor with a 1kV ignition system.
You need something with a gap at least 10 times as big as the spark-plug gap, otherwise the current will take the easiest path (don't forget that the engine is pressurised to about 9 bar).
Is there a lower voltage winding on the magneto?
 
Last edited:

Thread Starter

dandy1

Joined Sep 30, 2017
178
Really need to know more about the 50Hz drive circuitry on the left to know how your transistor switched to ground will perform.
Also, does the signal on the right go somewhere? Are we missing the connection to MCU input?
The signal is being sampled at the Diode clamp. This is the waveform:
IMG_20230811_202144.jpgbut is conditioned by the Diode for easier sampling:
IMG_20230816_185749.jpg
The signal comes from the primary side of a generator ignition coil which produces its own energy via a magnet passing the coil.
 

Thread Starter

dandy1

Joined Sep 30, 2017
178
You need something with a gap at least 10 times as big as the spark-plug gap, otherwise the current will take the easiest path (don't forget that the engine is pressurised to about 9 bar).
Is there a lower voltage winding on the magneto?
Not sure what your getting at here but the system is designed to short the primary side the same as I am trying to do only there's a mechanical switch to do so.
 
Top