I'm trying to look for a transistor that does a very specific task

Thread Starter

Fearless Z

Joined Jun 27, 2019
10
I'm trying to find a transistor that when a small positive current is applied to the gate. No current (positive or negative) can pass from the source to the drain. When there is no positive current applied to the gate current (both positive and negative) can pass from the source to the drain.

If you are able to complete the same action but, using two transistors that is fine.
 

panic mode

Joined Oct 10, 2011
1,829
gate has high impedance so there is no current there.
"positive" current is POSITIVE. it cannot be both positive AND negative.
 

MrChips

Joined Oct 2, 2009
21,070
You are looking for the opposite of how an NPN bipolar junction transistor (BJT) or an N-channel FET works, i.e. high input is ON, low input is OFF.
What you need to do is invert the logic level of the control signal, low input is ON, high input is OFF.
 

WBahn

Joined Mar 31, 2012
25,745
I'm trying to find a transistor that when a small positive current is applied to the gate. No current (positive or negative) can pass from the source to the drain. When there is no positive current applied to the gate current (both positive and negative) can pass from the source to the drain.

If you are able to complete the same action but, using two transistors that is fine.
When you use the term "gate" you are implying a FET, which are typified by extremely high gate resistance (essentially an open circuit).

Why do you want to apply a current as an input? Why not use a voltage signal?

Look up bilateral switches and transmission gates.
 

Thread Starter

Fearless Z

Joined Jun 27, 2019
10
i want to be able to connect the source to VCC and have a positive voltage come out the drain. and if i were to connect the source to GND i want it to be negative at the drain.
 

WBahn

Joined Mar 31, 2012
25,745
i want to be able to connect the source to VCC and have a positive voltage come out the drain. and if i were to connect the source to GND i want it to be negative at the drain.
How are you going to get a negative voltage out of the drain if you connect the source to GND???

What is the problem you are trying to solve? There may be a better way.
 

AnalogKid

Joined Aug 1, 2013
8,443
i want to be able to connect the source to VCC and have a positive voltage come out the drain.
So far, OK. This is the action of a p-channel MOSFET or a PNP bipolar transistor. There will be a positive voltage at the drain when the gate is pulled down to GND. If that is the wrong control polarity, an additional transistor acting as an inverter will correct it. If you want the gate to be above the source to turn on the transistor, then the transistor type changes to an n-channel.
and if i were to connect the source to GND i want it to be negative at the drain.
And now things fall apart.

What is "it"?

The drain cannot go below GND (be negative) unless the circuit has both positive and negative power supplies. Please restate your requirement more clearly.

This will go a *whole* lot faster if you post a sketch of the other circuit elements, or a table showing the various operating conditions.

ak
 

Thread Starter

Fearless Z

Joined Jun 27, 2019
10
I want to control 2 shift registers with 1 set of inputs so i need to stop signals from going to one shift register while sending signals to another. The problem i was having is you need to sent both high and low signals to a shift register. but i think that a bilateral switch will will allow me to restrict the signals going to the shift register.
 

AnalogKid

Joined Aug 1, 2013
8,443
I want to control 2 shift registers with 1 set of inputs so i need to stop signals from going to one shift register while sending signals to another. The problem i was having is you need to sent both high and low signals to a shift register. but i think that a bilateral switch will will allow me to restrict the signals going to the shift register.
If the "inputs" are logic-level signals, they you don't need a switching transistor - you need the correct arrangement of logic gates. Can you create a table with the desired output states for each combination of input states?

ak
 
Last edited:

djsfantasi

Joined Apr 11, 2010
6,412
What shift register are you using? I don’t know of any that uses a bipolar supply and/or a negative voltage input?

Usually, you can simply control a single shift register with one control line and an inverter.

To switch between two, I’d use a 1:2 line Decoder/Demultiplexer. Or build one from discrete logic gates. That link is the first explanation I found.

I think you are confusing ground with a negative voltage. A logic input of Gnd is a logic 0 or false. A logic input of Vcc (or somewhat lower) is a logic 1 or true.

A negative voltage is less than Gnd or 0. And is not used in logic circuits.
 

Thread Starter

Fearless Z

Joined Jun 27, 2019
10
https://www.onsemi.com/pub/Collateral/MC74HCT595A-D.PDF
That is what shift register i am using

The problem with using a 1:2 decoder is there are 3 inputs required to use the shift register and in my final design there will be up to 10 shift registers that all need to have seperate outputs but use the same input

I need someway to block the ones i am not using while sending signals to the ones that i want to use so they all can have different outputs
 

djsfantasi

Joined Apr 11, 2010
6,412
https://www.onsemi.com/pub/Collateral/MC74HCT595A-D.PDF
That is what shift register i am using

The problem with using a 1:2 decoder is there are 3 inputs required to use the shift register and in my final design there will be up to 10 shift registers that all need to have seperate outputs but use the same input

I need someway to block the ones i am not using while sending signals to the ones that i want to use so they all can have different outputs
So use a 2:4 decoder.

Or use separate decoders on each input line...

I haven’t performed the exercise, but my gut says this is a simple logic design problem.

Those are pretty standard shift registers. What are the three inputs you’re using?
 

AnalogKid

Joined Aug 1, 2013
8,443
This is a very simple logic problem, but you are not helping by keeping so many details secret. It took 11 posts to find out that not only do you not need a transistor as you originally described, it can not work for what you want to do. And 14 posts to find out that there are three "inputs".

Please tell us about how many control inputs, how many data sources, how many devices to select, any clocks, etc. Are the shift register selections mutually exclusive (only one shifting at any time), are they all shifting at the same time, or what?
I need someway to block the ones i am not using while sending signals to the ones that i want to use so they all can have different outputs
By "block", do you mean disable shifting? Are all shift registers driven with the same clock, and you want to control which one are shifting and which ones are parked?

The 595 has five input signals, but you say you are using only three of them. Which three? That's kinda important, wouldn't you agree?
ak
 

Thread Starter

Fearless Z

Joined Jun 27, 2019
10
I need to be able to individually control up to 10 3 state shift registers (the amount may vary) using the same 3 inputs. according to my shift register one is a shift clock one is a latch clock and one is the inputs. There are a few others but they aren't important.

My best guess on how to solve this problem is to single out one shift register by locking the 3 inputs on all the other shift registers but I'm not experienced enough to know what part i need to get. I've tried searching on google but i don't really know what to look up.

by locking i mean that while a shift register is locked i won't be able to change its state in any way until it is unlocked.

Yes, i want to be able to freely control which shift registers are locked and which ones are free to change
 

MrChips

Joined Oct 2, 2009
21,070
Thinking in terms of switching signals with analog switches is leading you astray.
This is a problem that a digital solution can solved. You do this with digital gates, decoders and multiplexers.

Start over by telling us exactly what you are trying to do and someone will provide you with a proper solution.

Here is a starting point.
In order to select one of 10 devices, you need at least a 4-bit address.
4-bits will allow you to select 1-of-16 devices.
5-bits will select 1-of-32 devices.

Don't worry about how to "lock" a shift register. We do this by a "select" input to the shift register or to the clock input to the shift register. This is called an AND gate. As a circuit configuration, we call these decoders, selectors, multiplexers, depending on the specific function being performed.
 

AnalogKid

Joined Aug 1, 2013
8,443
Now we're getting somewhere.

First, just to be clear, even though the shifter you have chosen has three-state outputs, the outputs will be permanently enabled. yes / no ? I'm asking this because the output enable is one of the control inputs you are not using.

To enable / freeze any combination of devices you need a 1-bit latch of some kind for each device. If you have a microcontroller running things, and it has 10 output pins, you could assign one to each shifter circuit and do the latching in software, but somewhere there has to be something that enables/disables the shifters in any combination. So ...

What control lines / bits / signals are there to tell each shifter whether it is running or frozen (enables / disabled)? There are several different ways to achieve this circuit, depending on how you want to control it.

ak
 

MrChips

Joined Oct 2, 2009
21,070
This is a common situation where the TS has a concept in mind but does not have the knowledge, tools, background and vocabulary to know how to ask the right questions. No put down intended, but the TS focuses on their own solution based on their limited knowledge.

The proper place to begin is to give us the overall picture on what you are trying to do. Refrain from offering your own solutions.
After we see the big picture then we will venture into the details and arrive at a proper solution. Without this approach we end up with 101 Q&A that lead to nowhere constructive.
 
Top