Help: TTL to Inverted TTL with two Outputs

Wendy

Joined Mar 24, 2008
23,415
This is a basic differential amplifier, which is also the root for all op amps and comparitors.

Q1 is a basic constant current source. The current is split between Q2 and Q3. If one of the transistors is turned on, say Q1, then Q2 doesn't get any current, so the outputs are opposite on the collectors.

This is a non-saturated amplifier (usually), which means its speed is very fast. Op amps add a lot of gain, and have to throttle it back with internal feed back to keep themselves from oscillating. The gain of this differential amplifier is limited to the transistors Q1/Q2.

http://en.wikipedia.org/wiki/Differential_amplifier

The design is very old, and was used in the days of vacuum tubes.
 

Thread Starter

Management

Joined Sep 18, 2007
306
This is a basic differential amplifier, which is also the root for all op amps and comparitors.

Q1 is a basic constant current source. The current is split between Q2 and Q3. If one of the transistors is turned on, say Q1, then Q2 doesn't get any current, so the outputs are opposite on the collectors.

This is a non-saturated amplifier (usually), which means its speed is very fast. Op amps add a lot of gain, and have to throttle it back with internal feed back to keep themselves from oscillating. The gain of this differential amplifier is limited to the transistors Q1/Q2.

http://en.wikipedia.org/wiki/Differential_amplifier

The design is very old, and was used in the days of vacuum tubes.
In iCircuit i was able to see the voltages and flow of current. :) Although its a little "twinky" :p hehe. Seriously though thank you for the details. It helps a lot. You guys' commitment to helping people is second to none.

So why is there a fear of Q2 and Q3 (especially) the latter of oscillating?
 

Wendy

Joined Mar 24, 2008
23,415
Because it is a high gain circuit, even if the transistors set the max gain. It has all the ingredients, in a large mix, to make a dandy oscillators. IC chips have to use something they almost never like, a capacitor, as part of the IC to limit this tendency to oscillate. With a circuit made with discrete parts this internal feed back is not there.

Think about it, the inputs are high impedance (a function of the circuit, think common collector), there are inverting and non inverting inputs, and inverting and non inverting outputs. Give this circuit any excuse at all (in the form of feedback) and it will squeal.

I have personally trouble shot ECL chips that reliably toggled at 40Ghz, and this is very old tech nowdays.
 

Ron H

Joined Apr 14, 2005
7,063
High Ft (unity gain bandwidth) transistors will oscillate, even if not in a high gain circuit. I have had emitter followers oscillate. Oscillation can occur because the resistance looking into the base can be negative at certain frequencies. Combine this with a little inductance in series with the base, and a little stray capacitance on the emitter, and you have an oscillator. I ran a sim of the differential pair, with a little base inductance and a little emitter capacitance. See the attachment.
 

Attachments

Thread Starter

Management

Joined Sep 18, 2007
306
In the previous sim, I added 10nF from the base of Q3 to GND, and the oscillation did not occur.
So I built this with 0402 parts in the lab and tested it and it was not working for me.

Many of the voltages in SPICE are the same except for the those at the output with constant input voltage. I haven't tried the input switching pulse. But I am concerned and would like help identifying where the issue could be. Any help would be appreciated.
 

ErnieM

Joined Apr 24, 2011
8,377
The 0V & -5V will be driving a SPDT switch. So hence I need a driver circuit. The original driver circuit using the IC in my last post introduced too much delay.
A "switch" usually refers to a mechanically adjusted connnection device and requires no "driver" beyond your hand.

I suspect you are referring to some other device. Does it have a part number or schematic?

Note:depending on what this device is you may be able to generate your -5V direct from the TTL outputs by sticking some wires into the correct places.
 

Ron H

Joined Apr 14, 2005
7,063
So I built this with 0402 parts in the lab and tested it and it was not working for me.

Many of the voltages in SPICE are the same except for the those at the output with constant input voltage. I haven't tried the input switching pulse. But I am concerned and would like help identifying where the issue could be. Any help would be appreciated.
If you built the circuit in post #44, it probably won't work. This circuit has parasitics added to "test" their effects on operation. Attached is a more complete schematic of what I would build if I needed this function. It may still have oscillation problems. You should build it on a board with a ground plane, with very short wires and traces.
 

Attachments

Last edited:

Thread Starter

Management

Joined Sep 18, 2007
306
If you built the circuit in post #44, it probably won't work. This circuit has parasitics added to "test" their effects on operation. Attached is a more complete schematic of what I would build if I needed this function. It may still have oscillation problems. You should build it on a board with a ground plane, with very short wires and traces.
Hi Ron,

I didn't build the circuit with the parasitic elements. I built one almost the same as what you just posted. I am not getting oscillation I don't think but I know I don't get the voltages I thought I would at the output, i.e. -5V & 0 when powered up and 0V at the input.

I have all that you have except for R6. One of my more important questions is why did you add R6 and what is the value "Rc"? What does it do for you?
 
Last edited:

Ron H

Joined Apr 14, 2005
7,063
Hi Ron,

I didn't build the circuit with the parasitic elements. I built one almost the same as what you just posted. I am not getting oscillation I don't think but I know I don't get the voltages I thought I would at the output, i.e. -5V & 0 when powered up and 0V at the input.

I have all that you have except for R6. One of my more important questions is why did you add R6 and what is the value "Rc"? What does it do for you?
Oops. I forgot to take R6 out. I was running the simulation with different values, to see if it affected the output. You can replace it with a short circuit.
 
Top