Differential Amplifier design

bertus

Joined Apr 5, 2008
20,869
Hello,

Did you notice the smily Johny posted?
Perhaps he hinted that there might be a problem finding the transistor

Bertus
 

dannyf

Joined Sep 13, 2015
2,197
Your ask is a little extreme.

Beta greater than 100 is desirable, beta greater than 50 is workable. You will need a to220 or beefier device. And that's not difficult to find, among singles, Darlington or MOSFETs.

It is a fairly straight design. You just happen to be on the wrong path.
 

Thread Starter

Rumination

Joined Mar 25, 2016
74
Your ask is a little extreme.

Beta greater than 100 is desirable, beta greater than 50 is workable. You will need a to220 or beefier device. And that's not difficult to find, among singles, Darlington or MOSFETs.

It is a fairly straight design. You just happen to be on the wrong path.
Maybe I am a fool, but I can still not understand how to determine the rest of the resistances. Does it means I have to redesign the circuit?
 

dannyf

Joined Sep 13, 2015
2,197
The original ask is a little extreme in that 1. It has a class a output stage delivering substantial current and 2 it asks for low open loop gain. The first issue is typically addressed with a class b output stage and the second issue with closed loop.

Anyway, the key is to maximize your gain in the VAs, assuming the loading from the output is minor.

You will then pick the VAs idle current so it can deliver sufficient current to drive the ops. This will depend on the assumed beta of your ops device.

Once the VAs idle current is picked, Rc can be easily calculated.

Re is 2kohm time half of the input stage tail current minus vbe, divided by VAs idle current.

From RC and Re, you calculate the gain of VAs. From that you know how much gain you have to get out of the input stage, thus those two Re resistors value.
 

dannyf

Joined Sep 13, 2015
2,197
So some examples. The max load current is about 150ma. So ops idles at 200ma. That means it's emitter resistor is about 22/0.2 ohm.

Assuming a beta of 100, ib is 2ma. We then idle VAs at 4ma. That means VAs collector resistor is (22 +0.7) / 4 or 5kohm.

Voltage over the re resistor is 2k x 1.8ma - 0.7v, or 3v. At 4ma, a resistance of 750ohm.

Thus VAs has a gain of 5/0.7, or 7x.

That means the input stage has to have a gain of 15x. So re resistor there should have a value of 2000 / 15 / 2, or 70 if my math is right.

Because it is open loop, the above will not actually produce a 100x overall gain (likely lower) but the general gist is the same.
 

dannyf

Joined Sep 13, 2015
2,197
Using the value calculated above, I got a gain of 20x -> because I had assumed single-ended input signal when you are using differential. So the Re resistors for the input stage should have a value of 140ohm vs. 70ohm above. That brought down the gain to 10x.

OPS used a 3055 -> beta is 100x.
 

dannyf

Joined Sep 13, 2015
2,197
The output doesn't sit at 0v at idle, but below ground due to the fact that we did not count for the loading of OPS on the VAS. The cure is to increase the VAS idle current.
 

Thread Starter

Rumination

Joined Mar 25, 2016
74
I still can't understand. :( But thanks for the help. The assignment deadline is tomorrow night. So thank you all for having patience with me, because I find this analog design very difficult.
 

Jony130

Joined Feb 17, 2009
5,186
Did your teacher or professor give you any information about transistor parameter? Q3 is a VAS stage (Voltage amplifier stage).
And how you are gonna to check if the amplifier meets the design requirements?
 

Thread Starter

Rumination

Joined Mar 25, 2016
74
Did your teacher or professor give you any information about transistor parameter? Q3 is a VAS stage (Voltage amplifier stage).
And how you are gonna to check if the amplifier meets the design requirements?
He haven't given any information about transistor parameter.
 

dannyf

Joined Sep 13, 2015
2,197
"I still can't understandunderstand"

That's why I always tell people that you get to your destination faster if you take a step at a time. Understanding the basics goes a long way getting you to work comfortably with complex projects.
 

Bordodynov

Joined May 20, 2015
2,696
How did you and Darlington are going to get a temperature-stable zero volts on the output of the amplifier? This theoretical research or layout will be made? I will offer its own version of a push-pull output and temperature stabilization (post #29). In my case it is not necessary a large output transistor quiescent current and do not need a radiator-cooler.
 
Top