Transistorized DC amplifier

[mercurial]

Hi

I understand that there are topologies like Common Emitter, Common Base and Common Collector for AC amplifiers. I'm sure that can also be used for DC input.

I was looking at converting a 0 to 3V DC input to 0 to 12V output using transistor amplifier stages.
Are there any alternative topologies and guidelines for dc amplification?
Any general rules to approach the design of operating point and calculation the output?

Or are the same trio of topologies as given above suited for dc amplification.

[ppTRN]

If all you need to do is a logic level shift you can use a single transistor circuit used as a switch, meaning that the transistor is only in cut-off or saturated, never in linear region.
Of course you need to have available the 12V supply. If you want to generate 12V you need a boost converter. Still based on a single MOSFET, but it needs to be understood and managed by a microcontroller or a boost controller

[Picuino]

https://www.falstad.com/circuit/circuitjs.html?ctz=CQAgjCAMB0l3BWcMBMcUHYMGZIA4UA2ATmIxAUgpABZsKBTAWjDACgAXcDQ8QqlChp8qVCExTQU2UjQRgakBPOKLhMQlmIo8NMMU14l2cmDggAJgwBmAQwCuAGw5sAblBCD1tPB6qLwJFEoaAQ2AHcPLxBCYWjICM8hJLjksB4oNgAnFPAdXJpfKkoEnIVhMHzoyqLPeDhE-19owszI6qq0mnVE8rzfPrB+NvBukW5eIdE3GNThWPGxWjFQ8FXRUMSF6J3k0tnxweGxesSO5owUfszFeh1fA1ohxfBsAH0at8g3gA88M0qbwQX1g8DIhAQhBQwJYbxQcLeTGIeA+73eMHqNFI2JxuIMXy2FWGCxqN0g9GeJLGpLEwM+3z+APhwIxYJ4kOhiLACPhTBwH2B6NBcCxeLF+ISQA

[Kleinstein]

One has the same types of basic circuit also for DC.  If a bit more precission is needed one may want to consider the differential amplifier / long tailed pair configuration as this has compensation of the DC drift of VBE that make the basic 3 circuit a bit drifty.

[Terry Bites]

I'm not sure if you want to obain a linearly scaled output or a hard step.
In the linear case you'd essentially build an opamp from discrete parts- that would be crazy of course!
If you must.... https://community.element14.com/members-area/personalblogs/b/frank-milburn-s-blog/posts/op-amp-made-from-discrete-components

[Picuino]

https://sound-au.com/project07.htm

[Ian.M]

Section H of Op Amp Applications Handbook has a history of early OPAMPs including schematics for Philbrick's discrete silicon BJT OPAMP modules.   

Its certainly possible to achieve the OP's spec. of x4 gain over an input range 0-3V with relatively few discrete parts, but the result is far too sensitive to supply rail variations, component tolerances, transistor gain and temperature for most practical uses.

Here's a LTspice sim of a four transistor OPAMP which may be suitable.  N.B the long-tailed pair transistors must be well matched.
 

[dobsonr741]

There are lots of color to "amplification".
First oder of business: Is the output proportional and linear, or 0V in -> 0V out and 3V in -> 12V out?
In other words, are we amplifying a linear or a digital signal?

Later on:
If linear, how accurate linearity you want? 10%? 0.1%?
What power rail is available, supposedly higher than the 12V max output?
How close you need to be to 0V on the input and output? 1mV? 100mV 500mV?

[jasonRF]

I would add one more question: what is the load? 

If it is a linear DC amplifier then I agree with other folks here that you would basically build a discrete opamp.  I don't think this is crazy - it can be a fun learning exercise and I would recommend doing it.  But of course a $1 opamp will perform better in general. 

jason

EDIT: for fun I played with a simulation of yet another simple linear x4 DC amplifier.  Something like this could work as long as you are okay with 100mV-class errors in the output voltage.   I did not do a stability analysis of this so picked a 'large' 1nF compensation capacitor.  If you have loads smaller than 50 Ohms then you would need a beefier output transistor and/or a heatsink; with loads >=100 Ohms you could get away with a pn2222 or even 2n3904.   

[mercurial]

If all you need to do is a logic level shift you can use a single transistor circuit used as a switch, meaning that the transistor is only in cut-off or saturated, never in linear region.
Of course you need to have available the 12V supply. If you want to generate 12V you need a boost converter. Still based on a single MOSFET, but it needs to be understood and managed by a microcontroller or a boost controller

sorry I was not clear, the output is proportional to input, 0 to 3V input 0 to 12V output so for 1.5V input I would get 6V.

I'm not sure if you want to obain a linearly scaled output or a hard step.
In the linear case you'd essentially build an opamp from discrete parts- that would be crazy of course!
If you must.... https://community.element14.com/members-area/personalblogs/b/frank-milburn-s-blog/posts/op-amp-made-from-discrete-components

Proportional output is needed.

There are lots of color to "amplification".
First oder of business: Is the output proportional and linear, or 0V in -> 0V out and 3V in -> 12V out?
In other words, are we amplifying a linear or a digital signal?

Later on:
If linear, how accurate linearity you want? 10%? 0.1%?
What power rail is available, supposedly higher than the 12V max output?
How close you need to be to 0V on the input and output? 1mV? 100mV 500mV?

a. 1%
b. 15V power rail
c. 0.1V is fine.

[mercurial]

https://www.falstad.com/circuit/circuitjs.html?ctz=CQAgjCAMB0l3BWcMBMcUHYMGZIA4UA2ATmIxAUgpABZsKBTAWjDACgAXcDQ8QqlChp8qVCExTQU2UjQRgakBPOKLhMQlmIo8NMMU14l2cmDggAJgwBmAQwCuAGw5sAblBCD1tPB6qLwJFEoaAQ2AHcPLxBCYWjICM8hJLjksB4oNgAnFPAdXJpfKkoEnIVhMHzoyqLPeDhE-19owszI6qq0mnVE8rzfPrB+NvBukW5eIdE3GNThWPGxWjFQ8FXRUMSF6J3k0tnxweGxesSO5owUfszFeh1fA1ohxfBsAH0at8g3gA88M0qbwQX1g8DIhAQhBQwJYbxQcLeTGIeA+73eMHqNFI2JxuIMXy2FWGCxqN0g9GeJLGpLEwM+3z+APhwIxYJ4kOhiLACPhTBwH2B6NBcCxeLF+ISQA

This looks good and its single stage too. Could you please explain how you calculated those resistor values.

[coppercone2]

I would watch out for transistorized electronics because its still a emerging technology. IMO stick to thermionic circuits, at least until silicon devices become affordable, rumor has it they are quite reliable compared to germanium.

[mercurial]

Hi,
Quite a number of circuits in this thread use a number of transistors and also differential input stage.
What are the advantages of those vs the one topology that Picuino presented , which used a single transistor and few resistors.
His circuit is here.

https://www.falstad.com/circuit/circuitjs.html?ctz=CQAgjCAMB0l3BWcMBMcUHYMGZIA4UA2ATmIxAUgpABZsKBTAWjDACgAXcDQ8QqlChp8qVCExTQU2UjQRgakBPOKLhMQlmIo8NMMU14l2cmDggAJgwBmAQwCuAGw5sAblBCD1tPB6qLwJFEoaAQ2AHcPLxBCYWjICM8hJLjksB4oNgAnFPAdXJpfKkoEnIVhMHzoyqLPeDhE-19owszI6qq0mnVE8rzfPrB+NvBukW5eIdE3GNThWPGxWjFQ8FXRUMSF6J3k0tnxweGxesSO5owUfszFeh1fA1ohxfBsAH0at8g3gA88M0qbwQX1g8DIhAQhBQwJYbxQcLeTGIeA+73eMHqNFI2JxuIMXy2FWGCxqN0g9GeJLGpLEwM+3z+APhwIxYJ4kOhiLACPhTBwH2B6NBcCxeLF+ISQA

[soldar]

That single transistor circuit is going to have poorer linearity, be affected by power supply voltage variations, by the individual characteristics of the transistor, etc.

That is why op amps were invented.

[Ian.M]

Also, its inverting (increasing input causes decreasing output) which is an unavoidable characteristic of all single transistor voltage amplifiers,  has a large uncontrolled DC offset, and gain and offset will vary a lot with temperature.

If you want 0V in to give 0V out, you either need a fairly fancy RRIO (Rail to Rail Input & Output) OPAMP, (as most single supply OPAMPs do poorly driving right down to ground), or a negative supply rail.   The simpler the circuit the more stringent the supply requirements are.  e.g. my circuit above needs both +15V and -15V rails which must be well regulated.   To make it work well with poorly regulated supplies, or a reduced negative supply, I'd have needed at least double the number of transistors, and most of the extras would have needed to be matched pairs.

[MarkT]

I was looking at converting a 0 to 3V DC input to 0 to 12V output using transistor amplifier stages.

This is normally done with rail-to-rail opamp, which would only need a 12V single supply.  With discrete transistors you'd probably need a negative supply and +15V positive supply to give enough headroom.

Opamps are much simpler to use than transistors.

[Picuino]

In general, to avoid drifts and improve linearity you will need a feedback amplifier, which is not easy to calculate. It can be an operational (differential input) or another topology, but the most common and where you will find more examples is with an operational. If you don't need a lot of speed try one like the LM324.

https://www.ti.com/product/LM324

https://www.ti.com/lit/ds/symlink/lm324.pdf

[Picuino]

Another feedback topology more linear.

https://www.falstad.com/circuit/circuitjs.html?ctz=CQAgjCAMB0l3BWcMBMcUHYMGZIA4UA2ATmIxAUgpABZsKBTAWjDACgAXcDQ8MFEChQ0+AqhBbZo2MDRpgElYQkL4RMBNmIo8kBBkgk8NfEhSUQAEwYAzAIYBXADYc2ANyiCTnsLypVvBU9-aAQ2AHdPYSpCEWioCJ9eMB5wP0TYtPFU3393EEz4zNyfWnFQ8AqQsMjC7yE470hEzAFWAWw0UQSTemxsXl1aYrlPDAB9fjxxyHGADzwwSH5xhBnYeDJCFXNxlnGUA73iaYnsdfhIGlIb27viQhmMkRLinR7IPoFi1XB3qmIkx0M3mi2WhzWMEuWx2a32h0OTBw40B5yh8Gu9yxDyeXAaIAeggwAkJVBYfGgeEIKGwGBUeB4hEIAzwICYMGZBgwxGwNCEVJU5CWVGs9mcrgAToJhNxePj4uI4C0ZSk5Sr-i1ibLpS9Us1Iq0QENDb4Egatab8abmlKrWqRKTkPA2JZte03b8BKLHC42FLVWk3fwfM7IgHLSr0raLcGTXqys0gA

https://tinyurl.com/2xjqa4uf

[mercurial]

Hi Picuino,

What literature are you referring to for these transistorized designs?
Could you let me know how you are calculating the resistors for a given input and output?.

[Picuino]

The resistors of the output are a voltage divider that determines gain.
Vout = Vin * (400+100) / 100 = 5 Vin


Analysis and Design of Analog Integrated Circuits, 5th Edition 5th Edition
by Paul R. Gray (Author), Paul J. Hurst (Author), Stephen H. Lewis (Author), Robert G. Meyer (Author)
https://www.amazon.com/Analysis-Design-Analog-Integrated-Circuits/dp/0470245999

[Picuino]

https://tinyurl.com/23mxnx7h
With 4x amplification and small signal cutout at the bottom.

[mercurial]

https://www.falstad.com/circuit/circuitjs.html?ctz=CQAgjCAMB0l3BWcMBMcUHYMGZIA4UA2ATmIxAUgpABZsKBTAWjDACgAXcDQ8QqlChp8qVCExTQU2UjQRgakBPOKLhMQlmIo8NMMU14l2cmDggAJgwBmAQwCuAGw5sAblBCD1tPB6qLwJFEoaAQ2AHcPLxBCYWjICM8hJLjksB4oNgAnFPAdXJpfKkoEnIVhMHzoyqLPeDhE-19owszI6qq0mnVE8rzfPrB+NvBukW5eIdE3GNThWPGxWjFQ8FXRUMSF6J3k0tnxweGxesSO5owUfszFeh1fA1ohxfBsAH0at8g3gA88M0qbwQX1g8DIhAQhBQwJYbxQcLeTGIeA+73eMHqNFI2JxuIMXy2FWGCxqN0g9GeJLGpLEwM+3z+APhwIxYJ4kOhiLACPhTBwH2B6NBcCxeLF+ISQA

Hi Picuino ,
I was referring to this design and the other feedback one that you posted, the more linear one.
How does one calculate the component values for these?

[Picuino]

This circuit with one single bjt is more difficult to calculate.
But you can try values in order to obtain good results.

10k is the input impedance.
200k is the feedback resistor (from output voltage to input current)

If the gain of the transistor were infinite (is only 100 in the simulation), the gain of the stage would be equal to:
Vout = Vin * 200k / 10k = Vin * 20

To calculate other parameters it is necessary to make much more extensive calculations.
The simplest is to calculate the amplifier parameters without the 200k resistor and then use the formulas for the feedback amplifiers.
But that approximation is not very accurate. For more accuracy it is necessary to use the transistor model and that is more complexity than I would like to write.
I don't think you will have a hard time finding literature on that.

For now with the formulas I have given you and a transistor with a good gain (250, higher than the 100 I have put in the simulation), you can make approximations with the simulator until you get what you need.

[Picuino]

I will recalculate it better, with formulas and adding polarization.

Forward gain:

ΔIc = ΔIb * 250
ΔVout = - 1000 * ΔIc

A = ΔVout/ΔIb  = -1000*250 = -250000

Adding Feedback resistor:

β = ΔIb/ΔVout = 1/50000


Total loop gain:

https://en.wikipedia.org/wiki/Loop_gain

ΔVout/ΔIb = G =  A / (1 - A*β) = -250000/(1+250/50) = -250000/(6) = -41666 [V/A]


Adding input resistance: (Out of feedback circuit)

Rin = 10k

ΔIb = ΔVin / 10k

ΔVout / ΔVin = -41666 / 10k = 4.16


Adding polarization:
Testing in the simulator until the output signal is zero at the low point and is not clipped, a resistance of 4k comes out.


Final circuit:

https://www.falstad.com/circuit/circuitjs.html?ctz=CQAgjCAMB0l3BWcMBMcUHYMGZIA4UA2ATmIxAUgpABZsKBTAWjDACgAXcDQ8QqlChp8qVCExrQ82bHjAJSJMInoxCCbAkLY62DDQxg0xECkogAJgwBmAQwCuAGw5sAblFNCPYXqNpikPxgENgB3D0FhQmFIqDDPGK9YsB44gCcE8BQ8TJ9veDYMsBphIxzk7I9KAvCxXmTfeIryr2LhSHi2rJyun1FOkpFuXj6492jMidGxf3BoJDB5j2D4ieTB2I6Mqf5wQenkGtzKsxHKjppIelOQYl5ZUt2xbAB9MpfIF4APOVYUF4QH1g8DI6kIZheLBe-3+TB4b1erxg8EgNFI6IxmLuH1WjyoUyMcUu9GwKFoPV2f28gPenx+rCMAKBKNBWghUJhkJwb0BSOBcDRWKF2I64S6sQlXi2ezxMsyVEucDYQA

[Picuino]

In reality the gain has a feedback factor (1+A*B) which is relatively low and will not linearize the transistor behavior too much. The transistor has a behavior that is non-linear because it is working in a wide range of currents, even going into saturation with what that means of gain loss. A more exact formula would be even more complex.

[mercurial]

I will recalculate it better, with formulas and adding polarization.

Hi
You use the term polarization, is it same as biasing a transistor?

[Picuino]

Yes.

[Picuino]

Some theory about feedback amplifiers:
 * https://testbook.com/physics/feedback-amplifier
 * https://www.jntua.ac.in/gate-online-classes/registration/downloads/material/a159298350984.pdf
 * https://www.geeksforgeeks.org/feedback-amplifiers/

Although the best way to study these concepts is a good book like the one I attached above (Gray - Meyer).

[Zero999]

Another feedback topology more linear.

https://www.falstad.com/circuit/circuitjs.html?ctz=CQAgjCAMB0l3BWcMBMcUHYMGZIA4UA2ATmIxAUgpABZsKBTAWjDACgAXcDQ8MFEChQ0+AqhBbZo2MDRpgElYQkL4RMBNmIo8kBBkgk8NfEhSUQAEwYAzAIYBXADYc2ANyiCTnsLypVvBU9-aAQ2AHdPYSpCEWioCJ9eMB5wP0TYtPFU3393EEz4zNyfWnFQ8AqQsMjC7yE470hEzAFWAWw0UQSTemxsXl1aYrlPDAB9fjxxyHGADzwwSH5xhBnYeDJCFXNxlnGUA73iaYnsdfhIGlIb27viQhmMkRLinR7IPoFi1XB3qmIkx0M3mi2WhzWMEuWx2a32h0OTBw40B5yh8Gu9yxDyeXAaIAeggwAkJVBYfGgeEIKGwGBUeB4hEIAzwICYMGZBgwxGwNCEVJU5CWVGs9mcrgAToJhNxePj4uI4C0ZSk5Sr-i1ibLpS9Us1Iq0QENDb4Egatab8abmlKrWqRKTkPA2JZte03b8BKLHC42FLVWk3fwfM7IgHLSr0raLcGTXqys0gA

https://tinyurl.com/2xjqa4uf

Current feedback is a good idea. It's also possible to design such a circuit with a high bandwidth.

The gain is a little off. Remember A_V = 1+R_f/R_g

I would replace the diode with an emitter follower to increase the input impedance.



 3V to 12V discrete.asc (1.43 kB - downloaded 16 times.)

Hi

I understand that there are topologies like Common Emitter, Common Base and Common Collector for AC amplifiers. I'm sure that can also be used for DC input.

I was looking at converting a 0 to 3V DC input to 0 to 12V output using transistor amplifier stages.
Are there any alternative topologies and guidelines for dc amplification?
Any general rules to approach the design of operating point and calculation the output?

Or are the same trio of topologies as given above suited for dc amplification.

Just one question: why?

Use a single supply op-amp, such as the LM358.


A_V = 1+R_f/R_g

For a gain of 4:
R_f = 30k
R_g = 10k

[Picuino]

https://www.falstad.com/circuit/circuitjs.html?ctz=CQAgjCAMB0l3BWcDoBYDsAmdA2dBOdMMADhzEnxAUmpFQGZqBTAWmICgAXcXcMTCEyZU-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

[Picuino]

https://www.falstad.com/circuit/circuitjs.html?ctz=CQAgjCAMB0l3BWEBmAHAJmgdgGzoRmACzICcpkORICkNIJNApgLRhgBQAblCOlul5FUvOkTpgkdadAQcA7r36CqfAVAVCR6dNWEbFqsDgnoRx6dxCrl16hd7IGE2eFcy5h6rZ3f1kTVtiamR0CSJqAPEnUIkwQQjzXF4cAH141FTIVIAPVHZ41IQs2HhSXAQ8YrZU9FrUlixkVLTmmHhIInJunt7SNICvcBM7cDMNaJQVexGM3kyMrNz89jri9rKKqoawerrG5sy20rguvvP+rI5xUnBk4LVBB6cYHFFYdGQwDp+OwXRNHMHEDxoMxuZxmB7qDAro7m9fI8DHw4asUXoRAEAE5Ih4+SCYlDwDg4-HadS2CTExSInwU-yBCkE+F8ZlgqFvYHJMKYjhAA

[Picuino]

The operational scheme is better at reaching the limiting voltages and is more linear, but its response is much slower for high frequencies.
The scheme with two transistors can amplify without problems frequencies of 1MHz or more, which would be very difficult to achieve with the operational.

[Zero999]

The operational scheme is better at reaching the limiting voltages and is more linear, but its response is much slower for high frequencies.
The scheme with two transistors can amplify without problems frequencies of 1MHz or more, which would be very difficult to achieve with the operational.

It depends on the op-amp and the slew rate.

What we have here is a type of current feedback amplifier, as I hinted in my previous post. The speed is limited by the turn-off time of Q3, so one can expect it to have a fast rise time and slow fall time. It has a wide bandwidth for low amplitude signals, but will suffer from slew rate distortion, with large amplitude, high frequency signals.

Pulling Q3's base up via a current source and reducing the resistor values will speed the circuit up, at the cost of more parts, higher power consumption and a lower input impedance.

A real current feedback operational amplifier has a push-pull output and the transistors are pulled up by current sources, giving a fast slew rate for negative, as well as positive output swing.

I've done an AC analysis. There's a peak around 10MHz, which indicates some resonance/oscillation.

 3V to 12V discrete AC.asc (1.59 kB - downloaded 4 times.)

The transient analysis shows a bit of over/undershoot. Note the fast rise time and sluggish fall time.

 3V to 12V discrete tran.asc (1.51 kB - downloaded 2 times.)