general transistor amplifier questions

Hey all,

I have some general questions about designing and making a simple transistor amplifier that I can't seem to figure out from reading various tutorials and forum posts.

Ive gotten transistors working as switches so far, just not as amps yet.

What I want to do is amplify the DAC output of the arduino DUE at low frequency (near DC). So, if i set the DAC output to a constant 1.5v, I want to amplify this to say 15v with higher current capability. I already have some pnp power transistors that I'd like to use if possible.

I'm guessing I will need something like this

so, a lot of circuit designs Ive seen have a cap in series with the Vin and Vout points in the circuit, but this cap is rarely explained other than to say it is there to "remove the DC component of the signal". That makes sense, but since I want to operate more or less at DC, can I just throw out these caps?

also, if both voltage and current are amplified, does that mean current capability is amplified, so the load will draw its rated current when at its rated voltage? or will the amp operate as a constant current source and force current across the load?

finally, does anyone know of a good tutorial on amp design, how to choose appropriate biasing resistors and such? Im finding a lot of good explanations on how they work, but nothing good enough that I feel I could successfully design an amp with any confidence that it would actually work.

any help would be appreciated! thanks!

The simplest approach would be to use an operational amplifier, which will amplify DC and AC voltages. Depending on the output current (i.e. if greater than about 10 milliamperes) you may need to use a power op amp. You'll also need a suitable power supply for the amplifier circuitry (higher than 15V). Google will show you many possible circuits, but it need not be complicated.

so, these standard transistor amps are not good for amplifying DC voltage?

I will use a 20v power supply for the external circuitry and probably need .5 amps, so higher power/voltage parts will be used.

Transistors are fine. You need to bias them to operate in their linear region of operation, that's what the unmarked resistors in your diagram do. The cap then lets the AC component of your signal be amplfied as you noted.
You could try it as set up here:

If you're going to be driving potentially 7.5 W, you're going to want switching, not linear control. Linear control will waste way too much power, which will cause your components to heat up. A lot.

What are you driving that needs half an amp? There may be a better way to do it, like PWM.

That makes sense, but since I want to operate more or less at DC, can I just throw out these caps?

You can but you will fry your Due output if you do. This is because you have a PNP transistor and so the bias voltage is going to be up close to your Vcc which is way over the 3V3 needed to fry your output.
You need the NPN equivalent of that circuit.

so, these standard transistor amps are not good for amplifying DC voltage?

They are, but designing a DC amplifier of the correct gain is much more taxing than using an op amp.

I like with the op-amp idea. A DC amplifier made from discrete transistors is going to be tricky for a beginner, and you may need multiple transistors. I haven't built an amplifier from transistors since the 1970's, and personally I'd try to avoid it.

If you can't find an op-amp with high-enough power requirements, you may be able to find a schematic that uses an op-amp plus a pair of transistors. (Putting the transistors inside the op-amp's feedback loop will simplify the design and give you better precision.)

If you need to get accurate amplification of low voltages (maybe below 1 volt or 1/2 volt) you'll generally need positive and negative power supplies.

That makes sense, but since I want to operate more or less at DC, can I just throw out these caps?

With a DC amplifier you can't have capacitors in the signal chain, but you'll probably have to make some other design changes (or use a different design). Simply "throwing out" (bypassing) the capacitors will likely throw-off the transistor's biasing to the point where the amp won't work.

What are you driving that needs half an amp? There may be a better way to do it, like PWM.

I guess I was thinking total current for the whole external circuit, not neccesarily the current through the load on one of these amps.

Im trying to make a piezo actuater that will allow for x,y,z positioning on the nano scale. So I will need one of these amps for each dimension. I'm thinking of it as, if I apply 1v across the piezo it will expand XXX nm, the finer the applied voltage resolution I have, the finer the resolution in positioning I should have, hence the 12 bit analog out on the due. I want to be able to move from any given x,y,z value to another, and hold at any given x,y,z if need be.

honestly, I have no idea how much current the piezo will draw, or how much it will move per volt, I just want to make sure I have enough current available and I can calibrate distance per volt in situ. But I dont think there is a way to convert PWM to to DC that would be better than just amplifying the DAC output.

I haven't built an amplifier from transistors since the 1970's, and personally I'd try to avoid it.

thats all i needed to hear. thanks.

If you need to get accurate amplification of low voltages (maybe below 1 volt or 1/2 volt) you'll generally need positive and negative power supplies.

ok I do want to get as much precision as i can, without getting too elaborate/expensive of course. I have a 20v power supply that I thought i would be able to split with a virtual ground, +10 and -10...of course, that's kind of further down the line and was hoping to avoid it if possible. right now was just trying to make a simple amp that took my 0-3.3v in and made it 0-15v

honestly, I have no idea how much current the piezo will draw,

Sod all

or how much it will move per volt

Sod all

But I dont think there is a way to convert PWM to to DC that would be better than just amplifying the DAC output.

PWM to DC is dead easy and much simpler than amplifying a low level DC output.

I think you should do some measurements with just your piezo connected directly to your power supply to get an idea of the amount of movement you get.

Piezos act as capacitors and do not draw significant current in the steady state. When used as actuators on the nano scale, typically one utilizes a programmable power supply/amplifier combination capable of several hundred volts output, accurate and stable to 0.01 V or better. You might look at the specs for a commercial piezo driver, such as this one: http://www.piezosystem.com/piezo_actuator_nanopositioning/piezo_controller_and_amplifier/piezo_controller_1_channel_version/30dv50_oem_digital_controller/

You can but you will fry your Due output if you do.

i think this is the same as saying, "no, no you can not."

This is because you have a PNP transistor and so the bias voltage is going to be up close to your Vcc which is way over the 3V3 needed to fry your output.

I dont really understand this. ive used this transistor just fine as a switch in saturation, full on and full off. it was on my UNO, so 5v instead of 3.3, but base current draw and limiting seemed to be the biggest thing to watch out for. Operating below saturation would seem to be safer?

PWM to DC is dead easy and much simpler than amplifying a low level DC output.

I looked at this, seemed like just a low pass filter, but the consensus seemed to be that the DC signal is not very clean/stable. I guess I could spend time researching filtering and cleaning signals or spend time researching amps. I went with the latter...

Piezos act as capacitors and do not draw significant current in the steady state.

this is the assumption I was working under.

typically one utilizes a programmable power supply/amplifier combination capable of several hundred volts output, accurate and stable to 0.01 V or better.

Basically Im going for a hacked version of this. I believe some of these actuators are on the order of an inch in length, hence the huge voltage. I plan to just use the ceramic peizo out of a buzzer and make it move with only 20v across it, also with (hopefully) with 10mV precision.

im roughly following this project , their amp provided more utility for them than i needed, so i was trying to simplify it down, and use components i had on hand (the pnp's) from responses here it seems the power op amps are going to be a better idea though...

This is because you have a PNP transistor and so the bias voltage is going to be up close to your Vcc which is way over the 3V3 needed to fry your output.

I dont really understand this.

What is the voltage on the base of the PNP transistor when it is fully on?
It is 0.7v lower than the Vcc. If you have that Vcc at 15V that puts it at 14.3V going into the output pin of your arduino.
Using a NPN transistor the base is just 0.7V off ground. That is why you have done switching before with an NPN.

make it move with only 20v across it, also with (hopefully) with 10mV precision.

You will need a DAC with at least 12 bits of resolution to do that. Adafruit has one, just follow it with an op amp amplifier. https://www.adafruit.com/blog/2012/07/24/new-product-mcp4725-breakout-board-12-bit-dac-wi2c-interface/

It is 0.7v lower than the Vcc

my understanding is that the base has to be at least 0.7 lower than Vcc to be in a conducting state. if Vcc is 15, you can still put the base at 1v, so long as there's enough current flowing out of the base, into the arduino pin at ground, to put the pnp in saturation.

is that wrong?

You will need a DAC with at least 12 bits of resolution

yup, i went with the due because it has 2x built in 12 bit DAC. thanks for the link though, i was looking for 12bit DAC before switching to the due. I even looked at adafruit, can't believe i didnt see that. i was looking more at their audio Shields i guess.

That is right when the arduino is switching the transistor on.
But when the arduino tries to turn it off it puts 5V on the base which is still enough to send the transistor into saturation and also send your arduino to silicon heaven.

ohhhhh yea...i had to use an npn to turn on a pnp....forgot about that little detail haha. thanks.

i followed along with this. its all laid out very clearly there....