EEVblog Electronics Community Forum
Electronics => Beginners => Topic started by: rwgast_lowlevellogicdesin on June 16, 2024, 12:13:27 am
-
I have been in to electronics for a while but mostly digital stuff and radio stuff, when I need to amplify something I usually use an opamp or an MMIC or some other purpose built chip like an lm386, or even a fet. The only way I know how to really use a BJT is as a switch or an inverter. I've been interested in building this little thing called the TAPIR E-Smog detector, which is a kit Elektor sells that amplifies EM radiation and outputs it as audio. There are plenty of other ways to accomplish the same thing using opamp's or a micro's ADC, but I really want to use this schematic to finally figure out how to use a BJT as more than a switch.
[attachimg=1]
So just looking at this schematic it all seems pretty simple. just connect an antenna or inductor to the input of a high gain amplifier and send the output to a speaker. Here are some of the things i don't understand about this circuit though:
#1 Are the transistors connected in what you would call a cascade? Apparently when googleing you dont series transistors but looking at this all the bases are in series..
#2 When using a transistor as an amplifier dont you need to bias the base with a voltage divider? How is t1 being turned on with virtually no current on it's base?
#3 Why does this circuit switch the type of transistor from 3904 for T1 to a BC847 for T2, and T3, looking at the data sheets there both ~100 hfe with a 300mhz speed, why not just use all 2n3904s?
#4 I was always under the impression the Collector current would be amplified to Beta X the Base current. How does the voltage get amplified in this circuit if transistor work like a current valve controlled by the base?
#5 What is going on between the base of T1 and the audio ouput jack? I dont understand the 100k 100k voltage divider with a 10uF to ground in between them.
-
It's just an amplifier with immense gain.
-
It's just an amplifier with immense gain.
Immense gain, immense distortion, ... just amplifies immense noise with lots of distortion. If used near an AM radio station it will probably rectify the signal and you can hear the radio.
I see no useful purpose to building this.
-
#1 Are the transistors connected in what you would call a cascade? Apparently when googleing you dont series transistors but looking at this all the bases are in series..
“Cascade” in common sense of the word, but not any specific “cascade” circuit. Just three common emitter amplifiers one after another.
#2 When using a transistor as an amplifier dont you need to bias the base with a voltage divider? How is t1 being turned on with virtually no current on it's base?
The transistors here operate as simple switches (common emitter (https://en.wikipedia.org/wiki/Common_emitter)). Any stronger noise on input goging above 0.4–0.5 V threshold(1) makes T1 conduct for a moment. Then it goes back to 0 V (or even slightly negative), turning it off again. This also answers #4.
Actually you can connect them in a similar fashion and use a LED as an output, switching it with nothing more than moving your finger close to the base of the first transistor. It’s that sensitive.
#3 Why does this circuit switch the type of transistor from 3904 for T1 to a BC847 for T2, and T3, looking at the data sheets there both ~100 hfe with a 300mhz speed, why not just use all 2n3904s?
BC547/847 have considerably higher gain than 3904s. In particular those of higher binning are rated for up to hFE ≥ 800, while 3904 are in 150 range. The schematic very explicitly calls for 847B, which is the middle binning and goes up to 500 (200–300 for at higher IC). But I have no idea, why would anybody use one 3904 instead of just going for three BC847.
(1) Remember BJT’s switching action, similar to diode’s, is not: completely off until 0.6 V and then completely open. It actually starts conducting much lower than 0.6 V. And with that amount of amplification you can detect this.
-
So what does r3, r4. and c3 do?
-
T1, T2, & T3 are DC coupled common emitter amplifiers.
r1, r4. and c3 set the DC bias of the output by providing negative feedback. Should end up with slightly above 0.7Vdc at the collector of T3.
-
You can download the Elektor article on this circuit here:
https://www.elektormagazine.com/magazine/elektor-201207/19936 (https://www.elektormagazine.com/magazine/elektor-201207/19936)
An E-smog detector is designed to detect the ‘radiant misbehaviour’ of nearby electronics. The TAPIR — short for Totally Archaic but Practical Interceptor of Radiation — is a simple design capable of detecting, and audibly pinpoint, any source of electric or — with the appropriate antenna — magnetic field.
Attached is a little on what it says about the circuit's operation and component selection.
-
#1 Are the transistors connected in what you would call a cascade? Apparently when googleing you dont series transistors but looking at this all the bases are in series..
Let's call it a sequence of three common emitter amplifiers. No rocket science to it, you just get 3x the gain and 3x90° phase lag at high frequencies. The lag could be a problem with feedback, but there is none here except at DC - see #2. The individual stages are inverting, so the overall amplifier is inverting as well, at least at low frequencies free of additional phase lag.
#2 When using a transistor as an amplifier dont you need to bias the base with a voltage divider? How is t1 being turned on with virtually no current on it's base?
It gets base current through R1,R4 from the output. Remember, inverting amplifier so negative feedback. If the output goes too high, it drives the input in such way to reduce the output. T3 collector voltage automatically becomes what's necessary to bias T1 base to sink most of its collector load current and put T2 in linear region. The R4,C3 lowpass prevents signal frequencies from feeding back and causes the amp to work open loop for maximum gain. It also avoids instability by filtering frequencies above the amps bandwidth where phase lag starts to accumulate and crosses 180° at some point.
The same feedback process automatically biases all other transistors to the Vbe of their subsequent stages. This is a linear amplifier and not a detector of 500mV peaks (for starters, the input signal is clamped to ground with schottkys so T1 would never turn on without R1).
#3 Why does this circuit switch the type of transistor from 3904 for T1 to a BC847 for T2, and T3, looking at the data sheets there both ~100 hfe with a 300mhz speed, why not just use all 2n3904s?
Note sure what's the point of BC847, they don't seem critical and any jellybean should work. Maybe because it's a European design? OTOH, 2N3904 may have been chosen for lower noise but I'm not sure. In such case, perhaps 2N2222 would be better still.
#4 I was always under the impression the Collector current would be amplified to Beta X the Base current. How does the voltage get amplified in this circuit if transistor work like a current valve controlled by the base?
Voltage gain of common emitter stages comes from collector current flowing through a fairly high resistance, which develops a voltage swing potentially much greater than the one driving the base. Note that β doesn't even appear in this equation, but it does determine input impedance of the stage.
Regarding base voltage swing, it is very well determined because transistors sort of follow the diode law: Ie increases exponentially with increasing Vbe, and under sensible circumstances (high β, no saturation) Ic is about equal to Ie. See Ebers-Moll.
-
#1 Are the transistors connected in what you would call a cascade? Apparently when googleing you dont series transistors but looking at this all the bases are in series..
Let's call it a sequence of three common emitter amplifiers. No rocket science to it, you just get 3x the gain and 3x90° phase lag at high frequencies..
3x the gain in dB..
Below a sim for several batt voltages..
The voltages and currents in the schematics are for 1.5v batt (the higher the batt voltage the higher the gain).
Added 100uV/10kHz amp input..
-
#1 Are the transistors connected in what you would call a cascade? Apparently when googleing you dont series transistors but looking at this all the bases are in series..
#2 When using a transistor as an amplifier dont you need to bias the base with a voltage divider? How is t1 being turned on with virtually no current on it's base?
#3 Why does this circuit switch the type of transistor from 3904 for T1 to a BC847 for T2, and T3, looking at the data sheets there both ~100 hfe with a 300mhz speed, why not just use all 2n3904s?
#4 I was always under the impression the Collector current would be amplified to Beta X the Base current. How does the voltage get amplified in this circuit if transistor work like a current valve controlled by the base?
#5 What is going on between the base of T1 and the audio ouput jack? I dont understand the 100k 100k voltage divider with a 10uF to ground in between them.
#1 Don't confuse "cascade" with a "cascode amplifier"; they are very different beasts.
#2 It isn't on/off. The relationship between Ib and Vbe is exponential. See #5
#3 It probably would work. If the circuit description doesn't say why, then it is not a very good example to learn from.
#4 Vout=Ic*Rc. Hence a higher Rc means a higher voltage swing but clipping with a lower input
#5 They are negative DC feedback, which provides current to set the bias.
Overall this is a crap circuit since it has a poorly defined behaviour, but then it is a crap application so that doesn't matter.
My suggestion is to get a textbook that is designed to teach how BJT circuits work. Search terms: "common emitter/base/collector amplifier" and "small-signal" and "large-signal" operation.
-
Ok so yesterday I watched tons of different videos by different creators explaining transistors and the difference between common Emitter/Base/Collector circuits and how to bias them. Things are starting to get a bit more clear but I am still confused about somethings. I have been using a program called EveryCircuit to sim things out, it is probably time I learned to use LTSpice or MicroCap, I can not set any parameters like FT or Vbe etc.... here is a pic of the transistor settings.
[attachimg=1]
I'm not even sure what saturation current means it will only let me select units in pA, fA, and aA, I guess that is femto and pico im not sure what aA is. My guess was that was like max base current, but the fact there is no millaAmp doesn't make a lot of sense. This sim also has zero documentation.
Anyways I ran a sim of a basic transistor circuit where the only objective is to switch the emitter on and observe the voltage and current changes. The sim i ran makes sense for the most part except for the voltage on the base.
[attachimg=2]
If im pumping 5v in to the base across a 150 ohm resistor, my expectation is that I would subtract the transistor junction at .7V from 5V leaving me with 4.3V on the base and 28.6mA across R1. Why is the sim telling me the base voltage is 2.53V, which would set the base current to 16.4ma. Am I missing something here when solving ohms law?
-
Saturation current is a parameter of Ebers-Moll, which linearly scales Ic at any given Vbe.
1.59A flowing through 1Ω emitter resistance shifts the internal emitter potential up, hence the weird base voltage. I guess.
-
I think the simulation is wrong, that base voltage is strangely high, or the transistor model used is very small and we are not seeing the smoke it's dissipating over 5W lol. Ouch.
You could try lower currents by 10-20x and see what results but I also expect VBE around 0.7V for math.
Maybe try the Falstad circuit simulator (https://www.falstad.com/circuit/) to compare results. People are happy with it once the learning curve is met.
I think the OP Elektor circuit goes back to late 1960's TRF AM radio receivers, which then led into the ZN414/MK484/TA7642 IC (but it has extra gain and emitter-followers).
1966 pic from cool386.com ZN414 page. (https://cool386.com/zn414/zn414.html) I was experimenting with similar circuits for lightning detection.
-
I love the use of negative feedback just for bias voltage. You see a negative feedback path, think "oh this will reduce distor.... capacitor wot?".
Why is the sim telling me the base voltage is 2.53V, which would set the base current to 16.4ma. Am I missing something here when solving ohms law?
You're using the transistor well outside what its model is designed for. It can't handle several amps of collector current, IRL it would blow up instantly and release scented perfumes. The 1R resistor on the collector is way too small.
Ohm's law with a fixed resistance is a simplified model that generally only works with resistors. Transistors don't have a constant resistance, it changes depending on what's happening with the other terminals, and can even go negative under certain conditions. It's generally not worth using ohm's law on a whole transistor BE or CE junction, it's too complicated to make sense of.
I have been using a program called EveryCircuit to sim things out, it is probably time I learned to use LTSpice or MicroCap, I can not set any parameters like FT or Vbe etc.... here is a pic of the transistor settings.
You might be confusing parameters of the transistor model and parameters of how you use the transistor in your circuit.
A common NPN transistor being used as a common emitter amplifier (like we're doing in all the pictures so far) has a Vbe vs Ib curve that looks a bit like a diode. This means the most common Vbe you will see is about 0.6 to 0.7V. But you can make it other voltages too by changing how much current goes into it (via resistors), it just turns out it's not typically useful unless Vbe is about 0.6-0.7V. If you see a Vbe above this on an ordinary NPN then it's your circuit that's wrong, not the transistor model.
FT is a parameter of the transistor, yeah some sims don't let you change this directly (instead you have to edit other parameters).
-
FT is a parameter of the transistor, yeah some sims don't let you change this directly (instead you have to edit other parameters).
The quality of the models is paramount. You have to be aware of what the model does and doesn't model, and whether or not that is relevant to your circuit and simulation. The quality (or otherwise) of the models is a solid reason for choosing one simulator over another.
For example, fT also varies with Ic. Sometimes that matters.
(https://entertaininghacks.wordpress.com/wp-content/uploads/2024/03/ft-ic.jpg)
-
Well there is nothing in the transistor parameters about max voltage or current, I used 1ohm because it was easy to measure. There are power transistors out there that can pass an amp. I tried the sim again using a 100ohm at the collector and it tells me the point between the base and the 150ohm resistor is 942mv, and its barely doubling the current at the collector.
[attach=1]
-
What NPN transistor model is it using?
Beta/HFE/gain isn't a constant number (despite datasheets often proudly marketing it as so on the front page), it collapses when you use a transistor towards the top of its range. An HFE of 2 could be legit at these current levels. Change you base resistor to 1K or 10K and you will probably see higher gain.
-
I understand ft hfe and other parameters change based on gate current. Im trying to figure out why my base voltage keeps changing. To calculate base current i would use 5v - .7 to get 4.3 then use ohms law on chosen resistor to calculate the current im limiting the base too. Im trying to figure out why my base voltage is all over in these sims and why its not 4.3v. How can i even begin to try and figure out more complicated amplifiers if i cant figure out what voltage/current my base sees. Am doing something wrong or is this sim doing something out of the ordinary.. as i showed above there arent to many parameters i can feed it
-
A transistor is not producing energy. It can’t pull current higher than the circuit can supply. On the collector end there is a 100 Ω resistor. If the transistor was a literal 0 Ω switch, it could do not better than pulling 50 mA. The 48.9 mA you observe indicates the transistor is already at saturation, almost completely pulled to GND. It will not magically multiply the base current of 27 mA to become 2.7 A, if there is nowhere for this 2.7 A to come from.
0.6 V is the voltage at which the P-N junction between base and emitter “starts to conduct”. In quotation marks, because this is using a model as simplified as trying to see a diode completely blocking everything below 0.6 V and then being a perfect short after reaching 0.6 V. In practice the change is not abrupt. Typically NPN transistors are operated a bit above 0.6 V. But this shouldn’t be understood as they being completely off at 0.5 V. Neither that with higher currents they don’t go above 0.6 V.
If you look at a simplified diagram (https://commons.wikimedia.org/wiki/File:NPN_BJT_(Planar)_Cross-section.svg) for a BJT transistor, note that base-emitter current goes through the same area as collector-emitter current. While the ralation is not simple, that should already give you a hint, that with higher base current or collector current you must see higher base-emitter voltage.
In attachment there is an excerpt from OnSemi’s BC847 datasheet. See how much the voltage changes depending on what happens in the circuit. But this is only an example for a specific device. What you see in the simulation depends entirely on how they modelled the transistor and which parameters were chosen.
-
Im confused if we replaced the 100ohm with a 0ohm jumper, your saying the C couldnt pull more than 50ma? I dont understand why in this case. Im under the impression the collector current should be the base current multiplied by hFE or 100 in this case, and should output be ~2.7amps. There doesnt seem to be any maximum power limitations in the options for this transistor, a big old metal can to-3 could pass 2.7 amps. I know a 2n3904 cant, but literally all the options for this transistor look fairly simple I dont understand why any amount of current running through it would cause an issue. Which parameters in the menu picture I posted above dictate the collector can only supply 50ma? Im very confused about the actual parameters of this transistor apparently, i thought it was like a sim of some kind of ideal transistor or something with very limited parameters you can change.
If I put this same circuit in LT or microcap and used a tip120 for instance, would I pass 2.7 amps?
Im also most confused about why my base voltage is not just 5v-Base Junction drop, is the collector current somehow effecting what happens on the base?
Im sorry if im sounding dense im just getting frustrated and really determined to figure this out. My next steps are to either breadboard this or use a known transistor model in a better sim and read the BJT section of the art of electronics again! Not understanding how to use transistor to make a predictable amplifier has been something that has bugged me for a long time!
-
I understand ft hfe and other parameters change based on gate current. Im trying to figure out why my base voltage keeps changing. To calculate base current i would use 5v - .7 to get 4.3 then use ohms law on chosen resistor to calculate the current im limiting the base too. Im trying to figure out why my base voltage is all over in these sims and why its not 4.3v. How can i even begin to try and figure out more complicated amplifiers if i cant figure out what voltage/current my base sees. Am doing something wrong or is this sim doing something out of the ordinary.. as i showed above there arent to many parameters i can feed it
You are trying to bias a transistor. You're probably hoping for intuitive results, but unfortunately that's not how these circuits work. You need to copy an existing (good) biasing scheme's equations/approximations/rules-of-thumb and then tweak from there.
DON'T copy the bias circuit of the original circuit this forum topic starts with (resistor from VCC to base). It's garbage, it doesn't work 90% of the time, unless you also copy the clever feedback circuit they use (but it looks like you're not doing that).
To be more precise: biasing an NPN with a single resistor is very unstable. Only a tiny, narrow range of resistor values will work. Anything above or below these won't bias the transistor usefully. In real life you'd only be able to do this with trimpot, but even then a simple temperature change would ruin your bias levels (because transistor HFE depends on temp).
There are lots of methods of "biasing a common-emitter amplifier", all with different tradeoffs. Hackaday (https://hackaday.com/2018/05/04/biasing-that-transistor-part-1-the-common-emitter-amplifier/) has some examples, but I don't like the lack of emitter resistors.
Sidenote: Part of my favourite common-emitter-amplifier biasing method involves placing a resistor above the collector AND below the emitter. Now the gain is mostly controlled by the resistors (gain~=Rc/Re), NOT the hfe of the transistor, so temperature changes are no longer as evil. You still then have to bias the base (eg with a two-resistor divider) but it will now be easier to recreate the sim results in real life.
-
Well that just explained to me why most biasing schemes include that emitter resistor! Couldnt figure out for the life of me why youd have both a collector and emitter resistor if you didnt stricktly need them both!
-
Im also most confused about why my base voltage is not just 5v-Base Junction drop, is the collector current somehow effecting what happens on the base?
Yes, it is, because that's how you have set your model parameters:
1.59A flowing through 1Ω emitter resistance shifts the internal emitter potential up, hence the weird base voltage.
In the second simulation with 100Ω collector load you have forced the transistor into saturation (only 111mV collector voltage) so its β collapsed to almost nothing. What golden_labels meant is that even if you replace the transistor with a short, the 100Ω resistor still won't conduct more than 50mA and neither will the shorted transistor of course.
-
Ok well there is only like 6 options i can change for the transistor parameters. I guess im not sure how to "set the model". I thought that basically involved setting the parameters in the sim to those in the datasheet for whatever transistor you are using? The parameters this sim uses like base resistance and saturation current are not defined in the data sheets for a 2n2222 2n3904 or tip120. How can i tell what the max currents are for the transistor in the sim? That way i can keep my examples with in the limits, or how can i translate datasheet values to the list of parameters supported in the sim?
-
Ok well there is only like 6 options i can change for the transistor parameters. I guess im not sure how to "set the model". I thought that basically involved setting the parameters in the sim to those in the datasheet for whatever transistor you are using? The parameters this sim uses like base resistance and saturation current are not defined in the data sheets for a 2n2222 2n3904 or tip120. How can i tell what the max currents are for the transistor in the sim? That way i can keep my examples with in the limits, or how can i translate datasheet values to the list of parameters supported in the sim?
Find a textbook that explains/uses the Ebers-Moll BJT model.
Look at the V-I curves for a BJT in a datasheet. Older datasheets tend to have more explicit information.
Then, if you feel like it, use LTSpice (or similar) and the model of a defined BJT, e.g. 2N3904. Play around looking at various V-I and I-I curves.
-
Well that just explained to me why most biasing schemes include that emitter resistor! Couldnt figure out for the life of me why youd have both a collector and emitter resistor if you didnt stricktly need them both!
What Whales said is really important. Emitter resistors allow you to set gain according to controllable external parameters. The local negative feedback will also significantly reduce distortion. Very good to know.
Biasing schemes will be covered in most integrated/discrete analog analysis textbooks.
-
Ok well there is only like 6 options i can change for the transistor parameters. I guess im not sure how to "set the model". I thought that basically involved setting the parameters in the sim to those in the datasheet for whatever transistor you are using? The parameters this sim uses like base resistance and saturation current are not defined in the data sheets for a 2n2222 2n3904 or tip120. How can i tell what the max currents are for the transistor in the sim? That way i can keep my examples with in the limits, or how can i translate datasheet values to the list of parameters supported in the sim?
Six parameters is not a lot, typical SPICE implementations have a dozen if not more and they often come with a library of models of common transistors you can use right away.
For basic simulations of small signal circuits these things shouldn't matter, though. Typical base resistance is tens to hundreds ohms, emitter resistance a fraction of an ohm. They are too low to matter except in very precise logarithmic converters and power transistors conducting lots of current, as you saw yourself. Saturation current determines Vbe for any Ic, so for BC847 you could tweak it to get 660mV at 2mA as is the typical spec. Again, it rarely matters because you must be doing something unusual if you care strongly whether it's 600mV or 700mV. And a 50°C difference in operating temperature suffices to shift Vbe by 100mV.
-
rwgast_lowlevellogicdesin: user magic has already answered the question about the 100 Ω resistor, so I will skip that.
I have a felling, that you’re trying to understand transistor through equations that model it. To calculate behavior using maths you already know, and use that as the gateway to grasping transistors. I was through that too and so were many others. This is a dead end.
See some basic transistor circuits. Blindly believe they just work. Play with them, see how they behave as black boxes, get accustome to that behavior. At this point start asking questions, why they behave this way. If you are in a particular state, what happens if you shift voltage up or down? What happens if you increase or decrease a value of a given resistor? What happens, if — with rising temperature — some parameter changes? How, operating in the same conditions, another transistor behaves? Yes, this does require leaving the comfort zone. But it is much more straighforward and easier way of grasping the idea.
-
I have a felling, that you’re trying to understand transistor through equations that model it. To calculate behavior using maths you already know, and use that as the gateway to grasping transistors. I was through that too and so were many others. This is a dead end.
The OP is trying to understand based on his equations (and mental models) that are so grossly oversimplified that they are useless. As various people have put it "Everything should be made as simple as possible, but not simpler". (Various people include Einstein and Zukofsky, but it goes back to William of Ockham".
The OP needs to recognise that he needs to understand - to some level - the more complete mathemetical models of the physics.
-
Thanks to the OP and all who have responded. This has been a great thread to learn from. Reminds me of the "what you know, what you know you don't know and what you don't know you don't know" chart. :clap: