Author Topic: Transistor, i hate you!  (Read 2794 times)

0 Members and 1 Guest are viewing this topic.

Offline phizaTopic starter

  • Contributor
  • Posts: 13
  • Country: it
Transistor, i hate you!
« on: October 11, 2019, 12:01:45 pm »
Finally i registered on this forum!

Bonjour,

i'm nerd, a passionate student of ee and i quit job two months ago to close my self in my bedroom/lab trying to make something beautiful. (I know it sound a bit crazy but it's true and i'm start to think that maybe  i've made a mistake :-DD)

I spent some money on RS, Mouser, TME, Amazon and eBay to get parts and tools that i needed for this specific project.

I hate transitors, i think that i've never understand 100% how they work. I mean i can use them for simple task, sometimes they are essential. But generally speaking if i can choose between two different design and one involves transistors the other no, i choose the second one. Even if that choice change radically the design i.e introducing an ADC, i'm ok with that, i mean it's my comfort zone.

First of all no beta matching problems and no need of temperature compensation.

Well but that isn't the biggest problem by my point of view...i mean, for example, it's about two week that i try to use the Ebers-Moll equation to get, with good approximation, an exponential response of the collector current Ic by varying the Vbe in certain circumstances. It's easier to me to get an exponential response of the Ib (i use only one transistor, regardless of temperature).

Sometimes it work, sometimes don't (even on Multisim or LTspice). I can't write a standard project method. Seems that to achive a good design with transistors you have to try all possible combinations and finally get a special secret formula and it look like a Chili recipe.

What do you think about it?

cheers   :D

some links:
https://schmitzbits.de/expo_tutorial/index.html
https://www.allaboutcircuits.com/projects/diy-synth-series-vco/
https://northcoastsynthesis.com/news/exponential-converters-and-how-they-work/
https://xonik.github.io/theory/vco/expo_converter_1.html
"Resist the temptation to take refuge in equations as a substitute for understanding how a circuit really works." The Art of Electronics
 

Offline Zero999

  • Super Contributor
  • ***
  • Posts: 20363
  • Country: gb
  • 0999
Re: Transistor, i hate you!
« Reply #1 on: October 11, 2019, 12:26:03 pm »
Hello and welcome to the forum.

Which type of transistor are you struggling with? There are two main types: field effect and bipolar? I assume you're talking about the latter, as you've mentioned beta and the Ebers-Moll equation.

Please post some example circuits which you've struggled to understand.

What job did you quit? How are you paying the bills now? Savings? Living with your parents? I think you've made a mistake if you're planning to make a living by building something in your bedroom and selling it.

Do you have any formal qualifications in electronic engineering? You'd be better off studying, than building something in the lab and not working./
 
The following users thanked this post: phiza

Offline phizaTopic starter

  • Contributor
  • Posts: 13
  • Country: it
Re: Transistor, i hate you!
« Reply #2 on: October 11, 2019, 12:51:59 pm »
Hi

many thanks for the quick reply

I'm using BJT (2N3904, BC547 mostly)

I'm a part-time student in EE at the University of Pisa and have a high school diploma in Electronics and Telcom.

Savings mostly and small jobs for the neighbors, i was a PLC developer always away on FCA Plant. Anyway i don't like it, i was also a IoT Hardware developer for a small company in Rome, but they paid me less. That's really was a mistake, quit a nice job for money.

Thanks.

PS:
the second transistor is not correctly polarized.
« Last Edit: October 11, 2019, 01:08:02 pm by phiza »
"Resist the temptation to take refuge in equations as a substitute for understanding how a circuit really works." The Art of Electronics
 

Offline Dave

  • Super Contributor
  • ***
  • Posts: 1356
  • Country: si
  • I like to measure things.
Re: Transistor, i hate you!
« Reply #3 on: October 11, 2019, 01:51:14 pm »
Why do you assume the second transistor is not correctly polarized? :-//
It has enough base current to go into saturation, everything checks out.
<fellbuendel> it's arduino, you're not supposed to know anything about what you're doing
<fellbuendel> if you knew, you wouldn't be using it
 

Offline phizaTopic starter

  • Contributor
  • Posts: 13
  • Country: it
Re: Transistor, i hate you!
« Reply #4 on: October 11, 2019, 02:31:27 pm »
Because second transistor Vbc is >0, to be in forward-active region shouldn't be <0 ? (and Vbe>0)

[Please don't kill me :D]
"Resist the temptation to take refuge in equations as a substitute for understanding how a circuit really works." The Art of Electronics
 

Offline atmfjstc

  • Regular Contributor
  • *
  • Posts: 121
  • Country: ro
Re: Transistor, i hate you!
« Reply #5 on: October 11, 2019, 02:46:36 pm »
i'm nerd, a passionate student of ee and i quit job two months ago to close my self in my bedroom/lab trying to make something beautiful. (I know it sound a bit crazy but it's true and i'm start to think that maybe  i've made a mistake :-DD)

I for one am not gonna mince words, it IS crazy and you HAVE made a mistake :) While we are all familiar with stories about a bunch of guys building the next big thing in their garage (e.g. Steve & Woz), what's less often mentioned is that such people a) were already at expert level when they started and b) had loads of connections in the industry. Just look at Dave himself, he has his reasonably successful multimeter project, but that's after decades in the industry and direct access to legendary experts (just look who he's had on his show).

Nothing wrong with learning about electronics in your spare time, but you should get a job to sustain yourself in the meantime. Trust me when I say it's going to take you many years until you can make money off of your own stuff. Also be aware that a major recession is imminent, and you'd best get your foot in the door soon, before companies panic even more and freeze hiring altogether.
 
The following users thanked this post: phiza

Offline rstofer

  • Super Contributor
  • ***
  • Posts: 9964
  • Country: us
Re: Transistor, i hate you!
« Reply #6 on: October 11, 2019, 03:28:11 pm »
You get exactly one shot at living your life - do what you want!

If you think about a saturated NPN transistor, the collector-emitter voltage will be around 0.2V and the base-emitter voltage will be around 0.7V so the base is always higher than the collector in a saturated common-emitter circuit.  I wouldn't spend too much time worrying about it.

What I would do is take your data and stuff it into a spreadsheet.  It would be interesting to plot Vbe vs Ic.  I assume Beta was a given in your circuits but I don't know that.  If it wasn't a constant, plot it!  In fact, duplicate as many of the datasheet graphs as possible.

It is possible to design around variations in Beta by a process known as emitter degeneration.  Put a resistor in the emitter connection and the overall gain of the circuit becomes essentially Rc / Re

https://electronics.stackexchange.com/questions/451441/transistor-design-with-beta-variation
https://www.electronics-tutorials.com/amplifiers/emitter-degeneration.htm

Clearly, transistors work!  We've been using them since around 1953 and somehow have managed to work around the issues.

Just Google for 'emitter degeneration'.

Watch w2aew's video on the subject:



« Last Edit: October 11, 2019, 03:30:43 pm by rstofer »
 
The following users thanked this post: phiza

Offline Rigolon

  • Regular Contributor
  • *
  • Posts: 70
  • Country: 00
Re: Transistor, i hate you!
« Reply #7 on: October 11, 2019, 05:06:49 pm »
Because second transistor Vbc is >0, to be in forward-active region shouldn't be <0 ? (and Vbe>0)

[Please don't kill me :D]

Don't worry, no one will kill you  ;D

Ok, I'm not that knowledgeable as many in here so I might be wrong. But here I go  ;D

Your assumption with Vbc having to be <0 is wrong.
You have to see what is the value of Vbe and Vce(sat) for the transistor you are using. For the 2n3904, Vbe = 0.65V and Vce(sat) = 0.2V
As long your Vce > Vce(sat) you are in the active region (Of course you could also be in the cutoff region but that's not the point here)

So let's use Vce = 0.5V
With Vbe = 0.65 we have Vbc = Vbe - Vce = 0.15V

As long your Vce > Vce (sat) you are in the active region or cutoff region.
To use as a equation like you intended, I guess you could say:

Vbc < Vbe - Vce(sat) is on active region. Can't put values because Vbe and Vce(sat) depends on the transistor.

 
The following users thanked this post: phiza

Offline virtualparticles

  • Regular Contributor
  • *
  • Posts: 149
  • Country: us
Re: Transistor, i hate you!
« Reply #8 on: October 11, 2019, 05:34:40 pm »
If you are trying to characterize collector current vs base to emitter voltage then you're in for some real sport. A transistor is much better characterized by collector current vs base current.  Within reason, you should see a fairly linear relationship there and you can use LTSpice to verify it. You want to use FETs if you want sort of linear drain current vs gate voltage.
 

Offline phizaTopic starter

  • Contributor
  • Posts: 13
  • Country: it
Re: Transistor, i hate you!
« Reply #9 on: October 11, 2019, 06:50:52 pm »
The circuit i try to design is an exponential converter, so i don't care a lot about linearity. What i'm trying to achieve is a circuit that do the following thing: if i change Vbe of 17.5mV (this is Vt*ln(2)) then Ic doubles. I don't have other requirements, even if it can be achieved with a different value of delta-Vbe. Even in a small range. Then i rearrange the value with an op-amp. It seems possible according to the model  |O
« Last Edit: October 11, 2019, 08:10:35 pm by phiza »
"Resist the temptation to take refuge in equations as a substitute for understanding how a circuit really works." The Art of Electronics
 

Offline rstofer

  • Super Contributor
  • ***
  • Posts: 9964
  • Country: us
Re: Transistor, i hate you!
« Reply #10 on: October 11, 2019, 07:01:07 pm »
 

Offline phizaTopic starter

  • Contributor
  • Posts: 13
  • Country: it
Re: Transistor, i hate you!
« Reply #11 on: October 11, 2019, 08:36:06 pm »
nostalgia circuits

You got it right  :-+

Well i can actually use an anti-log amplifier or any ic with an exp converter but i think that it's possible do it with only one bjt. (or I have to understand why it is not possible, careless of temperature)
"Resist the temptation to take refuge in equations as a substitute for understanding how a circuit really works." The Art of Electronics
 

Offline ratatax

  • Regular Contributor
  • *
  • Posts: 134
  • Country: fr
Re: Transistor, i hate you!
« Reply #12 on: October 11, 2019, 08:52:14 pm »
I also quit my job to do something crazy, I think it's totally worth if you focus on your goals and if you can find a new job easily in case it goes wrong.
You'll learn so many things, it's unvaluable. Just create some product that solves a specific problem better than someone else does, and you may (*MAY*) succeed
 
The following users thanked this post: phiza

Offline phizaTopic starter

  • Contributor
  • Posts: 13
  • Country: it
Re: Transistor, i hate you!
« Reply #13 on: October 12, 2019, 02:03:12 pm »
It would be interesting to plot Vbe vs Ic.  I assume Beta was a given in your circuits but I don't know that.  If it wasn't a constant, plot it!  In fact, duplicate as many of the datasheet graphs as possible.

So I take your advice and plot the "familiar" graph of Vbe vs Ic then i change the scale from lin to log and it's interesting notice how the "exp conversion" change, careless of Rc. Between 510 mV and 646mV Ic doubles every 27mV pretty well, after seems to be a bit slower.

But why 27mV? I set the simulation temperature to 20°C...  I think it's the 2N3904 model parameters on multisim, I'll try to check it.

Thanks everyone :)
« Last Edit: October 12, 2019, 02:59:05 pm by phiza »
"Resist the temptation to take refuge in equations as a substitute for understanding how a circuit really works." The Art of Electronics
 

Offline rstofer

  • Super Contributor
  • ***
  • Posts: 9964
  • Country: us
Re: Transistor, i hate you!
« Reply #14 on: October 12, 2019, 03:24:06 pm »
It would be interesting to plot Vbe vs Ic.  I assume Beta was a given in your circuits but I don't know that.  If it wasn't a constant, plot it!  In fact, duplicate as many of the datasheet graphs as possible.

So I take your advice and plot the "familiar" graph of Vbe vs Ic then i change the scale from lin to log and it's interesting notice how the "exp conversion" change, careless of Rc. Between 510 mV and 646mV Ic doubles every 27mV pretty well, after seems to be a bit slower.

But why 27mV? I set the simulation temperature to 20°C...  I think it's the 2N3904 model parameters on multisim, I'll try to check it.

The article I linked above shows the doubling every 18 mV.  Different devices?  Clearly different models.

Those graphs are very interesting.  I didn't expect such a straight line in the log graph.
 

Offline rfeecs

  • Frequent Contributor
  • **
  • Posts: 807
  • Country: us
Re: Transistor, i hate you!
« Reply #15 on: October 12, 2019, 06:02:42 pm »

But why 27mV? I set the simulation temperature to 20°C...  I think it's the 2N3904 model parameters on multisim, I'll try to check it.


The Ebers-Moll model equation on the pages you referenced is not quite right.  The more accurate equation for collector current similar to the Shockley diode equation:



Notice it has an n in the exponent.  That's the ideality factor.  It varies between 1 and 2 depending on the material and even depending on the current.

For the SPICE model, I think this parameter is Ne or NF.  Note the default LTSPICE BJT model value for Ne is 1.5.  This gives a change of 27mV per octave.

Look up "Gummel Plot".

« Last Edit: October 12, 2019, 06:11:28 pm by rfeecs »
 
The following users thanked this post: phiza

Offline phizaTopic starter

  • Contributor
  • Posts: 13
  • Country: it
Re: Transistor, i hate you!
« Reply #16 on: October 12, 2019, 08:24:05 pm »

Look up "Gummel Plot".


yep! This is the right name for that thing.

I investigated the model a bit and there are two more options about temperature, i set everything to 20°C.


For the SPICE model, I think this parameter is Ne or NF.  Note the default LTSPICE BJT model value for Ne is 1.5.  This gives a change of 27mV per octave.

I also found NF that is the "forward current emission coefficient". I set it to 1 and now Ic doubles every 17.5mV at lower voltages.
Looking at the formula we see that a lower value of "NF" or "n" increases Ic for the same amount of Vbe.

Maybe with some math it's possible to get the best value for the scope, i mean some kind of approximation about the optimal ranges of Vbe or things like that.


Again thanks, i go back to study now, turning off simulations for a while.  ^-^
« Last Edit: October 12, 2019, 09:33:53 pm by phiza »
"Resist the temptation to take refuge in equations as a substitute for understanding how a circuit really works." The Art of Electronics
 

Offline Jan Audio

  • Frequent Contributor
  • **
  • Posts: 820
  • Country: nl
Re: Transistor, i hate you!
« Reply #17 on: October 13, 2019, 01:48:17 pm »
You hate the software, not the transistors, lets not kill him.
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf