Following tutorial, got a working pre-amp but don't fully understand how to get the proper values

Thread Starter

Kain6622

Joined Apr 10, 2021
8
Hello everyone, First time post of the forum, so here goes...

I'm currently trying to follow the Mosfet amplifier tutorial on this site, I understand parts of it and have not clue about others. I have made a working circuit but with a lot of estimations and running it through ltspice first:

1624897512923.png

Where I am failing to understand is how to calculate the 'K' value to help calculate the voltages/ resistor values etc, I've searched over google and through forums but there is no answer that really helps:

K=12μCoxWL is a formula that comes up quite often but no explanation on how to use it.
even in the tutorial that I am following, it makes the assumption that the K value is already known, making it near impossible to calculate the values for Vgs.

The assumptions I made above where in the understanding that Vgs >Vth (for the mosfet i'm using its between 1-2.5vth IRLB3034), Vg > Vgs, in the tutorial they chosen Vg = 1/3Vdd. I'm trying to use a 9v supply so chose Vg = 3v, knowing Vgs < vg, set it too 2.5v (the max threshold voltage), meaning Vs is 0.5v, this allowing me to get the values in the circuit above... and though it works on these assumptions, I don't believe I have it working at it's optimum due to lack of understanding of this tutorial, I've build the circuit and it works as a preamp but I'm hoping to try improve it but can't find the resources to help me understand the biasing in saturation etc better.

Does anyone have any advice on this/ know of good material for a hobbyist of electronics to read that might help me solidify my knowledge to be able to get the maximum performance out of this circuit.

Thanks
 

Papabravo

Joined Feb 24, 2006
17,318
It is a value that is used in the design and fabrication of silicon chips. It has absolutely no bearing on circuits that you can build. You would use it in simulating a device to be made in a silicon foundry. AFAIK there are very few individuals with a personal silicon foundry, which is not surprising considering the expense.

Instead of showing us a solution you don't understand, why not tell us what problem you are trying to solve and ask how we might go about it.

One more thing: why is there a missing component to the right of capacitor C3?
 

Thread Starter

Kain6622

Joined Apr 10, 2021
8
Thanks for the reply Papabravo,

Ok, I've re-read the tutorial to make sure I know what I'm unsure of so here goes...

In the tutorial, it says that the drain current can be obtained through Id = K(Vgs - Vth)^2, but nowhere in the tutorial does it help one obtain the values of K or the value of Vgs without the other. In their example they pre-define the value of K, which is fine for a text book question but in real life circuit design how does one go about obtaining this value? They late say that Vgs is obtained via Sqr(Id/K)+Vth, again the mystery K value. From what I understand about the K value is that it changes FET to FET even on the same part number (A bit like the Hfe of a bjt transistor?).

I've looked online and there are some references to the Vgs to Id graph to extract the value and some that use the gm value (which this tutorial does show how to get at the start) but exactly how to obtain the K value from the data sheet is still unclear (since if this method is true, the graphs are generally exponential), What i'm needing help with here is to understand is the best method of biasing a E-NMOS to be used as an amplifier, and this tutorial leave the questions:

1: How to extract the K value from the datasheet (if it's needed)
2: An alternative method if the K value is not needed that can be applied to the mosfet of our choosing for biasing for amplification
3: Am I missing something completely here and just making problems for myself?
4: They use in the example Vg = 1/3Vdd, they are using 15vdd for making Vg = 5v, which is above Vth so all is good but if we change the Vdd to a lower value then it becomes a problem, in my case 1/3(9v) = 3 (very close to Vth), is the choice of 1/3 just design choice? and if so what rule of thumb would they go by to make sure that it's value is not too low or too high?

I'm slightly worried that I might be over complicating this problem and missing something obvious completely, and hoping that someone here might be able to help shed some light on the matter so that I can properly bias a NMOS Pre-amp for any mosfet I may choose (without having to test all the colors of resistor rainbow to see what works) and be sure that I am getting the best or at least as near best output from it.

I hope this explains what i'm confused about and hope someone might be able to enlighten me with some new knowledge in this area.

Kind regards

(With regards to the missing component next to C3, it was a resistor put there to emulate the impedance of the cable, the design was crudely done to emulate a guitar input following this post (I only skimmed over this post as I'm more concerned with understanding the amplifier problem first)).
 
Last edited:

BobTPH

Joined Jun 5, 2013
4,056
Are you trying to write a spice model for a MOSFET? Normally you would use one provided by the manufacturer.

Bob
 

Thread Starter

Kain6622

Joined Apr 10, 2021
8
Are you trying to write a spice model for a MOSFET? Normally you would use one provided by the manufacturer.

Bob
I did get the model for the mosfet from the manufacturers site, but was in .spi format, so to use it in ltspice, I used the spice directive to include it in the simulation (was the best method I could find to get the model in the program), it will look different as I removed the comments in the file to make the image smaller.
 

Papabravo

Joined Feb 24, 2006
17,318
Thanks for the reply Papabravo,

Ok, I've re-read the tutorial to make sure I know what I'm unsure of so here goes...

In the tutorial, it says that the drain current can be obtained through Id = K(Vgs - Vth)^2, but nowhere in the tutorial does it help one obtain the values of K or the value of Vgs without the other. In their example they pre-define the value of K, which is fine for a text book question but in real life circuit design how does one go about obtaining this value? They late say that Vgs is obtained via Sqr(Id/K)+Vth, again the mystery K value. From what I understand about the K value is that it changes FET to FET even on the same part number (A bit like the Hfe of a bjt transistor?).

I've looked online and there are some references to the Vgs to Id graph to extract the value and some that use the gm value (which this tutorial does show how to get at the start) but exactly how to obtain the K value from the data sheet is still unclear (since if this method is true, the graphs are generally exponential), What i'm needing help with here is to understand is the best method of biasing a E-NMOS to be used as an amplifier, and this tutorial leave the questions:

1: How to extract the K value from the datasheet (if it's needed)
2: An alternative method if the K value is not needed that can be applied to the mosfet of our choosing for biasing for amplification
3: Am I missing something completely here and just making problems for myself?
4: They use in the example Vg = 1/3Vdd, they are using 15vdd for making Vg = 5v, which is above Vth so all is good but if we change the Vdd to a lower value then it becomes a problem, in my case 1/3(9v) = 3 (very close to Vth), is the choice of 1/3 just design choice? and if so what rule of thumb would they go by to make sure that it's value is not too low or too high?

I'm slightly worried that I might be over complicating this problem and missing something obvious completely, and hoping that someone here might be able to help shed some light on the matter so that I can properly bias a NMOS Pre-amp for any mosfet I may choose (without having to test all the colors of resistor rainbow to see what works) and be sure that I am getting the best or at least as near best output from it.

I hope this explains what i'm confused about and hope someone might be able to enlighten me with some new knowledge in this area.

Kind regards

(With regards to the missing component next to C3, it was a resistor put there to emulate the impedance of the cable, the design was crudely done to emulate a guitar input following this post (I only skimmed over this post as I'm more concerned with understanding the amplifier problem first)).
This is a bit more context. In the original post you had K as a function of the constant 12μ, and Cox which is a capacitance associated with the chip fabrication geometry times W (width) times L (length). W & L are measurements associated with the particular MOSFET process. You may have assumed that these two K's are the same. That may or may not be the case. A simple dimensional analysis could show that they are not if the units don't match, but would not be conclusive if they did.
 

Thread Starter

Kain6622

Joined Apr 10, 2021
8
This is a bit more context. In the original post you had K as a function of the constant 12μ, and Cox which is a capacitance associated with the chip fabrication geometry times W (width) times L (length). W & L are measurements associated with the particular MOSFET process. You may have assumed that these two K's are the same. That may or may not be the case. A simple dimensional analysis could show that they are not if the units don't match, but would not be conclusive if they did.
Yes, in the first post I was partly quoting from post of similar people asking the same question, where one answer mentioned the function constant, which has confused me further, Looking further, some sites seem to infer that there is a relationship between Gm and K, see here. I need to sit down and read through this properly but again getting confused with in the link i provided in this reply of the selected values for their examples.... they say 'Assume Vgs(off) = ##, and Idss = ##' and work through the example, are these values that we can select ourselves or is it something that one needs to extract from the datasheet for proper operation? mind I have never seen vgs(off) in the data sheet so I assume that means that Vgs < the min Vth value at least, if not <zero.

I did find an alternative method that auto biases the circuit but I'm not a fan of it as I'm trying to understand how it works, so i can design/operate the circuit myself, (it make me thing of an AI driven car, with a person in it not able to drive, if the AI breaks or needs to pass control to the person, the person has no knowledge to control the car).

Am I making things for complex that they need to be? the tutorial in this post infers that JFET, DE & EH MOSFET amplifiers are pretty much designed the same with a few minor differences, the transconductance and how it is determined and also the differing biasing requirements. I think i'm heading in the right direction but any additional hints would be greatly appreciated.
 

Thread Starter

Kain6622

Joined Apr 10, 2021
8
The Link I provided in the previous link was very helpful and is actually part of a whole free book called "Semiconductor devices: theory and application" which seem to be very informative.

Though I still had problems with the best way to plot the Vgs vs Id graph, Over complicating things as I have been for this whole problem I eventually found this video, which does into testing the transfer characteristics of a mosfet and also videos on how to perform this in ltspice also which allows you to then perform the equations to get the values you need (The main issue I found for myself was not using an appropriate Mosfet for the task... with a continuous drain current of 195A doesn't make the Vgs vs Id graph look very nice when you only need an Id of about 10mA but it's all learning). Sometimes the obvious is not so obvious, so hope this will help anyone that runs into the same issue as I did.

Kind Regards
 

Papabravo

Joined Feb 24, 2006
17,318
The Link I provided in the previous link was very helpful and is actually part of a whole free book called "Semiconductor devices: theory and application" which seem to be very informative.

Though I still had problems with the best way to plot the Vgs vs Id graph, Over complicating things as I have been for this whole problem I eventually found this video, which does into testing the transfer characteristics of a mosfet and also videos on how to perform this in ltspice also which allows you to then perform the equations to get the values you need (The main issue I found for myself was not using an appropriate Mosfet for the task... with a continuous drain current of 195A doesn't make the Vgs vs Id graph look very nice when you only need an Id of about 10mA but it's all learning). Sometimes the obvious is not so obvious, so hope this will help anyone that runs into the same issue as I did.

Kind Regards
Gathering information from divers sources can lead to confusion. I applaud your efforts to get to the underlying explanations for the things you are seeing. When I was an undergraduate I struggled with a number of concepts at the intersection of physics and mathematics until a professor, teaching a required course in a subject in which I had scan interest, introduced me to dimensional analysis. That is analyzing the units of each term and factor on both sides of an equation. Whatever else is going on in the physics or the mathematics, the units on both sides of any equation must be the same. This is a subject which is worthy of some attention as you try to navigate the wealth of information from different points of view.

I'm actually proud of your discovery that the constant 'K' and the transconductance, 'gm' could potentially be related. I though about mentioning it, but I wanted to avoid throwing in a confusion factor by introducing a "new" concept. Keep doing what you are doing, because it seems to be working.
 
Last edited:
Top