Is there any bare ready to use mosfet available to buy? Like the one used in the storage devices to store 1 bit of data.
No. You can "store" one bit in a flip-flop, but you'll loose the state if you power-off and the input (write) pin(s) are different from the output (read) pin(s).
A single transistor/MOSFET doesn't have any memory. And, if you think about it "philosophically" you need at least one data pin and one read/write pin.
The transistors used for storage in that way have the gate insulated, and use various techniques to inject charge into it to write it. For non-volatile storage, why not use an EEPROM, and get hundreds or thousands of bytes for less than the cost of a cup of coffee, and 2 wires to read/write? I mean, if you told us what you're trying to do, we could make some comments on what the best ways might be.
Sure, Mouser, Digikey, Jameco, etc. all supply MOSFETs -- ready to buy. But, since I don't know what you're referring to when you say, "the storage devices to store 1 bit of data", I'm probably not giving you the answer you're looking for.
You would probably do a better job of helping us help you, if you read: How to use this Forum
And on the subject of a One Bit Memory: You could put a capacitor across the Gate and Source, and charge the capacitor to some voltage that will turn the MOSFET "on" [depends on the MOSFET, and on what it will be driving}. That would constitute a 1 Bit memory. How long it will hold that state, is unknown. In fact, within a flexible bracket of time, one could merely charge the gate capacitance on a MOSFET, and call that a 1 Bit memory.
All depends on the point of all of this 
Sorry for not explaining it properly.
Actually I would like to create the transistor/mosfet based 1 bit non volatile memory(even if it stores data for 1 day is enough) I know I can use Relay or Ferrite bids or just a bare switch to do the same thing but I like to do it only using any type of transistor but it has to be transistor only it can have supporting elements like resistor or Cap to store the data.
After reading about FGMOS I was thinking that there may be already similar device available to do this.
Do you have an application for this odd storage device, or is it a 'can it be done out of curiosity'
project?
If the second you've had a few suggestions, and if the first I'd like to know what it is - there's probably a better way .....
Allan
allanhurst:
Do you have an application for this odd storage device, or is it a 'can it be done out of curiosity'
project?If the second you've had a few suggestions, and if the first I'd like to know what it is - there's probably a better way .....
Allan
Its a Curiosity only. 
bkrenderman:
...but I like to do it only using any type of transistor but it has to be transistor only it can have supporting elements like resistor or Cap to store the data.
It would help if you learned how to use punctuation. The following sentences are quite different:
- ...it has to be transistor only, it can have... means it must be done solely with a transistor, with no supporting components.
- ...it has to be transistor, only it can have... means it has to be a transistor, but it can have supporting components.
The way you wrote it, one is left to wonder if you meant no supporting components, but accidentally wrote "can" instead of "can't".
bkrenderman:
Is there any bare ready to use mosfet available to buy? Like the one used in the storage devices to store 1 bit of data.
No, you have to buy 256 million at a time!!
Seriously the circuitry to program EEPROM/Flash memory which I think is what you mean is built into every
EEPROM/Flash memory chip and generates high voltage rails on the chip to charge the floating gate by
quantum tunneling, and sequences the write and erase cycles for you.
My English is not that good! sorry for the pronounciation.
I know the EEPROM is already there but I dont want to use it, also I don't want any complex circuitry to write or read data from few Mosfets, but I want to build it with Discreet component and transistor of any type just to show the students and other interested people how the single transistor or few of them can store data even after it has been turned off.
by using EEPROM people assume it as a some kind of magical black box that does the job. when we can use transitor for every storage why cant we have single transistor to show how it store non volatile data the same way? why do we have to find alternatives like Relay etc. for demonstration.
again I am asking is there any FGMOS that exist?
bkrenderman:
My English is not that good! sorry for the pronounciation.I know the EEPROM is already there but I dont want to use it, also I don't want any complex circuitry to write or read data from few Mosfets, but I want to build it with Discreet component and transistor of any type just to show the students and other interested people how the single transistor or few of them can store data even after it has been turned off.
by using EEPROM people assume it as a some kind of magical black box that does the job. when we can use transitor for every storage why cant we have single transistor to show how it store non volatile data the same way? why do we have to find alternatives like Relay etc. for demonstration.
again I am asking is there any FGMOS that exist?
Well, it still isn't quite clear what you want. it sounds like you want a demonstration circuit that mimics the functionality of a what? EPROM? FPGA?
I know of no single bit EPROM or single gate FPGA. But, what would that achieve, anyway. It really wouldn't illustrate the functionally to an audience. But, maybe a MOSFET, with a some sort of setup to charge the Gate, or a capacitor across the Gate. And a power source, that drives an LED in series with the MOSFET's channel. Along with a mechanical switch that allows you to turn the power source off, then on, thus demonstrating the non-volatile nature of the device.
- You would charge the Gate capacitance, which would turn the LED on [via the switching action of the MOSFET].
- Then turn the power off, and then back on again to show that the LED is still lit.
- Then, discharge the Gate capacitance, which would turn the LED off.
- Then switch the power off, then back on, to show that the LED remain off, even with the power switched on. In both cases, the state of the LED would be preserved, thus demonstrating non-volatile storage.
Is that [or something like that], what you have in mind?
ReverseEMF:
Well, it still isn't quite clear what you want. it sounds like you want a demonstration circuit that mimics the functionality of a what? EPROM? FPGA?I know of no single bit EPROM or single gate FPGA. But, what would that achieve, anyway. It really wouldn't illustrate the functionally to an audience. But, maybe a MOSFET, with a some sort of setup to charge the Gate, or a capacitor across the Gate. And a power source, that drives an LED in series with the MOSFET's channel. Along with a mechanical switch that allows you to turn the power source off, then on, thus demonstrating the non-volatile nature of the device.
- You would charge the Gate capacitance, which would turn the LED on [via the switching action of the MOSFET].
- Then turn the power off, and then back on again to show that the LED is still lit.
- Then, discharge the Gate capacitance, which would turn the LED off.
- Then switch the power off, then back on, to show that the LED remain off, even with the power switched on. In both cases, the state of the LED would be preserved, thus demonstrating non-volatile storage.
Is that [or something like that], what you have in mind?
How long the gate capacitance will hold the charge approximately? and how can i keep it safe from emf noise.
bkrenderman:
How long the gate capacitance will hold the charge approximately?
That is difficult to determine, because it's temperature dependant, and because it involves leakage current of reversed biased protection diode(s), those currents can vary from device to device [i.e. it's not a well defined parameter].
I took a look at a few MOSFET datasheets and discovered that a typical value for Gate-Source leakage current is ±10µA!! I did find one [2N7000] with a leakage current of -10nA. But, even at that lower current level, the timing is pretty bad.
Using the absolute value of 10nA, and assuming a capacitance of 1µF, and a voltage sag of 1V, and using the following formula to get a ballpark value:
T = CV/I = 1µF(1V)/10nA = 100 seconds!
That can be extended several ways. Say we use a MOSFET with a max Gate voltage of 20, and with a low Gate threshold, such that, say, the gate voltage can drop to 3V before the MOSFET turns off, and charge the capacitance to 15V, and increase the capacitance to 10uF [making sure it's a low leakage capacitor]:
T = CV/I = 10µF(15V - 3V)/10nA = 12,000 seconds = 200 minutes
So, not a very good non-volatile memory. You could increase the capacitance even more, but you'll start running into leakage current in the capacitor that will, likely, work against the goal.
The thing an EPROM or FPGA has going for itself is no need for protection diodes, thus MUCH smaller leakage currents, on the order of an insulator!! That's why they can work, whereas using a MOSFET just isn't the same!
Boy did I learn something!!
One possibility, though: with lower voltages on the Gate, that leakage current may be quite a bit less, so using a low threshold MOSFET with low voltages on the Gate, might make enough difference for this to at least serve as a teaching device. Only experimentation will tell.
bkrenderman:
...and how can i keep it safe from emf noise.
And, when you say "Keeping it safe from EMF", I think you mean ESD [ElectroStatic Discharge]. And, that's the rub -- the built in static discharge protection is the very thing that makes this not work so well. Also, adding a large capacitance across the Gate-Source will also protect it from ESD.
ReverseEMF:
That is difficult to determine, because it's temperature dependant, and because it involves leakage current of reversed biased protection diode(s)
MOSFETs don't have protection diodes as there are no defined power rail voltages to protect to. Some
have gate-source protection zeners however.
MarkT:
MOSFETs don't have protection diodes as there are no defined power rail voltages to protect to. Some
have gate-source protection zeners however.
Isn't a zener a diode?
ReverseEMF:
Isn't a zener a diode?
"Protection diode" refers to a diode to ground or supply from a pin. These are almost ubiquitous on microcontrollers (3.3v microcntrollers with "5v tolerant" pins, of course, don't have them on the positive side), and protect pins by preventing the voltage on the pin from exceeding supply or going below ground. While a zener diode on the gate of a MOSFET serves a similar purpose, it is used in reverse biased configuration, so there is a much higher voltage drop (somewhere around or above the maximum Vgs spec). But these will pose a similar problem, since they have non-zero leakage too.
The key thing that makes flash memory work as well as it does is that the floating gate is totally isolated, and tiny (so it won't act like an antenna). The charge is injected into it via tunneling with significantly higher voltage than is used for reading. A discrete transistor, on the other hand, has an antenna (the pin) connected to the gate and a diode (with the leakage that entails) connected to it.
I don't think you can make truly non-volatile memory with discrete components... At least not ones that exist - I'm sure someone could put a flash-style floating gate transistor into a package, but I can't imagine anyone doing so. (the floating gate in a modern flash cell is also not quite what you think; there is a floating gate, but it's used a bit differently - have you read the wikipedia article on flash memory)
Your best bet is to use old CD4000UBE series chips (unbuffered), perhaps a teflon pcb and some
ultra low leakage conformal coating. Some of these old CMOS devices had very low gate leakage
with no protection diodes I think. And of course they were extremely static sensitive.
Sorry for resurrecting an old post, but I was looking for an answer to the exact same question, and noticed that no-one had addressed the OPs (and my) question, so deemed it prudent to reply once I'd found the answer, for anyone looking for such info in the future.
The question: do any discrete FGMOS (Floating Gate MOSFET) transistors exist that I could experiment with?
The answer: I'm not sure about single discrete devices, but there exist DIP and SOIC packages containing two or four individually accessible FGMOS.
For example, the ALD1108E, or any of the devices listed at ALD EPAD High Precision MOSFET Arrays
Whilst this company calls them something else, a read of the Application Notes file at the top confirms that they are indeed Floating Gate MOSFETs.
As the device numbers are listed, you could then just do a search for them at on the web or at some of the usual suppliers (mouser, Digi-Key, etc)
And now for a little rant:
Yes, this is only my first ever post here, but I have to say, I was motivated to sign up to respond here when it became clear that people (who I'm sure thought they were being helpful) decided that because they didn't know why this question was being asked, that they knew better than the person who asked it, and that also meant they didn't need to read the OPs questions and posts properly.
Floating Gate/FGMOS was mentioned in both the subject line, and explicitly in a post, yet we get replies such as
ReverseEMF:
But, since I don't know what you're referring to when you say, "the storage devices to store 1 bit of data", I'm probably not giving you the answer you're looking for.You would probably do a better job of helping us help you, if you read: How to use this Forum
And another one later on saying that the misunderstanding was because of the OPs lack of grammar skills!
I'm not trying to single out anyone here, almost all replies questioned why this person would want to do such a thing and therefore didn't answer the question.
So just a friendly reminder, (yes, from a noob), to take questions at face value of you can. If you don't know why someone is asking something, re-read the question before answering. Sometimes our desire to be "helpful" can leave pretty basic questions unanswered. If the topic and question were re-read, without judging the OPs intentions, it would be obvious what was being asked. Even if you disagree with the premise for the question answer it anyway. Do this before questioning the OPs intentions.
2 Likes
Taking questions blindly at face value is how the xy-problem exists. xyproblem.info
If storing charge only for a day, consider a low leakage capacitor. Likely film. You will have to look at dielectric absorption when you choose the capacitor type. Also however you "set" it and "read" it will have to be very high impedance when there is no power.
At an extreme, look at supercaps they can hold a charge for a significant period of time. Not sure how one would charge / discharge it quickly.
John