scott wurcer said:
No problem Scott. If I came across as abrasive or obnoxious, I apologize and assure you and all that it was unintentional.
I wasn't trying to go off into the deep end of things, but we all got sidetracked. I agree with you that quantum physics is nice to know, but it isn't mandatory when developing useful circuits at the mAcro level. Some did, however, claim that the OEM (TI, Fairchild, On Semi, etc.) got it wrong at the mIcro level and that atomic physics does not support the CCCS view. So I delved into the atomic end of it and got carried away.
For the record, I agree with you that the 3 terminal black box, or mAcro viewpoint works very well. We are intersted in using devices to achieve functional performance that benefits mankind. Most real world electronic products don't require that we go down to the QM (quantum mechanics) level of understanding. Again, forgive me for getting a little carried away.
For the record, I have a patent on a transimpedance logarithmic photodiode amplifier topology, U.S. no, 5,670,775, filed 1995, awarded 1997. If you get a chance, I realize that you're busy, let me know what you think. Others are also welcome to comment. It utilizes a p-n junction diode as the logging element. I get better results. A bjt is not as good a logging element as a diode.
I spent years doing research & development on logging amps because I was the hardware dude for 11 yrs. for a company that made money machines. These are the currency validators found on vending machines that accept your dollar bill, scan and verify it, and issue credit enabling you to buy a product from the vending machine. I did a lot of optical recognition using linear and log amps. I'm very familiar with them. Thanks for posting.
Best regards.
Claude Abraham said:
This topic seem to be really important to you
Why don't you post your equations on a physics specialized forum? They'll address your equations and understanding of stationary, time-invariant, quasi-static, electrostatic, electrodynamic, etc... better than anybody here (me included) is able, cares or bothers to. Or even better, why don't you publish these in a peer reviewed journal?
syn08 said:
This topic seem to be really important to you
Why don't you post your equations on a physics specialized forum? They'll address your equations and understanding of stationary, time-invariant, quasi-static, electrostatic, electrodynamic, etc... better than anybody here (me included) is able, cares or bothers to. Or even better, why don't you publish these in a peer reviewed journal?
These concepts have already been published in peer-reviewed journals & textbooks. What I've put forth is common knowledge. Ask anybody in the EE or physics community who works extensively with e & m fields and you'll find that they have similar views. When dealing with E & H, I & V, in no specific order, it is impossible to say that this causes that.
I agree with you that taking it down to this level is a bit too deep for this forum. After all, on an audio forum, the quest should be to share info that advances our enjoyment of music. The CCCS view of the bjt & the VCCS view of the FET are not exactly correct at the quantum or mIcro level, but they are good enough at the mAcro level. If we keep that in mind, we can discuss transistors and their associated circuitry ad infinitum w/o a problem.
But somebody always reminds us that the CCCS view of the bjt does not hold at the quantum level. I agree with that. But they then proceed to insist that the VCCS is the correct view at quantum levels. I just jumped in to say that it is the QCCS (charge controlled) view that is correct at quantum level.
If you would rather avoid the issue at the deep quantum level, I agree with you. The CCCS for the bjt, and VCCS for the FET work well at the mAcro scale 1st order approximation. At the quantum level, the QCVS is what is used for both bjt & FET.
Once again, publishing my views is redundant. What I've posted is common knowledge since the 19th century. Best regards,
Claude Abraham said:
At this point, you lost me completely. I don't understand what you mean by micro, macro, quantum levels (and please, don't make me pull my credentials guns), what model do you support for the BJT and why.
Following your same dual logic, a MOSFET is a current controlled device, because you would need a current to charge the gate, which creates a voltage across the gate (of course, in a finite time), which in turn creates an electric field (or the other way around?) which controls the drain current (also in a finite time). Does it make any sense?
If you need to have the last word, go ahead, I'm not going to revisit this topic, but otherwise please leave it here.
syn08 said:
Of course a MOSFET needs current to charge the gate. But should we provide that current with a current source or a voltage source? With a dc current source, if we drive the gate with constant current, the voltage will ramp up until the voltage exceeds the breakdown value, punch through will occur, and the device is destroyed.
With a constant voltage source, the current is provided by said source, but once the gg-s voltage approaches the source value, the charging diminishes. By controlling the source value to less than the breakdown value, we charge the gate safely.
"Voltage drive" provides both I & V. So does current drive. But a FET works best if we directly control V, and let I be defined by V and the circuit elements.
A bjt OTOH, should not be driven from a constant voltage source. We control its emitter or base current, then let V be defined by the circuit. If a constant current source is not available, placing a large enough resistor in series with ther voltage source provides control of the current.
he terms current controlled and voltage controlled at the mAcro level refer to which quantity is *directly* driven, and the other is incidental. Both are needed, but only 1 can be the controlled quantity.
Does this help?
I think that it is useful to note different models of active device operation, however it is not really necessary to be a successful design engineer.
Normal discrete circuit designers do not design the devices they use. What they need to know is the performance of what devices they have to work with, and how to implement them most successfully. Because of this, you can often think of an active device such as a bjt or an fet as a 'black box' with specific characteristics.
While some of us like to delve into subtle characteristics and exceptions in standard design or active components, it takes a lot of time and effort, and just isn't necessary to make a working and probably successful circuit.
In my life experience, I have found that 'transistor theory' is not that important, compared to transistor performance. I don't make the darn things, I just use them. That is what separates design engineers from device engineers. An IC designer might have a different point of view.
Normal discrete circuit designers do not design the devices they use. What they need to know is the performance of what devices they have to work with, and how to implement them most successfully. Because of this, you can often think of an active device such as a bjt or an fet as a 'black box' with specific characteristics.
While some of us like to delve into subtle characteristics and exceptions in standard design or active components, it takes a lot of time and effort, and just isn't necessary to make a working and probably successful circuit.
In my life experience, I have found that 'transistor theory' is not that important, compared to transistor performance. I don't make the darn things, I just use them. That is what separates design engineers from device engineers. An IC designer might have a different point of view.
Such details not necessary need to be calculated every time, they just sink into subconscious mind, and when you start thinking of some design on "macro level" you can imagine how it works more precisely.
You see, feel, hear processes, predict behaviors of different variants of the solution even before drawing schematic and calculating values.
Like, when you are managing a team you don't need to look always on complete dossier for everyone, you just assume how to organize them, knowing well their unique characters.
It is my approach.
You see, feel, hear processes, predict behaviors of different variants of the solution even before drawing schematic and calculating values.
Like, when you are managing a team you don't need to look always on complete dossier for everyone, you just assume how to organize them, knowing well their unique characters.
It is my approach.
Lumba Ogir said:Wavebourn
you mean no analysis whatsoever at micro level?
I mean the Mother Nature created the best computer whatsoever. The Brain.
There is one exercise, called scaling. Imagine something very familiar. For example, a bicycle. Think of how it works, how it looks. Think of macro details -- wheels, pedals, frame. How it looks, each of them, how it works. Go down, deeper and deeper, as deep as you can, to processes and details that perform that processes. Now, go up. Build a bicycle in your imagination. Build a team of bicyclists. And so on.
Sometimes you don't know in depth how some detail works, you experiment observing it's behavior. Sometimes you suddenly see some unknown yet habit of some part, now the picture of the detail stored somewhere in the brain is more complete.
It is necessary to analyze on micro level, like when you learned to ride a bicycle you analyzed each detail of the process, you've experimented, did trials and errors, but now you take a bike and ride, without paying attention on all details, but your subconscious mind remembers everything, and each and every time you ride your bike it learns something new.
The same with electronics design, if you learn how to design, instead of how to use books with examples, magazines, forums, and software simulators. However, they are useful tools. To learn. But not to design. You can't ride a bike reading instructions and simulating it in computer software reality.
Lumba Ogir said:
Do you mean breadboarding, measurements, listening tests? Sure!
If something goes wrong (like oscillations, or a hole in a chair burned by a resistor, output transistors survived shorted output, but a filter cap that is after a fuse blew up), deep level analysis is needed. Sometimes deeper, sometimes less deeper is enough.
Lumba, I am not trying to ignore your original questions about part numbers of low noise devices, but I can't easily verify them at the moment. The Hitachi devices had a standard number, but I can't remember what it was. The Rohm devices seem to be what you suggested. There are many other devices that were ultimately developed for pre-preamp input stages, about 30 years ago. Unfortunately, most are discontinued, today.
Low noise j-fets are another device series, altogether.
Low noise j-fets are another device series, altogether.
Wavebourn,
I think the problem rather was that you could not foresee the event, or maybe you thought that it just could happen to others, which is a very common mistake. Actually, that category of accidents are avoidable, simply by making allowances for every contingency and taking all necessary precautions.
I think the problem rather was that you could not foresee the event, or maybe you thought that it just could happen to others, which is a very common mistake. Actually, that category of accidents are avoidable, simply by making allowances for every contingency and taking all necessary precautions.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.