Please excuse me if this has been asked before...
There are certainly good reasons to keep the quiescent current constant. However if one targets the Iq*Re=Ut condition, the current must rise with temperature as Ut is proportional to (absolute) temperature. Are there any improvements in crossover distortion with rising Iq with temperature or penalties if the current is kept constant?
There are certainly good reasons to keep the quiescent current constant. However if one targets the Iq*Re=Ut condition, the current must rise with temperature as Ut is proportional to (absolute) temperature. Are there any improvements in crossover distortion with rising Iq with temperature or penalties if the current is kept constant?
Your wrong, no you can't do that, it will never work, Oh my God what an idiot, and on it goes. Those comments never make any friends and never really add to a conversation.
I'm sure many of us have had those thoughts and mainly at work this pops in our silly little heads and then if we are smart we shut our mouths and listen to what is really being said. Sometimes or first reaction is correct and the idea is just crazy, sometimes those crazy ideas are the answer to a problem nobody is seeing.
When we find something isn't right and we have the answer why not, it is almost always helpful and appreciated by others who leave their ego out of the conversation Helpful interactions are remembered and appreciated and this often comes back to you later when you also need some of the same type of help. Give and take is what works, always being on the attack comes back to bite you later when others make comments about someone and how negative they always seem to be. Those with a cheerful helpful attitude often rise to the top, when the opposite happens most people want to jump ship, it isn't worth it to have to put up with being attacked on a regular basis.
Moral of the story is as always it is much better to not say anything if it isn't requested, you have nothing good to say, or you actually have nothing relevant to add to the conversation.
Back up your comments with facts or add something relevant to the conversation or add to it and you will be remembered as a confident and helpful person who can be relied upon to add rather than negate what others are saying or believe.
My 2C's worth of useless advice I expect.
I'm sure many of us have had those thoughts and mainly at work this pops in our silly little heads and then if we are smart we shut our mouths and listen to what is really being said. Sometimes or first reaction is correct and the idea is just crazy, sometimes those crazy ideas are the answer to a problem nobody is seeing.
When we find something isn't right and we have the answer why not, it is almost always helpful and appreciated by others who leave their ego out of the conversation Helpful interactions are remembered and appreciated and this often comes back to you later when you also need some of the same type of help. Give and take is what works, always being on the attack comes back to bite you later when others make comments about someone and how negative they always seem to be. Those with a cheerful helpful attitude often rise to the top, when the opposite happens most people want to jump ship, it isn't worth it to have to put up with being attacked on a regular basis.
Moral of the story is as always it is much better to not say anything if it isn't requested, you have nothing good to say, or you actually have nothing relevant to add to the conversation.
Back up your comments with facts or add something relevant to the conversation or add to it and you will be remembered as a confident and helpful person who can be relied upon to add rather than negate what others are saying or believe.
My 2C's worth of useless advice I expect.
Hi Juergen,
You are correct. I think this may have come up earlier. The magic 26mV will increase at higher temperatures. It is what is called PTAT - proportional to absolute temperature.
So, one could argue that a 30 degree increase in junction temperature would call for approximately a 10% increase in bias current, or a 2.6mV increase in Vre. So if the idle junction temperature was 55C, this would make sense. However, junction temperature is not really constant when the amplifier is delivering significant signal current to a load. Put another way, many of us would be very happy in the first place to be able to maintain Vre to within a couple of mV of where it ideally should be.
Cheers,
Bob
Perhaps MIT is good enough company. Also, the calculation of ft is given a physical interpretation.
http://ocw.mit.edu/courses/electric...-fall-2005/lecture-notes/lecture23annotat.pdf
Last edited:
Firstly, thanks to Bob for the helpful, as always, reply.
It's not at all how I think of the behaviour of FETs so I didn't see your point, but it's clearer now.
I would still expect in typical audio amplifier circuits that the base resistance (internal + any stoppers) would dominate and the calculated equivalent Ft would be of only theoretical interest.
But, as I said, I don't have much experience with FETs yet so I will find out.
That's really the point of this current project, not only are FETs a nice solution but that the new perspectives help to understand other electronics more deeply.
No. The FET discussion has helped me understand other aspects of electronics. Probably similar for most people who want to learn more.
Best wishes
David
It's not at all how I think of the behaviour of FETs so I didn't see your point, but it's clearer now.
I would still expect in typical audio amplifier circuits that the base resistance (internal + any stoppers) would dominate and the calculated equivalent Ft would be of only theoretical interest.
But, as I said, I don't have much experience with FETs yet so I will find out.
That's really the point of this current project, not only are FETs a nice solution but that the new perspectives help to understand other electronics more deeply.
No. The FET discussion has helped me understand other aspects of electronics. Probably similar for most people who want to learn more.
Best wishes
David
Last edited:
Thank you for the links.
To pre-empt Waly's objections, I note these are for GHz IC FETs and perhaps of limited relevance to audio power FETs.
Base resistance does not seem to be even modelled.
Bob was a fast communications IC expert, I recall.
Perhaps he can tell us how much is materials related, they are not simple silicon tech, and how much size or other factors.
Best wishes
David
Last edited:
This is a general treatment, and using the gate current means the resistance is
not a factor in the calculation. Also, we have used gate current drive for some
applications in switching power devices since some years ago, particularly to control EMI.
Last edited:
On p. 23-5 the source is shown as a Vs, not a current source.
So I think resistance should be a factor, if it were modelled.
Yes, I have seen this and it came to mind but I have no practical experience with it.
Best wishes
David
The calculation is done in terms of the input current. The resistance is a factor only if the input voltage
is used in the calculation, instead of the input current. A controlled gate current drive can be done instead
of the conventional controlled gate voltage drive, which is always through a small resistance.
Last edited:
Hi Dave,
I understand, it is a bit non-intuitive to think in terms of driving a MOSFET with an ac signal current. A different way to look at it is to drive the properly-biased gate with an ac signal voltage of varying frequency and measure the gate current. One then measures the ac current in the drain under short-circuit conditions. This gives the "current gain" at ac for the MOSFET for purposes of looking at ac frequency response as influenced by the MOSFET's equivalent ft. In this experiment, the RG is not a significant factor.
Note, of course, that the "current gain" of a MOSFET really only has significance at higher frequencies, as it tends toward infinity at low frequencies.
In reality, we usually drive MOSFETs from a finite or quite low ac impedance. In that case, you are absolutely right that the end result may be dominated by the time constant of RG against Cgs. This is more of an issue for laterals because (if I recall correctly) they use a polysilicon gate, while a HEXFET uses a metal matrix to drive the gates locally. Its been awhile since I studied the device geometries, though.
Cheers,
Bob
The bottom line is pretty much the same for Si and HEMT 3-5 JFETs in terms of the concept of ft, but of course its significance in the GHz applications is much stronger. About 20-25 years ago I did have some work directly with research on Indium Phsophide FETS and BJTs. At the time the big thing was what kind of base doping to use, e.g. Beryllium vs. Carbon vs. something else, and the level of difficulty of achieving that doping. Carbon was more difficult because I think it needed a special MBE machine at the time.
We briefly held the record for ft of an Indium Phosphide BJT at Bellcore that was carbon-doped and achieved ft = 200 GHz - long since surpassed. It was cool to have our own little research fab. We were making OEIC transimpedance amplifiers for 10GHz optical receivers at the time, where the photo-diode and transimpedance amp were monolithically integrated. It was for multi-wavelength applications, so there were 8 receivers on each chip, with an 8-fiber ribbon to be ultimately be attached to it.
Cheers,
Bob
This is essentially what the MIT paper does.
Uses a V source but does the calculation based on the current.
I didn't want the discussion with "Rayma" to turn into quibbles, but it still seems awkward to me, if you want a current based calculation then use a current source.
This reminds me of the discussion of heat transmission with D.Self.
Theoretically the speed in metals is set by the enormously fast electron velocity, but it still takes five minutes to heat the handle of my cast iron frypan.
Best wishes
David
You could have used advanced hetero-junction fabrication but applied it to super-performance audio transistors rather than waste it on merely faster telecommunications applications
Best wishes
David
Hi David,
"Electron velocity" should not be confused with wave propagation speed. Indeed, the latter is very fast. In metals about 60% of the speed of light. But the propagation speed of the electrons itself, also known as drift velocity, is much much lower. In a copper wire, for example, with a length of 10cm, a diameter of 1mm and at a DC current of 1A, it takes almost three hours to travel from one side to the other.
I wonder, has this something to do with the low heat propagation speed in solids?
My
Cheers, E.
See also: https://en.wikipedia.org/wiki/Drift_velocity
