Electrical Engineering - The modern circuit.

How many of you guys have an electrical background, or a robotics back ground? Discuss anything about electrical and robotics engineering.

I've attached some work at .25 micron MOSFET technology. We design our analog and digital circuit based around the MOSFET transistor.

e;
The first is a cascode op amp which is designed to operate at 600V/V amplification. (We designed op amps for medical equipment).
The second is a full layout for a inv-nand-nor circuit. You could send that to china and have it manufactured as is...
The last is the circuit for the layout in number two. Pretty simple, as said just a inv-nand-nor circuit..

.25 micron is PDG. Moore's Law is still in effect. Even with our most high tech fab we are hard pressed to get .20. Good for you!

Right now, I have a brush less motor project. I'm working on a power board. MOSFET but not H-bridge. I don't to reverse the motor.

Also, I have documented in my Journal my attempts at automation in the grow room.

Doer said:

.25 micron is PDG. Moore's Law is still in effect. Even with our most high tech fab we are hard pressed to get .20. Good for you!

Right now, I have a brush less motor project. I'm working on a power board. MOSFET but not H-bridge. I don't to reverse the motor.

Also, I have documented in my Journal my attempts at automation in the grow room.

Click to expand...

Well actually Intel is building a Fab here in Chandler AZ that will manufacture 14nm mosfets. Yes, 14 nanometer. .25 micro is 250nm. But, I appreciate the compliment.

I don't actually manufacture any chips, rather just design the circuitry that goes into custom chips. We can design at any technology level (transistor size), however we normally design in the current minimum fab size (14-22nm).
My bachelors is in Automation and I love it. I haven't checked your journal, but I have seem some of the Arduino setups on here (cool shit). I myself have an Arduino that I use, however have never used it to automate my grow. I've thought about it though.

Lastly, just to state again that yeah .20 is really good, however the size matters only slightly in the design. If you scale your mosfets down, all you need to do is scale the other properties down accordingly.

edit: removed

Sorry, read too fast. Thanks for the correction. I was thinking Nm, not mm, of course.

Designed a bunch more shit:



1- 2stage OpAMP. Differential amp to a common drain unity gain amp.
2- Folded cascode amp, essentially the same amp as 1, but with more transistors for larger rail swing, higher gain, and larger driving capability; it had to drive 10uF (pretty damn large capacitance for a 360nm long mosfet technology.)
3- Transient output of 2
4- Closed Loop (feedback) transient output of 1
5- Open Loop transient output of 1
6- Output parametric transient of 2. Basically shows I can go rail to ~1.75V. Rail to rail would be 0-2.5, but it's not possible when you cascode

Just built some mindfucking current mirrors, otherwise known as "constant current sources" or just "current sources". My bias circuitry is good enough to bias ANYTHING. I have modeled and designed my circuits so that they are temperature, noise and process independent. Meaning, in an unideal world, I have managed to design and create an "ideal" current source!

e;
essentially, what can bee seen is that I varied Vdd (the supply voltage; what powers the circuit) to show that I can take up to .5V (or more) of noise (20%) to my signal before I generate a current that is different than 10uA. Futhermore, 10uA is a baller bias point because at 2.5V, the power used in this circuit is (2.5V*10uA) = 25uW... That's awesome. As growers we can appreciate that as we regularly use 1000W lights. 25uW is 25,000,000,000x smaller power use than a 1000W light.

Furthermore, in the last shot you can see that I also created a bias Voltage at about 570mV, which also is constant and can be used to bias other nmos in the circuit. Everything is rock solid

Another example of my work:

Another Folded Cascode which had to meet specs given:





Wont let me upload half my shit, so yeah, this is a bunch of a random 3 folded cascode designs lol. Enjoy cuz im not in the right mind to label all of them atm; might edit later.

Designed a Class AB amplifier (best audio amp besides vacumn tubes)

Essentially, most amplifiers get "killed" (lose all their gain) by low resistance loads, due to the fact that all the current flows to the load, not to the the amp. This Class AB amplifier can drive 4ohm loads (approximately equal to a very good speaker). These s/s's show a Class AB amp which had to meet certain specs. I had a time restriction so I just missed one spec, but for the time frame, that's a hell of an amp.

"best" is a relative term. Best as to what feature? I'm not an engineer, but I do know that sonically, a class A amp will not have any crossover distortion so will sound much better than AB in certain situations.

mindphuk said:

"best" is a relative term. Best as to what feature? I'm not an engineer, but I do know that sonically, a class A amp will not have any crossover distortion so will sound much better than AB in certain situations.

Click to expand...

Only class AB, class A and class D amplifiers can be classified as "good" audio amps, due to the drive load (very low resistance). Consequently, you need an amplifier which provides large bandwidth and gain, as well as a low output impedance, so not to "kill" your amp with a low resistance load. Class A amps are always on, and provide the cleanest output versus input, but they are EXTREMELY inefficient. Class B is the opposite of Class A. Class AB is the best of both worlds, and Class D are digital amps. Class A amps in theory are cleaner than AB, but we now design class AB amps that are only minutely less clean than A. The same principle can be applied to Vacuum Tubes, which are still cleaner than any MOSFET amp, due to inherent noise in transistors and silicon.

The most simple way to understand it, is to think of how parallel resistances work; when there are 2 parallel paths a current can flow through, each having their own resistance, the overall effective resistance to the current is LOWER than the LOWEST resistance. If your amp has a high impedance and the output node, the low resistance speaker will effective pull all the current to it, leaving very little current for the amplifier.

http://www.hifivision.com/amplifiers/174-types-amplifiers-class-class-b-class-ab-class-d.html

So in this case, "best" describes a convolution of:
Power Consumption
Heat Creation
Lack of Signal Distortion
Gain
Efficiency in terms of everything not mentioned above

Put those qualities all together, and the best overall is the Class AB amplifier, which is why it is the one that is commonly manufactured. But, you're correct, class AB amps are "best" relatively speaking with regard to what I mentioned above.

The thing that I was told, which I have no idea if it is true or not, is that the distortion we get with pure class A is second order harmonics whereas class AB has odd order, which is less pleasing to listen to. This is the reason many of us that listen to music prefer the more musical sounding tube amps, even though they are not 'clean' their distortion is even order harmonics. I had an old B&K MOSFET amp that actually had some tube-like characteristics. Since class A is so inefficient and get so hot, they are sometimes used in a bi/tri-amp setup, running just the tweets, which need less power and may be more efficient drivers than mids or woofs and maybe more acoustically important as to creating a nice, clean sound. Personally, I think the mids are most important since they carry the majority of the vocals. I have heard some incredible amplifiers over the years, some of them have been tube, some solid-state but the pure class A almost always beat any AB. The problem is and always has been the extreme inefficiency and heat buildup.

Have you heard of Jeff Rowland? http://jeffrowlandgroup.com/ClassicAmps.htm
If you look at some of the older amps, there are some links to "technology." Most of this is over my head but you might find it interesting. He talks about the design of the power supply as well as the signal circuit and how they differ from 'traditional' tech at the time. His amps were some of the most expensive, $7-10,000 and this was in the '80's.

mindphuk said:

The thing that I was told, which I have no idea if it is true or not, is that the distortion we get with pure class A is second order harmonics whereas class AB has odd order, which is less pleasing to listen to. This is the reason many of us that listen to music prefer the more musical sounding tube amps, even though they are not 'clean' their distortion is even order harmonics. I had an old B&K MOSFET amp that actually had some tube-like characteristics. Since class A is so inefficient and get so hot, they are sometimes used in a bi/tri-amp setup, running just the tweets, which need less power and may be more efficient drivers than mids or woofs and maybe more acoustically important as to creating a nice, clean sound. Personally, I think the mids are most important since they carry the majority of the vocals. I have heard some incredible amplifiers over the years, some of them have been tube, some solid-state but the pure class A almost always beat any AB. The problem is and always has been the extreme inefficiency and heat buildup.

Have you heard of Jeff Rowland? http://jeffrowlandgroup.com/ClassicAmps.htm
If you look at some of the older amps, there are some links to "technology." Most of this is over my head but you might find it interesting. He talks about the design of the power supply as well as the signal circuit and how they differ from 'traditional' tech at the time. His amps were some of the most expensive, $7-10,000 and this was in the '80's.

Click to expand...

To answer the beginning part of your post:

I posted my Harmonic Distortion plot (the 11th upload) (the one with a lot of red in it, and you can clearly see how the harmonics are coming in. "3rd harmonic is -61.2dB under 1st harmonic")
That is a plot of exactly how my amplifier will perform given the parameters I modeled it as, in terms of Harmonic distortion.As can be seen, all harmonics are being rejected at over 50dB (pretty decent). If you get that up to 60db+ you won't ever be able to notice the distortion at all, which is what I achieve before my 3rd harmonic.

Secondly, yeah, Rowland amps are very well made and designed amplifiers. We were designing an amplifier that had to output 2mW or less, as can be seen by my 547uA current draw for my design, so class AB was the obvious choice.