Small geometry MOSFET flicker noise, how bad is it?

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.

tk@halmi

Well-known member
Joined
Jun 3, 2004
Messages
999
Location
Oregon, USA
I was playing with some small geometry MOSFETs in the simulator. To my surprise PSpice spits out noise figures similar to J112 and PN4393 JFETs (~3nV at 1KHz) when using a Zetex part, the ZVN3310A, for the same bias conditions in a simple opamp. Also, it seems that the 1/f corner is extremely low compared to JFETs. Is this possible or my Spice model is incorrect?

Thanks,
Tamas
 
[quote author="tk@halmi"]I was playing with some small geometry MOSFETs in the simulator. To my surprise PSpice spits out noise figures similar to J112 and PN4393 JFETs (~3nV at 1KHz) when using a Zetex part, the ZVN3310A, for the same bias conditions in a simple opamp. Also, it seems that the 1/f corner is extremely low compared to JFETs. Is this possible or my Spice model is incorrect?

Thanks,
Tamas[/quote]

Sounds a bit unbelievable, but processes have been steadily improving. In the old days the 1/f corner was at about 10MHz! Now you can look at some of the Nation*l MOS amps and see 1kHz e sub n pretty low.

I think it would be big news and touted by the manufacturer though, if they had gotten into 4393 etc. territory.
 
> Is this possible or my Spice model is incorrect?

Always assume that SPICE LIES.

Well, sometimes SPICE gets the "right" calculations from bogus models (still wrong answer).

Noise parameters are unlikely to be right. About every small BJT has Rb=10, which can't be right.

These are what, $0.25 parts? It would be good to kill the PC and set up a breadboard, do some real-life tests.
 
[quote author="PRR"]> Is this possible or my Spice model is incorrect?

Always assume that SPICE LIES.

These are what, $0.25 parts? It would be good to kill the PC and set up a breadboard, do some real-life tests.[/quote]

Amen :thumb:
 
I have 17 breadboards all full with various jigs. One more won't hurt. I have a hunch that it will be a disapponting experiment.
 
Sam,

I am using ZVP and ZVN parts in place of JFETs and BJTs for VAS and output purpuses successfully. I was wondering if I could use them as input devices as well. They allow higher voltage rails, more current, better dissipation, no seconday breakdown and no thermal runaway. In some cases they have more gain than the best JFETs. I like that they can be found at many supliers too.

Cheers,
Tamas
 
There seems to be a debate about using MOSFET transistors in an output stage, at least as far as linearity, transconductance, and ON-resistance is concerned. I'm not that experienced here, I only know what I've read.

From what I've learned about them, the improved HF response is an advantage because of the potential for a higher second pole than a BJT would have. In theory, this should allow you to raise your dominant pole while still maintaining stability, thereby allowing you to increase NFB (which would be needed to compensate for worse open-loop distortion).

Have you found any difficulty with linearity under load at all with these devices, Tamas (or anyone else)? I'd love to hear about your findings.
 
Raising the dominant pole is a tricky business as it tends to make the amp susceptible to capacitive loading problems. I had better results with two-pole compensation.

Samuel
 
[quote author="featherpillow"]
Have you found any difficulty with linearity under load at all with these devices[/quote]

In my experience, the gain droop under load is absent in MOSFETs, and it is very appearant in BJTs. Of course, there are other mechanisms that contribute to non-linear transfer. There are volumes of literature on this topic and there are legions of products using either BJT or MOSFET (or both) with great success.
 
One of the key insights for me was noting the MOSFET source-follower self-bootstrapping of the large gate-source capacitance and how it is affected by the load. The absence of a d.c. drive requirement is misleading to some---the important effects pertain to capacitances and also to the intrinsic high-frequency figures of merit.

But having no d.c. involved is handy, and the absence of saturation effects in heavy conduction compared to bipolars can simplify drive requirements as well.

Against this you have the disadvantage that rarely will a batch of parts be consistent enough to use "out of the box" unless the circuit is very clever and/or has adjustments. The distribution of threshold voltages is just way more broad than bipolars. As well, the mobility of holes in silicon is way less than electrons, so the P parts are much worse than the N for comparable geometries. Thus, the circuits will be hard-pressed to work symmetrically in an output stage: either gm of the P device will be lower than the N, or the capacitances will be higher. The result, for a straightforward complementary output stage design, will be a gain asymmetry and consequent odd-order distortion. (EDIT: brain fart thanks PRR--even-order)

There are some bipolar power transistors with pretty decent parameters that have been around for a long time, with f sub t's in the 30-60MHz range and pretty good safe operating area, and fairly constant beta over a wide range of Ic's. But if you really want to go as fast as possible and have really high loop gain the DMOS stuff in a good design will beat them I think. As Samuel says (about), anything that tries to go that fast will be load-sensitive---although there are techniques where the loop compensation ends up being partly controlled by the output loading to make the amp less prone to instability. There's always the old trick of the L-R series network to isolate capacitive loads, too.
 
It may be obvious that I'm reading some of Douglas Self's work, and as you may surmise, he doesn't really care for MOSFETs in an output stage. But he does discuss a few advantages, including those that I've listed.

It's interesting to see different approaches to the same problem. Each time I think I have an understanding of how to approach a design, I run into a different opinion or approach.

As one of my friends used to say, "Chill out Captain Queeg, there are no f@cking strawberries."
 
> the mobility of holes in silicon is way less than electrons, so the P parts are much worse than the N for comparable geometries. .....result, for a straightforward complementary output stage design, will be a gain asymmetry and consequent odd-order distortion.

Even-order, no?
 
[quote author="PRR"]> the mobility of holes in silicon is way less than electrons, so the P parts are much worse than the N for comparable geometries. .....result, for a straightforward complementary output stage design, will be a gain asymmetry and consequent odd-order distortion.

Even-order, no?[/quote]

Yep---corrected :oops:

Here is a timely reference about the original topic just handed to me from a EDN Analog e-letter:

http://web.mit.edu/klund/www/CMOSnoise.pdf
 
Back
Top