I still need a little clarification on how a SiC JFET behaves. A Si JFET is on with zero bias Gate to Source. Is it correct that a SiC JFET is off until the Gate is pulled positive by a volt or two (but sill less than the SiC bandgap potential)? Assuming yes, the SiC JFET would measure and open circuit from Source to Drain. Correct?
Might it be that some of these questions guide you in the directions you might want to research? There is a "missing link" in the understanding between the microphone and the brain.I should not be surprised that I have stimulated questions before I laid out my thoughts in a cogent article. That will teach me to post a what-I'm-going-to-do-before-I-do-it post. We'll call it testing the waters. I'm glad you all are interested. And why not? Making an amplifier "sing softly" is what you all do so well. If a SemiSouth JFET helps, then you are by definition interested. If it doesn't help, this should be a thread of one poster (me). Again, we can thank Nelson for bringing to all of us the fact of the usefulness of the SemiSouth JFET for these types of amplifiers. The question of why they are useful is what I want to write about. And in that context, my background as a practicing engineering professional and educator gives me a starting point for my thinking. But, as I said, I am not an amplifier design expert so I will be very upfront in my article to say what I am not considering in the analysis. The reader will need to evaluate this information (and point of view) in the bigger picture, and I see that process has already begun.
We suspect that an undistorted waveform as a source would pass thru our ideal amplifier and connect to the perfect tranducer in our ergo acoustic listening rooms and no one would ever leave there home again. Except to go to the record store You may have figured out by now how much disagreement is involved in that last statement, along with all the objective and subjective (singing amplifiers?). Although we are in this thread for the ideal amplifier part of that, we actually don't know where the missing link is. We can sometimes make small tweaks to a very "Hi End" amp and see improvements but, N.P. has driven into our heads that it is the gain devices to be blamed for the majority of the non-linearities, distortions and other less than acceptable sounds.
It may be that an amplifier with the proverbial what goes in, comes out larger transfer curve is not what we want? Personally, it is easy to listen to the type of amplifier N.P. likes to make. It must be the focus on low level harmonics created by simple non-linearities? Or? Amplifiers measuring the current trendy THD+noise and whatever else don't seem to agree with that philosophy but, the majority of people listening to his amps will!
Dr. Mazzola. US 6,767,783 talks about SIT; which like R100A is very dear to, and coveted by DIYers. Thanks to Mr. Pass, and Mr. Rothacher who introduced it, and further pioneered its use for DIYers. SIT is impressive. But; its specialized use in radars may have made it [inadevertently] costly, and rare for diy audio. Maybe another SIT can be made to operate in the MHz range instead; like power BJTs. Thus it may have a lower cost that will be agreeable to the growing diyaudio population. In the interim, audio DIYers continue to rely on, and are basically stuck using old faithful; the MOSFET.
Best regards.
Best regards.
Might it be that some of these questions guide you in the directions you might want to research? There is a "missing link" in the understanding between the microphone and the brain.
You may have figured out by now how much disagreement is involved in that last statement, along with all the objective and subjective (singing amplifiers?). Although we are in this thread for the ideal amplifier part of that, we actually don't know where the missing link is. We can sometimes make small tweaks to a very "Hi End" amp and see improvements but, N.P. has driven into our heads that it is the gain devices to be blamed for the majority of the non-linearities, distortions and other less than acceptable sounds.
It may be that an amplifier with the proverbial what goes in, comes out larger transfer curve is not what we want? Personally, it is easy to listen to the type of amplifier N.P. likes to make. It must be the focus on low level harmonics created by simple non-linearities? Or? Amplifiers measuring the current trendy THD+noise and whatever else don't seem to agree with that philosophy but, the majority of people listening to his amps will!
FLG I think you have hit the heart of the matter in several ways. First, I am definitely reading what you all are posting as a learning experience for me. Keep "guiding."
I had beers with my business partner this evening talking about all of this. It is definitely more interesting that some other stuff I have to do to earn a living. Second, I agree about the "goal." From reading many of the posts in the Pass Labs forum, I now see that the goal is low, low distortion and simple harmonics from what is left, primarily second and third. This is both practical and explanatory. I have, like many others, heard that the tube amplifier sound can have "warmth" and so I understand that it is about leaving a very small amount of what otherwise is a poison. Many things are like that.
And so that is what I have been studying vis the R100 transfer and output curves: The magic offsetting non-linearity that keeps the distortion low while leaving small amounts of simple harmonics. It works in real life, obviously, but it also works in simulation. That is, the objective part of it. I can't speak to the subjective part of it because I don't have a Nelson clone yet. My business partner, who is a gifted circuit board layout artist, is working on correcting that part of our experience.
Regards,
Mike
Jeff Casady developed the SiC SIT for radar before coming to Mississippi State University and co-founding SemiSouth. He had "SIT" on his brain when we did that patent disclosure (lol!). In Nelson's language, the SIT is more like a triode while the JFET is more like a pentode. Channel design allows for the variation. SemiSouth made JFETs because the goal was the power switching market, not the military radar market. I plan to show in a future article how to make your SemiSouth JFET act more like a SIT.
Ah yes, the normally on versus normally off issue.
One of the things I took special pleasure from was shaming my academic colleagues into accepting that a normally off JFET is technically feasible in SiC. SemiSouth proved this by selling it commercially. This is not feasible in silicon because of the reason you allude: the built in potential of the silicon junction is too low (0.7 V typically). It's just not practical to design a channel to turn on in less than 0.5 V. It's like putting a rocket motor on a drag racer: you can expect the driver to show up in the Darwin awards. With 4H-SiC we have at least 2.5 V to play with. The threshold voltage range for the R100 is 0.75 to 1.25 V. Naturally, the transconductance is high, but manageable.
Yes, for SemiSouth's enhancement-mode products the channel is fully pinched off at zero gate bias, so the drain-source will show open; but not the gate-source or gate-drain, they act like SiC pn diodes. It is also perfectly feasible to make a depletion-mode JFET in SiC like the silicon counterparts. Those parts were also sold. Among DIY folks, we know of one of them as the SJDP120R085, which is a normally on JFET. Another which was popular in power switching was the SJDP120R045. Basically, this part is the same thing as a R085 except the die area is twice as large.
I will be trying the f2J with the SJDP and a little mix of ZV9 thrown in. No spice of course. It doesn't sit well with my stomach.
Dr. Mazzola. I visited your company's website www.impowersystems.com . It is clear, crisp and highly informative. I hope that other DIYers do the same. The service, devices and experience/expertise which you provide are in part attendant to SiC JFETs. A window shows a prominent SiC JFET module [1200V at 600A] next to a larger-sized IGBT power management module of one half the power capability. This SiC JFET module may make the Mother of All Audio Power Amps; and best done by Mr. Pass. Your website implies [speculation on may part] that this SiC JFET technology is currently [or can be] alive and well, and that devices are being made or may be manufactured by a possible authorized toll producer. This is great news. There is hope for obtainium SiC JFETS [and more]; when for several months the audio diy community believed them to be dead and gone.
Best regards
I suppose some of us audiophools should get specific about what we'ld like to see in the future papers authored by fellow member Mike M, but, it"s a big complicated subject. Putting all this down in clear consise... wait a minute I'm not workin! Speaking of work. What do you know about GaN? their not letin me play with it but, everybody including customers thinks it hopeless after years of marketing bla anyway... like some of the other SiC players I suppose...
Back to the subject at hand...I'd like lower exponectial C increase and at a lower Vds! what's the deal, these things look worse than MOSFETs in that regard??? As long as we're here these things have the C and leakage also
I'd like to see the Vds vs. Ids transfer curves from the data sheet below the Ids of the data sheet (lower than Vgs1.5V for R100 and greater than -2V Vgs for R085), like 1-5A and at least up to 6-8Vds.
I'd like too see a SOA graph, transconductance vs. Ids(@0-5A) graph and real Typ Rthj numbers??? The Tc of the Vgs sounds Negative but excelent from what I've seen, read, heard???
I'm just dumpin whats in my head so I can think up more stuff but, these are all important factors to have in one of our output parts. Since we're not swicthing 100s of Volts with inductors and Cs that stuff dosen't make the datasheet???
I know I missed a bunch of important stuff, maybe someone will pick it up? I've had all kinds of electronic expeirience from Moto OpAmp device guy in the 80's to ASICs. Wafer Processing, Test (Mostly) and Assembly. SMPs server Apps now. But, it don't matter. If your into audio, you aren't learning this stuff in school or work and most of the PhDs I work with, can't even understand what I'm taling about??? and won't bother, and, they think I'm some nut? But I never saw my CTO do anything but take the money they could be paying me? One of these days I think I'll challenge our classD group to a contest of Class A to Our class D. Probably best done after I have another job...
Back to the subject at hand...I'd like lower exponectial C increase and at a lower Vds! what's the deal, these things look worse than MOSFETs in that regard??? As long as we're here these things have the C and leakage also
I'd like to see the Vds vs. Ids transfer curves from the data sheet below the Ids of the data sheet (lower than Vgs1.5V for R100 and greater than -2V Vgs for R085), like 1-5A and at least up to 6-8Vds.
I'd like too see a SOA graph, transconductance vs. Ids(@0-5A) graph and real Typ Rthj numbers??? The Tc of the Vgs sounds Negative but excelent from what I've seen, read, heard???
I'm just dumpin whats in my head so I can think up more stuff but, these are all important factors to have in one of our output parts. Since we're not swicthing 100s of Volts with inductors and Cs that stuff dosen't make the datasheet???
I know I missed a bunch of important stuff, maybe someone will pick it up? I've had all kinds of electronic expeirience from Moto OpAmp device guy in the 80's to ASICs. Wafer Processing, Test (Mostly) and Assembly. SMPs server Apps now. But, it don't matter. If your into audio, you aren't learning this stuff in school or work and most of the PhDs I work with, can't even understand what I'm taling about??? and won't bother, and, they think I'm some nut? But I never saw my CTO do anything but take the money they could be paying me? One of these days I think I'll challenge our classD group to a contest of Class A to Our class D. Probably best done after I have another job...
Last edited:
flg; Your past and rich experience will let you delve in [and enjoy reading the patents of Casady et al [inclusive Dr. Mazzola] on SiC devices. Audio DIYers also need an abundant and affordable supply of SITs and JFETs. I asked myself why is the price of the Tripath TA 2020 [class D chip] less than $5 when it appears [to me] more complicated than a solo SiC SIT or JFET. I hope that I am wrong in this speculation. No doubt, there are reasons to account for this different pricing.
Best regards.
Best regards.
Thank you Antoinel for visiting the iMPOWER website. Frankly, it needs an update and my partner and I are planning on it in conjunction with a roll out of new products.
The major variable regarding price is: your wafer cost, package cost and Volume. Basically, without volume you won't have revenue unless you already have umptine other parts going out the door with a good profit margin. It is a big deal to have a processing line, assembly line, test and all with quality. I think SS went to contractors? Mike? and did not own all the necessary equip to do it themself, but just what I heard/read..
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.