I did put together a test of the C2M1000170D at the same time. It was a
little disappointing in that the transconductance was quite low, about .5 S,
but the capacitance is also low. I had rather hoped it would make a super simple
Zen amp with some snap to it, but I think it'll require a little more effort.
😎
Hmmm. I don't have any SiC parts at present other than the Semisouth 085 jFETs, which sound nice but require some hands-on treatment due to their high gate leakage. Maybe Semisouthfan could chime in with a reason for this. One would think that given the band gap of SiC, leakage wouldn't be an issue. The similar United Silicon Carbide part shows almost identical behavior (at least in the data sheet), so it's not just a Semisouth thing....
At the other extreme end you have this one with plenty of transconductance.
http://www.wolfspeed.com/downloads/dl/file/id/161/product/8/c2m0025120d.pdf
What do you think of this one?
http://www.infineon.com/dgdl/Infine...n.pdf?fileId=db3a304341e0aed001420353f03a0e4b
Depletion Mode
Decent Transconductance lavels
Decent Capacitance levels
Zero temperature coefficient
Thanks for giving them a try Papa..... If anyone could make them work magic it would be you! I popped the lower IRFP 240's out of my first Amp Camp Amp build and substituted the C2M1000170D as an easy test. The gain of the amp went way down, but I left them in anyway. The parts did sound pretty good to my ears despite the lower output. I appreciate you trying them and letting us know your thoughts. Merry Christmas to you, and to all here in this amazing place!
Hi 2pD
Yes I think so too, but in a single stage amp or as an output stage device the low transconductance seems to hurt performance. I've been thinking of a circuit I ran across in some old app notes that was called a "Transconductance multiplier". It married a bipolar transistor and another device to provide greater transconductance but keeping the characteristics of the original device. One example they showed used a triode tube as the main device.
Yes typical LARGE SiC photodiode is e-15 A with reverse resistance of e17 Ohms. Have you seen any papers on actual structures, I have access to the IEEE library?
You can achieve a similar result with cascoding. You just need to select each device appropriately. I can't comment which method is better, but I tend to prefer dumb solutions, so I like cascoding.
That is discouraging to hear. I had high hopes for the C3M's. Haven't got around to trying my quad of them out yet, but was originally thinking about trying them in an F6 circuit.
Both the SemiSouth variety and the United Silicon Carbide variety of JFET are vertical channel, i.e., trench types. Certain issues with forming the vertical junction in the trench are the culprit for leakage current, not the theoretical properties of a wide bandgap pn junction (this is an example of the difference that sometimes occurs between simple theory and actual practice). Having said that, it is my experience that the actual reverse gate leakage current of both genuine SemiSouth R085's and the United Silicon Carbide equivalents are usually quite low and in general not a factor in the amplifier. In other words, the data sheets include lots of topside margin.
The Infineon SiC JFET is a lateral channel JFET with completely different construction. They are not equivalent to the relatively low absolute value of threshold voltage characterizing commercial vertical channel JFETs. Vertical channel JFETs typically have higher transconductance compared to the Infineon or the Cree transistors, although this should be evaluated at linear amplifier bias conditions, rather than switching conditions used in data sheets. Higher transconductance helps the negative feedback used to one extent or another in all of these amplifiers.
The main attraction of the C2M1000170D was the low capacitance, which
allows a Schade type input feedback at a higher input impedance so no need
for a buffer. I have lots of big Fets if I can put up with the capacitance.
I'll play with it some more.
😎
It may have its place in a tube emulation scheme, especially push-pull, where a 1200V breakdown would provide some nice safety margin against mischance.
What about using C2M1000170D as the cascode device with another device of higher transconductance mosfet?
You get the transconductance you want with the capacitance you want.
Or do you feel cascoding is a step backwards in some other way (Less Zen)?
Edit: C2M0280120D appears to have quite a bit better transconductance and current rating with similar capacitance to C2M1000170D. No idea whether that might be of interest to you.
http://www.wolfspeed.com/power/products/sic-mosfets/c2m0280120d
Last edited:
Cascodes can have snarky oscillation problems if one is not careful, and even if one is - I have been bitten in sand state and vacuum state applications.
Snarky oscillations aside, you also see higher capacitance at the lower Vds
seen in cascode oscillation. Sometimes you can make it work, sometimes
not so much...
😎
seen in cascode oscillation. Sometimes you can make it work, sometimes
not so much...
😎
I have a pair of lowish impedance Amplimo toroidal p-p tube output XFMRs loitering around in search of gainful employment. Maybe they will get hooked up with some SiC in a typically perverse Wrenchone implementation.
I'll be putting in some sort of Digi-key order over the holidays, if only to replenish stock on one of my fave Ixys depletion mode mosfets used in all my shunt regulators. Adding a few Cree parts to the mix might happen then.
I'll be putting in some sort of Digi-key order over the holidays, if only to replenish stock on one of my fave Ixys depletion mode mosfets used in all my shunt regulators. Adding a few Cree parts to the mix might happen then.