Yes; actually the 0.1A-2.5A zone for class B to class A, with peaks to 5A or 10A.
https://www.mouser.com/datasheet/2/827/UNITED_SILICON-10-03-2018-UJ3N065080K3S-1487513.pdf
Vto (for 20mA!) is -6V to -14V! Whereas in a MOSFET amp we must drive the upper gate ~~5V above the upper supply rail, for these depletion devices we must drive the upper gate a dozen volts *below* the lower supply rail. We may sometimes overlook the 5V offset of a MOSFET; this dozen-Volt offset kinda demands boosted rails for the driver.
The large negative bias for linear zone "does remind us" of vacuum tubes. But the likely signal swing is very much less than a tube. The transconductance looks like 3S to 5S, a tube would be like 0.005S.
We always like to run power amps as "Cathode Followers" if we can swing the drive. 3S of gM is like 0.3 Ohms which promises good gain and linearity into 8 Ohm loads. But the over-the-rails gate swing is awkward.
The gate capacitance does not look that brutal to me. Many MOSFETs are similar. Many BJTs effectively have large base capacitance due to hFE fall-off above the audio band. There is a fundamental limit here and I think many recent devices approach that limit, in their several ways.
Attachments
that is why the NPass Camp Amp continuous to amaze....in our local builds, sound was at par with the 300b.....in a comparison listening test, the audience asked to change amp to the 300b, but the operator merely changed the program source, the audience did not realise it, but knew it when the operator told them.....
It is not about power, it is about how the gate is insulated; either by a dielectric (or an oxide), or by a PN junction.
I love normaly open
80mΩ - 650V SiC Normally-On JFET
https://www.mouser.es/datasheet/2/827/UNITED_SILICON-10-03-2018-UJ3N065080K3S-1487513.pdf
80mΩ - 650V SiC Normally-On JFET
https://www.mouser.es/datasheet/2/827/UNITED_SILICON-10-03-2018-UJ3N065080K3S-1487513.pdf
Just received four of these from Mouser. I am thinking of going for a ciclotron. As has been mentioned, an extra, or possibly two supplies are indeed required for the bias, and these must ramp up before the bridge voltages.
Using the manufacturers' spice model, open-loop damping factor calculates around 10 for the 2x24V rails ciclotron, biased at 2.5A or therabouts. I wonder if this will be any good driving conventional two-way medium to lowish sensitivity speakers, such as my proacs, limited to moderate levels of course.
Another proposition would be the quasi-complementary.
Using the manufacturers' spice model, open-loop damping factor calculates around 10 for the 2x24V rails ciclotron, biased at 2.5A or therabouts. I wonder if this will be any good driving conventional two-way medium to lowish sensitivity speakers, such as my proacs, limited to moderate levels of course.
Another proposition would be the quasi-complementary.
Last edited:
Nice, I'm waiting for payday
I was thinking of trying an F2 first, then flip it round and try a beefed up B1 (rev 1), then...
PLH? J2, F6, lots of options
never enough time Spice simulated SIT vs MOSFET: 0-1
Details:
Devices: UJ3N065080K3S vs IRFP240
Spice models: SIT: UnitedSIC provided / MOSFET: LTSpice standard
Iq: 2.5A both
Topology: Common-drain ciclotron
Load: 8 Ohms resistive
Rails: 24V x2
Po = 5.5W & 20W (two tests)
Results:
The mosfet circuit wins due to lower harmonic power, by 5dB to 15dB depending on harmonic, all the way from the 2nd to the 19th.
Only in the lower power test does the SIT win only at the higher harmonics. In the higher power test, MOS is better throughout.
But, let us not despair as HD is not easy to model in spice. I guess it is measurement time.
Details:
Devices: UJ3N065080K3S vs IRFP240
Spice models: SIT: UnitedSIC provided / MOSFET: LTSpice standard
Iq: 2.5A both
Topology: Common-drain ciclotron
Load: 8 Ohms resistive
Rails: 24V x2
Po = 5.5W & 20W (two tests)
Results:
The mosfet circuit wins due to lower harmonic power, by 5dB to 15dB depending on harmonic, all the way from the 2nd to the 19th.
Only in the lower power test does the SIT win only at the higher harmonics. In the higher power test, MOS is better throughout.
But, let us not despair as HD is not easy to model in spice. I guess it is measurement time.
Maybe not the same device category name, but according to wikipedia "The SIT is a type of JFET with a short channel length."
Plus the Semisouth SJDP120R085, widely regarded as a SIT in this forum, is termed "Trench Silicon Carbide Power JFET" in the datasheet, plus its transfer characteristics are wickedly close to those of the UnitedSIC's device.
Those here who have delved into such devices can perhaps comment on similarities and differences. Me, I am happy with the UJ3N06580 transfer characteristics
Last edited:
Now that is interesting. Was that in simulation or real life? Were comparative measurements performed? Listening tests?
Looking at the datasheet one could swear that the Cree device is an enhancement power jfet, rather than a mosfet, as the saturation region is pushed quite far away.
Last edited:
I did a listening test with some 24V+ regulated supply and 3A bias.
Also dist measurement at like 1W 7ohm 1kHz (my load is 21/3ohms) did not show consistent differences to IRFP240.
It was around the same.
Would love to do more tests with a different setup though.
Also dist measurement at like 1W 7ohm 1kHz (my load is 21/3ohms) did not show consistent differences to IRFP240.
It was around the same.
Would love to do more tests with a different setup though.