This is match to play with burning hockey puck 😉
Congratulations have a mosfet winner game
Attachments
Thanks for the heads up! I did a copper ground plate for a refurb recently and it was painful but worthwhile in the end - I have to look at it everyday and if I didn'g do it "right", probably kick myself over and over.
The first job is to get the heatsink config nailed, before deciding wire topology. I am waiting a quotation from Austerlitz on their KS216, since a 50cm piece of it would dissipate the 200W this beast makes by -30°C. And it'd double nicely as the back of the amp "case", which is actually being conceived as a speaker stand for my 15" tall speaker.
I'll be definitely referencing that circuit at some point in a future build although heavily scaled down. Probably 3 pairs at output biased at 1A each with 22 to 30V rails. Maybe throw Cree devices at the output. Toshiba Jfet input.
Or maybe even throw a single pair of hockey pucks at it
Last edited:
There seems to be a slightly improved sweet spot while maintaining same distortion character by using R drain or around 0.22 Ohms and Rsource of 0.18 Ohms
The voltage reference not included in schematic is going to be interesting. These parts are super sensitive to changes in Vgs
The first is done with Nelson's circuit with IXYS hockey puck.
The second has an added cascode to the IXYS hockey puck to reduce capacitance.
Both results are exceptional really. You'd have to build both to judge with your ear.
These are simulated results, I just added a simple voltage reference so that the circuit biased up at 3Amps
Please disregard those results. I just realised the feedback is different from Nelson's design. I will redo it right now.
I am using this one from IXYS not exact but SOT227 with similar transconductance and capacitance.
Good enough to get a feel for the amp
.model IXFN90N30 VDMOS(Rg=3 Vto=3.6 Rd=13.2m Rs=9.9m Rb=17m Kp=90 Cgdmax=30n Cgdmin=10n Cgs=100n Cjo=30n Is=720p mfg=IXYS Vds=300 Ron=33m Qg=360n)
Good enough to get a feel for the amp
.model IXFN90N30 VDMOS(Rg=3 Vto=3.6 Rd=13.2m Rs=9.9m Rb=17m Kp=90 Cgdmax=30n Cgdmin=10n Cgs=100n Cjo=30n Is=720p mfg=IXYS Vds=300 Ron=33m Qg=360n)
For the cascode I just used IRFP240, but a low capacitance Cree device as the cascode would turn the Hockey puck into a bloody super device.
I only cascoded the common source device not the follower
I cascode to source pin not to ground (actualy no difference in this case, source is at ground).
I only cascoded the common source device not the follower
I cascode to source pin not to ground (actualy no difference in this case, source is at ground).
Last edited:
(4^2/4)/2 = 2Watts avg into 4Ohms
I can measure at higher watts.
At 5 Watts avg into 8 Ohms the cascoded version starts looking quite a bit better than stock, but most of what you hear is happening in the first watt, so hard to quantify the real improvement when you think of it from that view point
Last edited:
- Status
- Not open for further replies.