Toshiba MOSFETs with a twist

[abraxalito]

Pretty much all power MOSFETs have that reverse diode between D and S. I took a look at the datasheet, fig5 gives me the heebiejeebies because the output impedance is so low. This means the gm is varying substantially with Vds. I prefer devices with nice flat lines meaning higher output impedance and hence better PSRR.

[jkeny]

Ok, thanks - Do you mean Fig 6 - this one:

DS resistance.jpg (49.95 KiB) Viewed 662 times

[nige2000]

more off topic thought ramblings

these toshiba parts been discussed
can they be used as the active parts in a line level speaker crossover transferred into mini class a amps for each speaker driver?

[abraxalito]

Ok, thanks - Do you mean Fig 6 - this one:

No, fig6 only shows Rds(on) which applies when the FET is fully enhanced so for very small voltages between the drain and source. I was referring to the Vgs stepped plot here :

IPD060N03.png (28.95 KiB) Viewed 645 times

Notice that for low Vds the lines look about parabolic but then as Vds increases, there's a kink and the plot then goes east-north-east. That part after the kink is showing the output impedance looking into the drain of the FET. If the line's horizontal, the FET is acting like a current source (infinite Zout) but here there's a very much non-infinite Zout. The output current increases with increased Vds even as Vgs is held constant.

With bipolar transistors there's a parameter called the 'Early voltage' which is related to this. If you extrapolate the sloping lines backwards towards the Y axis and beyond, they converge at a point which represents a negative Vds. The Early voltage of this FET looks to be in single digits volts - whereas a good bipolar transistor will have an Early voltage in the high tens or hundreds of V.

Contrast this plot with the equivalent from Toshiba :

TK13E25DVds.png (39.32 KiB) Viewed 645 times

You can see there's a slight slope for higher drain currents but we're operating below 1A (Vgs < 3.5V) where the Zout looks impressively high.

[abraxalito]

these toshiba parts been discussed
can they be used as the active parts in a line level speaker crossover transferred into mini class a amps for each speaker driver?

The answer's yes, I am already using a CCS-loaded MOSFET buffer as a filter stage after my DAC's passive filter to EQ away the NOS droop. If you'd like a schematic I can put one up. Essentially build the active filter out of Sallen-Key unity gain buffer stages, each being a 2nd order section.

[nige2000]

these toshiba parts been discussed
can they be used as the active parts in a line level speaker crossover transferred into mini class a amps for each speaker driver?

The answer's yes, I am already using a CCS-loaded MOSFET buffer as a filter stage after my DAC's passive filter to EQ away the NOS droop. If you'd like a schematic I can put one up. Essentially build the active filter out of Sallen-Key unity gain buffer stages, each being a 2nd order section.

all going well ill try this speaker amp and gain stage and then the new version i/v on the tda1387 then the nos stuff, hopefully nos correction for soekris too

i saw that Nelson pass and pure audio project were doing an active line level crossover for open baffle speakers
http://www.pureaudioproject.com/pap-c1- ... en-baffle/
and thought it was interesting

i thought these toshiba parts would be quite suitable to make the crossover and the following amps for speakers with high/mid/low drivers
do you think its could justifiable in comparison to regular speaker crossovers

[jkeny]

I'm nearly finished a differential version of a DAC + headphone amps
Will report on it here when done!

[nige2000]

Differential as in DAC with differential output or two outputs pos and neg then amplied to have push pull outputs to headphones?

Hmm headphone jack has common GND or return that won't work

Output trafo?

All wrong guesses I think

[jkeny]

DAC with differential outputs - a differential version of PCM5102
I rewired my Superlux headphones to be differential so no common ground

[nige2000]

DAC with differential outputs - a differential version of PCM5102
I rewired my Superlux headphones to be differential so no common ground

yea that is what i thought would be optimum

its an awful shame all headphones arent wired differential