Toshiba MOSFETs with a twist

[Fran]

you'd need quite a bit of delicacy to handle those.

[nige2000]

you'd need quite a bit of delicacy to handle those.

There about the same size as the fets in the GB amp

I'm optimistic they will pop on these pads easier than they would on correct size pads

I've about ten sot 23 type parts to place per channel for the gain stage
Ill be well practiced by the time I'm done

[nige2000]

richard

i misordered bc857c instead of bc857b will that work?

[nige2000]

thought id start by laying down the transistors and the fets, the two k339s and the bc847b are in so far, wire jumpers are underneath
looks messy but i think it will work

[abraxalito]

i misordered bc857c instead of bc857b will that work?

Sure.

[abraxalito]

That's the build I'm focussing on at the moment. I played around with filters and tried to source some ready-made high value inductors (10mH and up) for the AIF. The best shielded 10mH miniature inductor is a Coilcraft (LPS6235) but they're hard to get here and relatively expensive so I have moved the filter to a low impedance circuit for this prototype. Low impedance loading means smaller valued inductors (I have 22uH right now, there are a few choices) but the fly in the ointment is larger valued caps. I'm trying out X5Rs...

I got it playing properly last night, had an issue with my MOSFET source follower doing the power supply oscillating. Always use a gate stopper resistor! Oscillations sound like a kind of hum but it varies with cable positioning. Anyway good news is that putting the filter between the DAC and I/V stage sounds more natural. There's more ease, less indication that I'm listening to a recording. The X5Rs are slightly microphonic in that tapping the board gives a dull 'thunk'. Other than that, definitely its a keeper as its using cheap and compact Ls and Cs.

One curious thing is I find I no longer need any capacitance on pin7 for bass nirvana. This isn't something brought about by the filter though, it was also the case with the filter after the MOSFET. So I figure the function of the cap is improving the LF PSRR of the output current source. Feeding a virtual short though (under 2ohm in current incarnation) the PSRR becomes irrelevant.

Needing no cap on pin7 makes volume control work so much better - there's no lag.

pocket_dac2.jpg (109.73 KiB) Viewed 373 times

Here's the lash-up - on the board in the foreground is the filter, I/V stages and NOS output correction. The DAC board is behind - 4 * TDA1387 operating in anti-phase, though only outputs from the +ve phase are being used. Over to the right with the green LED is the headamp which consists of K339s in parallel.

[nige2000]

What value for gate resistors?
10-200r?
I guess should be used throughout?

Interesting about the pin 7 cap
I found larger caps a slight negative sq wise on pin 7 in my battery version
Think I left a microfarad film in
Anyway the voltage control would be almost unusable

Assume I used a variable voltage divider?

[abraxalito]

What value for gate resistors?
10-200r?
I guess should be used throughout?

I've not noticed any oscillation problems inside the gain stage. The issue I had was a FET as source follower passing a lot of current (almost 100mA) and having caps hung from its source as additional decoupling. That was what caused the oscillation, in conjunction with the gate being AC grounded (fed from TL431) with some inductance to gnd.

Interesting about the pin 7 cap
I found larger caps a slight negative sq wise on pin 7 in my battery version
Think I left a microfarad film in
Anyway the voltage control would be almost unusable

Assume I used a variable voltage divider?

I still haven't figured out the best way to do volume control - I'm leaning towards having a CPU in there and switching resistors between pin7 and 0V via some MOSFETs. All that's needed is a variable amount of resistance between pin7 and pin4, O/C giving max volume and short giving mute.

[nige2000]

if theres any pics you want me to host just send them on, looks cool

whats the chip under the big purple cap?

[abraxalito]

74HC86 doing buffering and inverting the DATA line to one pair of DACs. Above it is TL431 for its supply.