P and N characteristics

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
I'm sim'ing a preamp with MOSFET push-pull output stage. The P side's current swing is much larger (although the RMS is the same).

I examined the spec sheets, and indeed the P side needs a higher Vgs to produce the same Ids.

The on-resistance curves are also different, which explains why the P side doesn't current-hog (and, I suppose what makes them complementary?). But this also means that the gate RMS voltages are a bit different. Is this to be expected?

Thanks,
Jeff.
 
Member
Joined 2011
Paid Member
Yes, FETs are much more prone to manufacturing process variations than BJTs.

Every BJT has the same gm-versus-Ic curve. Every NPN, every PNP, every small signal device, every BJT power transistor has the same curve.

On the other hand, gm-versus-Ids in a MOSFET, varies all over the map. It varies with carrier type (hole mobility is less than electron mobility, thus Pchannel MOSFETs have lower gm-versus-Ids than Nchannel MOSFETs). It varies with device geometry. Larger MOSFET die size has greater gm-versus-Ids than smaller die size. It varies with channel doping. It varies with gate oxide thickness. Since these last two things fluctuate in high volume manufacturing, gm-versus-Ids fluctuates too.

If you're going to design with FETs you're going to have to live with huge unit-to-unit variability and with huge differences between Pchannel and Nchannel. You might even have to fit your own SPICE models of the highest current unit AND the lowest current unit, then simulate with all four permutations.
 
Thanks, guys.


I wasn't sure what Nigel meant so I removed the feedback. Wow. The gate voltage delta goes from 5.3~5.6 to 2.2~8.2. I'd have been pulling my hair out if I had started there. ;)

Cheers,
Jeff.

I meant add feedback but now after seeing your circuit I can see it is already there. Its the type of front end transistors that cause the problem of different voltages on the two phases.
 
K170/J74 are not acting as buffer. Maybe the problem is in the mosfet spice models.

I guess the JFETs are performing as a VAS here? (What exactly is a "buffer" anyway? Just something that allows different input and output impedances?)

The SPICE model issue certainly wouldn't be the first time. Although I did get these from Fairchild (well, onsemi.com anyway).

Cheers,
Jeff.
 
Member
Joined 2011
Paid Member
Usually a "buffer" in audio circles, is a noninverting amplifier whose gain is very close to +1.0, and whose input impedance is high, and whose output impedance is low, and whose available output current is big. If you want to drive 10 volts RMS into a 600 ohm transformer (like Scott Wurcer's crazyman modified Pass M2), you're gonna need a buffer.

The "Sallen Key" type of active filter circuit, only uses buffers. So it is easily implemented with one device active circuits: cathode followers. Or emitter followers. Or source followers. Or, if you're Nelson Pass, two device source followers.
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.