you can run the mosfets in class a as well. just dial up the idle current. the good thing here is that mosfets don't suffer from secondary breakdown, are more rugged than BJTs, are much easier to drive than BJTs (especially 2n3055).
great topic!
i have a BJT version JLH headphone Amp which is really low effeciency and providing the warmest sound i 've heard,but it's has some defects:needing to be warmed up for 30mins and really very lack of bass(my one is with a 10k pot and 1uf wima coupling cap and 440uf output cap,i calulated the corner fz is enough low),what can i do to modify it for more bass?
i would like to modify your fet version into a portable ear amp,because i think jfet and class ab is more effective than bjt and class a,but after reading your posts above,it seems that it's not so good sound if i bias it into class ab....
thanks a lot ^-^
maxyeah: the MOSFET version is every bit as good sounding as the BJT version but I will leave others to judge. it is particularly suited to drive reactive loads or low impedance loads due to its current capabilities.
your issue may be due to too low of an input impedance. the original jlh doesn't have the best biasing scheme for the input BJT. you may refer to the one I posted earlier.
Also, you may consider using a p-channel MOSFET as the input transistor. Zetex has a line of them in to92 package (ZVP10xx series, or BS250p). Just make sure that Cgs is <200pf, and you use a 110 gate stopper there in case of oscillation.
i am assuming that your headphone amp isn't a portable application.
In that case, a full blown jlh is an overkilled and you can greatly simplify the design.
the following was posted somewhere a few days ago. it used to be called super-linear pair but effectively it is a JLH front end with a resistor load (you can envision the JLH upper output transistor to be a constant current source).
it has an input impedance of 10k (R9), gain of 5x (about r5/r4+1), Q1 idle current of 1.5ma, and M1 idle current of 500ma (6w dissipation). you can pretty much use any transistors - if you do use a high powered MOSFET, use a 110ohm gate stopper to slow it down. it can practically drive any low impedance headphone to hear bleeding levels and beyond.
I also included the swithcad simulation file for you to play around.
you may notice that R1's rating is quite interesting. it is not a standard 22ohm power resistor. It is not.
it is meant to be a 12v 5w miniature halogen light bulb. the kind that you will get from a lighting store. it is not much bigger than a peanut, and looks quite like a vaccum tube, .
I would also suggest that you try out tpa6020 (or ths6012) or similar current feedback amps. they are some of the best chip amps out there and are perfectly suited for (portable or desktop) headphone amp applications.
thank you so much!
but in my application the mini-fetJLH is supplied by 1 or 2 9v ni-mh batteries so that it's able to be portable...without an heavy transformer and regulator circuits with large caps and resistors.
do you think it's possible to driver the JLH by batteries ? actually,my headphone is AKG's K271s,low impedance and less bass....i remember that an amps sourcing lots of current but not gain so much is best for it,then i tried my BJT JLH ,but i am not really satisfied with it.BUT your posts made me recognize that JLHs are various,so i'd better to make a suitable one to feed my need....my best farvorite head amps are portable ones ,espeacially supllied by batteries,and i have succesed in making some part-discrete battery amps.Since i read your posts,the Keyword FET and CLASSAB like meaning "PORTABLE" "BATTERY".aHa....i am just a MINI AMP fever
maxyeah: that can be challenging. you have to competing priorities.
1) portability requires that it has to be efficient. that means that you want to run the mosfet as it works in Class AB.
2) 9v batteries means that you want the saturation voltages on the active devices to be as low as possible. that means that you do NOT want to use mosfets. Mosfet's usually have saturation voltages over 2-3v, and typically I budget for 4-5v for them. so if you drop 2-3v each device, you have lost 4-6v on the two mosfets and your output swings just 3-5v.
i googled your headphone and it has a 91db/mw sensitivity - low for a headphone. at 55ohm, you will need to drive 3+v peak or 7v swing in the output to generate about 100mw (110 db? very loud).
you can certainly find mosfets with low saturation voltages (Zetex's E-line for example) and they come in either to92 or d-pak so they can be made quite small. However, I think in your case, the answer may lie with an IC like tpa6020 or the likes as they are a lot more efficient and don't lose much in the output devices.
楼主
发表于 2007-11-3 20:58
.