[A3rd.Zero]

This is my first try at actually making something up so don't kill me for any stupid mistakes i might have made

, but I am interested in everyones thoughts. Its a loop with a complementary symmetrical Class AB emitter follower instead of the typical buffer.

Thanks for your help and criticism.

Milo

edit: changed title

 

[Budgie]

Are you looking to build this or just the fun of the designing?

Maybe you should add the resistors in and repost it, since it is looking pretty weird without them. (Right now the opamp output is tied to B+ as is the current source, and the outputs have no current limiting, etc.)

There is a cap from B- to ground, but not from B+ to ground?

 

[A3rd.Zero]

ok. Thanks.

Ill work on it. Right now I'm thinking more of just the idea rather than the actual implementation, but I understand that it's a bit unclear.

Milo

 

[jeffreyj]

Quote:

Originally posted by A3rd.Zero ok. Thanks. Ill work on it. Right now I'm thinking more of just the idea rather than the actual implementation, but I understand that it's a bit unclear. Milo

Yeah - "a bit unclear," is a serious understatement... I just about died laughing when I saw the "resistors omitted for clarity" in the schematic because omitting the resistors does anything *but* improve the clarity!

There are some serious issues going on with design, otherwise - the bias for the common emitter stage is wrong and attempting to drag the op-amp output into Class A while leaving the buffer in Class AB seems misguided. Also, I'm pretty sure you need the op-amp to be inverting, not non-inverting, for stability reasons.Take a look at Linear Technology's AN-18 Application Note to get some more info on adding discrete buffers to op-amps.

 

[A3rd.Zero]

I see. After budgie's comments the note seemed funny to me too. Ill look up the application note, thanks. Again this is my first run at this and I really have had no formal electronics education, its just me trying to wade through some good books in my free time. Thanks for the help.

Milo

 

[A3rd.Zero]

Ok so this may be a stupid question but why wouldnt you want to follow an Op amp biased into class A with a class AB output stage?

Milo

 

[00940]

I guess it goes like this :

You believe in the benefit of class A since you biase the opamp. So, why do you go with a "lesser" class for the buffer ?

 

[A3rd.Zero]

sure. makes sense. Correct me if I am wrong but a Class A topo is superior because it introduces no crossover distortion, generaly. AB topos introduce some corssover distortion but have a greater potential "power output". So the Idea is leave the Op amp in Class A so that it dosent introduce distortion and live with the fact that only the output stage introduces distortion very little distortion but gives a greater power output.

Actually after writing this it seems to me that it may be better to have the output stage in Class A and the opamp in Class AB but Im sure that its potential power output is less than that of the previous model.

NOTE: It may be obvious that I am still in the very early stages of understanding this stuff so there are likely great gaps in my arguement. Please expose them!

jeffreyj, I read the app note and it made a little sense to me (sorry). I see how an inverting setup can help controll oscillations but why are the oscillations occuring in the first place? ( I know, "why is the sky blue" but bear with me please). Thanks for your thoughts.

Milo

 

[KTpG]

IMO it would save you some trouble if you just followed the Class A biased opamp with a MOSFET running class A... like an IRF510 in something like a Szekeres amp. Then you are Class A through and through... and it is a tad simpler since you skip the bipolar transistors (which can be a pain to work with) and use only one simple to implement MOSFET.... which could drive speakers if you wanted it to.

 

[kheldar]

I know your looking to design it yourself, but here is a discrete buffer based meta42 thats pretty simple to put together.

http://headwize2.powerpill.org/ubb/s...ys=20&srch=42;

sounds good too =)

 

[jeffreyj]

Quote:

Originally posted by A3rd.Zero sure. makes sense. Correct me if I am wrong but a Class A topo is superior because it introduces no crossover distortion, generaly. AB topos introduce some corssover distortion but have a greater potential "power output". So the Idea is leave the Op amp in Class A so that it dosent introduce distortion and live with the fact that only the output stage introduces distortion very little distortion but gives a greater power output.

In a general, vague sort of way you are right: Class A is nominally superior to Class AB or Class B except when it comes to power output for a given supply voltage/current capability. The real answer, like most things, is much more complex. An interesting book that goes into great detail about Class A, AB and B amplifiers is Douglas Self's "Audio Power Amplifier Design Handbook." He is a tad opinionated and heavy-handed at times, but the book is an excellent reference on solid state audio amplification.


Quote:

jeffreyj, I read the app note and it made a little sense to me (sorry). I see how an inverting setup can help controll oscillations but why are the oscillations occuring in the first place?...

The first three and last four pages of AN-18 are the most important, but there's a lot of information to absorb in them so re-read them until it starts to make sense.

Oscillation occurs in an amplifier circuit when the feedback signal stops opposing the input signal and starts reinforcing it. More technically, it's when the phase shift of the feedback signal is between 330 and 360 degrees relative to the input signal. The rolloff of gain with increasing frequency in an op-amp implies that there is either shunt capacitance or series inductance present, and wherever there is reactance there is phase shift. The trick is to reduce gain to unity before phase shift exceeds 330 degrees (this is a rule of thumb, btw, not a rigorously derived principle). Keep in mind that an inversion is the same as 180 degrees of phase shift and it might be a little more clear as to why it is important at times to use one configuration over the other

BTW - I highly encourage you to pursue designing your own buffer, but I also encourage you to build a couple of estabilished designs for reference, at least!

 

[A3rd.Zero]

OOOOh boy! its like Christmas all over again!

Quote:

In a general, vague sort of way you are right

Hooray !!!! (I know that doesnt really mean that I had any idea what I was talking about but I can live with that).

Quote:

re-read them until it starts to make sense

Re-reading has become my banner ever since I got the Art of Electronics

Thanks for the explination of oscillation it cleared up some things well.

Yeah I know I should build some established designs first (I have built some by the way) but Im just so damn excited to do something even vaguely original, even though I may not have the skills.

Ok well back to the drawing board. Ill try this again in a few months.

Thanks again all you have been a really big help.

Milo

 

[blipblop]

why would anyone omit resistors for clarity. I think he was trying to omit them to preserve some secrecy regarding his design!

 

[A3rd.Zero]

Quote:

I think he was trying to omit them to preserve some secrecy regarding his design!

<evil laugh> muhaaaaaaaa </evil laugh>

yes it is a secret, BECAUSE ITS A BIG LASER!!!!!!

No I omitted them because I had seen a few schems that had resistors inserted but with no value so I figured that I could just not bother drawing them (it was easier because I draw in AutoCad). Also I had not figured out exactly how to implement the idea and was hoping that some one could help me, which everyone did.

BTW: Does anyone use autocad for drawing schems? I know its sort of like using a Porsch as a golf cart but I thought Id ask.

Milo

 

[00940]

for such simple schematics, expressSCH is sufficient, at least for me