Class B distortion

[meow2222]

is nasty crossover distortion. Now, the nfb is taken from the circuit's ou tput terminal, not from the opamp's output, so surely an ideal opamp should correct for the 0.6v Vbe drops. Question is why doesn't it?

owers the output directly from -0.6 to +0.6v

a tr pair, plus 2 nfb resistors. That's it.

the rest of the time.

that's easy, there is no bias network

by people with no testgear or knowledge of how to use it.

on, unless it can be arranged without increasing part count, which seems un likely.

If there's any way I can help it there will be no biasing, no requirement f or thermal tracking, no polyfuse and no anything that isnt essential. Anyth ing and everything will be sacrificed to the god of BOM on this one. I'm ai ming for the 2nd most minimal audio amp here.

My most minimal was just 2 trs, 3 Rs, 2 caps, but class A has limited appli cations. Now I want to go class B for battery use.

I'm aware the problem can be reduced simply by going higher voltage; upping the rails from +-5v to +-15 would reduce it to a third the amount. But I n eed this to be compatible with common speakers, and 15v 0.1A = 150 ohms i sn't.

NT

[Johny B Good]

The problem I'm alluding to (where the supply rails are used to drive the output transistors) has nothing to do with biasing, it's the issue of where that drive current is sourced or sunk from which is the op-amp's output terminal in the case of the second example which uses a 10nF cap when it aught to be a simple wire link (or possibly a low value resistor with the 10nF cap across it). That cct as shown, can't possibly work as it stands.

In essence, you appear to have gone for the first cct sans the biasing network. If your needs can be met without using complementary darlington power transistors, that gives you a 'dead zone' transition crossover region total of about 1.3v Pk to Pk.

A very high speed op-amp and output transistors may well suffice to reduce the crossover spikes to inaudability but I'd expect you'd need a suitable LPF between the output and the speaker terminals to minimise RF emissions from the speaker cabling.

If you need the extra gain of darlington pairs, the dead zone crossover region will be very nearly doubled to something like 2.5v Pk to Pk which is a sizeable transition voltage swing on quieter passages of music playback. There's a very good reason why I regard this technique as being just a little bit "shitty".

The second example which drives the output transistors via the supply rail connections does have the virtue of allowing the op-amp itself to directly supply the load during the very quiet passages without calling upon the assistance of the output transistors avoiding the need for high speed transitions of the 'dead zone' at volume levels where this form of distortion product would be at its most noticable.

Furthermore, the half volt bias reduces the size of the transition spike voltages to around 150 to 200 mV for each be junction involved compared to the 650 to 700mV in the case of your proposed cct.

IMO, your best option would be a design based on using the op-amp's supply rails to drive the output transistors as per the second example (but with the 10nF cap replaced by a shorting link to correct the error).

[Dave Plowman (News)]

Anti-aircraft air rifle?

[damduck-egg]

What are the rules on on spraying slurry using a powerful jet nowdays? Even if it managed to survive flying through the cloud of ordure it might not be the nicest thing to carry home after such an encounter. Or just have an irrigation Gun spraying randomly. A lot of the cheaper ones from Maplin etc won't be too strong and a good water jet would damage them.

G.Harman

[Tim Lamb]

In message , "Dave Plowman (News)" writes

I have .22 rim fire but her majesty's finest get very stressed by *off licence* use.

>

[Tim Lamb]

In message , [e-mail] writes

Nice thought but I'm not allowed to spread high nitrate product between October and February.

[tony sayer]

We had one at school or was it a X12? It did work, around half an hour from coming back from Sinclair's from repair, buts thats about all...

[Ian Jackson]

In message , [e-mail] writes

I think that the simplest answer starts with realising that NFB works best when there's lots of open-loop gain available (ie when there's no NFB). The closed-loop gain is then determined solely by the feedback ratio, and not by the vagaries of open-loop gain (primarily frequency amplitude and phase distortion).

However, in the crossover region, there IS no gain, so the feedback simply doesn't work, and therefore can't correct the crossover distortion kink in the waveform.