Minimum hfe

I've installed on this runs.

OK which linux distribution and which version?

?-)

Use AC bias. It is only audio, so just make a 30 kHz oscillator and bias it ike they used to bias tape. this removes alot of vaiables, and if you give it just a little bit of inductance on the output lik=ne, will be very forgiving on devoce parameters.

Think soft.

Oh, and if Thinksoft hasn't been used for a name yet and you use it you better take care of me or I am comning over there with my gun....

LOL

it like they used to bias magnetic tape. This moots alot of vaiables, and i f you give it just a little bit of inductance on the output line, will be v ery forgiving on devce parameter variations. Think soft. Oh, and if Thinkso ft hasn't been used for a name yet and you use it you better take care of m e or I am coming over there with my gun.... LOL

goes good in debian (tested upto debian 7.5)

if you have a 64 bit debian heed the the message when you try to run wine and install the stuff it requests.

Prolly makes no difference on voltage rating of the power transistor; besides you _did_ say "random".

Pray tell, what the heck does AC bias have to do with the price of rice in China?

Its Avlinux 6.0.1, debian based. But its faulty somehow, won't run anything I install.

Yesterday I tried to get PC no2 running it, but other problems there. Currently working on an old laptop that's sat unused for years, hopefully that will do it.

NT

It fluctuates Peak to Peak?

Jamie

OK fixed an ancient laptop running 98 and got LTspice on it. Will post the schematic when its finished. Trouble is, the thing won't sim! It goes thru its calculations for however long, but doesn't plot anything on the graph, its just a black black area with axes. :/ What is it with computers!

NT

Did you click on the node of interest?

rice in China? "

If it is for audio, the design can be made tolerant of alotmore variance of specs of the transistors.

Unless you want to run pure class B.

be built from random electronic scrap. There's a darlington pair class A s tage where I need to know the minimum hfe we can count on with 95% confiden ce. The darlington pair consists of a small random npn followed by a 2nd tr that must dissipate 1w in hot weather with a small crude heatsink. We can eliminate high voltage trs from the picture.

as it can. Peak tr current is anything upto 1A.

gling to answer that question.

OK, good news. Circuit first:

-------| |-----+-----(_) | | Cin | |e _|c | in 3.3u R2 1k2 +--(_) Tr2 | | |e |

---------------+------------+---------+---o 0v

Spice reckons it works nicely. Quiescent output voltage (bottom of loudspea ker) is 2.8v, which will drop a bit as tr2 gets hot.

Just one thing I'm not following. Spice says it has a V gain of over 6 - I' d expect far less given R1&2. Why so much gain?

Maybe I can answer the question. The signal source is much lower impedance than the feedback/bias resistors, so it rules. It would explain the visible nonlinearity. Adding 47R in series with signal input gives better linearit y.

Much thanks to everyone for raising potential issues. I'll build it IRL whe n I can, check it works sufficiently well in the real world.

Its crude but it works :)

NT

Z = Sqrt(XL^ + 47^) That would be the two components, not the over all Z input.

If you run an AC analyst on it with a signal input and set the sweep 20 to 20Khz you can get the phase error and response across the spectrum.

If you've the resources to recover scrap, you've got the resources to sort the scrap by gain....

RL

........and pin-out.....

RL

I have heard of peak oil, but never about peak rice...

Many a time it is the operator...

There's nothing but scrap. Maybe one could built a meter using LEDs but why make life hard if its not needed? Time is not in generous supply, people work long hours for food.

NT

ias it ike they used to bias tape. this removes alot of vaiables, and if yo u give it just a little bit of inductance on the output lik=ne, will be v ery forgiving on devoce parameters.

Heard of peak dung? :)

NT