Common Base Biasing

Yes, that's more or less it. As you say, there are usually other considerations, notably input impedance, bandwidth, and power dissipation. If the source has a lot of shunt capacitance, the layout is crappy or the transistor is really fast, you might need a base resistor to suppress oscillation.

Cheers

Phil Hobbs

--
Dr Philip C D Hobbs
Principal
ElectroOptical Innovations
55 Orchard Rd
Briarcliff Manor NY 10510
845-480-2058

email: hobbs (atsign) electrooptical (period) net
http://electrooptical.net

Phil Hobbs expounded in news:Zr-dneuHGbGRSgTQnZ2dnUVZ snipped-for-privacy@supernews.com:

Thanks Phil. I think I'm finally getting a handle on this, after all these years.

Reading Ch 2 of the Art of Electronics made the BJT biasing of emitter follower, common emitter (and implicitly) common collector configuration biasing easy. I only wish I had known that 30+ years ago.

Everything else I've ever read seemed to make it so complicated with h-parameters etc. Finally some common sense prevails for a simple guy like me.

Warren

Phil Hobbs has noted input impedance as another consideration; this comes from the fact that your input impedance of that emitter is about 25 mV/I_e, and the operating current is most likely decided by a target input impedance. If your need is a low-Z sensing amplifier (like, for a large- capacitance photodiode), this is the cheapest way. Coupling transformers and op amp trickery are the next things to try...

One reason phototransistors have a bad name, is that the transistors aren't connected in common-base to amplify the signal. This kind of amplifier offers amazing improvements in the bandwidth of an optoisolator or photodiode.

whit3rd expounded in news: snipped-for-privacy@glegroupsg2000goo.go oglegroups.com:

Yes of course but if your base is DC connected to ground (as I stated), there isn't much room for flexibility in RE. If OTOH you make the base AC grounded then of course other options are possible.

I've been busy with an audio phase shift oscillator (using one BJT instead of an op amp). Finicky beasts they are.

With higher frequencies they seemed to model as I expected. But I had much difficulty trying to get f down to 2-12 Hz, that I need for a guitar tremolo circuit. It seems that as you move into higher caps or R, you need increasing amounts of gain (and I am aware of the min gain requirement). I think given the amount of trouble, I'm going to use an op amp solution instead for that.

I ran into that when I did my AVR based MIDI module. I had to be careful about the opto isolator used (the part number escapes me now).

Warren

The input resistance generally won't depend much on R_E at all, because the small signal resistance of the transistor's emitter is so very small. Your negative supply would have to be within 25 mV of -(V_BE) for R_E to dominate. Input impedance depends predominantly on collector current.

A phase shift oscillator requires a minimum voltage gain of exactly 29, iirc. (Proving that is an interesting exercise.) If you aren't careful with the impedance levels of your RCs, you could load down the poor BJT and not have enough gain available.

Phototransistors are basically just BJTs with light shone on them, and they're almost always used without feedback. They're completely beta-dependent, and there's no way to extract charge out of the base, so they just keep conducting till all the stored charge gets swept out of the base region. You'd never design a circuit like that with an ordinary BJT, so nobody should be surprised that phototransistors are a bit slow and vague.

They're also very noisy, because almost none of the light gets detected--even with a beta of 300, you still get current transfer ratios of 0.1 or something, which makes them roughly 30 dB noisier than they need to be.

Cheers

Phil Hobbs

--
Dr Philip C D Hobbs
Principal
ElectroOptical Innovations
55 Orchard Rd
Briarcliff Manor NY 10510
845-480-2058

email: hobbs (atsign) electrooptical (period) net
http://electrooptical.net

The CB is non-inverting...

That looks good, I always look at that type of config as voltage divide with a little loss to it. They are good how ever at converting and semi isolation of signals.

Jamie.

Well that would work, where that current is chosen for the GBW desired, or noise figure.

Would that also be the case for the Fairchild FPT100 (or FPT101) where the emitter is a very dinky area in one corner of the die? I think i might be able to find one or two to send if you want to play with them..they seem to be one or two orders of magnitude more sensitive than the just-an-ordinary-transistor (common) types. ..and who sez one cannot play games with the base?

Nobody, but most of the PT packages are 2 pins. At least the SMD ones, and lots of TH others are in LEDs like package...

--
Thanks,
Fred.

Fred Bloggs expounded in news: snipped-for-privacy@o20g2000yqk.googlegr oups.com:

To be clear, I'm not using CB configuration in the phase shift oscillator!! They are two completely independant "projects".

I'm going to post my phase shift oscillator separately for comment. :)

Warren

You can do things with the base, if it's pinned out--which it usually isn't. The primary photocurrent (before beta is applied) is so very small, though, that you can't get any speed anyway.

I can't find a datasheet for the FTP100--Quest has five (5) of them, but no datasheet. In any case, there's nothing preventing someone from building a better phototransistor. You can do it yourself, by wiring a PD from the base to the collector of an ordinary BJT, and you'll have at least 100 times more photocurrent to work with.

Cheers

Phil Hobbs

I sometimes add a b-e resistor to a photocoupler to speed it up. That reduces CTR, which one makes up for by using lots of LED current. The resistor also increases the transistor's breakdown voltage, which is handy in my silly optocoupler-based high voltage amplifiers.

John

Robert Baer expounded in news:LZmdnZBN3shMWQfQnZ2dnUVZ snipped-for-privacy@posted.localnet:

..

In my case, I just used a 6N138 (or similar). You can see it used here:

If you repeat a midi signal too many times through optoisolators without regenerating it fresh, it can get too muddy to use. I only needed to receive it into my ATmega from the keyboard or computer.

Warren

Using a Darlington makes it slower and more vague. Sometimes that's okay in an optocoupler, but we were talking about standalone phototransistors earlier.

It seems like most of the time when I start designing in an optocoupler, I wind up designing it out again. Laser interlocks are one exception.

Cheers

Phil Hobbs

--
Dr Philip C D Hobbs
Principal
ElectroOptical Innovations
55 Orchard Rd
Briarcliff Manor NY 10510
845-480-2058

email: hobbs (atsign) electrooptical (period) net
http://electrooptical.net

Opto-fet SSRs are cool!

I've seen too many of these...

+5 | | R | | in------R-------+ +--------out | | | | A C K > B | E | | | | gnd gnd

as if the opto has some magical ability to eliminate noise, like using garlic to keep vampires away.

John

It's a low pass filter.

So is a resistor and a capacitor. Only cheaper and better controlled.

John

It does have the ability to eliminate noise--its bandwidth is about 100 kHz!

Cheers

Phil

--
Dr Philip C D Hobbs
Principal
ElectroOptical Innovations
55 Orchard Rd
Briarcliff Manor NY 10510
845-480-2058

email: hobbs (atsign) electrooptical (period) net
http://electrooptical.net