BWD160A function generator output amplifier

[IconicPCB]

Recently I purchased an old BWS160A function generator whose health was questionable. Price was right and i did not mind.

I reckoned i could find the manual on line and after all it was the seventies technology...how difficult could it be ( don't answer)?

So sure enough everything bar the output amp ( buffer) was OK. The output amp is a unity gain current amplifier designed to provide notionally 50 ohm output impedance.
In fact the only difference between BWD160 and BWD160A models is this buffer ( see the attachment) plus a few tweaks to the rest of the circuit ( looking at the BWD160 manual).

A simple enough affair and a simple fault to locate. Replaced the faulty FET (Q1) and the output started to sing...literally.

The function generator is designed to allow DC  to be added to the signal waveform, the old output DC offset facility.
So while the output is below the 0V DC level it is clean and stable. When the output is shifted so that any part of it becomes positive, high frequency oscillations become superimposed only in the positive portion of the signal.

I have not sprayed it with new transistors yet, I am hoping someone has seen this kind of thing and may shed some light on it.

TIA

[duak]

Is there actually a 1nF cap from the emitter of Q3 to V-?

It's not commonly known and actually hard to believe that an emitter follower can be made to oscillate with certain values of load capacitance.  It's hard to predict so if the 1nF cap should be there, it might be as a suppressor and may be bad.  OTOH, if it's just stuck on, perhaps it was an attempt to cure the oscillation.  If the circuit was revised, maybe the manufacturer discovered this and made some changes.  I beiieve a resistor in series with the base is one solution.

What's happening essentially is that the load capacitance is being transformed into an inductance as viewed from the base and because these elements are in series you can get an oscillation.  If memory serves, Dennis Feucht (formerly of Tektronix) wrote up something on this condition.  Search for "Feucht emitter follower oscillation". 

Cheers,

[IconicPCB]

Hello Duak,

I was aware that th emitter follower is not unconditionally stable.

I might try a ferrite bead on the base lead. It may be sufficient to kill the oscillations.

And thanks for the lead for the article.