ft tester

[DrStrangelove]

I realise it's not really indicating ft.

It's just a rough & ready bit of kit. Very rough. None of the British made transistors did anything.

I may make an additional lower frequency coil if I get around tuit, though the wafer switch in question only has 3 poles.

Since there's a shedload of said 3 way switches down in the shed (see wot I did there?) I could use two and expand the number of coils to 5.

[Mike Watterson]

Certainly many early UK small signal transistors only did to IF frequencies, hence one of the Bush sets having a valve.
Some UK car radios only used the transistors for audio output and ran the valves with 13.8V HT. A few had transistor inverters replacing vibrator packs.
A lot of the Russian transistors even in late 60s & early 70s had terrible RF performance so they used them for an HT inverter and audio, the rest of the military R/T using Rod Pentodes. Only one non-military radio, for forestry work, was produced. They were also in satellites and the Sputnik (but really that was military too, developing ICBMs).

I wonder how feasible it is to add an attenuator and IC buffer amp in front of the transistor base or after the collector? These tend to be low impedance so would work best at which ever side is lower impedance.

I did use a LW loading coil for a MW aerial as low frequency coil on my Ambit FET based "GDO". It certainly extended the range, but useless as a dip meter. I added a J310 FET buffer and an SMA socket to use the Dip meter as small portable signal Generator from 400 kHz to 200 MHz.
In general the design is poorer sensitivity at lower frequencies and there was a revised version with a totally different oscillator that is much more sensitive as a Dip Meter. Handy to adjust "traps" on aerials.

[Craig]

One of the RF gurus on UK vintage Radio, and an ex Hewlett Packard RF instrument designer, said "Oscillator circuits can use transistors quite close to their Ft value. If well designed, they don't need a great amount of device gain. So it'll be a handy indication."

But most, for example, Germanium transistors generally conked out in the low MHz. The exception being the OC44 family that went to 18MHz. MESA germanium transistors like the AF139 (in a 4 legged TO18) have an fT of 550MHz and a maximum frequency of oscillation of 2.7GHz, which is pretty exceptional for Ge.

Vanilla silicon transistors can get up to the hundreds of MHz, with surface mount devices getting well into the GHz range. Good luck measuring them on an fT tester! Or the AF139 for that matter.

Craig

[DrStrangelove]

Yes it would seem unlikely to be much use for RF transistors.

There's a tin of those magic "unmarked untested" Bi-Pak TO-92s in a tobacco tin next door.

I was disappointed with their low gain before I found they sang away quite happily at 500MHz in my hfe tester.

However I'm quite pleased to have found the motivation to have progressed as far as I have after all this time.

Maybe I'll progress the electric field mill a bit more whilst I'm at it. But I won't hold my breath.

We all know it's not really measuring ft, just giving an indication that the device will do something.

Or in the case of a lot of ancient germaniums: nothing at all because the leads have fallen off.

[ppppenguin]

hfe (current gain) is often misunderstood. Good circuit design should never rely on hfe, except that it must be high enough that there isn't too much base bias current. Of course not all design is good. There may also be exceptions for germanium devices where hfe was often very low. I've also had problems with line output transistors such as BU208 which can have incredibly low hfe and thus need a lot of base drive to get them to saturate.

[Craig]

This is the spec sheet for the Germanium AF139 https://www.silicon-ark.co.uk/datasheets...iemens.pdf . I have some that I bought from Langrex at £10 for 5.

Quite apart from the low frequency specification being on the money on my Peak tester, there is no leakage current. The last gasp of Germanium mesa structure transistors were quite amazing.

But hfe is rather important for RF oscillators.

[Craig]

Goofed - it is fT that is important for RF oscillators, NOT hfe.

[DrStrangelove]

Goofed - it is fT that is important for RF oscillators, NOT hfe.

Don't worry about it: my online goofs would fill a book.

Like designing a double tuned transformer with k=1 then wondering why it was so weird trying to tune it up.

[Mike Watterson]

Goofed - it is fT that is important for RF oscillators, NOT hfe.

Well, both, depending on the Q and overall loss of the circuit.
So a low loss high Q design will oscillate closer to fT. A circuit less well matched to transistor input and output impedance will need more device gain.

There are GaAs transistors, some as low as 1.5V supply, that are up to 400 GHz. LNBs are a cheap domestic product with transistor that have to give gain at 12 GHz for Ku and 22 GHz for Ka (I have one Ka LNB).

[DrStrangelove]

Well I didn't know that: a x100 probe will kill an oscillator.

And poking about with a dvm will spur it back into life.

The 2nd coil is wound and eventually oscillations were achieved.

The marked BCY70 seemed reluctant, however some Bi-Pak unmarked untested could be anything in TO-92 sang quite happily.

And I've got a tobacco tin full of them. ROTF.

I think they were out of a pack of a 1000 I bought some time in 1973.

The only remaining question being: what did I want 1000 unmarked transistors for in the first place.