UK Vintage Radio Repair and Restoration Powered By Google Custom Search Vintage Radio Service Data

Go Back   UK Vintage Radio Repair and Restoration Discussion Forum > General Vintage Technology > Components and Circuits


Components and Circuits For discussions about component types, alternatives and availability, circuit configurations and modifications etc. Discussions here should be of a general nature and not about specific sets.

Closed Thread
Thread Tools
Old 15th Nov 2011, 9:48 pm   #1
Join Date: Apr 2003
Location: Blackpool, Lancashire, UK.
Posts: 4,089
Default AF11x transistors, 'tin whisker' problems and solutions.

The Germanium AF11x transistors and in particular the AF117 are a constant source of trouble and thus a frequent topic in these forums. As a result, this little guide has been cobbled together as a one-stop reference to the basic ins and outs, dos and don'ts, ifs and buts...and so forth.

Before we get down to reviving that dirty old radio you found at a boot sale last weekend, however, let's find out what this so-called "tin whisker" problem is all about. Take a look at the NASA website where you'll see some excellent images accompanying the text.

Aside and on the subject of dirty old radios found at boot sales, some evidence - or, rather, suspicion - exists to suggest the 'tin whisker' problem is more likely to occur with a radio that's been disused for a very long time. Fact or fiction, it certainly appears to be true that an AF11x fitted radio which has been in regular use is less likely to suffer the problem. Anyway, we'll move on now.

Snipping of screen leads

This is what you'd call a 'quick dodge' and it may or may not work; sometimes it can cause a radio to become unstable with whistles during tuning, as well as the tuning becoming way 'off cue'. That said, it can be a handy method of determining which particular transistor(s) are faulty in a totally dead radio via the expedient of snipping the screen lead of each appropriate transistor in turn and testing the radio after each snip. Sod's Law dictates, however, that the duff transistor will be the last one to have its screen lead snipped. You never know though, after snipping the first lead the radio may burst into life so keep your fingers crossed!

Make sure you snip the correct lead though. On the AF11x types, the screen lead is the central one ie on the axis of the transistor can.

Be aware, too, that the snipping dodge only applies if the short concerns the screen lead. As described via one of the links in the above post, shorts can also occur between other leads; e to b, b to c or even e to c or a combination thereof.

Zapping the short circuit

This involves blowing out, using an electric current, the shorting whiskers but is really only practical when the short involves the screen lead. Short-to-screen is, however, the most common scenario.

The method is to connect together the e, b and c leads (Removing insulating sleeves, if present, and twisting the three leads together is adequate) and moving the screen lead clear to ensure it doesn't touch the now-united e, b and c leads.

Next, connect one lead of a power source to the combined e, b and c leads. The other lead of the power source should then be briefly but repeatedly mated with the screen lead; tapping it on and off, if you like, in the same manner as you would when tapping your F8 key to boot your PC up in Safe Mode.

The presence of shorting whiskers will cause a spark as you apply the power source to the screen lead. When the sparking ceases, you've ridded the transistor of the short circuit.

The power source choice varies between individual people. Some people prefer to discharge a capacitor of, say, 30uF or 50uF across the transistor leads, others make do with just a 9V battery or a series pair to give 18V. The battery method is not always effective though.

I myself use a very ancient but beefy model railway transformer which has three jack sockets for three different outputs. The highest of these spits out 18V at a whopping, in this context, 4A and does the job very nicely. This voltage and current isn't crucial, of course, but it gives an idea of the sort of power which will do the job. Perhaps needless to say, that sort of current will cause the leads the glow red hot and probably kill the transistor completely; hence the need to apply the power source, whatever it may be, to the transistor leads quickly and intermittently until sparking ceases.

It's a fine balance. Too low power and the fix won't work, or will only just break the short circuit to the extent where it will soon 'grow back' again. Too high power risks wrecking the transistor for good. The right amount of power, wisely applied, will remove the whiskers adequately enough to allow the transistor to give a good deal more reliable service.

The soldering iron trick

This dodge involves holding a soldering iron on the transistor can to the point where the heat melts away the shorting whiskers. Time required? up to ten seconds and certainly no more than fifteen (see next paragraph), assuming the iron to be up to temperature. The trick works for 50%-60% of attempts but will only melt the whiskers at their junction with the transistor can, so the fix is likely to only be short lived*. If successful, the transistor will not begin to function again until it's cooled down. Germanium does not actually like heat, but these transistors seem to survive soldering iron attacks alright.

Perhaps needless to say, applying the soldering iron for too long will cause the grease inside the can to vapourise. If this combines with the melting of the soldered joint at the base of the transistor can, the can will eventually been forced off - rather like gases in a bottle of plonk forcing out the cork. This rather conveniently brings us to the next section, which you may as well follow if your soldering iron has gone too far and 'popped the can'.

*(Some sources suggest the whiskers grow from the can inwards which, although unproven as far as I'm aware, makes sense. If true and if the soldering iron dodge works, the transistor should then, in theory anyway, function normally for some considerable time until fresh whiskers grow and once again cast their wicked spells.)

Cleaning out the can

As explained above, the can of the AF11x can be removed by unsoldering. Bear in mind, however, that although the germanium gubbins inside the transistor are surprisingly resilient, excessive heat application can cause irreversible damage.

These transistor cans are filled with silicone grease. Once the can is removed, clear out all grease from the can. This will take most of the whiskers with it. Then, very carefully, remove as much grease as possible from the internals of the transistor. Very gently use of a small, soft and clean artists paint brush is ideal for this. Having done all that, refill with fresh, clean, grease and refit the can.


As a convenient follow-on to the previous section we come to the restuffing option. This is ideal if you wish/need to replace a failed AF11x with something more reliable but desire to retain an appearance of originality.

Remove the can as mentioned in the previous section, thoroughly remove the grease and the transistor internals but make sure there's something remaining onto which the legs of the replacement transistor can attach. Snip the legs of the replacement transistor just enough to allow that component to fit inside the AF11x case, then solder the legs of the replacement to the stubs of the original legs atop the glass base. Finally, refit the original can and, of course, it's not necessary to refill with grease.

Some words of caution now. Before fitting your replacement transistor inside the by-then-empty AF11x, ensure the replacement is in fact in working order. The job is rather fiddly and you don't want to spend time doing the transplant only to find it was a wasted effort. Also, when installing a silicon replacement bear in mind that pinouts can vary according to manufacturer. You don't want to assume a certain silicon transistor is, say, e b c and solder those legs to the e b and c of the original AF11x only to later find the silicon pinout was something else, such as b e c. If you do make this mistake you can, of course, refit the Frankenstein AF11x to allow for the mistake but it can become mighty confusing - especially for any future repairer.

Substitute transistors

Discussed many times in the forums are suitable alternatives for the AF11x series. At one time the use of Silicon substitutes was rigidly dismissed as it was assumed that Silicon (Si) cannot replace Germanium (Ge). Not so, as forum threads and the following information will reveal:
  • Ge OC169/170
These are effectively earlier versions of the AF11x types and as such are 'drop in' replacements. However, they're now difficult to obtain and can suffer the same whisker problem. That said, if you've got some to hand they can be retained as standby replacements, zapped if necessary - as indeed can ex equipment AF11x types.
  • Ge AF125/6/7
For those who want rid of the AF11x types but don't want to go down the Si route, these are the usual choice. Shop around though as prices can vary from reasonable to OK-ish to downright expensive.

The AF12x is physically much smaller than the 11x types and thus can be fiddly to fit. The pinout is totally different to that of the 11x so you must pay attention to what you're connecting to where. Leads are also relatively close together and the original linear layout of the 11x leads may require you to fit the 12x types somewhat contorted. You should therefore ensure there is no possibility of the 12x leads shorting to each other.

Incidentally, reference is made throughout the forums as well as here to 'legs', 'leads' and 'pins'. All refer to the same thing - the wires for connecting the transistor.
  • Ge AF239
There is at least one reference in the forums to this transistor being a suitable substitute for the AF11x types. I cannot, however, comment on the validity, or otherwise, of this. Comparing transistor data, readily available online these days, will reveal the truth.
  • Si BC213/4 types
These two examples, especially the 214, are known to be suitable as AF11x substitutes. They are also cheap to obtain. The most obvious difference with Silicon types compared to AF11x types is the omission of the screen lead
  • Si BF450/3/4
Same comments as above, excepting that the BF45x is a little more expensive and not quite so easy to obtain.

In addition, some of the threads linked to in the next post will suggest other substitutes. One even suggests the Ge OC45 can replace an AF11x although I've not tried this myself.

Silicon transistors in general

Other Silicon types, over and above those few listed above, may be listed in the forums as being suitable, either currently or in the future. Much information on suitability is down to experimentation and experience, as well as trawling through data as mentioned earlier. You could, if feeling experimental, even try virtually any small signal transistor providing it's PNP and not NPN. All transistors mentioned in this thread are, by the way, PNP.

It's worth mentioning that many transistors, especially Si types, may have a suffix attached to their coding. For example, BF450RA. This might be a manufacturer code, an indication of gain* or a variation of pinout. If unsure, ask in the forums or conduct a web search.

*It might be of interest at this point to step back, for a moment, to the Ge types and the AF11x's as a random example, be it AF115, 116, 117 or whatever. These are all essentially the same component at the manufacturing stage, but are then grouped - and thus coded - according to gain. Anyway, moving on......

As with the AF12x types mentioned earlier, fitting Si types can also be fiddly for the same reasons. The legs of Si transistors are often flat, as opposed to round, in cross section. Sometimes it's therefore necessary to slightly open out the mounting holes in the PCB. Use a pin chuck for this and operate from the solder side (track side) of the PCB.

You should further be aware that the pinout of what is ostensibly the same type of transistor can vary. This variation occurs occurs due to different manufacturers and case styles. For example, looking towards the pins of one, say, BF450 with it laying on the flat of its 'D' section case, pinout might be: bottom left e, top centre b, bottom right c. Take another BF450 from a different manufacturer and that layout might be totally different. You can always ask in the forums if you encounter such problems, or check online.

For some mysterious reason which I, personally, cannot explain, on occasions the replacement of a Ge with an Si at the oscillator stage will cause the oscillator to cease functioning. This 'fault' appears to apply to only a few makes/models of radio. That said, don't assume the PCB junctions at the oscillator transistor position are the same as those at the IF positions - they can differ. So double check your work before assuming you've got the dead oscillator problem. If you have indeed got the problem then you're advised to either replace the transistor again (with a known good Ge example) or create a forum thread with the relevant details.

Sometimes a Ge to Si change necessitates adjustments to biasing and/or even alignment. In most cases, however, you'll get away with it albeit, perhaps, with a slight loss of gain. As previously, if you do encounter the need for such adjustments and feel unable to deal with it yourself then please create a thread for the purpose.
Darren-UK is offline  
Old 15th Nov 2011, 9:50 pm   #2
Join Date: Apr 2003
Location: Blackpool, Lancashire, UK.
Posts: 4,089
Default Re: AF11x transistors, 'tin whisker' problems and solutions.

This post with its links, intended as a follow-up to that above, may duplicate some of the information given in post #1. However, rather than sticking to digesting only one person's waffle, it's often interesting to read about the opinions, experiences, solutions and outcomes posted by other members. The links given below therefore lead to just a few existing threads on the subject. As this guide was originally produced back in 2007, many more threads have appeared since the list below was compiled.

To read more about the 'tin whisker' problem see here. Please note that this link more or less centres upon the NASA research as already linked to in post #1.

To read more on the 'repair' of faulty AF11x transistors see here or here.

To read more on the substitution, including with Silicon, of AFxxx transistors see here, here, here or here. The latter link discusses the Silicon BF450/RA.

This thread may also be of interest, especially the posts concerning short circuits not involving the screening ie between e, b and c or any two of those.

As stated in post #1, if any members have a question or problem not covered via any of the information given in this thread, please create a new thread in the relevant forum section.

There are two more points perhaps worth mentioning.

Firstly, with a typical, for the time, AM radio the AF11x transistors, or any other type for that matter, will be visible on the chassis/PCB. However, other radios - in particular but not exclusively - those with VHF/FM - will very likely make use of screened modules. These modules may also contain AF11x transistors. To save yourself a lot of work, and grief, you are advised to check via the service data before commencing your repairs.

Secondly, if you fancy experimenting with different transistors for the replacement of AF11x's yourself, obtain some old wreck of a Bush TR82 series radio (but not, of course, that silly TR82/97 imitation thing). Obviously, make sure it's a version which uses AF11x's though as some used OCxx type transistors and, if you're a transistor newbie, it might be wise to avoid the VTR103 - the VHF/FM version of the original TR82. These TR82's have several advantages for this work; quick and easy access without (hopefully) the need to remove the chassis, metal chassis so no risk of PCB damage from soldering iron heat and, best of all, the transistors are easily installed and uninstalled due to being mounted on little metal posts.

To conclude, this guide is not intended, deliberately or inadvertently, to be the very last word on the subject. That said, myself and the other moderators hope the guide will be of interest and also will, hopefully, prove useful.

Good luck and happy repairing and dewhiskering .
Darren-UK is offline  
Old 11th Feb 2018, 3:35 pm   #3
AC/HL's Avatar
Join Date: Nov 2003
Location: Idle, Bradford, West Yorkshire, UK.
Posts: 5,932
Default Re: AF11x transistors, 'tin whisker' problems and solutions.

See also this post for another method:
Bill, BVWS member
AC/HL is offline  
Closed Thread

Thread Tools

All times are GMT. The time now is 5:51 pm.

All information and advice on this forum is subject to the WARNING AND DISCLAIMER located at
Failure to heed this warning may result in death or serious injury to yourself and/or others.

Powered by vBulletin®
Copyright ©2000 - 2018, vBulletin Solutions, Inc.
Copyright ©2002 - 2018, Paul Stenning.