Mullard Lockfit transistor failure survey - UK Vintage Radio Repair and Restoration Discussion Forum

ukcol · 2nd Mar 2018, 6:08 pm

By the time (in the dying days of the radio & TV servicing trade) that I had move on to other things, I had seen lots of equipment fitted with Mullard Lockfit transistors. My experience was that their failure rate was very low and so I had no reason to suspect that they had a problem.

My only experience of Lockfit transistors in recent years is of those in sets in my collection and I have not seen any failures at all in the few sets in my collection that employ them.
I gather from other members that Lockfit failure rate is in fact quite high but perhaps not as high as the failure rate of transistors in the "tin whisker" group.

I suspect that my experience of Lockfit Transistors being reliable is because:

1] They have only started to fail in recent years, before which time they were reliable..

2] The number of sets I have in my collection that have Lockfit transistors fitted is very small and as a result my sample population is too small to be of significance.

Judging by the amount of discussion about lockfit failure compared to that of the AF11x series, failure of the former is the less serious problem of the two.

I thought it would be interesting to get some feedback from you folk out there about your experience of Lockfit failure in terms of numbers and of the failure mechanism.

paulsherwin · 2nd Mar 2018, 6:17 pm

The failure rate certainly seems to be accelerating. A decade ago I had never encountered a bad lockfit.

It's difficult to quantify the failure rate as they generally don't fail in a clear and unambiguous manner. The first one I encountered was a BC465 which was causing quite dramatic crackles and changes in current which could easily be seen using a simple DMM. More commonly they just become a bit noisy though. If I find one bad one I tend to change the lot, as it's a cheap and easy job with direct or near equivalents readily available.

The failure mode appears to be completely different from the AF11x tin whisker problem.

G6Tanuki · 2nd Mar 2018, 6:24 pm

I've had them fail in plenty of 1970s-era radios - the likes of Roberts R707s - where they go noisy/low-gain. I don't know the failure mechanism, just that presented with a R707 with low volume and/or noise-through-the-speaker-which-isn't-present-at-the-volume-control my initial response is to replace the Lockfits.

Some Roberts radios (later R606 for example) also use a pair of what I describe as "Fat Lockfits" as the push-pull output transistors - BC465A/464A - with a rather ineffectual heatsink arrangement clamped to the plastic outer cases. These really don't handle prolonged 'turn the volume up to 11' duty well. There's different opinions on what to replace them with - I've in the past used Texas Instruments "TIP"-series TO202 ones (connected using flying-leads) with success; at least they can be bolted to a decent heatsink.

paulsherwin · 2nd Mar 2018, 6:32 pm

Up until now I've replaced BC464/465 pairs with BC327/337s, which are a bit underrated on paper but haven't failed yet. I now have some 2SB772/2SD882s which are a bit meatier, but I haven't had cause to try them yet.

Studio263 · 2nd Mar 2018, 6:34 pm

In my experience they have never been reliable. The "BF" ones seem much worse than the BC ones, but if there's been one in the area of a fault I've always made a beeline for it. GEC colour sets, Pye hybrids (mainly in the decoder), BRC monochrome portables, its always the Lockfit transistors that seem to cause bother. B&O liked them too for a while, the slightly disappointing Beocord 1200 and 1600 open reel recorders contained quite a few. These were odd, as sometimes normally shaped transistors were fitted in one channel and Lockfits in the other. Background noise? Channel imbalance? Replace all the Lockfits and, as if by magic, all the faults disappear. The Beomaster 3000 used some too, in its IF stages. I think they were BF194s, just about the only component I've ever had to replace in this section of the receiver.

I remember being told in the 1990s that the reason they were no good was because the plastic they were made from soaked up water from the air, leading to a reduction in gain. I don't know if this is true, but I don't doubt that they are little horrors which are better out than in!

For those unfamiliar, the BC ones follow the same series as the more familiar types. Just swap the "1" for a "5" and you get a much more sensible alternative, e.g.

BC148 = BC548
BC157 = BC557

ukcol · 2nd Mar 2018, 7:22 pm

Quote:

Originally Posted by Studio263

In my experience they have never been reliable.....

Here is a good case for canvassing experience from lots of sources. I saw lots of the TVs you mention and although I changed plenty of Lockfits it was never enough to suspect a problem.

The same with B&O equipment. I was a service manager for a B&O dealer from 1976 to 1986 and didn't find a particular high rate of Lockfit failure in B&O equipment. I must say though that we didn't stock or sell any B&O R2R machines.

It would be interesting to know the mechanism of failure, it must be something particular to the lockfit package.

paulsherwin · 2nd Mar 2018, 7:47 pm

I suspect it's something to do with the legs, which are inflexible pins rather than wires. Maybe the plastic seal around the pins breaks allowing moisture ingress, or maybe the legs flex slightly eventually causing microscopic internal cracks in the junction.

HamishBoxer · 2nd Mar 2018, 7:54 pm

The good thing is the replacements are pretty cheap too.20 for £2.50.

kalee20 · 2nd Mar 2018, 8:14 pm

I've always thought it surprising that the TO92 epoxy package works so well, but the much larger Lockfit doesn't. With more plastic, I'd have expected better protection, not worse...

Epoxy packages are not hermetic anyway, and allow moisture ingress along the leads. But silicon is pretty immune to moisture damage. In some atmospheric conditions it would be a no-no, but for non-aggressive conditions the TO92 seems fine!

Germanium is attacked slowly by water vapour, which is why these have to be hermetic. It was a lesson learned early on.

paulsherwin · 2nd Mar 2018, 9:10 pm

Quote:

Originally Posted by HamishBoxer

The good thing is the replacements are pretty cheap too.20 for £2.50.

You can do it for a lot less than that if you're so inclined. In radios, the AF stages are just used as general purpose NPN and PNP transistors, and just about any GP type you have to hand is likely to be OK. As an exercise I even replaced the BF194/5s in the RF/IF stages of an R606MB with 2N3904s and it was absolutely fine, though it would be better practice to use a proper VHF type with characteristics more similar to the originals. I normally use MPSH10s or 2SC2999s to replace these.

MotorBikeLes · 2nd Mar 2018, 9:14 pm

I have had quite a few failures, and I think it is purely mechanical. Tap a suspect and expect the fault to get worse or improve, then just replace it.
Les.

Mike Fulton · 2nd Mar 2018, 10:58 pm

I didn't succumb to the 'dying days of the radio & TV servicing trade' like so many others and still have TV and audio repairs past my armpits!!
I do, however, remember having to replace lockfits with the 5 types as a matter of course in the 70s & 80s due to their unreliability. I worked for a Philips ASD who was also a main Hacker & Roberts dealer so we probably saw more than our fair share of the blighters.
Still have some NOS individually blister packaged BF & BC lockfits - don't think I'll use them though

Maarten · 3rd Mar 2018, 2:14 am

Strangely, I'm not aware of the lockfits really causing trouble in continental sets. If a survey is to be held, it might be interesting to mention the type and date code of the transistor that failed, the way in which it failed and the type of set it was mounted in.

Nuvistor · 3rd Mar 2018, 8:14 am

The transistors were not that old in the 70’s when I was in the trade, I don’t remember them giving much trouble then, nothing that would have raised concern. The AF11x range were definitely giving problems with shorts etc, now I know it was tin whiskers, I didn’t then, I just replaced them like for like.

Argus25 · 3rd Mar 2018, 8:26 am

Quote:

Originally Posted by Maarten

Strangely, I'm not aware of the lockfits really causing trouble in continental sets. If a survey is to be held, it might be interesting to mention the type and date code of the transistor that failed, the way in which it failed and the type of set it was mounted in.

I agree, I have not in fact found a single faulty lock fit yet. Though I have found a number of other transistors in the same black resinous package go faulty with noise and erratic collector current in vintage transistor radios from the 1970's. Yet I have not seen this problem in vintage transistors with metal cases filled with liquid resin, say for example like the classic BC107, AF125 or AF178.

My favourite signal transistors are mil spec types with metal cases and resin filled,like 2N2222A, and gold plated leads. I have never seen one fail yet. I remain at best suspicious of the black resin encapsulated transistors and I wonder if microscopic defects in the resin propagate through the junction area damaging the transistor crystal.

When resin encapsulated or "plastic" transistors arrived in the late 1960's they were regarded as a revolution in economy, nobody said anything about reliability. Early Fairchild transistors have a ceramic body with a black resin top, generally these were pretty good.

KeithsTV · 3rd Mar 2018, 9:23 am

I posted a Mullard Technical note on Lockfit transistors here.

https://www.vintage-radio.net/forum/...d.php?t=142907

Keith

G6Tanuki · 3rd Mar 2018, 9:52 am

Quote:

Originally Posted by Argus25

When resin encapsulated or "plastic" transistors arrived in the late 1960's they were regarded as a revolution in economy, nobody said anything about reliability. Early Fairchild transistors have a ceramic body with a black resin top, generally these were pretty good.

Reliability is, to be honest, just fine: I'd guess that over 99% of the Lockfits ever manufactured were still functioning well-enough at the time the equipment they were fitted-to was scrapped. Thing is, just as with AF11x TO7 transistors, wax-capacitors etc us vintage-types are now working with gear that's well down the 'long tail' of service-life so failures that would never have been contemplated by the original equipment-designers start to reveal themselves.

The Fairchild ceramic/epoxy 'jelly-bean' transistors are some of my favourites: same goes for their early-60s 'uL' series of RTL logic-chips. I've only had them fail when I've done something stupid to them.

kalee20 · 3rd Mar 2018, 10:32 am

Quote:

Originally Posted by Argus25

My favourite signal transistors are mil spec types with metal cases and resin filled,like 2N2222A, and gold plated leads. I have never seen one fail yet. I remain at best suspicious of the black resin encapsulated transistors and I wonder if microscopic defects in the resin propagate through the junction area damaging the transistor crystal.

When resin encapsulated or "plastic" transistors arrived in the late 1960's they were regarded as a revolution in economy, nobody said anything about reliability. Early Fairchild transistors have a ceramic body with a black resin top, generally these were pretty good.

I've never encountered a resin-filled 2N2222a, all that I have found necessary to hack open have been full of emptiness. Same with other TO18's.

Yes I like the ceramic package with black epoxy top-filling, it looks really nice.

If there is a latent problem with epoxy-on-silicon (or other things which epoxy fails to protect against), what about the millions of epoxy plastic DIP IC's? One would expect them to be even more vulnerable!

Nuvistor · 3rd Mar 2018, 10:53 am

Quote:

Originally Posted by G6Tanuki

Thing is, just as with AF11x TO7 transistors, wax-capacitors etc us vintage-types are now working with gear that's well down the 'long tail' of service-life so failures that would never have been contemplated by the original equipment-designers start to reveal themselves.

Unfortunately the AF11x series and wax capacitors were failing well within the service life of the equipment but I agree, if the item lasts longer than the service life of the equipment it’s fitted to it has done its job.

On the other hand those AF11x transistors and wax capacitors kept me in employment.

Argus25 · 3rd Mar 2018, 7:21 pm

There are two distinct types of black epoxy material. For example early TTL IC's from the late 1960's made by Texas were a very hard resinous type where the material was like a cured resin under pressure almost looking identical to Bakelite and super temperature resistant, can't mark it with a soldering iron. The IC's were cut apart with a small circular saw. I have some, they are very rare now, I could post a photo later.
Latter the material changed to a softer and lower temperature resin, but still pretty hard like all current packages.
Companies like Texas Instruments, Signetics and Fairchild thought that ceramic packages were superior to plastic. All of their extended temp range mil spec IC's were ceramic not plastic and I'm convinced this it the superior package but it's more expensive.
It might be there are two or more types of epoxy package still in use For example modern BC639's have a hard type like the old lock fits, but some modern IC's have softer material it would appear.

2nd Mar 2018, 6:08 pm	#1
ukcol Rest in Peace Join Date: May 2003 Location: Harlaxton, Lincolnshire, UK. Posts: 3,944	Mullard Lockfit transistor failure survey By the time (in the dying days of the radio & TV servicing trade) that I had move on to other things, I had seen lots of equipment fitted with Mullard Lockfit transistors. My experience was that their failure rate was very low and so I had no reason to suspect that they had a problem. My only experience of Lockfit transistors in recent years is of those in sets in my collection and I have not seen any failures at all in the few sets in my collection that employ them. I gather from other members that Lockfit failure rate is in fact quite high but perhaps not as high as the failure rate of transistors in the "tin whisker" group. I suspect that my experience of Lockfit Transistors being reliable is because: 1] They have only started to fail in recent years, before which time they were reliable.. 2] The number of sets I have in my collection that have Lockfit transistors fitted is very small and as a result my sample population is too small to be of significance. Judging by the amount of discussion about lockfit failure compared to that of the AF11x series, failure of the former is the less serious problem of the two. I thought it would be interesting to get some feedback from you folk out there about your experience of Lockfit failure in terms of numbers and of the failure mechanism.

2nd Mar 2018, 6:17 pm	#2
paulsherwin Moderator Join Date: Jun 2003 Location: Oxford, UK Posts: 27,969	Re: Mullard Lockfit transistor failure survey The failure rate certainly seems to be accelerating. A decade ago I had never encountered a bad lockfit. It's difficult to quantify the failure rate as they generally don't fail in a clear and unambiguous manner. The first one I encountered was a BC465 which was causing quite dramatic crackles and changes in current which could easily be seen using a simple DMM. More commonly they just become a bit noisy though. If I find one bad one I tend to change the lot, as it's a cheap and easy job with direct or near equivalents readily available. The failure mode appears to be completely different from the AF11x tin whisker problem.

2nd Mar 2018, 6:24 pm	#3
G6Tanuki Dekatron Join Date: Apr 2012 Location: Wiltshire, UK. Posts: 14,007	Re: Mullard Lockfit transistor failure survey I've had them fail in plenty of 1970s-era radios - the likes of Roberts R707s - where they go noisy/low-gain. I don't know the failure mechanism, just that presented with a R707 with low volume and/or noise-through-the-speaker-which-isn't-present-at-the-volume-control my initial response is to replace the Lockfits. Some Roberts radios (later R606 for example) also use a pair of what I describe as "Fat Lockfits" as the push-pull output transistors - BC465A/464A - with a rather ineffectual heatsink arrangement clamped to the plastic outer cases. These really don't handle prolonged 'turn the volume up to 11' duty well. There's different opinions on what to replace them with - I've in the past used Texas Instruments "TIP"-series TO202 ones (connected using flying-leads) with success; at least they can be bolted to a decent heatsink.

2nd Mar 2018, 6:32 pm	#4
paulsherwin Moderator Join Date: Jun 2003 Location: Oxford, UK Posts: 27,969	Re: Mullard Lockfit transistor failure survey Up until now I've replaced BC464/465 pairs with BC327/337s, which are a bit underrated on paper but haven't failed yet. I now have some 2SB772/2SD882s which are a bit meatier, but I haven't had cause to try them yet.

2nd Mar 2018, 6:34 pm	#5
Studio263 Octode Join Date: Apr 2007 Location: Hampshire, UK. Posts: 1,578	Re: Mullard Lockfit transistor failure survey In my experience they have never been reliable. The "BF" ones seem much worse than the BC ones, but if there's been one in the area of a fault I've always made a beeline for it. GEC colour sets, Pye hybrids (mainly in the decoder), BRC monochrome portables, its always the Lockfit transistors that seem to cause bother. B&O liked them too for a while, the slightly disappointing Beocord 1200 and 1600 open reel recorders contained quite a few. These were odd, as sometimes normally shaped transistors were fitted in one channel and Lockfits in the other. Background noise? Channel imbalance? Replace all the Lockfits and, as if by magic, all the faults disappear. The Beomaster 3000 used some too, in its IF stages. I think they were BF194s, just about the only component I've ever had to replace in this section of the receiver. I remember being told in the 1990s that the reason they were no good was because the plastic they were made from soaked up water from the air, leading to a reduction in gain. I don't know if this is true, but I don't doubt that they are little horrors which are better out than in! For those unfamiliar, the BC ones follow the same series as the more familiar types. Just swap the "1" for a "5" and you get a much more sensible alternative, e.g. BC148 = BC548 BC157 = BC557

2nd Mar 2018, 7:47 pm	#7
paulsherwin Moderator Join Date: Jun 2003 Location: Oxford, UK Posts: 27,969	Re: Mullard Lockfit transistor failure survey I suspect it's something to do with the legs, which are inflexible pins rather than wires. Maybe the plastic seal around the pins breaks allowing moisture ingress, or maybe the legs flex slightly eventually causing microscopic internal cracks in the junction.

2nd Mar 2018, 7:54 pm	#8
HamishBoxer Dekatron Join Date: Aug 2007 Location: W.Butterwick, near Doncaster UK. Posts: 8,935	Re: Mullard Lockfit transistor failure survey The good thing is the replacements are pretty cheap too.20 for £2.50. __________________ G8JET BVWS Archivist and Member V.M.A.R.S Last edited by HamishBoxer; 2nd Mar 2018 at 7:59 pm.

2nd Mar 2018, 8:14 pm	#9
kalee20 Dekatron Join Date: Feb 2007 Location: Lynton, N. Devon, UK. Posts: 7,088	Re: Mullard Lockfit transistor failure survey I've always thought it surprising that the TO92 epoxy package works so well, but the much larger Lockfit doesn't. With more plastic, I'd have expected better protection, not worse... Epoxy packages are not hermetic anyway, and allow moisture ingress along the leads. But silicon is pretty immune to moisture damage. In some atmospheric conditions it would be a no-no, but for non-aggressive conditions the TO92 seems fine! Germanium is attacked slowly by water vapour, which is why these have to be hermetic. It was a lesson learned early on.

2nd Mar 2018, 9:14 pm	#11
MotorBikeLes Nonode Join Date: Dec 2012 Location: Kirk Michael, Isle of Man Posts: 2,350	Re: Mullard Lockfit transistor failure survey I have had quite a few failures, and I think it is purely mechanical. Tap a suspect and expect the fault to get worse or improve, then just replace it. Les.

2nd Mar 2018, 10:58 pm	#12
Mike Fulton Pentode Join Date: Apr 2006 Location: 7 Miles North of Durham, UK. Posts: 130	Re: Mullard Lockfit transistor failure survey I didn't succumb to the 'dying days of the radio & TV servicing trade' like so many others and still have TV and audio repairs past my armpits!! I do, however, remember having to replace lockfits with the 5 types as a matter of course in the 70s & 80s due to their unreliability. I worked for a Philips ASD who was also a main Hacker & Roberts dealer so we probably saw more than our fair share of the blighters. Still have some NOS individually blister packaged BF & BC lockfits - don't think I'll use them though

3rd Mar 2018, 2:14 am	#13
Maarten Dekatron Join Date: Dec 2007 Location: Haarlem, Netherlands Posts: 4,203	Re: Mullard Lockfit transistor failure survey Strangely, I'm not aware of the lockfits really causing trouble in continental sets. If a survey is to be held, it might be interesting to mention the type and date code of the transistor that failed, the way in which it failed and the type of set it was mounted in.

3rd Mar 2018, 8:14 am	#14
Nuvistor Dekatron Join Date: Aug 2013 Location: Wigan, Greater Manchester, UK. Posts: 9,433	Re: Mullard Lockfit transistor failure survey The transistors were not that old in the 70’s when I was in the trade, I don’t remember them giving much trouble then, nothing that would have raised concern. The AF11x range were definitely giving problems with shorts etc, now I know it was tin whiskers, I didn’t then, I just replaced them like for like. __________________ Frank

3rd Mar 2018, 9:23 am	#16
KeithsTV Nonode Join Date: Aug 2007 Location: South Bradford, West Yorkshire, UK. Posts: 2,573	Re: Mullard Lockfit transistor failure survey I posted a Mullard Technical note on Lockfit transistors here. https://www.vintage-radio.net/forum/...d.php?t=142907 Keith

3rd Mar 2018, 7:21 pm	#20
Argus25 No Longer a Member Join Date: Oct 2016 Location: Maroochydore, Queensland, Australia. Posts: 2,679	Re: Mullard Lockfit transistor failure survey There are two distinct types of black epoxy material. For example early TTL IC's from the late 1960's made by Texas were a very hard resinous type where the material was like a cured resin under pressure almost looking identical to Bakelite and super temperature resistant, can't mark it with a soldering iron. The IC's were cut apart with a small circular saw. I have some, they are very rare now, I could post a photo later. Latter the material changed to a softer and lower temperature resin, but still pretty hard like all current packages. Companies like Texas Instruments, Signetics and Fairchild thought that ceramic packages were superior to plastic. All of their extended temp range mil spec IC's were ceramic not plastic and I'm convinced this it the superior package but it's more expensive. It might be there are two or more types of epoxy package still in use For example modern BC639's have a hard type like the old lock fits, but some modern IC's have softer material it would appear. Last edited by Argus25; 3rd Mar 2018 at 7:36 pm.