Transistor case type - UK Vintage Radio Repair and Restoration Discussion Forum

Neil Purling · 21st Feb 2017, 12:02 pm

Do the shiny black cement dome top transistors actually have a case name?
They are shown in the image below.
https://www.vintage-radio.net/forum/...7&d=1484686230
They are not absolutely reliable components, but I don't know if it is because they represent an early generation of silicon transistors.

MrBungle · 21st Feb 2017, 12:13 pm

Yes - they are TO-106.

They are usually a disc of ceramic with the transistor substrate bonded to it, wired in then sealed over the top.

TBH the most reliable packages are the more modern TO-92 plastic ones. Even more reliable than the TO-18/5/39 cans from experience.

Neil Purling · 21st Feb 2017, 2:12 pm

I don't know why these TO-106 devices often turn rogue: Low-gain, noisy or just dead.
Maybe Far East radio makers used reject components? I have certainly seen devices with no marking whatsoever which I have had to change.
The gear I see them in are generic radios that appeared with several brand names, made in Hong Kong and often sold by market traders or low-end retailers.

paulsherwin · 21st Feb 2017, 2:38 pm

They were widely used for a few years before the TO-92 standards became common. Philips made the BC108 series in TO-106 for a while. I haven't found them particularly unreliable, though you do come across the odd bad one as with any transistor type.

I think TO-106 was the first attempt to make mass market transistors without the expense of a metal encapsulation.

Craig Sawyers · 21st Feb 2017, 2:41 pm

They were fairly frequent back in the day, and included a TO5-sized version too, the TO105. Tektronix used them in various products in the 60's including the 453 'scope. And yes, even top-brand TO105 and TO106 housed devices used to randomly fail.

It was probably down to differential contraction between the ceramic header and the dome-like epoxy blob.

Craig

G6Tanuki · 21st Feb 2017, 6:50 pm

The TO-106 encapsulation was originally developed by Fairchild, as the delivery-platform for their silicon planar technology in the early-1960s.

The resultant transistors were cheap, and had remarkably stable and consistent performance-charactistics when compared to their Germanium predecessors.

Fairchild then went on to release their first series of "Resistor-Transistor Logic" [RTL] ICs based on the same technology - the uL900-series - in the very-early-1960s.

http://semiconductormuseum.com/Museu..._923_Index.htm

which used a larger version of the same encapsulation.

When you consider how old many of these first-generation silicon planar transistors are, does their failure-rate *actually* come out any worse than their contemporary tin-whiskery OC169/170/171 or the AF114-118 series??

21st Feb 2017, 7:50 pm

One of the most attractive packages I think.

Neil Purling · 21st Feb 2017, 9:00 pm

These packages seem to be around mid-late 1970's, from what equipment I have seen.
I wondered what material they were made of as the part with the lead-in wires was matte black & the top shiny stuff.
Was it usual for completely un-marked devices to be used for equipment assembly. Could they be reject or out-of-spec components sold cheap to the trade? I can understand the drive to cut the cost of every unit.

MrBungle · 21st Feb 2017, 9:15 pm

Ceramic with epoxy top

Neil Purling · 21st Feb 2017, 11:30 pm

I think that Craig Sawyers suggests a likely reason why this style of transistors do fail.
Maybe the construction is making slightly higher failure rates.
In one situation one component out of six of the same casing had to be replaced. The others are still fine.

21st Feb 2017, 12:02 pm	#1
Neil Purling Nonode Join Date: Jul 2005 Location: Hull, East Yorkshire, UK. Posts: 2,087	Transistor case type Do the shiny black cement dome top transistors actually have a case name? They are shown in the image below. https://www.vintage-radio.net/forum/...7&d=1484686230 They are not absolutely reliable components, but I don't know if it is because they represent an early generation of silicon transistors.

21st Feb 2017, 12:13 pm	#2
MrBungle Dekatron Join Date: Jun 2016 Location: London, UK. Posts: 3,687	Re: Transistor case type Yes - they are TO-106. They are usually a disc of ceramic with the transistor substrate bonded to it, wired in then sealed over the top. TBH the most reliable packages are the more modern TO-92 plastic ones. Even more reliable than the TO-18/5/39 cans from experience.

21st Feb 2017, 2:12 pm	#3
Neil Purling Nonode Join Date: Jul 2005 Location: Hull, East Yorkshire, UK. Posts: 2,087	Re: Transistor case type I don't know why these TO-106 devices often turn rogue: Low-gain, noisy or just dead. Maybe Far East radio makers used reject components? I have certainly seen devices with no marking whatsoever which I have had to change. The gear I see them in are generic radios that appeared with several brand names, made in Hong Kong and often sold by market traders or low-end retailers.

21st Feb 2017, 2:38 pm	#4
paulsherwin Moderator Join Date: Jun 2003 Location: Oxford, UK Posts: 27,947	Re: Transistor case type They were widely used for a few years before the TO-92 standards became common. Philips made the BC108 series in TO-106 for a while. I haven't found them particularly unreliable, though you do come across the odd bad one as with any transistor type. I think TO-106 was the first attempt to make mass market transistors without the expense of a metal encapsulation.

21st Feb 2017, 2:41 pm	#5
Craig Sawyers Dekatron Join Date: Nov 2010 Location: Oxford, UK. Posts: 4,984	Re: Transistor case type They were fairly frequent back in the day, and included a TO5-sized version too, the TO105. Tektronix used them in various products in the 60's including the 453 'scope. And yes, even top-brand TO105 and TO106 housed devices used to randomly fail. It was probably down to differential contraction between the ceramic header and the dome-like epoxy blob. Craig

21st Feb 2017, 6:50 pm	#6
G6Tanuki Dekatron Join Date: Apr 2012 Location: Wiltshire, UK. Posts: 13,998	Re: Transistor case type The TO-106 encapsulation was originally developed by Fairchild, as the delivery-platform for their silicon planar technology in the early-1960s. The resultant transistors were cheap, and had remarkably stable and consistent performance-charactistics when compared to their Germanium predecessors. Fairchild then went on to release their first series of "Resistor-Transistor Logic" [RTL] ICs based on the same technology - the uL900-series - in the very-early-1960s. http://semiconductormuseum.com/Museu..._923_Index.htm which used a larger version of the same encapsulation. When you consider how old many of these first-generation silicon planar transistors are, does their failure-rate actually come out any worse than their contemporary tin-whiskery OC169/170/171 or the AF114-118 series??

21st Feb 2017, 7:50 pm	#7
Guest Guest Posts: n/a	Re: Transistor case type One of the most attractive packages I think.

21st Feb 2017, 9:00 pm	#8
Neil Purling Nonode Join Date: Jul 2005 Location: Hull, East Yorkshire, UK. Posts: 2,087	Re: Transistor case type These packages seem to be around mid-late 1970's, from what equipment I have seen. I wondered what material they were made of as the part with the lead-in wires was matte black & the top shiny stuff. Was it usual for completely un-marked devices to be used for equipment assembly. Could they be reject or out-of-spec components sold cheap to the trade? I can understand the drive to cut the cost of every unit.

21st Feb 2017, 9:15 pm	#9
MrBungle Dekatron Join Date: Jun 2016 Location: London, UK. Posts: 3,687	Re: Transistor case type Ceramic with epoxy top

21st Feb 2017, 11:30 pm	#10
Neil Purling Nonode Join Date: Jul 2005 Location: Hull, East Yorkshire, UK. Posts: 2,087	Re: Transistor case type I think that Craig Sawyers suggests a likely reason why this style of transistors do fail. Maybe the construction is making slightly higher failure rates. In one situation one component out of six of the same casing had to be replaced. The others are still fine.