new bc109's/108's vs vintage?

[tacobender]

So Im planing on doing another Clyde McCoy clone and I was wondering if its worth the trouble to buy some vintage bc109's/108's, or just get new ones from Mammoth? I'm wondering how much of a difference there is?

[hymenoptera]

I'm curious, too.

I just got some BC108 from Tayda, but haven't tried them in anything yet, and have no idea what I just bought. It was sort of an impulse buy 

[R.G.]

Let's talk about what a BC108 is. Or what any transistor type number is.

There are no physical "BC108" transistors - or any other type number transistors. A transistor type exists only as a set of datasheet specifications, a set of numbers. A datasheet specification is a set of tests, a set of ranges of transistor properties.

When a manufacturer makes transistors, they try to make them all come out with certain properties - breakdown voltages, gain, frequency responses, all those numbers on the datasheet. Then they test them. The ones that meet all the requirements in the BC108 datasheet can be labeled and sold as type number BC108. The ones that don't quite make it are not usually thrown away. They're sold as a type number that is specified on some other datasheet with lesser performance numbers.

So when you get a physical transistor that is stamped "BC108" all that means is that the person/company doing the stamping is assuring you that they meet the tests on the BC108 datasheet at the time the person did the stamping. As we well know, this is worth no more than the person's/company's word. Counterfeits are a worldwide problem these days.

So whether something is a BC108 or not reduces to a question of whether it meets the ranges of numbers on the BC108 datasheet. Or **A** BC108 datasheet. Each manufacturer makes up their own datasheets.

But back at the question of vintage versus modern BC108s, presumably you're wondering how different vintage versus modern BC108s *sound*.
The answer is: there are no numbers on a datasheet that says how a BC108 sounds, so the question of how vintage versus modern BC108s sound has no meaning.
Worse: which manufacturer's "vintage" BCxxx versus which manufacturer's modern BCxxx? They made them in different semiconductor lines, and each semiconductor line has its own quirks, which change from time to time. There are engineers that do *nothing* but tend to the semiconductor process lines to try to keep them doing the same thing all the time.
It requires constant tinkering. This leads semiconductors to be slightly different on different days of the week from the same line. So a Monday BC108 could be different from a Wednesday BC108 from the same manufacturer in the same week, let alone a vintage device being different from a different manufacturer's BC108 fifty years later.

Bottom line: the question of vintage versus modern BC108s, BC109s or any other type number doesn't have any real meaning. A transistor meets published specs for the type number or it doesn't. They may sound different, but that's for reasons not on the datasheets. And anyone with a load of old transistors that were left over from some manufacturing run 50 years ago has a high need to convince you that there's magic in the fact that they're old, so you're going to read a lot of lovingly crafted advertising saying how they don't make them like they used to.

[LightSoundGeometry]

an overall system will function properly as long as the physics and math work out right, e.g. a transistor with same specs will function the same as any other transistor with similar said specs in an overall system?

How can o SI semiconductor sound "harsher" than a GE semiconductor if properties and variables are held constant across both systems like a fuzz box for example?

I ordered four bc108s from tayda, and without looking at my notes for exact numbers, two were in the 80's and two were 400+ for the beta but cannot remember the leakage. I am unsure how the leakage works, and how leakage works differently in SI & GE.

I would like to add upon RG's bottom line; that there is a psychological thing going on in the brain due to legend & myth mainly propagated by advertising. There is some inherent factor with humans staying in touch with the past, and history.. for example, vintage radios pretty much suck but the nostalgia of them and the history of the technology is there. I still listen to ball games etc on a 1960's era transistor radio. Thus, having some vintage parts in the guitar rig may make an individual connect with the overall system and play/perform better even though the parts may not effect tone whatsoever when looking at the music rig through the lens of pure physical science.

* week number two in EET school. So far so good. I can't believe how much I have learned already in such a short period of time.

[R.G.]

an overall system will function properly as long as the physics and math work out right, e.g. a transistor with same specs will function the same as any other transistor with similar said specs in an overall system?

Here's the problems:
1. A transistor type number is only a label that can legitimately be placed on any device fitting within the min to max specifications on the datasheet.
2. What the transistor does for items not mentioned on the datasheet can be **anything** at all and it's still legitmately one example of that type number.
3. Very often, the subjective performance (i.e. sound or tone) produced through a circuit will depend either on the entire circuit's cooperative function, or on transistor characteristics that do not appear on datasheets.

Taken together, this means that while it's possible that some magical transistor properties exist, it's generally not possible to know this ahead of time. Nor was it ever. The vast number of real, original, vintage, no-fooling Fuzz Face units do not sound the way that most modern people think of as the premium Fuzz Face sound. They were simply made by ordering unselected parts, stuffing the boards and shipping the boxes to customers. Today's effects community knows light-years more about what makes a good pedal sound than was common in the 60s-70s "Golden Age"

So the answer is yes and now. Yes, the physics and math are the same. Mother Nature does not grant exceptions for parts made more than 50 years ago. And no, because no two transistors are ever the same in all possible ways, not even two adjacent transistors on a wafer (although that's a whole lot better than two transistors that happen to have been made in the same year!) Worse, the datasheets do not and can not specify all of the audio-critical things needed.

This leads rapidly to a school of design thought that says to design things so that the quirks and inconsistencies of the transistors don't matter much. I had a whole semester on designing circuits to be device-tolerant, so that even a 3:1 or 5:1 change in many of the datasheet numbers would not make the circuit not work properly.

How can o SI semiconductor sound "harsher" than a GE semiconductor if properties and variables are held constant across both systems like a fuzz box for example?

Frequency response is one of those things that might change how it sounds. Data sheets have sneaky numbers, like only specifying a maximum or a minimum. Hfe can be specified as something like 100 minimum. That means a device with a gain of 100,000 meets the spec too. Datasheets never specify MAXIMUM frequency response, only minimum. So one device can be much faster than another.

Leakage is a biggie. Germanium, by its very semiconductor physics, leaks about 1000 times what silicon leaks. That means that it's much harder to turn off, and does so less abruptly. Very often, what matters in "harshness" is how quickly the device clips - the sharpness of the saturation or cutoff edge. That is a convoluted product of the speed, gain, leakage, and external parts for a transistor.

I don't mean to be discouraging - just to let you know that the real issues are numerous, complexly interrelated, and sometimes subtle.

I ordered four bc108s from tayda, and without looking at my notes for exact numbers, two were in the 80's and two were 400+ for the beta but cannot remember the leakage.

The SGS-Thomson datasheet for the BC107/108/108 shows eight different min/typical/max entries for the BC108. The BC108 is further divided into BC108, BC108A, BC108B and BC108C, each with different min/max/typical entries, and two sets, one at 2ma, one at 10uA. The BC108 at 2ma is show as min/max of 110 to 800, a nearly 8:1 range. At 10uA, it's show only as "typically 110". The BC108A is shown as only "typically 90" at 10uA. That means that the maker does not really guarantee that it's gain is any bigger than 1. Or even zero. 
The BC108B is show with a gain at 10uA of minimum 40, typically 120. It might be anywhere between 40 and 800 and the manufacturer is justified in calling it a BC108.

And those are just a portion of the intricate story on gain at a couple of test points.   

I am unsure how the leakage works, and how leakage works differently in SI & GE.

Silicon leaks about 1/1000 of what germanium does, neglecting any possible high leakage from abuse or bad processing.

I would like to add upon RG's bottom line; that there is a psychological thing going on in the brain due to legend & myth mainly propagated by advertising. There is some inherent factor with humans staying in touch with the past, and history.. for example, vintage radios pretty much suck but the nostalgia of them and the history of the technology is there. I still listen to ball games etc on a 1960's era transistor radio. Thus, having some vintage parts in the guitar rig may make an individual connect with the overall system and play/perform better even though the parts may not effect tone whatsoever when looking at the music rig through the lens of pure physical science.

Dead true.

* week number two in EET school. So far so good. I can't believe how much I have learned already in such a short period of time.

Keep it up!

[hymenoptera]

All good stuff. Totally relevant to this discussion, I'm sure. And all very true!

But what I think the OP is asking is for direct experience and *opinions* with the various specific new (and available) parts. For example, did someone try type x from dealer y and found they worked great in pedal z? Possibly no discernable difference in their opinion in circuit z? Or did they stink? Threw 'em out? Or has someone tried manufacturer x and found that some are out of spec, wrong pinout, or...? Or has someone tried the new ones from source y and they all worked, but just didn't sound as (insert gratuitous adjective) as the original ones they painstakingly desoldered from that organ in grandma's basement and then hand-sorted into 17 different hfe ranges?

At least, that's what I'd like to know. If I can stock up on something I'd like to know if I'm wasting my time before I spend the $. For example, I bought some brand-new from China 2n2222a in TO-18 cans hoping to build a white/pink noise gen, negistor drones, stuff like that, only to find that you need the real-deal original parts (or one of many various functional parts) in order to get that funky zener effect that the originals were capable of. That's something that no datasheet can tell you. Oftentimes we're looking for something not written in the datasheet (like the LM308's specific use in a Rat... there's only so many opamps that can be used like that), and any old part just won't do.

So hopefully someone has had some recent and direct experience with the new BC108 parts out there. Certainly they're some modern epitaxial planar part that specs 'close enough', but how much of that is close enough for one project vs another? That's what I'd like to know.

[smallbearelec]

Re the BC10x:

It would be nice if there were a simple go/no-go test of this kind that would tell you whether an NOS BC10x is likely to curl your fur. As others have pointed out, however, there are too many variables involved for this to be possible. What I do guarantee is that I know the provenance of what I sell and warrant that it will be what I say it is--it's not trash that has been re-labeled by a schlock-house in Shenzhen. Much of my stock comes from surplus sources that don't advertise on eBay, can't deal with small orders and so do not want to be household names. We have always taken small orders, will continue to and are happy to be a well-known, continuously growing Cave.

Related issue:

I bought some brand-new from China 2n2222a in TO-18 cans hoping to build a white/pink noise gen, negistor drones, stuff like that, only to find that you need the real-deal original parts (or one of many various functional parts) in order to get that funky zener effect

I'm guessing that you may not have known that I had a part that would be good for this:

http://smallbear-electronics.mybigcommerce.com/2n2712-work-alike-noisemakers/

I had samples tested by one of the synth-DIY people. They are not cheap, but they will give good noise. Unsorted raw stock is also available.

http://smallbear-electronics.mybigcommerce.com/transistors-2n292x-house-number/

When you need something special, please ask. I might know a work-around, or I might know where to find what you want. Then I get a new line item, and you get what you were looking for. The higher price that I ask for obsolete and hard-to-find items reflects a few realities: I have a hearty appetite, and I need to be fed so that I can forage in basements for interesting parts. Likewise, my (9!) associates have to be fed and watered so that they can ship orders promptly.

Regards
SD

[Gus]

 You could buy or build a curve tracer and test the transistors.

[FuzzFanatic71]

I bought some brand-new from China 2n2222a in TO-18 cans hoping to build a white/pink noise gen, negistor drones, stuff like that, only to find that you need the real-deal original parts (or one of many various functional parts) in order to get that funky zener effect

http://smallbear-electronics.mybigcommerce.com/2n2712-work-alike-noisemakers/

I had samples tested by one of the synth-DIY people. They are not cheap, but they will give good noise. Unsorted raw stock is also available.

http://smallbear-electronics.mybigcommerce.com/transistors-2n292x-house-number/

It would be nice if there were a simple go/no-go test of this kind that would tell you whether an NOS BC10x is likely to curl your fur. As others have pointed out, however, there are too many variables involved for this to be possible. What I do guarantee is that I know the provenance of what I sell and warrant that it will be what I say it is--it's not trash that has been re-labeled by a schlock-house in Shenzhen. Much of my stock comes from surplus sources that don't advertise on eBay, can't deal with small orders and so do not want to be household names. We have always taken small orders, will continue to and are happy to be a well-known, continuously growing Cave.

Regards
SD

Yeah your transistors are great. I got some a few months ago for a Uni-Vox Unidrive clone I built. The shipping to Australia was a killer, worked out to be about $5 per transistor. The end result though was well worth it. Would get some from you again in a heartbeat?

[LightSoundGeometry]

I bought some brand-new from China 2n2222a in TO-18 cans hoping to build a white/pink noise gen, negistor drones, stuff like that, only to find that you need the real-deal original parts (or one of many various functional parts) in order to get that funky zener effect

http://smallbear-electronics.mybigcommerce.com/2n2712-work-alike-noisemakers/

I had samples tested by one of the synth-DIY people. They are not cheap, but they will give good noise. Unsorted raw stock is also available.

http://smallbear-electronics.mybigcommerce.com/transistors-2n292x-house-number/

It would be nice if there were a simple go/no-go test of this kind that would tell you whether an NOS BC10x is likely to curl your fur. As others have pointed out, however, there are too many variables involved for this to be possible. What I do guarantee is that I know the provenance of what I sell and warrant that it will be what I say it is--it's not trash that has been re-labeled by a schlock-house in Shenzhen. Much of my stock comes from surplus sources that don't advertise on eBay, can't deal with small orders and so do not want to be household names. We have always taken small orders, will continue to and are happy to be a well-known, continuously growing Cave.

Regards
SD

Yeah your transistors are great. I got some a few months ago for a Uni-Vox Unidrive clone I built. The shipping to Australia was a killer, worked out to be about $5 per transistor. The end result though was well worth it. Would get some from you again in a heartbeat?

you know you will get a part that will work for sure

thanks RG

[tacobender]

Thanks everyone for your input and expertise. But I guess this is just another science vs mojo question ....and also why I put sockets on my board so I can play musical chairs with my transistors:) ....got some old Motorola si caps, and I'm gona pull some more vintage si caps out of an old CB radio. So I guess it all comes down to whatever floats your boat, I'll let you all know what I find when I'm done tinkering, thanks-mike.

[hymenoptera]

I'm guessing that you may not have known that I had a part that would be good for this:

http://smallbear-electronics.mybigcommerce.com/2n2712-work-alike-noisemakers/

I had samples tested by one of the synth-DIY people. They are not cheap, but they will give good noise. Unsorted raw stock is also available.

http://smallbear-electronics.mybigcommerce.com/transistors-2n292x-house-number/

Steve, I've already got those on my short list for my next order from you, thanks!

When you need something special, please ask. I might know a work-around, or I might know where to find what you want. Then I get a new line item, and you get what you were looking for. The higher price that I ask for obsolete and hard-to-find items reflects a few realities: I have a hearty appetite, and I need to be fed so that I can forage in basements for interesting parts. Likewise, my (9!) associates have to be fed and watered so that they can ship orders promptly.

Regards
SD

Good to know, thanks!