Transistor manufacturers - favorites?

Okay, after reading the large "new law" thread, and the whole old vs. new transistor debacle, I was wondering what everyone's opinion is about different transistor manufacturers?

For example, many transistors can be sourced from any of a dozen or so manufacturers - there's obviously going to be differences between them.

For example, I can buy a transistor from STMicro or Fairchild that has the same specs and is often half the price of the same part from OnSemi. I know OnSemi has a good rep and all, but is there any reason to want an OnSemi MJE350 over a Fairchild MJE350, for example?

I'm really at a loss when figuring out which parts to order, as I can choose between Philips, General Semi, STMicro, Fairchild, OnSemi, Central Semi... and the list goes on.

Also, the same thing applies to MOSFETS, as there are often many 2nd sources for IRF devices other than International Rectifier (such as Fairchild)

Any thoughts?
 
Farorite Transistors

It depends on what type of transistor you are buying but these are the best I have used and are used in countless high end designs:

Toshiba: Jfets, small signal BJT, power BJT
Zetex: Mosfets, small signal BJT
International Rectifier: Mosfets, High speed diodes
Hitachi: Mosfets
Harris: Mosfets, High speed diodes
IXYS: Mosfets, High speed diodes

Analog Devices: Op Amps
Burr Brown: Op Amps

H.H.
 
That's quite useful, thanks Harry. :)

For example, however - if I wanted to buy a device such as BC550, MJE340, or other device that has many many sources, how does one know which ones are better than others?

And if one is better than the others, why? Quantitative evidence would be ideal, but anecdotal is useful too.

(I can imagine that some devices may be more rugged than others, despite what the SOA curves may show)
 
Hi Tingyness

you wrote : "For example, however - if I wanted to buy a device such as BC550, MJE340, or other device that has many many sources, how does one know which ones are better than others?"

Thats the problem. You can get BC550 from Philips,Onsemi,Fairchild and Infineon.
If you have chosen a manufacture and wants to run a production you may likely have to stick with the one and only.
The reason is that they do not produce the BJT the same way, and at this point they will never have the same perfomance. So if the are mounted in a audio signal path you could be able to hear a difference.
You have to test when you choose a new source.

If you take something like a SMPS chip called MC34063A made by Onsemi. You can get second source to this one but i have been unlucky. So after choosing another source my currentsense was not correct anymore ... Even when it was calculated from min and max values from the spec. sheet

I hope it shares some light?

Sonny
 
process stability and quality control

Sonny, I guess you put too much trust in the stability of the process. Quite often, the variation within the output of one manufacturer is bigger than the principal difference between two manufacturers.

I don't know how second sourcing works for transistors, but for other Semiconductors, the original manufacturer and the second source will actually either transfer the process including masks or at least check that all relevant (well, you never know what is relevant) parameters are compatible. In other cases, the second source will simply claim compatibility but will often change one letter in the part name. For transistors, there are probably multi-lateral bodies such as JEDEC that decides if parts may bear the same name.

Then there is also the issue of how seriously a manufacturer takes quality control...


Anecdotal evidence: a couple of years ago one could get simple power transistors (BD139/140, BD911/912) or standard regulators (7812) that bore either no brand marking or were made by east block companies like Tesla. One vendor was notorious at the time, in that you could never know what brand you got.

The problems I had:
BD139/140: case was rotten, i.e. contact plate not flat, hfe out of spec
BD911/912: hfe way too low, parts blew easily
7812: parts blew easily

Haven't seen any of those dubious parts in a while, not even from the once notorious dealer. I understand the Czech company Tesla now delivers consistently high quality.

Eric
 
Hi Eric

You could very well be right about me... and trusting the manufacture a bit to much...

;)

This is one reason why we stick to a few brands in our little company (3 person)... So we have ended up using brands who have a high level of qaulity and service ... Like TI,AD, Infineon,Zetex,BC components, Philips ... To name a few.

By the way, Eric what are you doing for living...

We make controls for hydraulic valves.

Sonny
 
Hi Sonny,

I am a physicist by training. Did my PhD in laser optics. Had the previlege to develop lots of electronics for metrology which couldn't be bought at the time.

Now I am with a big optics company. Used to do a lot of research on laser things. Now, I am mostly involved with project management, but I get to trouble shoot experiments occasionally and sometimes even electronics...

Where in Danmark are you situated?


Greetings,

Eric
 
Since I still haven't gotten a really concrete answer on this...

I'm looking to buy some transistors - mostly low/medium power ones - most are older 2n series devices, and I can get most of them made by either OnSemi, Philips, Fairchild, General Semi or STMicro.

General semi, ST and Fairchild tend to be somewhat cheaper than OnSemi or Philips - is there a good reason for this?

Does anyone have any suggestions or comments about any of the above companies with regards to quality, reliability, consistency, etc?

(if it helps, i'm getting some 2n5551, 2n5401, bc546/556/560/550, mje340/350, and a few other types)
 
ThingyNess said:
Since I still haven't gotten a really concrete answer on this...

I'm looking to buy some transistors - mostly low/medium power ones - most are older 2n series devices, and I can get most of them made by either OnSemi, Philips, Fairchild, General Semi or STMicro.

General semi, ST and Fairchild tend to be somewhat cheaper than OnSemi or Philips - is there a good reason for this?

Does anyone have any suggestions or comments about any of the above companies with regards to quality, reliability, consistency, etc?

(if it helps, i'm getting some 2n5551, 2n5401, bc546/556/560/550, mje340/350, and a few other types)

Well .. Fairchild, ST and General Semi produce parts who is equally good to Philips and Onsemi.
And i think fairchild has the widest program off them all. They have nearly everything in european/american BJT.

If they are sold in High Volume the price tends to be lower.

Fairchild is hard to get in denmark ... They will only sell in 10000 unit/order....

I would try them out.

Sonny
 
Hi Jocko,

I feel flattered! But seriously: I do lack systematic training in EE, hence I have many blind patches. For example, I have never designed anything digital with more than a dozen TTL chips. The idea of using rubber logic (i.e. PLDs) is appealing to me, but I've never had the time to venture into that...

And I don't think physicists are generally/necessarily smarter than engineers. I have, however, had EE majors who chose to do their masters externally and who drove me into dispair...

Eric
 
You've actually used a dozen TTL chips at once! Wow, beat me by 8 or so. But then, how many 70 MHz IF amps have you designed.

Part of my liking of physicists goes back to my days in "bootcamp". We had a grad student with a physics degree who knew more than all the idiot EE PhDs there that were trying to corrupt our brains.

I believe your lack of "systematic training" is probably an advantage. (I assume you seen my rantings on how much I criticize our engineering educational system.)

You didn't touch the fiber optic lead-in...........

Jocko
 
And he should have no trouble understanding why fiber optics over a 1 meter single-mode fiber is a joke.

Jocko,
as far as i understood my buddies Manfred and Thomas (both having built extraordiarily good CDPs from the scratch), the evil thing with optical cable is the SPDIF interface driving it. The SPDIF twisting clock and data and the clock generated by the drive and clocking the DAC instead of vice-versa how it should be. Both guys plead the optical cable itself as not guilty.

An optical cable sure does smear the pulses in the time domain but at which length? not at 5m ! Thomas has 4x 5m optical cable between DAC and drive and it is the best digital i ever heard.