NPN transistor that consistently comes in around 200 to 230hfe?

[Gus]

Transistor Hfe can fool you.  Say you "think" you need 200 hfe.  Some transistor testers only test at one collector current(Ic).  This current might not even be close to what your circuit works at.

The different number transistors can be made from different dies and masks and have the same hfe at say one Ic BUT they can have different slopes and curvature with hfe vs Ic.

So to really do this right, find the process number of the transistor die and the transistor number you found you like, there can be other numbers with the same die but sorted and/or having different pinouts and cases.

Then get a curve tracer and curve trace the transistor and compare and select OR build/buy a Transistor tester that tests the hfe at the transistor in circuit IC and then switch the IC up for another point and down for a third pint and then do a three point match. 

I often multi-point match things for a better match.

What meter are you using to measure the transistor.

Transistor circuit IC would be the DC voltage across the collector resistor divided by the collector resistor value. 4VDC and say 10K would be .4ma

[Toney]

  Well I guess however variable his multimeter (or whichever device he use to test hfe), so long as it is the same one he "found" the desired sweet spot, so long as he tests the new transistors with the same, he should be good to go.
I think you have a good point, though. I was surprised just how much hfe was effected with some devices, testing collector current at differing amounts.

How much does temperature affect hfe in a silicon BJT, such as a 2n3904?

[JDoyle]

How much does temperature affect hfe in a silicon BJT, such as a 2n3904?

The 2n3904 is generally considered a switching transistor so the datasheets tend to be slim on that type of information, but on the Rohm datasheet for the 2n3904 (available at datasheetcatalog.com) the bottom graph on page 2 gives you an idea.

At room temp (25 deg. C) and 1ma Ic, the 'typical' 2n3904 has a gain of about 120 (remember the scale of the graph is logarithmic); at 125 deg C it is a little less than 200 and -55 deg. C it is about 75. Which may seem like a non-issue considering the temp range, but when one is trying to find transistors with hFEs in a small range such as this - it matters.

[R.G.]

At room temp (25 deg. C) and 1ma Ic, the 'typical' 2n3904 has a gain of about 120 (remember the scale of the graph is logarithmic); at 125 deg C it is a little less than 200 and -55 deg. C it is about 75. Which may seem like a non-issue considering the temp range, but when one is trying to find transistors with hFEs in a small range such as this - it matters.

This is an important concept. As you note

**There is no such thing as a transistor with a single current gain or Hfe. Hfe is a variable.**

Even if you find a transistor that makes your meter reassure you that the gain is "200" or whatever, if the temperature or the collector current changes, the Hfe will change. If you use a different meter, the same transistor will give a different number. If you measure on a different day, or if your fingers warm the transistor, the Hfe will be a different number.

You can't get a constant Hfe. The best you can do is to get one that measures the way other ones you've liked have measured *at the same temperature *with the same meter *with the same treatment (holding it, gloves, etc,) *immediately after one another. All this does is give you similarity.

[Toney]

Having got used to the variability of Germanium transistors,  I never paid any attention to temperature as a factor with silicon ones.  More things to learn...
I have gone and had a look at the Onsemi 2n3904 datasheet with this in mind. http://www.onsemi.com/pub_link/Collateral/2N3903-D.PDF
They show the Ic hfe rising (nearly doubling) as the current moves from 0.1 mA to 1mA.
Then looking at the static characteristics, they show Ic hfe rising by roughly 1/3 as the temperature moves between 25c and 125c (am ignoring -55c) @1 to 10mA, but that's massively hot.
I would guess that 15c - 45c would cover the field life of any stompbox. I'd wonder what is the variation is with that.
I guess, I'll go and heat up some 2n3904s and look. lol.

[amptramp]

  They show the Ic hfe rising (nearly doubling) as the current moves from 0.1 mA to 1mA.
Then looking at the static characteristics, they show Ic hfe rising by roughly 1/3 as the temperature moves between 25c and 125c (am ignoring -55c) @1 to 10mA, but that's massively hot.
I would guess that 15c - 45c would cover the field life of any stompbox. I'd wonder what is the variation is with that.
I guess, I'll go and heat up some 2n3904s and look. lol.

15C = 59F and I can assure you that if you are touring, you will run into situations where you have to bring all your equipment in from the cold and then try to keep constant characteristics as the equipment warms up during the gig.  In the Toronto area, the temperature range extremes are -15F/-26C to 105F/41C.  Places like Chicago would be more extreme than that.  But the problem, as I have mentioned, is that once you are in from the cold, unless you have a means to heat the unit up to temperature, the characteristics are going to change during the gig.  In a circuit like a Fuzz Face, this is a disaster.

[Toney]

.

15C = 59F and I can assure you that if you are touring, you will run into situations where you have to bring all your equipment in from the cold and then try to keep constant characteristics as the equipment warms up during the gig.  In the Toronto area, the temperature range extremes are -15F/-26C to 105F/41C.  Places like Chicago would be more extreme than that.  But the problem, as I have mentioned, is that once you are in from the cold, unless you have a means to heat the unit up to temperature, the characteristics are going to change during the gig.  In a circuit like a Fuzz Face, this is a disaster.

Oh right I am in Australia -15c doesn't seem real to me.
Did you mean a silicon FF?
On a related note, I was measuring some electros they other day, just mucking about really, and I noticed them swing through their full +/- 20 when going from inside, more even. They were 10uF's that started of cold at ~ 8uF and ended up at ~10.5uF, in the sun 15 minutes  later.
This would obviuosly have a decent cumulative effect going from cold van to on-stage conditions too.

[amptramp]

Some manufacturers put out data sheets on their electrolytics and when you look at them, you wonder how you could ever expect anything that variable to work at all.  The capacitance appears to peak at 25C with sharp dropoffs in some cases to half value at 0C and 70C.  The ESR (equivalent series resistance) has a strong negative tempco on some of them, although some show a minimum near room temperature and an upward trend as you go away from that.  You may not be able to eliminate electrolytics, but you can specify them so that the worst-case characteristics are adequate.

Usually, electrolytics determine a low frequency turnover or circuit decoupling.  Selecting a very low frequency with other parts of the circuit llimiting response to frequencies well above that will permit the use of electrolytics with variable characteristics.  Paralleling bypass electrolytics with ceramic or film capacitors will help tame the frequency and value dependencies.

Electrolytics can be designed in so that their characteristics are not important.  You cannot say the same about transistors used in discrete distortion circuits like the Fuzz Face.  All variations are significant and alter the output sound significantly.

[stringsthings]

Oh right I am in Australia -15c doesn't seem real to me.
Did you mean a silicon FF?
On a related note, I was measuring some electros they other day, just mucking about really, and I noticed them swing through their full +/- 20 when going from inside, more even. They were 10uF's that started of cold at ~ 8uF and ended up at ~10.5uF, in the sun 15 minutes  later.
This would obviuosly have a decent cumulative effect going from cold van to on-stage conditions too.

if you take a look at the datasheets on your average capacitor, they come with a +- capacitance tolerance .... ( for example:  10uF -10%, 30% ) ... and they come with an operating temperature range ( for example: -55C to 105C ) ... as well as voltage rating ... and like most consumer products, you get what you pay for ... ( i.e. better specs = larger price ) .... so your choice of capacitor, resistor, transistor, etc. comes down to good-old economic factors ....