transistor theory problem

[Claude Abraham]

1/10 is used because that takes the transistor out of the active region and into the saturated region where Vces<Vbe, where the B-C junction is forward biased.

But I just explained that 1/10 may be overkill. If the worst case min value of beta is 80, for example, then the derated low temp value is around 50. Using 1/25 assures saturation. The "1/10" value is needed only for lower beta devices. If 1/25 or 1/40 assures saturation under all conditions, wouldn't that be a lot better than using 1/10? The 1/10 value should be used only when necessary. That is my point.

 

[audioguru]

Go to the electronics supply store and ask for a transistor that has more gain than the minimum for it.
They will laugh at you.
You and they don't know if the transistor you buy is a good one or a passable but poor one.

Don't you want to use the worst-case-values so that every single circuit you make works perfectly like every circuit I make?
The datasheets say the base current must be 1/10th the collector current for all of their transistors to saturate pretty well so I use it. I don't gamble with 1/25 or 1/40. I also do not want to pretest every transistor then throw out the ones with passable but low spec's.

 

[Claude Abraham]

Go to the electronics supply store and ask for a transistor that has more gain than the minimum for it.
They will laugh at you.
You and they don't know if the transistor you buy is a good one or a passable but poor one.

Don't you want to use the worst-case-values so that every single circuit you make works perfectly like every circuit I make?
The datasheets say the base current must be 1/10th the collector current for all of their transistors to saturate pretty well so I use it. I don't gamble with 1/25 or 1/40. I also do not want to pretest every transistor then throw out the ones with passable but low spec's.

Dude, you're not one to give lectures. I've seen plenty of data sheets. Many refer to 1/50, 1/25, etc. Some low voltage high gain devices spec assured saturation for 1/100. There is no need to pretest devices. The minimum beta value is on the sheet, as well as temp characteristics. A minimum worst case low temp value can be computed in seconds. No big deal at all.

Some applications can do just fine w/ the 1/10 value. If the driving source can supply the current, then it's not a problem. But if the signal source source specifies a limit on output current, then the 1/10 value may demand more base current than the source can supply. If 1/25, or 1/40 assures saturation under all conditions, then that value should be used. I've been doing this forever. I wouldn't offer advice unless I was sure that it is valid. This may shock you, but you are not the only EE in the world. Nothing personal, but you can afford to lighten up a bit.

 

[audioguru]

Dude, you're not one to give lectures. I've seen plenty of data sheets. Many refer to 1/50, 1/25, etc.

I was talking about "ordinary" transistors that are spec'd to saturate when the base current is 1/10th the collector current.
I know the "typical" curves are much better than the printed max saturation voltage loss but we cannot buy typical transistors.

American transistors (2N3904, 2N4401) are spec'd with 1/10th. European transistors (BC548) are spec'd with 1/20th. I have never seen a 1/50th or 1/100th spec for the max saturation voltage loss of a normal transistor. Maybe you mean a darlington pair of transistors?

The 2N5089 has about the highest hFE that I have seen, 400 to 1200. But its saturation voltage loss is spec'd at 1/10th when some of them are barely saturated (a max of 0.5V).

 

[Claude Abraham]

I was talking about "ordinary" transistors that are spec'd to saturate when the base current is 1/10th the collector current.
I know the "typical" curves are much better than the printed max saturation voltage loss but we cannot buy typical transistors.

American transistors (2N3904, 2N4401) are spec'd with 1/10th. European transistors (BC548) are spec'd with 1/20th. I have never seen a 1/50th or 1/100th spec for the max saturation voltage loss of a normal transistor. Maybe you mean a darlington pair of transistors?
The 2N5089 has about the highest hFE that I have seen, 400 to 1200. But its saturation voltage loss is spec'd at 1/10th when some of them are barely saturated (a max of 0.5V).

No, I was referring to single bjt devices. I design a lot of SMPS & dc motor drives using bjt parts spec'd as "low Vcesat". Please refer to attached data sheet NSS12500. It is a 12V-5A pnp device. At 0-3A of current, Vcesat is spec'd at Ib as 1/100th of Ic. But from 4-5A, Ib is 1/10th of Ic.

If you visit On Semi, & just do a general search under "low Vcesat" bjt devices, you will find more like this one. Obviously, as I stated, there are many applications where 1/10 is necessary, & less base drive would be too risky. But lower voltage bjt devices are often employed for FET/IGBT gate drives. They offer high beta at high current levels. Even w/ a base drive of 1/100, Vcesat remains low, even at 3A of Ic.

Of course when Ic reaches high values approaching device spec'd max, beta droop occurs. Then more base drive is needed to assure saturation. At 1.0A of Ic, the 1/100 value is fine, but at the full 5.0A, 1/10 is recommended. It's easy to understand why.

Anyway, that is all I was getting at. If the 1/10 value is needed, then of course I would never risk desaturation by reducing the base drive. But many high gain low saturation devices fully saturate with much less base drive. If the data sheet says 1/25, 1/40. or even 1/100 assure saturation, why waste power by using 1/10? If Ic = 2.0A, wouldn't you rather use 0.020A base drive, instead of 0.200A? Then again, at 5.0A, I would go w/ the 1/10 value.

If the vendor says 1/10, no less, I don't argue. But that 1/10 value is not universal. I'm sorry if I came across as hostile.

 

[audioguru]

The very unusal power transistor has a fantastically low saturation voltage loss. But I notice that its voltage rating is very low and its size is not much bigger than a speck of ground pepper.