Thanks for all your replies gratefully appreciated, although I don't understand all of them lol.

@Bordodynov I like the idea of this. It also keeps the circuit real estate down and keeps it tidy.

Why use switches on transistors when there are stabilizers with on/off input and polarity protection:
http://ww1.microchip.com/downloads/...urrent-Low-Dropout-Regulators-DS20005685B.pdf

Cheers,
Mike.

 

The regulator's circuitry will attempt to produce an output that is 5V more positive than the ground terminal, not the expected 0V.

I dont believe it can produce anything without GND, its not left floating, its left cut off. Its like disconnecting a - negative terminal from your car's battery.

 

I dont believe it can produce anything without GND, its not left floating, its left cut off. Its like disconnecting a - negative terminal from your car's battery.

@iimagine
Unfortunately I am not in a position to test with an actual device at this time. However, the regulator circuitry is like that of the LM317. The basic regulator attempts to (re)produce its 1.25V reference between the output and sense terminals; there is no ground terminal. If a resistor is connected from output to sense, and another resistor is connected from sense to ground, then the division ratio of the resistors sets the final output voltage, with the voltage from output to sense always maintained at 1.25V. As the value of the resistor from sense to ground increases, less current flows through the output-to-sense resistor, requiring that the regulator output voltage increase so that the output-to-sense resistor still has 1.25V across it. The specific device that the TS used simply has the 2 resistors built in to set the output to 5V, with the sense terminal no longer being accessible (replaced by the terminal of the sense-to-ground resistor). If the regulator were an LM7805 (but it is not), operation would be different and your circuit might well work okay--but that is only a guess; the LM7805 is not spec'd to work with no ground connection (the internal circuitry requires ground); the LM317 is spec'd to operate without a direct ground connection (the internal circuitry derives power from the current passing through the device with no dependence on a ground connection). Here's a simple block diagram of an LM317:
View attachment 184557
You can see that ground has no role in its operation. However, I do amend my earlier post: the output is likely to be a few volts less than the input voltage (e.g.12-3=9V), rather than "a bit less than 12V."

 

Yes, there's quite a difference.
Where did you get your Spice Model?

The LTspice simulation below, with this Diodes model, seems to fairly closely follow the data sheet graph.

View attachment 184540

View attachment 184562

I downloaded the Zetex set some years ago. Not sure why there would be so much difference unless they updated teh model since then/ I'll have to try to find the date on the one i used.
Where did you get the model you used? It's a little weird they reversed the month and day in the year in that model statement.

I still see a little discrepancy look at Ic=3 amps and Vc=0.5 instead of Vc=0.6v I am wondering if they used a different temperature.

It's also interesting to do the plot with linear scales as it looks linear too. So it looks like the majority of the Vc Sat voltage comes from the RE and RC values.

 

Where did you get the model you used?

From the Diode website as found here.

I still see a little discrepancy look at Ic=3 amps and Vc=0.5 instead of Vc=0.6v

My sim shows about 0.5V@3A and the data sheet shows about 0.5V@3A.
Which graphs are you looking at?

So it looks like the majority of the Vc Sat voltage comes from the RE and RC values.

I think that's typical of most BJT's saturation voltages.

 

From the Diode website as found here.
My sim shows about 0.5V@3A and the data sheet shows about 0.5V@3A.
Which graphs are you looking at?

I didnt see the other small graduation
I measure 507.07mv on your plot 500 on theirs, but that's good.
A little under 350 on yours, a little over 350 on theirs, not bad.

I think that's typical of most BJT's saturation voltages.

I think they vary a LOT. For example Zetex was known for exceptionally low Vsat on many BJT's due to good processing.
Compare theirs to some common transistors.

 

I think they vary a LOT.

True.
What I meant was, is that the curve of saturation voltage versus collector current usually follows an ohmic relation, at least at higher currents, as you noted for this transistor.

 

True.
What I meant was, is that the curve of saturation voltage versus collector current usually follows an ohmic relation, at least at higher currents, as you noted for this transistor.

Oh ok ha ha. Probably due to impurities.
Yeah so we might wonder why they use a log scale for current.

 

so we might wonder why they use a log scale for current.

Probably to get better resolution at the lower current levels.

Even better might be a log-log plot as below (voltage vertical, current horizontal):
That shows an intrinsic, fairly constant saturation voltage up to about 10mA which then transitions to a nearly linear (Ohmic) slope above about 200mA.

 

Probably to get better resolution at the lower current levels.

Even better might be a log-log plot as below (voltage vertical, current horizontal):
That shows an intrinsic, fairly constant saturation voltage up to about 10mA which then transitions to a nearly linear (Ohmic) slope above about 200mA.

View attachment 184589

Well try a perfectly linear one and see what you think.
Start from maybe 100ua.

 

Well try a perfectly linear one

One what?

 

One what?

linear vs linear plot

 

linear vs linear plot

Like this?
That doesn't show the transition from the nearly constant saturation voltage region to the Ohmic region.

 

Like this?
That doesn't show the transition from the nearly constant saturation voltage region to the Ohmic region.

View attachment 184605

@crutschow
Indeed, your linear plot totally hides the performance at <0.1A. Your point is made; a log plot reveals what is not even hinted at in the linear plot.

 

Like this?
That doesn't show the transition from the nearly constant saturation voltage region to the Ohmic region.

View attachment 184605

I believe that is a hyper critical view. But go ahead and zoom in on that "nearly constant" voltage and see what you get.
The log plot is deceiving in that way. Do a linear plot from 0 to maybe 100ma or something like that.

 

@crutschow
Indeed, your linear plot totally hides the performance at <0.1A. Your point is made; a log plot reveals what is not even hinted at in the linear plot.

The log plot is deceiving. Do a linear plot from 0 to maybe 100ma.
Also notice that there is a 'hint' of non linearity where we see the plot cross the Y axis. It doesnt matter though as you will soon see.
This is where the practical comes into play also.

 

I want to use a PNP Transistor as a switch instead of a mechanical relay.
All circuits are +12v
The switched +12v will be used to switch the transistor on, to allow a +12v to flow into the +5v regulator.

Why do not use circuit, recommended in datasheet?

 

The log plot is deceiving.

I see the log-log plot as showing detail that the linear plot doesn't.
If your opinion is, that detail is useless, fine.
Further discussion on that is pointless.

 

I see the log-log plot as showing detail that the linear plot doesn't.
If your opinion is, that detail is useless, fine.
Further discussion on that is pointless.

Oh well i just thought that looking at such a straight line is informative.
I was looking at the detail below 100ma and i saw that it stays very linear until it gets very low, and after all this is a 3 amp transisatator
So i guess i am not sure if the extra info is good to have it seems like anyone using this would do much higher than 1ma say.
Maybe log for low stuff and linear for high stuff...

 

Devices are not ideal, so some of the EXTRA components deal with the non-idealness.

Some quick things are parasitic capacitance and leakage currents. Bypass capacitors, usually 0.1 uF Ceramic capacitors at the power pins actually help compensate for the parasitic inductance of a PC trace. Without them, the part may oscillate. Pure CMOS chips cannot have their inputs floating.

hat leakage current needs a place to go. The leakage current combined with parasitic capacitance can cause an unconnected gate on a FET to suddenly turn on. Some parasitics are easy to identify, some not.

Take a piece of wire that sees a fan, so the wire can wiggle. That wire generates a really small current. A few pico-amps (1e-12) easily.
When your (me) setting up systems that need to measure near that level, it's something i have to take into account. Fundamentally, the wire is moving in the Earth's magnetic field, so it theoretically can generate a current and it does.

When dealing with high voltage, sharp points can be a problem.

Can wires have 90 degree bends in them. Sometimes yes. Sometimes no.