Base resistors needed for saturated transistors? 8

[mrkenneth]

Are resistors needed between the transistors and the PIC microcontroller in the diagram below?

Each transistor will be driving a variable load (actually, multiplexed LEDs) from 0 A to 1.25 A in saturation. The LEDs are connected in series to current-limiting resistors, so the current between the collector and the emitter terminals will be not be exceeded. I am worried about whether the base current would destroy the transistor, since there are no resistors... Does the base draw current when the transistor is "off" (Vc = Ve = +5V, while Vb = 0V)?

Thanks in advance!

---------------------------------------------------------
Operation Radiation:

http://www.mrkenneth.com

 

[IRstuff]

Still won't work. The low end of R1 will be tied to about 4V.

If the output has an active pullup, then just because you tell it to turn off does not remove it from the circuit. The drain connection to the p-channel device is still there.

TTFN

 

[sisif]

A simple and cheap solution is to use Darlington PNP's instead of simple transistors. To turn them on, you will need only 3...5 mA (usual gain is over 1K) instead of 25mA - which is the maximum sink current for the PIC output.
To turn them off, I'd suggest two things:
1 - Use the TRIS register to set the port to an input and leave the data register filled with 0's;
2 - You'll still have a leakage current of about 1µA, which gives around 0.12 mA on LED's. Not a lot, but they may still be visible in the dark (you can test it); to avoid this, add some 10K? resistors between base and emitter (R10 in the previous figure).
Resistor R1 (output->base) can be 470-1000 ?, depending of the gain of the darlingtons. For ß=1000 a 680? is a good choice.

 

[mrkenneth]

Thank you for all of the suggestions!

I have looked at some MOSFETs, but found that using a gate voltage of only 3.3 V would result in a very low drain-source current.

In the end, I am using sisif's recommendations of using Darlington transistors. Actually, I plan to use Darlington NPN transistors for the 1.2 A side and some cheap PN2907A PNP transistors for the 150 mA side, because using two Darlington transistors would result in too great of a dropout voltage. The light leakage should not be a nuissance.

Thank you all again!

---------------------------------------------------------
Operation Radiation:

http://www.mrkenneth.com

 

[cbarn24050]

I checkd the microchip site but I cant find the part no you quoted, however all the other pics have a diode between the output pin and vdd which means you will not be able to switch off the darlington transistors.

 

[mrkenneth]

Sorry, the actual part number of the microcontroller is PIC16F687 (datasheet:

http://ww1.microchip.com/downloads/en/devicedoc/41262a.pdf)

You are right about the clamp diode to Vdd and to Vss! I guess I will have to power the PIC from the same 5-V supply as the LEDs.

Thank you cbarn24050 for the reminding me of the diode. (Too bad I can only give each member one star...) I was already starting to design the circuit.

Thanks again to everybody!

---------------------------------------------------------
Operation Radiation:

http://www.mrkenneth.com

 

[IRstuff]

Oh, yeah... In addition to the substrate diode, the I/O pins would have the ESD protection diodes.

TTFN

 

[itsmoked]

Oh, here's that graph I did on the current loading of a PIC's output pin. You might find it useful in your calcs.

Keith Cress
Flamin Systems, Inc.- <

http://www.flaminsystems.com>

- kcress@<solve this puzzle>

 

[mrkenneth]

Thank you itsmoked! Very useful data! Too bad Microchip does not include these characteristics in their 300+ page datasheets. The voltage drop is more than I expected... I guess the curve would be similar for the DIP package and for the low output.

What happened after 25 mA?

Thanks again!

---------------------------------------------------------
Operation Radiation:

http://www.mrkenneth.com

 

[Yashu]

Check out TI's 'power+ logic':

http://focus.ti.com/paramsearch/doc...log&familyId=356&uiTemplateId=NODE_STRY_PGE_T

 

[zeitghost]

<What happened after 25 mA?>

Not a lot, the bond wire melted. )

 

[itsmoked]

That's what I was expecting ... They're like 60AWG..

Keith Cress
Flamin Systems, Inc.- <

http://www.flaminsystems.com>

- kcress@<solve this puzzle>

 

[MacGyverS2000]

It would be nice if MC would put a little star on their datasheet and a footnote mentioning something simple, like "Output voltage will drop approximately 60mV per mA of output current." , but that would be too unscientific

Dan
Owner

http://www.Hi-TecDesigns.com

 

[cbarn24050]

HI Macgyvers, my data sheet for the PIC16F870 has numerous graphs showing the output characteristics for various conditions. I can't speak for the others but I would be suprised if they didn't show them as well.

 

[mrkenneth]

Thank you Yashu! Using those would make my project so much easier and even reduce the cost! One of the TPIC costs less than eight discrete transistors/FETs, and I save the resistors too! Why is the TPIC listed under the category of Temperature Sensors and Control ICs? The ICs seem better suited for the Logic or the Interface category...

It looks like they only offer open-drain devices though, so I guess I would still have to use some PNP transistors or P-channel FETs.

my data sheet for the PIC16F870 has numerous graphs showing the output characteristics for various conditions.

This is the first time that I have seen such graphs from Microchip! All of the other PICs that I have used (PIC16F687, PIC18F4520, PIC18F4550, and others) just say:
"Graphs and tables are not available at this time." under the section DC AND AC CHARACTERISTICS GRAPHS AND TABLES. Shouldn't all of the PICs have similar characteristics?

Thanks all!

---------------------------------------------------------
Operation Radiation:

http://www.mrkenneth.com

 

[cbarn24050]

Well I guess I must have been lucky, you could allways download it to use as a guide untill microchip update the datasheets for your device. I didnt think there were any TPIC devices that would answer your problem, certainly not as cheaply, bear in mind that many of these devices are not made anymore.

 

[LionelHutz]

mrkenneth; have you got your circuit sorted out?

I think you're very close with the last schematic you posted. Just stick a driver IC (the TI part would work) between your PIC and the PNP transistors like you first questioned. All else can stay the same and this will compensate for the voltage difference.

The last display circuit I looked at was connected to a PIC16C77. It used a common anode display (the +ve terminal is common to all segments). There was a MMBT3906 transistor with a 1k base resistor connected to the PIC that switched the positive side. There were 8 - 220ohm resistors from each segment to the port pins for the negative side. This was a 3 segment display that was multiplexed and it worked fine. The PIC and display were both connected to the same 5V power supply.

 

[coconutalley]

Mrkenneth,
I have a suggestion to simplfiy your circuit:

Get rid of all the transistors.
Replace the 2903 (I actually worked on this part a long
time ago when I was at Allegro). It was designed for a printer source driver for the inkjet heads.
If you search

http://www.allegromicro.com

you'd find they have LED drivers that can sink a lot current. They are designed for driving LEDs and relays. They come with serial or parallel inputs.