CD4017 max output question

[EdStraker]

I have a quick question that I'm afraid I think I already know the answer to.
Powering 3.2v 20mA White LED's
I need to connect 4 LED's to each output in effect to simulate an old marquis style sign.
From what I gather on the datasheet, it seems that either way I will be slightly over volt or way over mA on a 9v circuit.

Am I going to be stuck with having to use output BJT's to handle this load? Or am I reading something wrong?

Thanks for any help in advance.

 

[rjenkinsgb]

You could add something like a darlington driver IC? that would give much higher current capability.

A ULN2803 is intended for 5V logic inputs, though you can add series resistors for higher voltages, or a ULN2804 is rated for 6 - 15V input.

The only difference is the built-in base resistor. A 2803 with eg. 4K7 resistors would be fine, if they are easier to get than the 2804.

https://docs.rs-online.com/4ada/0900766b8135fb06.pdf

There are also such as the ULN2003 / 2004 etc., which are similar but with seven drivers each, rather than eight - and often far cheaper.

 

[EdStraker]

The plan was to use 4 outputs of the 4017 each driving four LED's with a 32 LED total array. I have only a passing knowledge of the Darlington Array/Driver and never dove into how to actually use it.

I have entertained using a PICaxe 14M2+ with a Darlington Array for this but it quickly became and issue when trying to code it for a fine speed adjustment and the limited input voltage thus adding yet another component (voltage Reg.) so I have dismissed that way entirely.

Would the Darlington Array be easier/more forgiving to use in this application?

 

[danadak]

If you look at graphs, and package power limitations, and per pin output limitation,
clearly need a buffer.

https://www.ti.com/lit/ds/symlink/cd4017b.pdf?ts=1682597083014&ref_url=https%253A%252F%252Fwww.google.com%252F

A 2N7000 should do the trick with minimal interface :

https://www.st.com/resource/en/datasheet/cd00005134.pdf

If you series the 4 LEDs they will have same current and approximate same brightness. But would need
a 15V supply for the LEDs. Note R value has to be computed for actual application.

If you parallel them you would have to use a R in series with each in order to approximate same current
thru them to get best brightness match. But this allows you to use a lower V supply for the leds.


Regards, Dana.

 

[Tony Stewart]

I have a quick question that I'm afraid I think I already know the answer to.
Powering 3.2v 20mA White LED's
I need to connect 4 LED's to each output in effect to simulate an old marquis style sign.
From what I gather on the datasheet, it seems that either way I will be slightly over volt or way over mA on a 9v circuit.

Am I going to be stuck with having to use output BJT's to handle this load? Or am I reading something wrong?

Thanks for any help in advance.

Your choice of voltage must include battery or ACDC source and thus mAh or Wh. e.g. a typ. Alkaline P3 9V battery is only 550 mAh. and make wise choices.

 

[EdStraker]

If you look at graphs, and package power limitations, and per pin output limitation,
clearly need a buffer.

https://www.ti.com/lit/ds/symlink/c...83014&ref_url=https%3A%2F%2Fwww.google.com%2F

A glance at the datasheet originally I was under the impression that the 2804 could handle up to 200mA per pin. My requirement is 80mA per pin *scratches head*

Your choice of voltage must include battery or ACDC source and thus mAh or Wh. e.g. a typ. Alkaline P3 9V battery is only 550 mAh. and make wise choices.

No, just keeping the voltage as low as possible to avoid using regulators all over the place to drive the timer, 4017, etc. as this is not the only circuit to be connected to source power. I suppose I can bump it up to 12v if need be. Planned on using the ever ubiquitous wall wart.

If you parallel them you would have to use a R in series with each in order to approximate same current
thru them to get best brightness match. But this allows you to use a lower V supply for the leds.

This was the original plan. Which is why I thought that adding BJT's would solve the problem with minimal space requirement without a complete circuit redesign.

 

[danadak]

A 2804 would be fine. To drive 20 mA to 4 LEDs in parallel, each output
driving 4 in parallel, with a R in series with each LED. And use low voltage,
like 5V.

Package dissipation should be fine as P output rated for 1.1V @ 100 mA
which is roughly 80mA x 1.1 V = ~ 88 mW x 8 pins = ~ 704 mW total for
the package.

The R in series with each LED = (Vsupply - Vcesat - Vled) / 20 mA =~
(5 - 1.1 - 3.2) / .02 =~ 35 ohms. Yopu would of course recal all this with
worst case numbers and tolerances of parts, like LED V, R tolerance,
Vsupply tolerance.


Regards, Dana.

 

[rjenkinsgb]

Would the Darlington Array be easier/more forgiving to use in this application?

Yes, just one pin in from each 4017 output you are using & one pin out to drive the LEDs, for each circuit.
Plus a common ground.

Think of each section as a high gain transistor with a base resistor included, that can switch up to around half an amp at anything up to 50V..

 

[EdStraker]

A 2804 would be fine. To drive 20 mA to 4 LEDs in parallel, each output
driving 4 in parallel, with a R in series with each LED. And use low voltage,
like 5V.

Since the 555 and 4017 will both run happily on 9-12v directly if I remember right (15v peak). which would translate roughly into a 7-10v on the 4017 outputs. Would I need additional input resistors on the 2804? Or can it be used as is?

 

[rjenkinsgb]

Would I need additional input resistors on the 2804

A 2804 will work fine at that. A 2803 would need extra resistors.

 

[Tony Stewart]

A glance at the datasheet originally I was under the impression that the 2804 could handle up to 200mA per pin. My requirement is 80mA per pin *scratches head*


No, just keeping the voltage as low as possible to avoid using regulators all over the place to drive the timer, 4017, etc. as this is not the only circuit to be connected to source power. I suppose I can bump it up to 12v if need be. Planned on using the ever ubiquitous wall wart.


This was the original plan. Which is why I thought that adding BJT's would solve the problem with minimal space requirement without a complete circuit redesign.

OK Is there any reason why you cannot use 5V for everything?

This will work with PN2222A 's 80 mA out 1.6 mA ini If nec. adjust 300 Ohms for brightness. from 4017 at 5V only 0.5W on 22 Ohm.

or

All LEDs shud be from same batch for matched Vf.

 

[danadak]

Since the 555 and 4017 will both run happily on 9-12v directly if I remember right (15v peak). which would translate roughly into a 7-10v on the 4017 outputs. Would I need additional input resistors on the 2804? Or can it be used as is?

The 2804 already has input R in it :

Regards, Dana.

 

[EdStraker]

OK Is there any reason why you cannot use 5V for everything?

Well, yes. I have other circuits that need to be powered off the same source for this project.

This will work with PN2222A 's 80 mA out 1.6 mA ini If nec. adjust 300 Ohms for brightness. from 4017 at 5V only 0.5W on 22 Ohm.

This was my original idea to keep the part size at a minimum for space constraints.

The 2804 already has input R in it :

So then that's a yes. Ill have to breadboard these when I get some 2804's and compare which one will fit the bill best.

Thanks for the replies and the information.

 

[Tony Stewart]

The 2804 Darlington drivers may be modeled as Vo= 0.8 + I * 1ohm thus for 4 x20mA=0.08V rise above 0.8V. There is a wide tolerance on the 1 ohm but this is the typical value.

 

[EdStraker]

I just ordered a few of the 2804 still waiting for them to arrive.

 

[EdStraker]

Just received my DA's, minor foul up on my part. I got 2004 instead of 2804. Quick look at the datasheet...seems to indicate this will still work for my project. But out of the question for MCU's
https://www.ti.com/document-viewer/uln2004a/datasheet

Since it is a 9v driven circuit, this should still do the trick?

Since I am not driving inductive loads can I leave Pin 9 floating? Or should I use this anyway?

 

[rjenkinsgb]

Since I am not driving inductive loads can I leave Pin 9 floating? Or should I use this anyway?

You can ignore it, connect it to power, or have a switch to ground to give a "lamp test" function.

 

[EdStraker]

OK, thanks for the input.

 

[AnalogKid]

This will work with PN2222A 's 80 mA out 1.6 mA ini If nec. adjust 300 Ohms for brightness. from 4017 at 5V only 0.5W on 22 Ohm.

View attachment 141282

Are you suggesting that the 2222 be dangle-biased into operating as a linear amplifier; that is, not as a saturated switch?

View attachment 141287

All LEDs shud be from same batch for matched Vf.

Are you suggesting that the variation in Vf from one LED to the next will be so small that the corresponding brightness variation will be ok? Assuming that the LEDs do not have internal current limiting, direct parallel LED connections do not work well.

Hmmm ... That brings up a side question I've never considered before: Is the Vf temperature coefficient of a white LED positive or negative?

ak

 

[Tony Stewart]

Are you suggesting that the 2222 be dangle-biased into operating as a linear amplifier; that is, not as a saturated switch?

I hope you are aware that when Vcb approaches 0 =Vce is already in partial saturation. > I make use of this fact to operate the device with Ic/Ib=50 rather than 150 when Vce>=2V. I expect Vce =0.7 ~ 1V. This is certainly not precision regulation here and some experimentation to the Rb will be necessary due to the high RdsOn <= 300 Ohms. Apologies for a sloppy design suggestion, erring on the side of cheap parts. I would prefer to push no more than 10 mA to each LED and choose higher efficacy which exists. (10 Cd @ 10 mA) in 5mm. ( I have extensive experience on the limits of parallel LEDs using same batch parts.)