# Switching losses in ULN2003 driver

Discussion in 'General Electronics Chat' started by prodigyaj, Dec 20, 2008.

1. ### prodigyaj Thread Starter Active Member

Dec 11, 2007
48
0
Hey all !

I have an application where in the ULN2003 has to drive IR Leds and let a surge current through the IR led ! The high frequency that the uln will have to switch is between 100Khz to 50Khz (10uS to 20uS pulse width). How much voltage voltage loss can be see through the ULN2003. The uln will be given a Vcc of 3.3V and will also have input logic 1 = 3.3V.

No such information is given in the datasheet apart from Toff = Ton = 1uS Max

Kindly help

Regards

2. ### hgmjr Retired Moderator

Jan 28, 2005
9,029
219
Can you draw up a schematic and post it in this thread?

hgmjr

3. ### prodigyaj Thread Starter Active Member

Dec 11, 2007
48
0

Vcc and Input logic 1 are both 3.3V !
Depending on losses I have select appropriate value of the R7 resistor for a fixed 90-100mA current flow

Last edited: Dec 20, 2008
4. ### hgmjr Retired Moderator

Jan 28, 2005
9,029
219
With a Vcc of 3.3V, you probably won't have enough voltage to illuminate two LEDs in series. Can you provide the part number for the IR LED so that we may look into a datasheet for it?

hgmjr

5. ### prodigyaj Thread Starter Active Member

Dec 11, 2007
48
0
Yeah I have worked on it ! I am using SFH4503 (from OSRAM )and at 1.5V it gives me 100mA so total drop cos of IR Led is 3V. Only thing now I have to consider is ULN losses

6. ### hgmjr Retired Moderator

Jan 28, 2005
9,029
219
And what is the value of the series current limiting resistor that you are using?

hgmjr

7. ### hgmjr Retired Moderator

Jan 28, 2005
9,029
219
According to the datasheet, the saturation voltage could be as high as 1.1 volt. This will leave only 2.2 volts for your LEDs. That is going to fall short of what you need for both LEDs.

hgmjr

8. ### prodigyaj Thread Starter Active Member

Dec 11, 2007
48
0
Current limiting resistor will be decided according to the total voltage that I would be left after drop across LEDs and ULN2003 !

Which datasheet are you refereing to ? is it uln2003 or sfh4503 ?
The SFH4503 datasheet i use is here

According to the datasheet at 1.5V I get 100mA ( which is the current I require ) and thus for 2 LED's drop is 3V, so if there would have been no losses in ULN the remaining 0.3V/100mA would give me 3ohm current limiting resistance !
But that was assuming no switching loss , all depends on ULN2003 switching loss

9. ### hgmjr Retired Moderator

Jan 28, 2005
9,029
219
I was referring to the ULN2003 datasheet. Its saturation voltage can be as high as 1.1 volt.

hgmjr

10. ### Audioguru Expert

Dec 20, 2007
11,249
1,349
The max voltage of the LEDs is 1.8V at 100mA. So a ULN2003 might just barely drive one. Of course it can drive two in parallel if each has its own current-limiting resistor.

11. ### hgmjr Retired Moderator

Jan 28, 2005
9,029
219
You could feed the control signal to two inputs at the same time. Then you could drive one LED on each of the two outputs.

This echos what audioguru has suggested.

hgmjr

12. ### prodigyaj Thread Starter Active Member

Dec 11, 2007
48
0
Now does this saturation voltage hold at frequencies of 50Khz to 100Khz ?

13. ### hgmjr Retired Moderator

Jan 28, 2005
9,029
219
You may be able to reach the 50KHz but 100KHz is likely to be a stretch.

hgmjr

14. ### prodigyaj Thread Starter Active Member

Dec 11, 2007
48
0
I dint get you, can you please elaborate ! Saturation voltage does it depend on input frequency ?

15. ### hgmjr Retired Moderator

Jan 28, 2005
9,029
219
Saturation voltage is an indication of how close the collector can come to ground when the device is turned on. It is somewhat frequency dependent.

hgmjr

16. ### prodigyaj Thread Starter Active Member

Dec 11, 2007
48
0
Well I am little confused now ! while designing the PCb i had enquired with many ppl who have worked on this but every used to say 'Assuming no losses' the only difference between them and me being 3.3V and they working at 5V ! but even in a 5V application more than 20% loss cannot be accounted for assuming no loss ! Could you help me out ? Also SFH4503 takes 1.5V and not 1.8V as suggested ( 1.8V is when u keep it on for 20ms and the junction temperature increases and thus increases the junction voltage, here i plan not more than 20uS ) So again the question comes back to voltage drop across ULN2003 at 50Khz to 100Khz

17. ### SgtWookie Expert

Jul 17, 2007
22,201
1,809
You may not be able to use a ULN2003 at that high of a frequency. Internally, it's seven Darlington pairs. Darlington pairs provide high gain, but are very slow when compared to discrete transistors.

Why don't you try connecting up a ULN2003 to a 100kHz square wave like you plan on driving it with, and a pull-up resistor on an output to 3.3v? A wild guess suggests that you will observe about 0.9v at the collector of an output with a 100mA load, or 24 Ohms to 3.3v. Your mileage may vary significantly.

Instead of using a Darlington driver, you might go with discrete transistors like a BC547 or 2N2222 with a 470 Ohm resistor on the base. That way you'll have a good deal more bandwidth, and the saturation voltage may get down around 0.2v.

18. ### prodigyaj Thread Starter Active Member

Dec 11, 2007
48
0
But sir ! The datasheet of ULN2003 mentions that the Ton = 1uS and Toff = 1uS
So can you please clarify , because I understood from that piece of information that 20uS can run just fine !

19. ### Ron H AAC Fanatic!

Apr 14, 2005
7,012
681
You need logic level MOSFETs. Some are specified for Vgs as low as 1.5V. With 3.3V of gate drive, you can get Rds(on) as low as 40 milliohms. This would give you a saturation voltage less than 4mV with 100mA through it.
Example: http://www.diodes.com/datasheets/ZXMN2A02N8.pdf

There are probably others as good or better.

Jul 17, 2007
22,201
1,809