[iso] Heatsyncs for ICs?

8 views
Skip to first unread message

mike k

unread,
May 17, 2017, 1:10:06 PM5/17/17
to Physical Computing at ITP, nycresistormi...@googlegroups.com
Hey electron-lovers,

I'm helping a friend with her circuit, which uses a series of ULN2004s to brighten and dim DC incandescent bulbs (rated at 2A / 12V).  We are running the 7 darlingtons in parallel for a total of 3.5A in available current (there was no chart in the datasheet for max power or a graph of max current at different voltages).

Sometimes a bulb seemingly latches on and smokes the IC.  I'm assuming it's a heat/current draw issue.  Are there higher rated ICs we can use?  All of the darlington arrays I find are 500mA per channel.  If it's a temperature issue, can we paste a heat-sync onto the IC?  


Best,
Mike

--
Mike Kelberman
Interaction Technologist

James Marquardt

unread,
May 17, 2017, 1:32:06 PM5/17/17
to NYCResistor:Microcontrollers, phys...@lists.nyu.edu
Odd...the datasheet tells you on the first page the highest voltage, and on page 4 it says that the maximum PEAK current is 500ma per driver, not 500ma continuous.

http://www.ti.com/lit/ds/symlink/uln2004a.pdf

I would use a MOSFET to do this.  They have several that can connect directly to an Arduino through a 10 ohm resistor and have a very low Rds which in turn keeps the heat to a minimum.

Jim

James Marquardt

unread,
May 17, 2017, 1:34:11 PM5/17/17
to NYCResistor:Microcontrollers, phys...@lists.nyu.edu
Here is the one I use for this purpose of Microcontroller control:

http://www.infineon.com/dgdl/irlb8743pbf.pdf?fileId=5546d462533600a4015356605d6b2593


Jim

On Wednesday, May 17, 2017 at 1:10:06 PM UTC-4, mike k wrote:

Anders Nelson

unread,
May 17, 2017, 1:35:36 PM5/17/17
to nycresistormi...@googlegroups.com, Physical Computing at ITP
Why not just use a $0.10 MOSFET?:


For your voltage range (12v) there are many MOSFETS available that are physically small, can handle high currents and you can switch them using logic voltage, i.e. 3.3v.


=]




--
You received this message because you are subscribed to the Google Groups "NYCResistor:Microcontrollers" group.
To unsubscribe from this group and stop receiving emails from it, send an email to nycresistormicrocontrollers+unsub...@googlegroups.com.
To post to this group, send email to nycresistormicrocontrollers@googlegroups.com.
Visit this group at https://groups.google.com/group/nycresistormicrocontrollers.
For more options, visit https://groups.google.com/d/optout.

jpbar...@aol.com

unread,
May 17, 2017, 4:26:18 PM5/17/17
to nycresistormi...@googlegroups.com
With transistors (including darlingtons) it's important to look at VceSat which is around 1V.  If the transistor were a perfect switch, VceSat would be 0V.
At 3.5A you will be dissipating 3.5W.  A good heatsink may help you get by.  You could try paralleling another ULN2004, but all transistors may not share current equally.  A single part is always best. 

The FET that James suggested would be perfect.  It seems to also have a "logic level" input because it states an Rdson at 4.5V.

There are some low VceSat transistors, but it is this characteristic that makes FETs a better choice for many applications.

Joe
--
You received this message because you are subscribed to the Google Groups "NYCResistor:Microcontrollers" group.
To unsubscribe from this group and stop receiving emails from it, send an email to nycresistormicrocon...@googlegroups.com.
To post to this group, send email to nycresistormi...@googlegroups.com.

Anders Nelson

unread,
May 17, 2017, 5:29:46 PM5/17/17
to nycresistormi...@googlegroups.com
Mmm, are you sure your math is correct? I believe the heat dissipation (at least with FETs) relies on RDS(on). For my aforementioned nFET with an RDS(on) of 49mΩ and power equaling (i^2)(r), that equates to 0.6W when carrying 3.5A.

Do correct me if I'm wrong though! I learned only enough semiconductor physics to pass the final and keep my projects from exploding.

=]


On Wed, May 17, 2017 at 1:26 PM, jpbarbetta via NYCResistor:Microcontrollers <nycresistormi...@googlegroups.com> wrote:
With transistors (including darlingtons) it's important to look at VceSat which is around 1V.  If the transistor were a perfect switch, VceSat would be 0V.
At 3.5A you will be dissipating 3.5W.  A good heatsink may help you get by.  You could try paralleling another ULN2004, but all transistors may not share current equally.  A single part is always best. 

The FET that James suggested would be perfect.  It seems to also have a "logic level" input because it states an Rdson at 4.5V.

There are some low VceSat transistors, but it is this characteristic that makes FETs a better choice for many applications.

Joe


-----Original Message-----
From: mike k <thatsfa...@gmail.com>
To: Physical Computing at ITP <phys...@lists.nyu.edu>; nycresistormicrocontrollers <nycresistormicrocontrollers@googlegroups.com>
Sent: Wed, May 17, 2017 01:10 PM
Subject: [NYCR:Microcontrollers] [iso] Heatsyncs for ICs?


Hey electron-lovers,

I'm helping a friend with her circuit, which uses a series of ULN2004s to brighten and dim DC incandescent bulbs (rated at 2A / 12V).  We are running the 7 darlingtons in parallel for a total of 3.5A in available current (there was no chart in the datasheet for max power or a graph of max current at different voltages).

Sometimes a bulb seemingly latches on and smokes the IC.  I'm assuming it's a heat/current draw issue.  Are there higher rated ICs we can use?  All of the darlington arrays I find are 500mA per channel.  If it's a temperature issue, can we paste a heat-sync onto the IC?  


Best,
Mike

--
Mike Kelberman
Interaction Technologist

--
You received this message because you are subscribed to the Google Groups "NYCResistor:Microcontrollers" group.
To unsubscribe from this group and stop receiving emails from it, send an email to nycresistormicrocontrollers+unsub...@googlegroups.com.
To post to this group, send email to nycresistormicrocontrollers@googlegroups.com.

--
You received this message because you are subscribed to the Google Groups "NYCResistor:Microcontrollers" group.
To unsubscribe from this group and stop receiving emails from it, send an email to nycresistormicrocontrollers+unsub...@googlegroups.com.
To post to this group, send email to nycresistormicrocontrollers@googlegroups.com.

jpbar...@aol.com

unread,
May 18, 2017, 8:03:37 AM5/18/17
to nycresistormi...@googlegroups.com
I showed the calculation for the ULN2004 not a FET.  It contains bipolar transistors, which have a Vcesat spec and no Rdson spec.

Only FETs have an Rdson spec.  Your calculation is correct for a FET.

This is a good exercise for many to see how a FET can handle more current that a bipolar transistor for the same power dissipation and the lower the Rdson the better.
To unsubscribe from this group and stop receiving emails from it, send an email to nycresistormicrocon...@googlegroups.com.
To post to this group, send email to nycresistormi...@googlegroups.com.

Anders Nelson

unread,
May 18, 2017, 10:04:18 PM5/18/17
to nycresistormi...@googlegroups.com
Neat!


On Thu, May 18, 2017 at 5:03 AM, jpbarbetta via NYCResistor:Microcontrollers <nycresistormi...@googlegroups.com> wrote:
I showed the calculation for the ULN2004 not a FET.  It contains bipolar transistors, which have a Vcesat spec and no Rdson spec.

Only FETs have an Rdson spec.  Your calculation is correct for a FET.

This is a good exercise for many to see how a FET can handle more current that a bipolar transistor for the same power dissipation and the lower the Rdson the better.



-----Original Message-----
From: Anders Nelson <anders....@gmail.com>
To: nycresistormicrocontrollers <nycresistormicrocontrollers@googlegroups.com>
Sent: Wed, May 17, 2017 05:29 PM
Subject: Re: [NYCR:Microcontrollers] [iso] Heatsyncs for ICs?


Mmm, are you sure your math is correct? I believe the heat dissipation (at least with FETs) relies on RDS(on). For my aforementioned nFET with an RDS(on) of 49mΩ and power equaling (i^2)(r), that equates to 0.6W when carrying 3.5A.

Do correct me if I'm wrong though! I learned only enough semiconductor physics to pass the final and keep my projects from exploding.

=]
On Wed, May 17, 2017 at 1:26 PM, jpbarbetta via NYCResistor:Microcontrollers <nycresistormicrocontrollers@googlegroups.com> wrote:
With transistors (including darlingtons) it's important to look at VceSat which is around 1V.  If the transistor were a perfect switch, VceSat would be 0V.
At 3.5A you will be dissipating 3.5W.  A good heatsink may help you get by.  You could try paralleling another ULN2004, but all transistors may not share current equally.  A single part is always best. 

The FET that James suggested would be perfect.  It seems to also have a "logic level" input because it states an Rdson at 4.5V.

There are some low VceSat transistors, but it is this characteristic that makes FETs a better choice for many applications.

Joe


-----Original Message-----
From: mike k <thatsfa...@gmail.com>
To: Physical Computing at ITP <phys...@lists.nyu.edu>; nycresistormicrocontrollers <nycresistormicrocontrollers@googlegroups.com>
Sent: Wed, May 17, 2017 01:10 PM
Subject: [NYCR:Microcontrollers] [iso] Heatsyncs for ICs?


Hey electron-lovers,

I'm helping a friend with her circuit, which uses a series of ULN2004s to brighten and dim DC incandescent bulbs (rated at 2A / 12V).  We are running the 7 darlingtons in parallel for a total of 3.5A in available current (there was no chart in the datasheet for max power or a graph of max current at different voltages).

Sometimes a bulb seemingly latches on and smokes the IC.  I'm assuming it's a heat/current draw issue.  Are there higher rated ICs we can use?  All of the darlington arrays I find are 500mA per channel.  If it's a temperature issue, can we paste a heat-sync onto the IC?  


Best,
Mike

--
Mike Kelberman
Interaction Technologist

--
You received this message because you are subscribed to the Google Groups "NYCResistor:Microcontrollers" group.
To unsubscribe from this group and stop receiving emails from it, send an email to nycresistormicrocontrollers+unsubs...@googlegroups.com.

--
You received this message because you are subscribed to the Google Groups "NYCResistor:Microcontrollers" group.
To unsubscribe from this group and stop receiving emails from it, send an email to nycresistormicrocontrollers+unsubs...@googlegroups.com.

--
You received this message because you are subscribed to the Google Groups "NYCResistor:Microcontrollers" group.
To unsubscribe from this group and stop receiving emails from it, send an email to nycresistormicrocontrollers+unsub...@googlegroups.com.
To post to this group, send email to nycresistormicrocontrollers@googlegroups.com.
Visit this group at https://groups.google.com/group/nycresistormicrocontrollers.
For more options, visit https://groups.google.com/d/optout.
Reply all
Reply to author
Forward
0 new messages