TLV3201: TLV3201 Comparator Output Current Drive capability (Single package SOT-23-5)

The short-circuit current specifies how much current flows when you connect the output directly to a 5 V power supply, i.e., when the entire 5 V is dropped over the output transistor. At that current, there is no voltage left for your load.

The typical voltage drop for 20 mA can be found in the output voltage vs. output current diagram.

Are you talking about the TLV3201 or the TLV7031?

Do you really need 20 mA?

Scott

Clemens has provided some valid responses.  As he mentioned, the output voltage vs output current plots should be used to determine what the output response of the comparator is like.  Singles and dual will behave similarly. 

Can you share what sort of failures you have experienced?  It would be best if you could capture some screen shots on an oscilloscope if possible.

In regards to your input question, you do not need to current limit your 1.9V input voltage assuming the supply voltage of 5V is on prior to applying the 1.9V.  The input is CMOS; that is why the input bias current is in the pA range.

Chuck

I was looking at the TLV7031 earlier but i am  using TLV3201. What i found for the TLV7031 was that it had a absolute max time (10s) associated the short circuit condition  I take that to mean it will destroy itself in a short circuit situation.

Its not clear what happens to the TLV3201. According  to the chart I'm good for sourcing  well over 40mA and I  am sourcing around 15 mA continuous.. I do no see any max time conditions for the TLV3201 so i assume the output can operate short circuit continuously . I could probably decrease the drive to around 10mA and be fine. 

Ultimately I have had  three failures so far and looking for reasons why. 

Chuck,

Thanks for the input. I tried to share a schematic but it didn't take so what i can tell you is that there is an external board that provides this input.(1.9V) but right now i am just bench testing, driving the input with a power supply. I have no series resistor or zener  to protect the input currently. Just applying voltage direct to   In- . I don't have any scope captures as i am just using using dc voltages now and observing the output with a meter. The only thing i know is that the bias point changes (v+ input) after the failure and the output stops changing.  I did see something on the datasheet about max input current to (V+ and V-) but don't know how this would be accomplished given it a CMOS input with high input impedance.

The input current matters for out-of-range voltages, which are clamped to the power rails. One case where this can easily happen is when the device is unpowered.

Hi Scott,

to post a schematic click the "Insert File" button (paperclip) in the header of text box.

Kai

Thanks Kai for helping Scott out.

Scott

I am still a little uncertain about your setup and the problem you are experiencing.  Hopefully you will be able to upload a picture of what you have built and then we can comment.  In general, connecting a voltage directly to the input would be ok assuming that voltage is withing the recommended input common mode range as Clemens mentioned earlier.

Chuck

Trying again to attach the circuit

Scott

thanks for sharing your circuit.  At 20mA drive to the LED, the output drop predicted by Figure16. Output Voltage vs Output Current is approximately 1V.  This means when the comparator output goes low, the output voltage will be 1V above ground.  Likewise, when the output goes high, the output will be 1V below VCC (for your 5V application, the output will be 4V).  The 100mA short circuit is when the output is completely saturated and the output is stuck at the rail.  So instead of getting an output low of 0V, the output will be stuck at the positive supply rail.  You should not review the short circuit current when predicting performance for a DC application like driving an LED.

You mentioned having failures but what exactly is the failure?  You would need to capture a screenshot of your failure so we can comment.  My preference for your application assuming you need to turn the LED on with 15 to 20mA is to drive a n-ch MOSFET such as 2N7000.  This way the current that is needed to turn on the LED passes through the 2N7000 instead of the output stage of the comparator.  Your circuit would comprise of the 5V supply driving the 200 ohms in series with the LED and the 2n7000 connected to ground.  Turning on the 2n7000, the 15-20mA flows through the LED and the MOSFET.

Chuck

Chuck,

Thank you your help is appreciated. I believe I understand what you are pointing out. In summary because of space i was hoping to drive the LED's without buffering. I was trying to find a 5V comparitor with sufficient drive for the LED's . As i am needing inversion my original thoughts were to perhaps use an open Drain/Collector version but imagined a push pull pull output to be adequate to drive an LED. I may investigate this further as well. If there are any other devices you are aware of with better drive all input would be appreciated.

(As added information the LED's are bi-color (single LED) driven which have to be driven with separate  input control signals so i am locked in to a common cathode LED device to be able to drive the colors independently) .

Scott

I thought of something new to consider.  The LM393LV is a dual channel device and the sink current capability is quite good (higher than TLV3201).  You could also put the channels in parallel to further reduce the impedance of the output transistors being turned on.  It's a change to your circuit board but this may provide a nice option for you.

Chuck

I like this device however it not clear how much current i can comfortably sink. The data sheet claims a max of 200ma combined outputs. Not clear if this is a dual or quad.  Just wondering if I might comfortably sink 15 to 20ma per stage. I am doing a new layout but really hate to chance another.

Scott

It is always a good idea to prototype before doing a board layout. This circuit can be easily bread boarded to verify performance. 

I notice that you keep focusing on the abs max.  It is good to make sure that you are operating within this range, but for your application it is most important to review the output voltage versus sink current plots.  Anyhow, this spec is intended for the dual version but I would not anticipate the quad being much different.

I was suggesting that you consider paralleling two channels to sink 15 to 20mA to get the lowest possible impedance and thus minimizing any voltage rise when the comparator outputs is low.  Sorry if this was not clear in my previous post.

Chuck

Chuck,

I'm really not focusing on max rather the data that exists on the datasheet for output current capability. Data is rather sparse. I wish more was provided but there are no plots for this device (LM393LV) rather output short-circuit current spec (100ma) but this brings us back to where we were earlier and not necessarily something i can use to determine what i can actually use as comfortable drive current. 

Not sure just how paralleling comparitors works. Never heard of anyone doing it but If have to use two devices just to get the drive then there is diminishing space savings over using a single comparitor with added  discrete buffer such as the FET you suggested.

A note on prototyping: It used to be simpler with pin packages however these are rare these days and with SMT devices it not so simple. I can model with my layout package (Cadence) but TI is limited to older device models and is currently mostly supporting KiCad models which forces me to model with a second package.  This is not convienent and a lot more time consuming.

Yes its all a bit difficult to figure out. I really appreciate your help.   

Scott

I see what you mean.  Since LM393LV is in preview status, the typical performance curves such as output voltage versus sink current are not included yet.

Just in case you are still interested, i will see if I can find the data and post it tomorrow.  This device has one of our strongest output stages in regards to sink current capability.

Otherwise, best of luck on your circuit.  I hope you find something that will work for you.

Chuck

Chuck,

Would be very interested in the data. i am working on an circuit update this week. I would love to see if they source 10-15 ma with  5V supply and leave me with enough voltage to drive an LED. 

If you post the data would it be on this  feed?

Scott

here is the typical curve that will be added to the datasheet.

Chuck

sorry, the attachment didn't come through the first time

To be sure what i am looking at, is  this the voltage across device output ( drain to source of the open collector stage)? 

i.e.  if i put 5V to a 500ohm resistor  tied to the drain,  what current can i expect through the resistor (Vcc supply of 5V) .  Given this is ~10ma of current, and looking at  the plot i would expect to see  somewhere in the neighborhood of 100mV drop across the device?

If this is so i don't see the drop going beyond  1 volt leaving me with ~4V or so.  Are there any reasons you see  why  a single device would not be suitable to drive an  LED 10-15ma. 

Scott

Your understanding is correct.  I pointed you to this device because of the strong output stage, less voltage drop per current being sunk into the output during an output low.  You can also use single channel device by analyzing it in the same fashion to see if it meets your needs.  This device just has less voltage drop than most of our other 5V devices.

Chuck

Chuck,

Thanks again. I like this device with one concern,  it appears there are only sample qtys available at this time (perhaps it too new). I would like your thoughts on this and possible another device. 

The  TLV7256 datasheet (Rev. A) was suggested by apps as being a possibility for driving the 10-15ma. There are however no curves just that it sources 25-30ma with Output Voltage of 4.5V (@5V Vcc) This is a much bigger package VSSOP8 but if it provides a road to success I can make it work. As reminder I have to drive the LED Anode as its a bi-color Common Cathode  This makes the push-pull more convenient as it can source current  but  can configure an open collector as well.  

Scott

Sounds like you are continuing to look at dual channel devices.  Unfortunately I don't have the information you are looking for on this device.  However, it does seem slightly better when comparing 5mA capability. 

When devices are sampling only that means that it has not been fully released to production but it is coming soon (preview status). 

Chuck

Chuck,

I can use single or dual devices. Primarily looking for a good solution.I am finding that if i use a common pinout parts such as a dual t that at least i have opportunity to change devices if one does not work. I have to assemble prototypes so i am trying to stay away from leafless and small pin pitch parts. I can handle SOT23;s and SOICs ok,  I also have to consider availability as a risk factor. Thought this would be easy but apparently not.

I am curious to find out if there are possibly inverters with better drive that could be used to buffer the comparitor.

LVC logic gates (e.g., SN74LVC1G14, SN74LVC2G14) can drive 32 mA.

this looks like a reasonable suggestion

I believe that we have offered all the support we can provide from a comparator perspective, so I will be closing this post.

If there are any other comparator related questions, feel free to open a new thread or respond with the question.

Thanks

Chuck

Chuck, 

Thank you for your help. I am posting this reply I made to applications. This is not attempt to dismiss your help as you have been very helpful and it is appreciated. I contacted applications prior to the thread as you may know and have been trying to work with them. I shared this thread to applications and based upon your response they have elected to close this as well. 

I tried to load the new spice simulation (Cadence based ) and regretfully it has interfered with my Cadence layout  so i had to remove it and reload Cadence. 

This is my reply to applications. (Appologies for the typo's)

It looks as if your closing this based on the email thread. The e2e thread has me no real solution. 

Hey Scott,

Chuck and I are the applications engineers for the comparators group and we were unaware of the email thread. We have provided several solutions that would solve the problems you communicated. However, this thread has not resolved your issue so perhaps we can setup a phone call and I'll befriend you on the e2e forum. This way, we can solve your problem and communicate more efficiently. We look forward to getting you a solution in the near future.

Thanks,

Joe