• State
  • Locked Locked
  • Replies 12 replies
  • Subscribers 64 subscribers
  • Views 2325 views
  • Users 0 members are here
Support feedback
Options
Options
Related

This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

TPS22916: TPS22916 handling of normal vs PTC batteries

Mark Baum
Prodigy 155 points
Part Number: TPS22916
Other Parts Discussed in Thread: TPS22917

To whom it may concern,

I would appreciate some assistance on a weird problem we've encountered.

We are using the 22916 to switch from 3v CR2 battery to and from ext battery supply of 3.3v.

There are actually 2 issues that I need assistance on, but I'll go to the most important first.

There seems to be a difference on how the 22916 is handling power depending on the type of battery used. Specifically, if the battery has PTC protection or not. When we originally tested the 22916, we used a non PTC protected battery and the switching worked fine. Basically, battery is active, place ext power to it and the ext power supply picks up the load which is circa 100mA. Pull out ext power supply and battery picks up the load, all seamless and very nice.

Sometime has passed since we did the above test and today, instead of using a non PTC battery, I just grabbed one of the branded PTC batteries we have for testing, put the ext power supply and WHALA 300mA! showing on the ext power supply.

I freaked out at first and quickly removed the ext power supply out of fear of frying something. But put ext again and same reaction. So I realized the battery in use was different to the one used during earlier testing and I used the original, sure enough, perfect. Put ext supply and you see only 100mA being used.

Ok, by this stage I'm still unsure what's going on. The battery that gave the 300mA result is an Energizer and the one that works as expected is some non brand chinese. Ok, so we have lots of batteries and I decide to take another (Panasonic) and I put it in and 300mA! again. This is when I clicked the difference, both of the branded batteries are PTC protected.

So I grabbed a different (non PTC) battery brand and works just fine, only 100mA showing on the ext power supply when plugged in.

Since the 22916 is MCU controlled for some other features of the design (GPIO to EN) I decided to have ready the GPIO send EN low so that it would disable the 22916. I then plugged the ext power, see 300mA send the EN low command and from 300mA goes to 120mA. THEN I proceed to put EN high and with this PTC protected battery, now I'm seeing 100mA from the ext power supply.

Did all the above tests several times and always get same results. Last test, was to brave it and to leave the ext power supply plugged in despite the 300mA being used. What I saw was also super interesting, over the span of say 7mins the current usage went from 300mA down to 100mA, slowly, gradually, but steady down to normal.

So now, I'm wondering what's going on and how this problem needs to be addressed. Ideally want to use PTC protected batteries, but this current spike seems dangerous and when I touch the 22916 there's very little heat, so where is the current going to? back to the battery? - that's obviously undesirable.

Using the GPIO to disable the 22916 and re-enable is fine when the MCU is awake, but this is not possible when the MCU is in sleep mode and there's no control over GPIO. The MCU in sleep mode uses only 15uA, but the lower current use makes no difference to the behavior  described above.

Please, your assistance will be greatly appreciated.


Regards,
Mark

  • 0
    TI__Expert 5275 points

    Mark,

    Thanks for the detailed explanation.  The behavior you are seeing is expected for the TPS22916.  While enabled, the device uses a voltage comparator between the input and output to detect reverse current (typically 500mA) before switching off.  

    So when the 3.3V supply is connected, I suspect the higher current is flowing back to the battery.    For a typical internal battery resistance of 1ohm the (3.3-3.0V=0.3V) difference would result in 300mA of current flow which is not enough for the TPS22916 to detect it.

    When the device is turned off, it always blocks the reverse current which is why you see the drop when you disable it.  When you re-enable, it stays off because it already knows reverse voltage is present.   (On a side note the 20mA difference you are seeing is coming from the 150ohm discharge resistor on VOUT when the part is disabled - the TPS22916CN does not have the output discharge resistor)

    Now for the mystery of why battery type makes a difference.   I expect the batteries without the series PTC resistance would have a lower total internal resistance that allows the reverse current to exceed the 500mA threshold and active the RCB.   You can verify using an oscilloscope and/or by measuring the batteries internal resistances.

    Also, even more puzzling, why the PTC eventually goes down to 100mA.  When the reverse current is flowing through the battery, it will be heating the PTC and changing its resistance.  Since PTC resistance change is non-linear, the current initially increase and then trip the TPS22916 RCB, but this could cause the current to step down.  If the current gradually drops down, then the PTC may have passed it's trip point and heats for protection  (Although, I would expect it to eventually cool back down to a lower resistance)

    If you cannot tolerate any reverse current, the best approach is to actively disable the TPS22916 via the ON pin when the 3.3V supply is connected; either through the uC GPIO or some form of voltage detection circuit.

    Best Regards,

    Adam Hoover

  • 0 in reply to Adam Hoover
    Prodigy 155 points

    WOW - very useful info Adam!

    The 22916 would only be off for a moment and I prefer fast switch times so the CN would not do, but thanks for the suggestion.

    Yeah, we already have a voltage detection circuit and we are able to handle the ON/Off of the 22916 easy enough when the MCU is awake, the only trouble is when its sleeping..

    There's not much literature unfortunately on how much current lithium batteries can handle before it all goes awry.

    I am aware that they should receive no current, but my guess is that this should not occur on a regular basis, however, in our case it would happen once in a blue moon (when external power is applied, it will remain), would be nice to know if the damage is severe or very slight. So far no explosions! ha

    And just to clarify one last thing, the 22916 stops the current of Vout reaching Vin, but doesn't stop the battery's current traveling back to the circuit even when an external power source is also present on Vout, correct?

    When I was first researching this chip, I thought it blocked the battery all together when Vout > Vin, but that's not the case on my tests.

    It seems the 22916 acts more like a diode than a switch - which is actually fine. I guess it would have to have 2 Mosfets to be more switch like.

    Could the TPS22913B be more appropriate perhaps?

    Thanks again,
    Mark

  • 0 in reply to Mark Baum
    TI__Expert 5275 points

    Mark,



    I share your frustration on battery specifications. With many electronic components, the effect of overstress is a matter of reliability and statistics. The more frequently the stress happens, the sooner the component will fail, so it is reasonable to bound the number of stress events that are expected over the product lifetime.

    Also, when the TPS22916 is turned off, it will block current altogether in both directions just like 2 back-to-back MOSFETs. This can be triggered by the ON pin or by VOUT > VIN. I believe your misunderstand, and that of many others, is that the VIN>VOUT RCB comparator has an offset ~25mV to prevent it from false tripping with a very light load (0mV drop).

    It is for this reason we are now listing the activation threshold as a current (IRCB = 500mA), in addition to the voltage threshold (VRCB = 25mV), starting with TPS22916, in an effort to make this behavior more clear. Please share if you have any other datasheet suggestions or would like to see a different RCB behavior in future products.

    TPS22913B uses a similar VIN to VOUT comparator architecture, so the RCB behavior would be the same.


     
    Best Regards,
    Adam Hoover

  • 0 in reply to Adam Hoover
    Prodigy 155 points
    Hi Adam,

    Woohaa! - I thought I had it in the bag, but you've thrown me a curve ball :)

    "This can be triggered by the ON pin or by VOUT > VIN" ..but we are putting 3.3v externally to Vout while Vin is getting 3V from a battery.

    How big does the difference have to be for the 22916 power off?

    Because if it powered off, then there would be no current running back the battery - in the same way as I see it happening when EN is put low.

    Perhaps its because we're holding EN high? I thought EN was tied to GND internally and it had to be maintained high so that the chip would work. Is this perhaps the problem?

    Ok, I think I found the reason to my problem in the spec sheet! "Once the ON pin is deliberately driven high (≥VIH), the Smart Pull Down is disconnected to prevent unnecessary power loss"

    We are "holding" EN high thus the Vout power difference is not being automatically considered or perhaps the condition disregarded.

    This is what happens when you work with so many chips and don't pay attention to the details of each! ..terrible and I'm sorry. I think this must be the source and reason to all the above problems.

    Thank you again for your time!

    Regards,
    Mark
  • 0 in reply to Mark Baum
    Prodigy 155 points
    Hello Adam,

    I'm sorry, but I'm getting increasingly confused here - not sure what's going on.

    This is from your spec sheet "The switch ON state is controlled by a digital input that is capable of interfacing directly with low-voltage control signals. Both Active High and Active Low (L) versions are available. When power is first applied, a Smart Pull Down is used to keep the ON pin from floating until system sequencing is complete. Once the ON pin is deliberately driven high (≥VIH), the Smart Pull Down is disconnected to prevent unnecessary power loss."

    The 22916 I'm using is EN high to activate.

    From my last reply, I understood that once EN was set high and the chip powered on, that the pull down was disconnected and that meant the chip remained active without having to have EN pulled high. And I (thought) that this was the error because we were keeping EN high and so when Vout>Vin the chip couldn't power off automatically because we were holding EN high.

    So just a moment ago I tried the above, but that didn't work at all.

    The 22916 will (not) remain on after EN is set high - if EN is not "maintained" high the 22916 stops working and cuts out the battery supply coming from Vin.

    So back to one of the most recent questions.. how much higher does Vout need to be than Vin to indicate teh 22916 to power off as per your quote "Also, when the TPS22916 is turned off, it will block current altogether in both directions just like 2 back-to-back MOSFETs. This can be triggered by the ON pin or by VOUT > VIN."

    Thanks and regards,
    Mark
  • 0 in reply to Mark Baum
    TI__Genius 15780 points

    Hi Mark,

    Perhaps I can chime in here. As you mentioned, the TPS22916 will not remain enabled if the ON pin is driven high and then the driving signal is removed. By maintaining the ON pin high, the internal pulldown is disabled, but without a driving signal the ON pin voltage may sink and the pulldown may be re-enabled.

    As for the last comment you had, if the ON pin is driven low, then reverse current will be blocked regardless of the VOUT to VIN difference. If the ON pin is high, then it takes 25mV of voltage difference between VOUT and VIN.

    Thanks,

    Alek Kaknevicius

  • 0 in reply to Aleksandras_Kaknevicius
    Prodigy 155 points
    Hello Alex,

    Thank you very much for your response.

    Indeed, the only option so far is to keep EN high to maintain the 22916 enabled.

    So then, the real question is why when we are putting 3.3v on Vout when the battery is only reading 3V does the 22916 not disable? - that's a 300mV difference there so plenty over the minimum required.

    Or are there different types of "disable"?

    When EN is no longer driven high, the 22916 immediately cuts the battery and you see 120mA being used from the external power supply (20mA more than should be due to the resistor to GND that it has internally).

    But when Vout gets 3.3, the mA use on the external power supply shoots to 300mA. Is that it? there are 2 different ways it finds to disable itself? is that by design?

    If the device was always on, then it would be easy (just set EN low whenever ext power supply is applied - we have a voltage sense in our design) but when the MCU is asleep, we can't put EN low, thus we are kind of counting on the 22916 to disable itself when Vout gets 3.3v.

    Thanks,
    Mark
  • 0 in reply to Mark Baum
    TI__Genius 15780 points

    Hi Mark,

    I just grabbed one of the TPS22916 EVMs and ran a quick test. Keeping the device enabled, I applied a 3V source on VIN, then a 3.3V source on VOUT. The device turned itself off and prevented reverse current from flowing. How long does the 300mA reverse current last? Is it on the whole time VOUT=3.3V and VIN=3V? Are you limiting the reverse current to 300mA? If so, have you tried increasing the limit?

    Also, if you would like to prevent the constant current draw from the output of the device (~20mA as you described) you should take a look at the TPS22916CN. This version of the device does not have the quick output discharge (QOD) feature and will not draw additional current when turned off.

    Thanks,

    Alek Kaknevicius

  • 0 in reply to Aleksandras_Kaknevicius
    TI__Genius 15780 points

    Hi Mark,

    Are you able to use a multimeter and actually measure the voltage drop across the TPS22917 (positive probe on VOUT, negative probe on VIN)? I suspect that the limited 300mA current is not allowing the voltage drop across the switch to be the full required 25mV.

    Thanks,

    Alek Kaknevicius

  • 0 in reply to Aleksandras_Kaknevicius
    Prodigy 155 points

    Hi Alek,

    Thank you very much for conducting this test and for the results.

    However, there's a small but potentially significant difference in your test versus a real life test. Based on the info you gave, there was no load on your set up.

    Our circuit is drawing 100mA.

    If you could perhaps put a 100mA load and test again, that might be the difference.

    I suspect that if we take out the 100mA load (thus no load), that the 22916 will react as you've described.

    Also, as per the initial post, I don't have this problem when not using PTC protected batteries. If you place a PTC protected battery in your test (as well as a load) you should experience the same result. But having said that, a PTC protected battery shouldn't really have an effect, considering all that's needed is a 25mV difference between Vin and Vout to switch off the 22916 - thus voltage turns it off, not current as I understood it, and this is not happening unfortunately.

    Regards,
    Mark

  • 0 in reply to Mark Baum
    Prodigy 155 points

    I guess you guys gave up on me? :)

  • 0 in reply to Mark Baum
    TI__Genius 15780 points

    Hi Mark,

    Apologies for the delay here. As for my setup, the lower voltage (3V) power supply was set to sink up to 1A, so the 3.3V supply would have been able to push a lot more current if the TPS22916 allowed it to. I think the issue is that the series resistance for the battery is limiting the current across the load switch, and therefore the voltage drop across the switch is not the 0.3V you are expecting. Have you been able to use a DMM to check the voltage difference between VOUT and VIN?

    Thanks,

    Alek Kaknevicius