TPS7A8300: TPS7A8300 Current Limit FoldBack

Hi KQ,

We do not have a pin to pin device with a programmable foldback current limit. Could you tell me more about your application? We may be able to find a way for the TPS7A8300 or the newer version TPS7A83A to work for your application.

Could you provide a multi-channel scopeshot with Vin, Vout, and Iout so that I can see how your rails are interacting?

Very Respectfully,
Ryan

Hi Ryan

I tested it with a controlled environment using a active DC Load and set the current to 2A constant and then time when the 2A load applied and how long it takes for the power supply to turn off due to overheat. the problem is the TPS7A8300. It is not possible to reduce the 5V VIN which was shared.   The problem with the above thermal shutdown is the VIN also dip along with the output sometimes, so it was not just the output DIP (and recover quickly in less than 100msec) It was a insufficient thermal disspiation problem we just missed it by 20% margin.  But thks about the availability of the TPS7A83A  we can switch to this on the next design or a larger package LDO with larger metal tab for conducting the heat away.  If only the VOUT dip and not affect the VIN we are OK but unfortunately the VIN went down along with it. CH1 is VIN, CH2 is VOUT, CH3 is our Load Supply which is to be the same as VOUT. It is this VIN that went down along wit the output that is our problem.

Hi KQ,

Unfortunately your image did not properly attach to your post.  From your additional explanation it sounds like your concern is the thermal shutdown rather than the current limit.  Please note that when an LDO enters thermal shutdown, the pass element will turn off.  Due to your 2 A load, the output capacitors will quickly discharge bringing the output voltage rail to GND.  When the junction temperature of the LDO cools enough, the LDO will turn back on.  When this occurs you will see your 2 A load current and inrush current flowing through the LDO to charge the output capacitors.  

Since your conditions are intentionally allowing the TPS7A8300 to enter thermal shutdown, the LDO will continue to cycle in and out of thermal shutdown.  Your upstream supply that is powering the TPS7A8300 will see this as a load transient.  The dip you see in your input voltage will be the load transient response of the upstream supply as it attempts to source the 2 A load and inrush current through the LDO.  You can reduce the magnitude of this dip by increasing the input capacitance to the TPS7A8300.  When the LDO turns on, the input capacitor will help source the extra current to charge the output capacitors.

Please keep in mind that Thermal Shutdown is considered a fault condition for an LDO.  Care should be taken so that the junction temperature does not exceed 125 C during your application's normal operating conditions.  Continuously running into thermal shutdown will degrade the reliability of the LDO.

Please refer to the following for how to attach images to the E2E forums:

Very Respectfully,

Ryan

Hi Ryan

I had already solve the thermal problem. Now at 2A continoulsy the IC is only slightly warm and I can source it infinitely.  So it is not a problem now.  Now the issue is the occasional Overcurrent limit. I had simulated it on the board.  If the over current condition is not remove, the TPS7A8300 will divert the current internally (instead of removing the current) and the IC gets VERY HOT (much hotter than the heat generated by the overccurent condition itself where the IC is only warm). I simulated with 2.5A overcurrent load. This IC does not have foldback current limit that limit the output to 2A.   [Note that this is a worst case scenario for overcurrent continuously].  So it seem the only possibility is to change to a newer chip with same pinout and larger current out capability. Which IC is the direct pin to pin and at least 3A?  Thnks

I have a field problem where we experience mostly Spikes at the VIN of this TPS7A8300. we see very large spike twice the VIN in short duration of tens to 100msec. However it is difficult to duplicate back in workshop as the load is pulsating that is occasional peak pulse overcurrent around 3A for pulses where the average current is closer to make the IC goes into thermal shutdown. under what kind of condition (Overcurrent shutdown or thermal shutdown) do we gets positive spikes (our load is a pulsating load, not constant current).

Hi KQ,

I am glad to hear that you solved your thermal issue under the normal operating conditions of your application.  You are correct that the foldback current limit for TPS7A8300 will not engage until the current is 3.4 A (typically).  Unfortunately it sounds like during your fault conditions you are still dissipating enough power to cause the LDO to enter thermal shutdown.

Depending on which package you are using for TPS7A8300 you may be able to use TPS7A84A or TPS7A85A as a pin to pin option for a higher current; however, it is important to note that switching to one of these higher rated current devices will not solve the thermal issue during your fault condition.  The only way to reduce the risk of entering into thermal shutdown is to reduce the power being dissipated in the LDO.

Very Respectfully,

Ryan

Hi Ryan

I had managed to gecapture a real case of the random spikes issue at the current output and the voltage input of the TPS7A8300. . It wan't thermal shutdown or overcurrent (I had artificially created the overcurrent using DC load as previously i created the condition artificially using DC Load). The problem was the output current of the LDO  was ringging in few nsecs when the Load was turn on but yet the output voltage of the LDO was stable. However the input voltage of the LDO rings along with the LDO current output .. This high rise/fall time high current ringging looks more like inductive but the power output to the load does not have any inductor but a wide PCB Trace with 2 low EST Tantalum. The Yellow (CH1) Line is output voltage of the  TPS7A8300. The large ringging CH4 (100mV/Amp) Current sense voltage across a 0.1ohm resistor, directly at the output of the TPA7A8300 and this shows the peak current  is about 2.6A The top 2 waveform (CH2, CH3) is the voltage at the input of the LDO (pickup at some distance away from the LDO Input). So it seem the current drawn by the LDO was in very large and short spikes and this ONLY happens very very randomly. Is this a problem with the load (a radio module) or this is a problem of the TPS7A8300??  Prior to this ringging the output current of the LDO was close to 0 as both LDO and load was turn off. The LDO has a 27nF slow start up capacitor allowing 10msec startup.

Hi KQ,

TPS7A8300 will source whatever current is required by its load (up until current limit is reached). As such if the load is switching states, the erratic current is likely associated with the load itself.

Very Respectfully,
Ryan

Hi Ryan

We had change out the TPS7A8300 with a extneral small power mod (TPS54231) and there was no spike observable so far. So i can easily confirmed by changing out the TPS7A8300. Can i know the list of LDO that is Pin to pin compatible with TPS7A8300?? It speed up my search and changing out TPS7A8300 was the easiest option for us given our board size restriction which was the reason we pick TPS7A8300

Hi Kyle/Ryan

Actually we found our problem is not the current capability or thermal issue. it was the load step current that was ringging (not sure what cause this). we need to try a power supply with SLOW step load response in tens of usec or even 100usec (and not capable of rise time in nsec). any suggestion ?? (it is use to see if we can replace as a quick fix instead of fixing the load)/

Hi KQ,

So you're thinking that a reduction in the bandwidth (response time) of the LDO  will cause this to behave better? Is that correct? This isn't a request we get very often so I don't have a list off the top of my head but I've included search results for 3A LDO and filtered out some of the faster ones that I could filter out easily. You can use it to further filter by key specs for your application to narrow down the field of possible LDOs. 

Since you are looking for slower LDOs you should probably stay away from the following devices that may be in that list:

• TPS7A83
• TPS7A84
• TPS7A85
• TPS7A52
• TPS7A53
• TPS7A54
• TPS757
• TPS758
• TPS744

http://www.ti.com/power-management/linear-regulators-ldo/products.html#p451max=3;4.5&p1128=6;1710&p1498=Catalog

I hope this helps. 

Kyle Van Renterghem

Applications & Validation Manager

Linear & Low Dropout Regulators

Hi Ryan/Kylie,

I tried testing the transient response of the TPS7A8300. I put a very low ESR (0.5ohm so parallel would be 0.25ohm) 1000uF (so total is 2000uF) tantalum capacitor directly at the output of the TPS7A8300. The load is NOT turn on.  The TPS7A8300 input is +5V and there is only a small 10uF Tantalum capacitor at the input of TPS7A8300. I am trying to undersand the waveform why it behave (when TPS7A8300 is enabled, small DIP in upstream voltage and as the TPS7A8300  reach the set voltage, there is a dip and spikes.  pls explain to me the behavior of the spikes and dip it appears the TPS7A8300 . I already reduce the output CBULK and input CBULK, i just wants to understand the transient load hehavior. IT seem longest slow start up time is not necessary the best thing to do.

i mean reduce output CBULK and increase input CBULK to the TPS7A8300 so the above is mean to understand the transient load hehavior and I wish to explore more on testing the TPS7A8300 tranisent load behavior. That is one thing I do notice if the load transient rise time or fall time is too fast the spikes gets much worst.

Hi KQ,

As the LDO output approaches the regulation voltage, the input current to the LDO will decrease because there is no more inrush current.  Inrush current is the current in addition to the load current that is used to charge the output capacitor(s).

This decrease in current will be a load transient with respect to your upstream supply (the supply providing Vin).  The ripple that you are seeing on Vin appears to be the load transient response of your upstream supply resulting from the LDO's decrease in current as it reaches the regulated output voltage.

The transient response will be worse with a faster transient.  This is due to the fact that it takes time for a control loop to detect and respond to a transient.

Very Respectfully,

Ryan

Hi Ryan

Thank you yes it was the load transient issue as we had verified. Do you think adding more input Bulk capacitance Low ESR tantulum will help elimnate this spike? Right now it is already 1000uF on the input of the TPS7A8300. Or I need to change the upstream TPS54531 it seem not fast enough? any suggestion for pin-to-pin compatible TPS54231 / TPS54531 that has faster transient response.

Is there any application notes that on how to compute the requires input CBulk capacitor value, given a know transient voltage observed at the input voltage of the TPS7A8300? Also is there possibility to modify the compensation value of the TPS54531 so that it has the maximum possible load transient response? One final option then is to find anothe replacent to the TPS54531 with much faster load transient response. please recommend me a suitable SOIC8 if possible pin-topin compatible. So instead of fixing the TPS7A8300 I go fix the upstead power supply response or add more bulk capacitance to teh TPS7A8300 (i have not know how a 2000 or 3000uF output capacitance would affect teh existing TPS54531 assumign these addtional Bulk capacitor would cater for the load transient). Thanks looks like pretty close to solving my problem

Hi KQ,

Your latest scopeshot of your system only shows ringing on the input rail for the TPS7A8300 and does not show ringing on the output of the TPS7A8300. As a result, you already have adequate input capacitor for a stable output in your system. Additional bulk input capacitor will not improve the TPS7A8300 performance; however, it can help improve the system level performance such as the system level PSRR. This is because the input capacitor will act as a pre-filter to the LDO.

As your questions are now about the TPS54531, I am looping in the team that is more familiar with that device.

Very Respectfully,
Ryan