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.

SN6505B: Very low voltage across push-pull DC/DC converter

Part Number: SN6505B
Other Parts Discussed in Thread: TPS61175, SN6501, PCA9544A, ADS1115-Q1, ADS124S08, TLV61220, SN6505D, SN6505A

Hello all,

I've designed and built my first isolated power stage as part of a 3V3, 5V and 5VA supply, for an instrumentation front-end. Please refer to this TI datasheet for the inspiration for the design, section 9.2 'Typical Application'. The output voltage wildly differs from the input, given that I'm using a 1:1.33 winding transformer I would expect in the region of 5V peak measuring after the MBR0520L diodes in the diagram (shown below):

Instead the SN6505B driver puts this waveform in:

And I get this out after the diodes (note the Vmax):

I suspect something in my implementation is fundamentally wrong, and given this is my first work with transformers I wouldn't be surprised. Here is the schematic of my PCB, the components after the 5VA output of the TPS61175 are not fitted to the PCB currently, so the regulator is unloaded:

That's a lot of information, so thank you for taking the time to read. My main question is; can you see any mistakes in my implementation that could lead to a much lower power output across the transformer? If anyone has any suggestions as to design best-practices or any gut reactions that they think could be the cause, that aren't necessarily substantiated by the evidence, I'd be happy to investigate them - I am totally stuck!

Thanks in advance,

Louis

  • Hello Louis,

    thank you for using TIs E2E forum!

    I have a few more questions regarding your application.

    Is it the SN6505B you are using? On you schematic I see SN6505B on the node, but SN6501 above.
    What load is connected after the TPS61175?
    What transformer are you using? Do you want to boost the voltage to 5V or higher values?

    Would it be possible to disconnect the boost converter and test your sn6505 independantly first?

    Thanks for providing more details!

    Best Regards,
    Matthias
  • Hello Matthias,

    Thanks for your reply, I am using the SN6505B - the incorrect labeling is from before I understood the current output I needed for the downstream components.

    The load on the 5VA supply would be 400mA/2W, for comprehensiveness (in case I've misinterpreted consumption data) that is comprised of:

    • 3 * PCA9544A - 5VA @ 12uA = 36uA
    • 12 * FS2012 - 5VA @ 30mA = 360mA
    • 12 * NPR-101 - 5VA @ 3mA = 36mA
    • 3 * ADS1115-Q1 - 5VA @ 300uA = 900uA
    • 1 * ADS124S08 - 5VA @ 460uA = 460uA
    • 1 * TLV61220 - 5VA @ 5uA = 5uA
    • TOTAL = 0.397A

    However, for debugging purposes there are no components connected after the test point in the diagram as I ended up removing all components and building each power stage and testing individually.

    The upstream regulator that delivers power to the SN6505B is a TPS259230DRC in this configuration:

    The transformer in the diagram is a Wurth Electronik 750315240 which I believe to be driven above the minimum frequency of 150kHz by the SN6505B. My understanding is the coil has to be a 1:1.33 turns ratio due to energy loss between the cores, so if I need to isolate 5V -> 5VA I will need to provide slightly more power to the output side.

    I can definitely remove the TPS61175 and check the output if you think that will help, is the reasoning that the regulator might be exerting too much load on the transformer for some reason? Would it be viable to remove all passives between the Schottky diodes and the 'TEST_5VA' test point also, or do you think everything else is probably inconsequential?

    Thanks in advance,

    Louis

  • Hello Louis,

    Thank you for providing more details!

    For debugging the easiest way would be to test the device on its own. Then add you circuitry step by step to find the vault.

    Just to get it right:
    you are connecting your primary side to the 5V after your eFuse?

    What supply is are you using and is there a current limit installed? Do you see current flowing with no load connected?
    Also, what is the reson for adding the boost converter after the SN6505B? Do you want to boost up your voltage to a higher level?
    This can normally acheived by adjusting the transformer ratio.

    Regarding your scope shots?
    Are you referencing your probe to the respecitve ground portential?

    On the first picture the scope is connected to the two outputs of the SN6505D? So it is D1 and D2?
    The second picture you are referencing your probe to secondary side GND and probe before the diodes on the secondary side?

    Usually you should see a voltage swing on both sides. Please also double check that your transformer is connected properly, meaning the center pin is connected to GND on the secondary side.

    Thanks
    Matthias
  • Hello Matthias,

    Yes the 5V supply going to the SN6505B is coming straight out of the other side of the eFuse. The other test points (J19 and J17) show stable voltages of 5V and 3.3V respectively.

    The supply has a current limit of 3.7A, and in it's present state is drawing about 300mA, which I expect as there's no load on either regulator.

    I took some additional scope images to clarify. The first is referenced to GND and is a reading from pin 1 of the T1 transformer, one half of the signal from the SN6505B (image A):

    The next is on the other side of the transformer, pin 6, referenced to GNDA before diode D1 (image B):

    And finally this one is of the signal after diode D1 also referenced to GNDA (image C):

    I was working with a technician on this issue today and he was horrified by image A, am I right in saying this should be a clean square wave rather than one possessing all of those spikes? He suggested removing and replacing the SN6505B entirely to see if it is at fault, would you recommend that based on the scope trace?

    I'm also tempted to remove all of the components after the coil's diodes but I'd rather not if there's a more likely explanation than a downstream part failure, although I'm told that having low/limited load on the transformer could also result in this behavior, is there any truth to that?

    Thanks in advance,

    Louis

  • Hello Louis,

    Indeed that looks faulty..

    It looks like your device is not switching at all.. And there is just some noise on the output.

    Also 300mA seems high if there is no load connected. Are you sure that you connected all the pins as specified in the datasheet?

    You maybe flipped the device when you populated it?

    Can you the 5V stable on the center tab of the transformer?
    Is your enable pin tied to 5V? You might could measure that voltage as well?
    Is there any noise on the CLK pin? Do you have it floating?

    Are you sure that all the circuitry before the SN6505B works properly?

    The datasheet shows how the device should be connected and in figure 35 you see how the swing on the D1/D2 nodes on the primary side look like.

    It is hard to debug the circuit with everything connected. There is a chance that your component is just damaged as well.

    Normally I would recommend to build up your power design partially. A good way to start with the SN6505B is using the EVM:

    http://www.ti.com/tool/sn6505bevm          

    Otherwise I would recommend to replace the part and depopulate all the devices that are connected behind it. Also I am not sure why you are using a boost converter after the device?  A good way to get a stable and clean output is to use an LDO, as shown on the EVM using the TPS76350.

     I hope this will help you solving your issue.

    Best Regards,

    Matthias

  • Hello Matthias,

    I replaced the SN6505B on the board, and now D1 and D2 just output a constant voltage and no voltage is present on the opposite side of the transformer, nor is any current being pulled. However this was with a previously used part, I'm going to order some new ones today. The CLK pin was pulled permanently high, I'm not sure if this is normal behavior or indication of a fault/failure, but all other pins like the ENABLE, VCC and GND were as I'd expect.

    I don't believe there's anything wrong with my layout but I'll post it just in case, U21 is the SN6505B:

    One mistake I did notice very early on was that the regulators are all tied to the same 'power good' line via their EN pin, and that the isolation would prevent the operation of the 5VA regulator whose ground potential is related only to the transformer, does that make sense? To avoid this problem I tied the GND and GNDA together via a soldered patch lead.

    I will post further when parts have arrived.

    Best regards,

    Louis

  • Hello Louis,

    I wanted to check with you if your devices already arrived?

    There are also some more comments I would like to make to help you with this design.
    We can also follow up on this in an offline disucssion or have a call/webex session to run through your design and how to get it working.

    In the meantime, I thought you are maybe interested into further marterial on power designs, so I would like to point out our training page to you:
    On TI.com you can find multiple videos regarding power designs. We also have some videos to get you started off with isolatiosn:

    training.ti.com/power-tips
    training.ti.com/ti-precision-labs-isolation

    Looking forward to hear from you!
    Best Regards!
    Matthias
  • Hello Louis,

    I wanted to check if you still need support for this?

    Best Regards,
    Matthias
  • Hello Matthias,

    I received replacement parts and fitted them then captured these traces:

    (between driver and transformer)

    (after transformer)

    To summarise, it appears I'm getting two sawtooth waves shifted by -5V. I'm on a very tight deadline and I think I'm going to have to compromise on the input isolation, or change the design. My main question is, is there a precedent for a transformer circuit behaving this way? Could it be a downstream failure of some kind?

    Best regards,

    Louis

  • Hello Louis,

    What is the output voltage you measure at the outputs after the diodes?
    How do you probe this? with reference to the GNDA net?
    Did you depopulate the components behind the devie?
    I strongly recommend to debug this step by step. So start at where your input voltage is attached and work you way through towards your isolator and then pupulate the parts behind it. Still I am wondering why you are using a boost converter after you isolated supply? Typically you would adjust the transformer windings to acheive a higher output voltage and smoothen this using a LDO.

    If you want to we can take this discussion offline and have a more detailed look on your design. That might be easier.

    Best Regards,
    Matthias
  • Hello Matthias,

    Just a quick reply - I'm removing the downstream components entirely today and debugging without.

    I probed those traces directly from the transformer coil input and output, but forgot to also capture traces after the diodes. Irregardless the waveforms had not changed in amplitude after using a new SN6505B, which defeats the idea of it being a driver part failure at least for me.

    I'm using a boost converter because it's of the series recommended in the SN6505 application example, but is also capable of providing the ~1A current output required (whereas a lot of the LDOs I looked at were all <1A). If this is actually something to avoid, or just simply not a best practice, then I will change the design accordingly.

    I discovered something last week which may be relevant - the transformer I use is the Wurth Electronik 750315240, but section '9.2.2.5.3 Recommended Transformers' of the SN6505B datasheet actually recommends the SN6505A as a driver for a 5V -> 5V 1A supply. In case this is the issue, I've ordered some spare SN6505A.

    I will be in contact about a screen share, thank you for your time.

    Best Regards,

    Louis

  • Issue not closable.
  • Hi Matthias,

    Good news! I followed your advice, and removed all components downstream of the diodes - the power fault is due to the regulator.

    I have no idea why specifically the regulator does not work with my circuit, I'm hoping you or someone can shed some light on that, but I think it was simply the wrong type or topology to use in this instance.

    I stripped back all components and added just decoupling passives and the resistors used by the regulator to set current and voltage, and immediately read 6.3V on the point after the diodes/rectifiers. Now I'm into the redesign.

    My final questions, before resolving, are to do with the appropriate replacement to the regulator. Section '9.2.2.2 LDO Selection' of the SN6505 datasheet states that;

    "The required minimum input voltage preventing the regulator from dropping out of line regulation is given with:

    Equation 1. VI-min = VDO-max + VO-max"

    I've read that my LDO's current output should be higher than my peak, which is 1A, so I'm looking at the LP3852EMP-5.0/NOPB. By equation 1 my understanding is that the minimum input voltage must be 435mV + 5V, taking datasheet values from here, which should make the output stable at a wide temp range as my transformer is outputting 6.3V. Have I correctly understood these factors?

    Below is a draft of the redesigned schematic:

    All input capacitors use an X7R dielectric, and the output capacitor has an ESR of 1.2Ohms. SJ1 is a solderable jumper to aid in testing.

    Best regards,

    Louis

  • Hello Louis,
    I am happy to hear that you made progress with your debug!
    By regulator you are referring to the Boost converter that is shown on your schematic and connected after the isolated power supply?
    www.ti.com/.../support
    Please note that these products are managed by a separate product line at TI, so for more detailed question I would recommend using the Non-Isolated DC/DC Forum in E2E:
    e2e.ti.com/.../addpost

    Not looking into going into much details with that: You connected a boost converter behind your isolated supply, these devices are made to step up a lower input to a higher output voltage.

    However as you measure your transformer puts out a voltage at 6.3V, but the boost converter is set to 5V. This means the output regulation of your boost would try to regulate the output lower, however it cannot regulate it to lower outputs than the input voltage, as it has just the one switch. There are other devices, that can step up and down voltages, these have two switches. However the TPS61175 has just one switch and the diode D3, which means it cannot provide a lower output than the input.

    Your understanding of the minimum input equation is correct. You should receive a stable output, as long as your input is higher than your set output voltage + the maximum dropout voltage of the LDO. Please note that the output voltage you see at the transformer output of the SN6505 will drop with higher output loads, you can observe this behavior by connecting an active load a variable resistor and increase the load current, you should see your voltage drop by about 500mV.

    What is the typical load you want to connect behind the LDO?

    As well, in one of your images above it seemed like you have traces going from the primary side to the secondary side? I recommend not to have any physical trace from the primary to the secondary side in order not to violate the galvanic isolation.

    Thank you for your answers and using E2E.

    Best Regards,
    Matthias
  • Hello Louis,

    did you made any progress with your design?
    Can we support you in any other way?

    Best Regards,
    Matthias
  • Hello Matthias,

    I think that's it - redesign is complete with a 1A supply LDO, overall load on the 5VA is only 700mA, sending off for parts and PCB.

    Thanks a million for your help, this one was a real slog!

    All the best,

    Louis