Post Go back to editing

LT8490 charging will not initiate

Category: Software
Product Number: LT8490
Software Version: ...

Hello,

Problem: LT8490 is not initiating charging stages. It has a status code, and some SVRO pins that go high signifying an error. I designed the whole thing. Ask me anything.

Input Max = 52V, 35A

Output Max = 42.6V, 30A

I had an old board revision that the LT8490 worked, and I bolstered from a 15A to a 30A model. It doesn't even try now.

I will attach all the stuff needed to analyze the scenario. Schematic and also a measurement of all the pins with VIN  = 51V, Vbatt = 36V.

PDFPDF

*it wouldnt let me attached an altium board file. Let me know if that is desirable and we can figure that out. Can also email me Thomas@nwrec.us



I attached the spec sheets below. LT8490 is an LT8705 but with solar/battery logic added. I am well versed in these...

 PDFPDF

Details, notes; observations, discussion:

 

  1. SVRO_IOUT: blinks (roughly 1/10 second) then SVRO_FBIN follows -->
  2. SVRO_FBIN: is on long, about 1 second
  3. This signal repeats.


The FBIN resistor network uses 0.1% accurate components. I found the equations for solving those very intolerant of a 1% variance. The first equation to solve for R_FBIN1 also enters a negative region past 100kOhms (and is undefined at 100kOhms, and is why 1% variance was crazy). The table of factory values on that same page have R_FBIN1 values beyond 100k. Attempting to regenerate the factory table  is not doable, can someone advise me here? This could be the key to the problem. Because again, my R_FBIN = 93.1k because I want 52V max (solar input specs). But this value is smaller than the 40V and 60V R_FBIN values in the factory table; when I should be between them in R value, logically speaking. This is indeed a conflict.

Also, here is an Oscope of the FBIN pin. Interesting...


And FBIW

It is the same signal. Excuse the noise, my scope sometimes is goofy.


The SVRO IOUT pin going high is funny, because there is no output current yet. But in the pin measurements provided I drew what I saw, a periodic tiny square pulse in IMON_OUT reaching 2.8V... enough to trigger the SVRO (they trigger at 1.25V roughly).

 

LT8490 does get hot to touch: the higher the VIN, approaching 51V, the hotter it gets. At low voltages, 20V and down, the chip doesn't noticeably get warm. EXTVCC is fed with a 24V regulator; so LT8490 is not powered by VIN like it could be.. It does seem to be ok if I left it alone. (hot to touch is about 50C) I think it would hit 60C and stops increasing, but I do not leave it on too long cuz its an expensive repair if it burns up haha.


I swung input voltage from 6v to 51v, to see if it triggers anything. no change.
Same with output being upon 24V and 36V battery packs. No change in behavior. Same SVRO behavior as mentioned above, I have the four SVRO's on LED's.

 

LDO33 and INTVCC are not overloaded but please check them out if you like. I bottlenecked the power to the LED's with a single 1K (see schematic) which makes the current 3.3mA at any time, even if multiple LED's were to trigger, the 3.3mA will be divided and not exceeding 3.3mA.

  • I will admit, INTVCC could be overloaded when the very large FET's in the power stages trigger, but we aren't even there yet. That was an experiment I am trying to get too with this board; large FET gate capacitance, and paralleling FETs. None of the power stage attempts to start right now.


Temp sensor is working.. lol..

 

I do have a code from the STATUS pin and FAULT pin, and also some other Oscope readings I am also attaching. Can someone interpret the STATUS pin message? Apologies for scope noise, its an old sucker. It's not real noise on the board.

STATUS pin readings: 4 images.
The byte repeating.

 

This is the farthest zoom out, the byte repeats 8 times, then there is this longer byte at the end, then this little squirt that's lower voltage then the rest at the very end.


Just one of the bytes zoomed in a but more (the longest high bit I believe is the start of the message?)


Two of the repeating bytes, then the loooong bit in the middle for the start of the final byte of bits; with squirt included at the end.


And at last, the underwhelming and not too helpful fault reading... yah. This beep is very quick and repeats after the status pin ends its message. Perhaps that is the little squirt at the end of the status message because it is not 3.3V like the squirt above.

 

 

 Ok that's all for now. As I play with it I will say more. Please ask me questions if you need more info. This is not a new problem, I have been at it off an on for a few months, but now is crunch time to get it 'done'.

 Thanks,

~Thomas
Director of Engineering
nwrec.us

 

 

Parents
  • The input dividers on your schematic are not matching with the pin voltages you are reporting if 51V is the maximum input voltage. If you are using a power supply and not a solar panel then you need to pull VINR low prior to power up so the MPPT algorithm doesn't function. Please use the attached spreadsheet to determine correct resistor values and verify those are the actual resistors placed on the PCB. Based on the schematic and 51Vmax input, VINR should never be 3.3V and the same is true for FBIN. Also, ensure you have a battery on the output of the LT8490 and you may find it beneficial to disable the LT8705 load until you get the LT8490 working properly.

    Also note:  The LT8490 cannot output 42.6V at 30A continuously without special cooling. This is a Pout of almost 1300W. The DC2069A demoboard has a max output of 360W and this is reaching the limit using a standard FR4 board with no forced air. If you have better cooling you could get more power output. You can add LT8705s in parallel (more phases) to the LT8490 to get more output power.

    XLSX

  • Hello! Thank you for responding!

    1: The FBIN divider and the voltage:

    I did a quick double check in a simulation, and what simulator gets does match my measurement in real life. Perhaps you did not know that the FBIW digital pin does pull low (when it is not high) and sucks some energy away from FBIN? It is not a floating digital output pin when not set to high.



    So sim gives 1.16V, I measured about 1.2V at FBIN. Where if we omit the FBIW pulling low, then it would be 1.7V. Bu that pin is either push to 3.3V on FBIN, or pull to 1.2V-variable with input voltage.

    2: I have used power supplies, battery through a resistor choke, and solar input ( I have a lot of different solar panels on hand to try), but I have not tried telling VINR into power supply mode. Maybe that will make it try something. I will get back to you on that

    3: Thank you for noting my power size. This is an experiment where I expect cooling to be necessary. if it is too obnoxious I was going to go parallel again with an LT8705 connnected to LT8490.

    Now when you say the LT8490 can't handle X amount of volts/amps, are you saying the chip itself or the FET's? I assume the LT8490 is insignificant on what it is controlling and should never have handling issues inside its maximum ratings (like <80V on Vin and Vout pins).

  • (also playing with the excel sheet you gave)

  • Yes, for power limitation, I was referring to the MOSFETs on an FR4 board with no forced air cooling and no other heatsinking.

  • Ok thank you!

    Will get back to you on the excel and VINR being put in power supply mode. I am very curious.

    So far it seems when i put in my values to the xcell  sheet, i get what i put on my board (laughs and sweats a little xD )

    Am i doing it right?XLS haha

    This FBIN is important because The values I get just do not match the table in the design-spec sheets; using the eqations even in this exell.

Reply
  • Ok thank you!

    Will get back to you on the excel and VINR being put in power supply mode. I am very curious.

    So far it seems when i put in my values to the xcell  sheet, i get what i put on my board (laughs and sweats a little xD )

    Am i doing it right?XLS haha

    This FBIN is important because The values I get just do not match the table in the design-spec sheets; using the eqations even in this exell.

Children
  • I can confirm that pulling VINR low had no influence on the system.

  • I guess i am just stuck over the fact that these don't line up



  • Based on your pin voltages and your reported input voltage, the resistor values on the PCB do not match what is on the schematic. It could also be a layout issue, solder issue or a damaged LT8490.

    Divide and conquer. Start with a power supply powering the input and VINR low and a battery on the output in the range of voltage designed. Use a power resistor to GND in parallel with the battery as the load. Do not have any other loads.

    You could also get a DC2069A demoboard and verify operation then modify the FBOUT and FBOW resistors and caps to the desired output voltage. You will also need to change the M3 and M4 MOSFETs and output caps to higher voltage parts. Don't change more than you have to at each step to verify operation before making another change working toward your schematic.

    The input side could match exactly the DC2069A demoboard so compare resistor values on your board with the demoboard.

    The Rfbin1 equation and the following equations in the left column on page 14 of the datasheet are approximations. The directions on the right column instruct you to refine these values with the Vx1 and Vx2 equations. Once refined you should get something close to the values in table 2.

  • "Based on your pin voltages and your reported input voltage, the resistor values on the PCB do not match what is on the schematic."

    I showed an image above of my simulated values, can you show me your math or logic as to what voltages are not matching what resistors? I could be missing something. (I'm sorry, I feel like I am being annoying.)

    I have a working board (older version of this one) next to me whom I reference. And I have referenced everything quite a lot to fix this bug, they are nearly identical boards. (which were based off the demo board files too). Hmm..

    Now, you make an interesting point about the equation. That it is merely approximation and that I'm supposed to bump the results around with the Vx1 Vx2. How much do I do an interpolation of the table and cross check with Vx1 and Vx2? (I did do Vx1Vx2 check with my values and they are true to what's on the board, but I did not interpolate from the table, i went strictly off equations.)

    Would it help if I provide all my calculations? I lost the page but I can run through it again in a day and just post that. (I think I'll do that anyways, might help here)

    I feel that is imprecise though and that there are 3 methods I described and all produce different resutls. But I will do that and see what I get.

  • the parts that dont meet voltage requirements: the caps handle 100V, and the FET does 60V. Is 60 not sufficient for a 32V lead acid absorption? which is 42.6V