Post Go back to editing

LTC2980 as LTC3636 power monitor

Category: Hardware
Product Number: LTC2980-24

Hello,

I'm designing LTC2980 as power monitor for few LTC3636 POL and have some questions:

A) It is not clear how to select the Rsns (located after the inductor of LTC3636 and connected to the LTC2980 Vsense odd channels.

Also, If I'm working with very light load ~ few mA up to 0.3A is there any specific guideline or restriction ? (I didn't see such also at LTC2977 DS)

B) It is not clear regarding overall accuracy of the system after using the supervisor to fix the LTC3636 voltages

According to the 2980 DS it is mentioned -- Margin or Trim Supplies to within 0.15% of Target

It is not clear if this is the DAC accuracy that should be added to LTC3636 VFB accuracy of 1%

Also, when using the calculations of Rdac (R30) according to LTC2977 DS I'm getting very high Rdac for getting Vdc min and Vdc max within 1% error range

Here is an Excel to calculate the above for 2 domains for example (2.5V, 0.92V) with 1% desired accuracy

I didn't use the Ifb which is neglected and don't have typical value at LTC3636 DS (it is -30nA to +30nA so 0nA is probably the typical value) and anyway 30nA over 47.5K ohm doesn't change the results more than 0.06%

Vfb 0.6
Vout desired 2.500 V 0.920 V
Rtop (R20) Real E192 47500 47500
Rbot (R10) Real E192 15000 88700
Vout Real (nom) 2.500 V 0.921 V
Deviation from target 0.00% 0.14%
Domain DC accuracy 1.0% 1.0%
Rdac (R30) < 1140000 3611150.79
Rdac selected 1140000 3611150.79
VFs_Vdac > 0.000 0.000
VFs(1.38/2.65) 1.380 1.380
Vdc min 2.468 0.911
Vdc min accuracy -1.30% -0.97%
Vdc max 2.525 0.929
Vdc max accuracy 1.00% 1.00%

I can see that Vdc max is 1% (it happens when the DAC outputs GND so R30 is parallel to R10) 

Vdc min is higher than 1% - Is there a way that Full scale can be limited?

The Rdac looks very high - is it reasonable?

Also - Vdcmax and min are related to Vfb (LTC3636 Vfb is 1%) so does the total accuracy is poor than that?

  • Hi RSZL,

    I hope this reply is not too late.  Sorry for the delay.

    A) The LT3636 does not have an IMON pin, so you only have 2 options: inductor DCR sensing or use a discrete sense resistor.  There is an article posted on the analog.com site that covers the pro's and con's to each approach.  In your case, 300mA full load means that the DCR needs to be 50mohm to 100mohm to get a decent amount of 'signal' for the current sense amplifier to accurately report the load current.  100mohm (inductor DCR or discrete resistor) will give you 30mV of signal.  You'll have good accuracy for load currents of tens of milliamps and greater.  When the load current is a few mA, your signal will be just 0.3mV.  Offset and gain error will dominate the reading.  You also need to consider the IR drop of a discrete resistor which is in the output path.  Load regulation will suffer as the sense resistance increases. It's basic tradeoff.  Here's a link ... Current-sensing-article 

    B) We guarantee the error in one line item called TUE, total unadjusted error.  There's no need to calculate reference error and DAC error and ADC error.  All of these are included in the 0.15% spec.  The TUE is the error in the READ_VOUT register value relative to the actual sensed voltage at the Vsense pins. I looked at the spreadsheet numbers and yes, the DAC resistor values look a bit high. You might want DAC resistors that are in the 400k to 600k range.  It's not a critical value wrt to TUE.  There's a tradeoff in selecting DAC resistor that is too low vs too high.  If Rdac is too low, you will see the DAC lsb steps in the regulator output, and also you diminish the effects of the 10-bit DAC resolution.  If Rdac is too high, the DAC voltage needs to swing more and you run the risk of saturating the DAC if you were to margin Vout.  However each DAC lsb will have a small effect on Vout which is a good thing.  You might want to try the Resistor Selection tool in the Utilities pulldown in LTpowerPlay.  

    Regards

    Mike