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.

SLVA369 circuit in TPS54060EVM-590 positive output not equal to negative?

Other Parts Discussed in Thread: TPS54060EVM-590, TPS54060, TPS7A49, TPS7A30

Question about the TPS54060EVM-590 evaluation board.

I have high hopes for the clever circuit presented in application note SLVA369 using the TPS54xxx family to create split rail pos+neg supplies from a wide voltage input DC source. For our current design the Vin is 4.5 to 25 VDC and desired outputs are +5.5 and -5.5 VDC @ 100-200 mA each.

Using the TPS54060EVM-590 evaluation board it's R2 & R3 v-set resistor are removed and replaced with a 10k pot. With the side of the pot that replaces R3 set to ~9.3k and the outputs not loaded the desired +5.5 and -5.5 VDC are seen at TP7 & TP9. When the outputs are loaded with 56 ohm resistors to draw 100 mA from each of the +- outs the circuit misbehaves. Measuring the total output between TP7 & TP9 the difference remains 11 VDC but measureing TP7 to GND & TP9 to GND the 11v total is no longer evenly split +5.5 and -5.5 rather its more like +6.5 and -4.5.

Plugging the desired inputs and outputs into SWIFTPOSNEG-CALC-v3.xls with the device type changed to the TPS54060 the eval board is loaded with and the L1 values also changed to match the part on the board it tells me the circuit should be good for a total of 236 mA so my 100+100 mA load is close to max but not at or above it.

One difference between the circuits shown in SLVA369 (figure 1 pg 3 & figure 3 pg 9 also copied in SWIFTPOSNEG-CALC-v3.xls) and the TPS54060EVM-590 evaluation board realization (figure 12 pg 10 SLVU374) I would like to discuss. In the app note version the U1 IC's ground pin is connected to ground. In the evaluation board circuit U1 IC's ground pin is connected to net N18V the negative output of the supply. Comparing the two it is hard to see the SLVA369 version where the TPS54060 controller is referenced to ground can actually work. I hope I am missing something because it then appears the evaluation board version referencing the switching regulator device to negative output does not have a means to hold the ground point half way between its outputs.



  • Thats all somewhat complex.  We will look into it.  I'm not sure that I have an HPA590 on hand, so I may need to order one.  It may be a few days before we get back on this.

  • Hello John Tucker & other IT E2E readers,

    While investigating this issue you might want to contact one of your fellow TI employees Pat Hunter. In the March 2011 issue of Electronic Products he authored an article “Designing an ultra-low-noise supply for analog circuits”. This immediately caught my eye because “analog circuits” is in fact exactly where the split supply we are designing here is going to power analog amps & filters feeding a 16 bit ADC. Hunter’s circuit uses the TPS54060 in the circuit we are discussing to generate noisy split rails then uses TPS7A49 & TPS7A30 low noise linear regulators to scrub the noise off the switching supply’s output.

    In Pat Hunter’s circuit the TPS54060 is shown with it’s GND pin connected to the -18V output. With my initial experience of the plus and minus outputs not tracking I now note Hunter’s circuit is set to output +-18V from the TPS54060 block which the TPS7A49 & TPS7A30 knock down to +-12V for the final outputs. Dropping 6V twice across a pair of 200 mA linear regulators is a non-starter in my application which is battery powered with a strict power budget. I am now wondering if Pat is setting the source voltage so much higher than the linear regulator’s output set points  to allow them to cope with their plus and minus source voltages not being equal? The TPS7A49 & TPS7A30 are referenced to GND and have dropout voltages circa 250 mV so as long as the wandering outputs of the TPS54060 don’t cross ~+-12.300V Hunter’s circuit should work for a low parts count but not low power solution.


    Norman Tracy