LM5085 usage in inverted buck-boost topology

Hi Ludo,

I cannot read the whole description because the lines are truncated. Please try pasting the text part again. But from what I could read it seems the current limit is set low (~325mA). The output current is limited in buck-boost configuration tis reduced as mentioned in the app note.

Hi Vijay,

Please find below corrected copy of the question, followed by further feedback based upon your comments.

We have designed an inverting buck-boost converter )+6.5V input to -25.5V output, load 250mA max) based around the LM5085 buck converter.

LM5085 design:

In testing this design all is working fine up to around 25mA after which the voltage regulation is lost (reference feedback voltage cannot be maintained to internal 1.25V reference voltage - no issues with current peak limiting and feedback voltage ripple level have been noticed that should affect this behavior)

Based upon application notes and provided reference designs for such converter types using Webench (as identified below, based upon LM25576) we already identified that a further capacitance will be required between output and input voltage. Some basic testing (capacitor with relatively high ESR was used) showed that regulation would then be maintained up to a load of around 50mA. A properly specified component (capacitor value and ESR) and placement is likely to improve the performance, but prior to experimenting further we would like to understand the behavior better.

Webench reference design:

However, prior to enter this detailed optimization mode, we are wondering why the LM5085 is not listed in the available solutions list for the inverted buck-boost topology (and as a matter of fact the same applies for the TPS5430 for which application report SLVA257 identifies the use of this part as inverted bick boost within our required voltage range)?

Given that the applications (both LM25576 and TPS5430) are similar to what we've done with LM5085, can you advise if the LM5085 can be used in this inverted buck-boost topology? Side note here is that quiescent current of the circuit (when working OK) is quite significant and well above listed device quiesent current (approx 60mA measured at no or low loads).

In case the LM5085 can be used, can any further advise be given on the specification for input and output capacitor? Also, any further requirements for feedback circuit with respect to compensation requirements (which surely is a difference with respect to LM25576 reference design)?

Dear Vijay,

Thanks for your comments on the current limit which indeed will play a role here as the inductor peak currents will be quite significant:

di/dt = Vin/L = 43.33 A/s

D= Vin/(Vout-Vin) = 0.77

Fs (220k) ~ 750 kHz

ILpp = 45 mA

With a load current of around 25mA the actual inductor peak current is around 25mA/0.23 + 45mA/2= 131.5 mA

Given that 325mA current limit results in load of around 70 mA (including inductor peak current) and taking into account some tolerances and additional circuit loads (quiescent average current flow was close to 60mA --> output current ~14mA) this could potentially clarify the problem seen when the input/output capacitance was fitted.

In the condition where this capacitance wasn't fitted the level of around 25mA indicates transition from discontinuous to continous mode, which seems to suggest that system was stable when operated with lower currents, but became unstable for higher currents. The capacitor seems to have fixed that.

The above could be valid explanations for the different behavior seen in the design. 

Further to initial questions I expect that based upon the above and your initial feedback that LM5085 could be used for this purpose, but where possible please advise on recommendations for the input-output capacitance.

I'll likely will be able to make further modifications later this week, so I will keep you posted on results.

Best regards,

Ludo Gommers