has anyone used a LTC3532 buck boost regulator

[[email protected]]

I'm having trouble interpreting the section on Closing the Feedback
Loop in the datasheet.
Is there anyone out there with experience with this chip?

 

[PeteS]

I'm having trouble interpreting the section on Closing the Feedback
Loop in the datasheet.
Is there anyone out there with experience with this chip?

There's no difference between this device and any other voltage mode
converter - for fast transient response and to avoid the double pole
you need to add extra filtering.

The RHP zero (which would normally be designed such that it provides
phase boost during final gain roll-off) is at too high a frequency in
boost mode (usually) to act as phase boost for this rather critical
frequency region, so a type III filter is used instead to get proper
phase response *and* fast transient (loop frequency) response..

If you can live with slower loop response, then use a type I filter as
explained in the text.

As far as I can see, the text is perfectly straightforward. In
genereal, LTC's datasheets are superior to most for their power
converters.

Cheers

PeteS

 

[[email protected]]

Thanks Pete
The problem I'm having is that my last analog filter experience was 18
years ago at tech. All I really want from this device is a 3.3 V supply
from a 3.6v lithium cell to power my microcrontroller and bluetooth.
The current profile is 0.5mA (sleep), 30 mA (on), 60mA (on +
bluetooth). Since size is not critical I wanted to use the 300kz
switching frequency to get maximum efficiency. So the example circuit
looks ok but it uses a higher switching frequency. I think the data
sheet is very detailed but for all its detail it is very thin on
application examples that dont depend on extensive prior experience. It
would be greatly improved if there were a complete worked trhough
example which detailed all the assumptions made. A good example of this
deficiency is in the calculation of the inductance - there are two
formulae one for buck and one for boost, but there is no explanation of
what to do with the different results, i assume we use the maximum
value but its not clear at all to the new user. Also "The inductor
ripple
current is typically set to 20% to 40%" what are the implications of
using 20% as opposed to 40%?
So for a someone with primarily digital experience this is a hard part
to get to know.

otilija

 

[PeteS]

Thanks Pete
The problem I'm having is that my last analog filter experience was 18
years ago at tech. All I really want from this device is a 3.3 V supply
from a 3.6v lithium cell to power my microcrontroller and bluetooth.
The current profile is 0.5mA (sleep), 30 mA (on), 60mA (on +
bluetooth). Since size is not critical I wanted to use the 300kz
switching frequency to get maximum efficiency. So the example circuit
looks ok but it uses a higher switching frequency. I think the data
sheet is very detailed but for all its detail it is very thin on
application examples that dont depend on extensive prior experience. It
would be greatly improved if there were a complete worked trhough
example which detailed all the assumptions made. A good example of this
deficiency is in the calculation of the inductance - there are two
formulae one for buck and one for boost, but there is no explanation of
what to do with the different results, i assume we use the maximum
value but its not clear at all to the new user. Also "The inductor
ripple
current is typically set to 20% to 40%" what are the implications of
using 20% as opposed to 40%?
So for a someone with primarily digital experience this is a hard part
to get to know.

otilija

OK

What you need then is a power supply that has the following
characteristics:

1. Stable from no-load to about 500mA (the bluetooth module will slurp
current in pulses and the supply has to be able to withstand it).

2. Has to operate in buck-boost mode (although it's not a SEPIC
converter).

3. Needs fast transient response (bluetooth again).

Note that although the average current for RF devices is fairly low,
the peak currents can be quite high (GSM / GPRS for example can be 2A
peak pulses), so transient response is very important in this
application.

Give me time to have my dinner and I'll do my good deed [tm] for the
week and give you suggested values.

Cheers

PeteS