Understanding MAX1771

[shawnmk123]

I was designing a DC-DC converter with these requirements
Power available is 3mW.
Vin = 5V
Vout = 12V


1.Is operating the MAX1771 at 250uA load current (3mW/12V assuming 100% efficiency, Power in=Power out) a good idea? I am confused because on page 3, graph show that the efficiency drops below 80% if the current is less than 0.1A for bootstrapped operation for Vin=5V. On the other hand, the supply current vs supply voltage curve, shows that for Vin = 5V, input current = 0.4A --> load current will be lesser than 0.4A -->low efficiency. Please advise what these graphs mean ?

2. MAX1771 works on the PFM operation algorithm. After reading some material, I understand PFM seems to help avoid the operation of the
converter in DCM mode by operating in CCM till the converter reaches output voltage regulation. Yet, page 5 shows DCM operation for medium loads. With my specifications, I may be operating at light loads which implies deep saturation. Please explain.

3. Please suggest if design equations for the MAX1771 (PFM operation) is the same as a regular boost converter with PWM operation?

Thank you,

 

[BradtheRad]

Since power available is 3 mW, this implies your 5V power source can produce 0.6 mA.
It would have 8k effective impedance (internal).

So, assuming you can get 0.6 mA going through the coil (instantaneous or average, whichever applies)...

Notice with a boost converter you are adding the supply V to whatever is coming through the coil. It is coming through 8k impedance.

I have tried simulating a boost converter with these specs. I cannot obtain more than 6 V with any configuration, even using a coil of several thousand henries. (Low current is associated with a large henry value. And vice versa.) I tried frequencies down to a few Hertz, and various duty cycles.

I believe you will need to use a buck-boost converter. It could work because the coil discharges through low impedance (after having charged through 8k impedance).

It still will be a job to obtain 12 V output. I have not yet found a coil / frequency / duty cycle that achieves it.

Of course all my talk is just theory.

 

[shawnmk123]

yes, I think so too but I am unable to convince my project supervisors. I also tried to simulate a similar boost converter LT1619 using LTSPICE but it doesnt work out unless you redesign for the voltage and power specifications.

Please provide answers for the other questions too.

Thank you,

 

[BradtheRad]

I have no experience with the IC you mention. I understand there are problems which can occur when an IC tries to attain a desired output volt level.

Such as keeping the transistor on permanently and never shutting it off.

Or going into 'hiccup' mode, when it conducts on some cycles but not on others.

Seems to me if there is power to operate the IC, there ought to be power available to convert 5V to 12V without a problem.

It might pay to closely examine all its specs and description. Your supervisors may believe the IC is smart enough (or that you are smart enough) to get into a mode that will do what they are asking. They may be seeing how resourceful you are able to be, whether or not you succeed.

You may coax positive results out by trying uncommon configurations. For instance, I saw a buck converter (simulation) create a higher volt level than the supply V. I believe it was at a certain frequency which created resonance with the load capacitor.

page 5 shows DCM operation for medium loads. With my specifications, I may be operating at light loads which implies deep saturation. Please explain.

If a load is light enough it goes into discontinuous mode.

If a load is heavy it leads to continuous saturation.

The link below is a current thread about the load's effect on output of a switched-coil supply.

https://www.edaboard.com/threads/259380/

You may find the animated simulation comparing three buck converters to be of interest.

 

[DBPhillips]

MAX1771 boost circuits can be designed with the equations from the MAX863 data sheet. But, you'll need to use the tOFFmin and tONmax specs for the MAX1771.

Also, the MAX863 datasheet has a much better explanation of how the parts work. The operation is very similar except (1) MAX863 error comparator has no hysteresis and (2) MAX1771 has a dual-mode comparator so you'll need to put a third resistor from REF to FB in non-boostrapped applications to keep the FB pin from going too low on power-up. If the FB pin voltage drops to around 50mV or less then the MAX1771 looks to the VIN pin for feedback.

Generally, the best way to prototype with the MAX1771 is to hand-cut a board from double sided copped-clad circuit board using a Dremel rotary grinding tool and a broken carbide drill bit or carbide engraver. A board can be cut in about 30 minutes that way after penciling on the layout. The MAX1771 generally doesn't like solderless breadboards. If it doesn't like the layout, it usually bursts with giant output voltage ripple, which means either switching EMI is getting onto the FB pin or there is a lot of stray capacitance on the FB pin. Or it might glitch and turn the MOSFET off too early which means something is wrong at the ground, VIN, CS or SHDN pins -usually its ground bounce.

Most common mistake when using the MAX1771 is to select an inductor with an inadequate current rating. The inductor saturation current rating must be greater than the peak current set by the current sense resistor. Second common mistake is to use a slow diode. The diode must be ultra-fast with a reverse recovery time less than about 70ns. MUR series diodes work. Schottkey diodes also work. Third most common mistake is to use input or output capacitors with ESR that is too high.

A long long time ago, there used to be a design spreadsheet for the part. Try calling Maxim and have the tech support guy ask Sangalli if he still has it.