Hi AAC,

I'm using an LM2577 in the adjustable configuration as seen in the image below:

Except I modified the values of R1 and R2,

Instead of having a resistor as R1, I used a 50k MCP41HVX1 pins 11 and 12 which are P0B and P0W giving me a variable resistance which I can control with a microcontroller digitally with buttons or an encoder.

R2 is a 5.76k resistor, so that the output voltage goes from ~1.23V to ~12V.

The problem that I'm having is that the output Voltage drops when I start drawing more than 50mA of current in the output.

I tested the output current draw and the results are the following:

6 resistors in parallel of 100 ohms each

• Output voltage: 3.0
• output current: 100mA

5 resistors in parallel of 100 ohms each

• Output current: 82mA
• Output voltage: 3.11

4 resistors in parallel of 100 ohms

• Output voltage: 3.28
• Output current: 61mA

3 resistors in parallel of 100 ohms each

• Output voltage: 3.49
• Output current: 34mA

As an input I've been using a 2000mAh LiPo battery which I measured to ~4.0 Volts:
https://www.sparkfun.com/products/13855

Is is feasible to use the MCP41HVX1's as a resistor R1 for the LM2577-ADJ? If so, why is the output current dropping the voltage with just 100mA?

 

You cannot set the output voltage of a boost converter to less than the input voltage. You need to use a different circuit for that, such as SEPIC or buck-boost.

It looks to me like the circuit is not attempting to boost at all and what you are seeing is simply the battery voltage minus the voltage drop across the inductor (which should be very small) minus the diode forward voltage. Get rid of the digital pot for testing and replace it with a fixed resistor calculated to give you something in the range of 5-10 volts output.

From the datasheet for the pot: [Edit - this should not be an issue]
3.10 Analog Negative Voltage (V-)
Analog circuitry negative supply voltage. Must not have a higher potential then the DGND pin.

[EDIT this statement appears to be wrong:] You can't use that pot in the way you are attempting.

100 nF at the input is far too little. That capacitor should be as close as possible to the 2577 and a capacitor of 200 to 500 uF (good quality, low ESR electrolytic) should be within a couple of centimetres.

 

Can you elaborate on why I won't be able to use this pot?

I don't have the V- connected to anything as of now, just the low power pins to control the resistor network and the P0W and P0B pins for the high power pins. Also if I connect that Analog Negative Voltage pin to ground it wouldn't it have the same potential as the DGND pin not higher?

I'm trying to understand exactly what I'm missing. I'm trying to regulate the voltage up if possible, I wouldn't mind if the voltage goes from 4-10 Volts, I'm currently able to get 12 volts in the output if I adjust the potentiometer, but when I connect a load the power isn't great at that voltage.

I'll look into the SEPIC and buck-boost converters.

 

The capacitor value at the input is what they have in the datasheet also but if it's better to use a 200 to 500uF I'll try that. Appreciate the response ebp/

 

Capacitor: The small capacitor is for high-frequency bypassing. In a typical boost converter there is moderate "ripple current" flowing from the input supply. The magnitude of the ripple depends on the inductor chosen and the input and output voltages. Typically the inductor is chosen for peak-to-peak ripple current of something in the range of 20-30% of the average input current. It is a triangle wave with a frequency equal to the switching frequency. Because the impedance of the input power supply may not be low at that frequency and because there may be inductance in the connecting wires, a large capacitor is normally placed close to the switcher input to reduce the source impedance.

Digital pot: The datasheet for that thing is one of the worst I have ever seen, and I have seen thousands. The information is there, but the thing is horribly badly written. Note, for example, in the bit I quoted then, rather than than is used, but that is kind of trivial compared with the overall incompetence of the writer.
After suffering through the datasheet some more, I think I was wrong and you should be able to use the digital pot. Carefully read the requirement for the analog power supply pins V+ and V-. The voltage at the top end of the pot must be less than or equal to V+ and at the bottom of the pot greater than or equal to V-. V- can be connected to the switcher common (ground) which must also connect to digital ground. It appears that it should be OK to connect V+ to the output of the boost converter, but this is of course a varying voltage. You'll need to read that awful datasheet very carefully to make sure that this meets the requirements.

Do you have a datasheet or part number for the inductor you are using? Please post a link to the datasheet, if possible.

 

Here's the inductor that I'm testing with:
https://www.digikey.com/product-detail/en/bourns-inc/2100LL-101-H-RC/M1412-ND/725959

Digikey part number: M1412-ND

 

Here are some pictures:

I can also provide the full list of components that I'm using.

NOTE: Don't mind the buttons and that's at the other side of the breadboard.

 

I can't seem to get to the inductor data sheet at the moment, but it should be just fine.

While you're contemplating the digital pot: Don't forget that if for any reason it is open-circuit, such as at start-up (I don't know this will happen, but it needs to be considered), the feedback loop for the converter will be open circuit and the output voltage will rise uncontrolled. If this is an issue, one way to limit the voltage is to put a zener diode, possibly in series with a small resistance , in parallel with the pot. If you want to operate cleanly up to 10 volts, you might chose a 12 to 15 volt zener. Use one rated at 1/2 watt or less because it will have a sharper "knee" (look at current v. voltage curves - the "knee" is where the plot makes the sharpish bend of about 90 degrees) than a higher power type. Zeners start to conduct a little below their nominal voltage and there is quite a bit of difference among types. The 1N52xx series is mediocre, but there aren't many that are readily available that are much better in this regard. For lots of analog work it would be a great boon if you could buy zeners with ultrasharp behavior at the knee.

 

I'll take a closer look at your photos a little later. Trying to get a switcher to work well on a plug in breadboard can be quite a problem for several reasons. More on that later.

 

I was looking at https://www.digikey.com/en/articles/techzone/2014/aug/the-sepic-option-for-battery-power-management and http://cds.linear.com/docs/en/datasheet/3467afe.pdf and the configurations for the SEPIC look pretty similar to the configuration of the LM2577 which makes me suspicious that they might be pretty similar internally.

Also for this other LT3467 which is a SEPIC wouldn't I also need a digital potentiometer to make the output voltage regulated via microcontroller?

 

You could make a SEPIC converter with the IC you are using now, but there are significant advantages, and some drawbacks, to going to a higher switching frequency. SEPICs require two inductors, and there are advantages to having them wound on a common core, though that isn't essential. At higher frequency there are more choices of coupled inductors and the inductors are physically smaller. Note the capacitor in the power path. That cap handles significant current so the type used is important. Again, higher frequency gives more choice and allows lower value. High frequency converters will not behave well unless built on a printed circuit board, and all the newer parts are surface-mount only.

You would still need a digital pot. The SEPIC allows the output voltage to be either higher or lower than the input voltage and is simpler than a buck-boost converter, though there are some b-b converters that are quite easy to use at low power because all of the required switches are integrated - but SMD only.

 

Hey @ebp, can you show me how the zener+resistor configuration would look like in parallel with the R1 resistor and exactly how it would work, if you don't mind? Thanks a lot for all your feedback and help !! Greatly appreciated.

 

Hey @ebp, can you show me how the zener+resistor configuration would look like in parallel with the R1 resistor and exactly how it would work, if you don't mind? Thanks a lot for all your feedback and help !! Greatly appreciated.

As I said previously, pick a zener with a voltage rating somewhat above the normal maximum output voltage of your supply, but less than you can tolerate under gross failure. For example, let's say you intend to let the voltage go up to 10 V during normal operation. To avoid the problem with the zener starting to conduct below its nominal voltage, a 12 volt zener, such as a 1N5242B (rated for 500 mW). The cathode would go to the positive output of the supply. A resistor in series probably isn't strictly necessary, but might help prevent excessive current into the feedback pin of the controller under fault conditions. Unfortunately the LM2577 data sheet doesn't specify an absolute maximum input voltage for the feedback pin. We can't make the resistor too high because it will increase the voltage. I'd choose 1k and hope for the best.
So: zener cathode to output, zener anode to one end of resistor, other end of resistor to the feedback pin.

To verify that the zener behavior around the knee region isn't causing any problems with normal regulation, disconnect the zener and set the voltage for the highest you normally want to run at, with at least some load on the switcher. Shut down and reconnect the zener. Power up and check the output voltage. If it is now lower than it was, you may need to increase the voltage of the zener.

Because I have equipment that makes it fairly easy, I would put a load on the switcher output and connect the input to a current-limited bench supply with the voltage set to around what I normally intended to use for input and the current limit turned up. I'd set the voltage to what I wanted at the output, then turn the current limit down to a bit above what was being drawn (this is a bit tricky, because as soon as the power to a switcher is limited to less than it needs to power the load at the set point voltage it will "collapse" the input supply as it tries to regulate). Then I'd turn off the input supply, leave the zener in-circuit but disconnect the normal resistor from the output to the feedback pin, and turn the supply on. Assuming it started OK & I didn't need to turn up the current limit a bit more and try again, I'd check the output voltage to be sure the zener was limiting about where I wanted it to.

I'm sorry - I promised to get back to you with comments on layout & totally forgot. I will try to do that today. If I forget again, please yank my chain.