DC to DC converter advice
TC
Posts: 1,019
Hello all,
I am working on a little side project for the propeller. But I need to drop 5V to 3.3V, but using the smallest package I can find. After looking for hours on DigiKey I could not find a LDO that I liked. So I went on to DC to DC converters. I found something I liked, but I don't know if it would be any good.
That is why I need your help. Could you please tell me if this DC to DC converter is any good?
I am not driving any inductive loads with it. just a prop, EEPROM, (8) 5mA LED's, micro SD card, and a couple logic devices, maybe (if I have room) SRAM, and FLASH. I chose the 1A model so I would have extra current if I need it, and so under normal use the DC to DC converter can work more efficiently.
Thanks
TC
I am working on a little side project for the propeller. But I need to drop 5V to 3.3V, but using the smallest package I can find. After looking for hours on DigiKey I could not find a LDO that I liked. So I went on to DC to DC converters. I found something I liked, but I don't know if it would be any good.
That is why I need your help. Could you please tell me if this DC to DC converter is any good?
I am not driving any inductive loads with it. just a prop, EEPROM, (8) 5mA LED's, micro SD card, and a couple logic devices, maybe (if I have room) SRAM, and FLASH. I chose the 1A model so I would have extra current if I need it, and so under normal use the DC to DC converter can work more efficiently.
Thanks
TC
Comments
I don't get it? How can a DC-DC converter module take up less room than an LDO at this low voltage? How do you decide that the LDO is too big especially when you say maybe SRAM and FLASH (not small). I'm confused, isn't an SOT-23 pack more than adequate for this task and dropping 5V down to 3.3V is not exactly LDO as we know it. It is also unlikely that you will be using more than 100ma or so especially if you tie the LEDs directly to 5V. Well for red LEDs you might need to have a simple diode from 5V but the the Prop can drive the cathodes and not suffer any meltdown because the LEDs are going to drop at least 1.6V + 0.6V for the common diode so the highest voltage that the Prop pin will see in this configuration is 2.8V. Without the diode it is possible for the red LEDs to glow very faintly when they are turned "off", so the diode is a simple way to fix this. If you used any other color you would not need the diode.
How did you arrive at this figure of 1A?
1 amp and 94% efficency is awesome.
I thought of that, but just the transistor has a size of (SOT-23-3) 2.9mm x 1.3, then add a 0402 resistor, and a SOD-523 diode, and I am larger then the DC to DC.
I did a DigiKey search for 3.3V output @ 1A. the smallest package that came up that I could find was a SOT-223-6 (6.5mm x 7.2mm) The DC to DC has a package size of 3.2mm x 3.5mm
I just chose the 1A of current. because this will also provide power for some things on a breadboard. But most of the time, it will be only powering the stuff I listed. The SRAM and the Flash are only hopefuls. I don't know yet if I have the space to use them.
Most of the time I will be getting power from a USB jack. So I really dont have to worry about battery life.
So the 1A figure is just a rounding up to the nearest amp? I have complex configurations and they just don't draw all that much current. Also, rather than a single regulator you could distribute two small regulators in some designs. However, I think you are grossly overestimating your current requirements and you will find that even a little SOT-23 reg will handle this nicely. Even if the peak current was 300ma (somehow) you will find that the real limiting factor is thermal and the average current is much lower than this.
Basically
I think I am going to go with your advice.I found a 3.3V @ 300ma LDO in a DFN (2mm x 2mm) package. And the price is much better $0.55USD.
Thanks
Would you please post de p/n ? It may be useful for some of us
Thanks!
Alex
Knew I would forget something.
Here it is
I like the LXDC3EP33A-107.
In perusing the spec:
1. Since this is a switching regulator there will be ripple voltage. Possibly 50mV max or so, (although I can't find the max spec).
2. It looks like they want you to use a 5uF ceramic cap on the output. Not required but suggested. Probably very close to the output. I would also add a similar cap to the input.
3. Clearly, in most applications you will not be able to run at 1A because the chip will warm up. I would say keep it less than 500mA max or so.
4. Very cool! It operates down to 0mA of load. (Many LDOs can't do this, they usually need 5mA to 10mA min to keep in regulation.)
5. The regulater runs up to 4MHz. So there will be some EMI, ElectroMagnetic, radiation.
For most applications I would most likely run it in power saving mode.
Note, since there will be ripple voltage simple AtoDs using input pins will not give stable readings. Predictable but not stable.
External AtoDs don't like ripple either, but their reference voltages can be filtered.
That AP7331 will regulate down to 0mA load.
Prudent design would limit the expected output current to 200mA to 250mA depending on the temperature.
Duane J
I like that even better. And the package is really small.
Thanks
No pricing on the AP7331 yet.
Yea, that is one thing I did not like about it.
The AP7331 is going for $0.56 at DigiKey. I know that does not help you out that much.
The MCP1700 is one to consider if you don't need the shutdown feature or the rapid transient response. The SOT-23-3 package is not as tall as the SOT-25.
LDO's such as the AP7331 or the MCP17xx series use p-mosfet pass transistors. Keep in mind that there is a substrate diode pointing backward across the transistor from the output to the input. The AP7331 data sheet does not show it, but it is almost certainly there. It is just something to keep in mind.
Duane J
I was looking at the following diagrams, MCP1700 shows the substate diode, AP7331 does not...
Some symbols show a separate reverse diode symbol, as in the MCP1700, even though there is no extra structure for the extra diode. Since there seems to be no mention in the spec for the characteristics of the reverse diode I suspect it is just the body diode just like the AP7331. That extra diode symbol is just added for clarity.
This body diode is never forward biased under "normal" operation and can be ignored.
However, if for some reason the input power gets shorted to ground the body diode will discharge the output capacitors. A good thing.
In the past, some regulators, with bipolar transistors, could be damaged if the input was shorted. The output capacitance would be discharged through the regulator circuits. Bad news. They often recommended an external bypass diode for protection against this. These modern regulators are self protected with the MOSFET body diode.
One more thing. A lot of the modern high current MOSFETs do add a extra reverse diodes with fast switching characteristics specifically for use in circuits that have inductance. These reverse diodes switch much faster than the body diodes. Some also have zener characteristics.
Anyway, for our applications, I think they both have similar MOSFET pass transistors whatever the symbol they used.
Duane J
When I took my EIT exam (a prelude to a PE exam), there was a section on Engineering Economics.
Why so?
Cost is always a factor -- from the start -- in any project.
For the one-off hobby gadget, the most expensive and best performing chip might do. But if you are working on a 'for profit' project, you want to minimize costs while achieving appropriate performance.
Just to clarify - all _power_ MOSFETs have body diodes. Small signal MOSFETs/IGFETs do not, in particular MOSFETs in
CMOS chips don't have them and are symmetric (source and drain are equivalent). In power MOSFETs source and drain
are completely different in construction.
I agree that MOSFETs in CMOS circuits don't appear to have a body diode.
However, that appearance is illusionary. That body diode always exists, its just not always connected to the source. in this example of a NAND gate that middle FET doesn't directly connect to the power pins but its body diode is still connected to V<sub>SS</sub>.
You can test if a body diode is present on CMOS output pins with an Ohmmeter.
Short V<sub>DD</sub> to V<sub>SS</sub> to make sure all the FETs are turned OFF.
You will see 1 diode drop from the output pin to the power pins in either polarity.
If the body diode weren't there you wouldn't see the diode drops.
The Prop is made of CMOS and this test can be done on it also.
That pesky spec limiting the pin input voltage to less than V<sub>DD</sub> + .6V and greater than V<sub>SS</sub> -.6V and the current to less than .5mA. That current passes through the body diode. Note, this body diode is a bit weak so the current limit is only .5mA instead of something much higher.
OK, this may not be strictly true with the Prop as each pin is both an input and an output. CMOS inputs have their own protection diodes.
Duane J
The body diode is the big one in reverse from drain to source. The small one is the channel diode for normal operation of the fet. The picture is the p-channel element from the schematic of the mcp1700 LDO. The substrate of the p-channel device is a block n-type material that has two p-type wells dug in at the drain and source ends. The gate bridges over substrate material in between, insulated by the thinnest layer of glass.
When a negative voltage wrt the source is applied to the gate, the electrostatic field drives electrons away from the channel under the gate and creates a local p-type region that bridges across between the two p-type wells. The transistor conducts. The small arrow symbol indicates that there is now a channel diode, with a p-type channel under the gate, up against the n-type substrate. No current, except leakage, flows across the channel diode.
The large body diode on the other hand is the ordinary diode formed by the junction between the substrate and the drain. If the transistor is reverse biased, that diode conducts. Simple. In normal operation, that diode is reverse biased and contributes only the minority leakage current.
I think categorically, all mosfets have both. The body diode plays an essential role in some circuits, for example, the standard small signal mosfet circuit for translation of voltage levels (e.g. i2c interface 5V<->3.3V) makes use of the presence of the body diode. In some mosfets, such as CMOS analog muxes and RF mosfets, the substrate comes out as a separate wire which needs to be connected to the most negative voltage n the circuit (or most positive for pmos). But in most mosfets and ICs, the substrate is prewired to the source. The channel and the body diodes are essential parts of the construction.