Shop OBEX P1 Docs P2 Docs Learn Events
Low Noise 5V Regulator with 24V Input — Parallax Forums

Low Noise 5V Regulator with 24V Input

I'm working on a project which has a 24V input and we need to power both 5V and 3.3V devices.

We're using a current driver which can output a 5V source but it's limited to about 25mA of current. We want to increase the available 5V supply to about 50mA.

I don't have a lot of voltage regulator experience but I've used the LM1086 a few times. I think the main reason for using the LM1086 is that is what is used on the Propeller Proto Boards. From my reading of the LM1086 datasheet, it shouldn't be used with voltages higher than 20V. The LM1086 is also larger than what we'd like to use (board area is a bit tight).

We presently use a LM2476 switching regulator for the 3.3V supply which appears to be working well (there is a Propeller on the board).

Since we have a decent 3.3V supply, one suggestion was to use a boost regulator like the REG711.

http://www.mouser.com/ProductDetail/Texas-Instruments/REG711EA-5-2K5/?qs=sGAEpiMZZMtitjHzVIkrqfSWpcWTPe%2bS4QkMF7khHQM=

I don't know if my gut reaction to avoid a stepup regulator is justified. It just seems odd to have current pass through both a buck and a boost regulator.

While it would be nice to have 50mA at 5V available, we will likely only use 25mA of current at this voltage. I figure dropping the voltage from 24V to 5V with 25mA of current should generate 0.475W of heat if we used a linear regulator. This seems like an acceptable amount of heat for this application so I'm wondering if a linear regulator would be our best option.

One concern I haven't mentioned is we're using wireless interface (Bluetooth) so we'd like to minimize RF noise. I personally have experienced many RF issues when using switching regulators so I'm hesitant to add a second switching regulator to this project.

Am I worrying too much about the stepup regulator? Are there linear regulators which would work well in this application?

Comments

  • I think a more sensible approach would be to use a 24V->5V switcher, followed by a linear LDO regulator for the 3.3V supply.

    -Phil
  • LoopyBytelooseLoopyByteloose Posts: 12,537
    edited 2015-08-17 19:23
    Phil is right. Going from 24VDC to 5VDC with a linear regulator creates too much heat for a sensible construction. You might find the voltage regulator so hot that it melts the solder and slides right off the board if it doesn't do a thermal shut down first.

    There are lots of these tiny boards with switchers that will handle 24VDC going down to 5VDC and run very cool. Plus they can give your more useful amps at 5VDC.

    +++++
    At least try the switcher with the Bluetooth before you get too far into going the other direction.
  • With such low output current, consider the 78L05. It accepts up to 35v input, and could provide 100mA output at your differential. Comes in a variety of packages: SOIC and TO-92. Pretty low noise.
  • I think a more sensible approach would be to use a 24V->5V switcher, followed by a linear LDO regulator for the 3.3V supply.

    -Phil

    This approach has crossed my mind many times but our main current users are 3.3V devices. The 3.3V current is likely below 100mA most of the time but it can peak above 200mA.

    As I stop to punch some numbers, I see we can pass 279mA from 5V to 3.3V for the same heat as pulling 25mA from 24V down to 5V. I suppose a 5V switcher and 3.3V linear does make the most sense.

    Thank you for your input Phil.

  • Duane DegnDuane Degn Posts: 10,588
    edited 2015-08-17 19:55
    Going from 24VDC to 5VDC with a linear regulator creates too much heat for a sensible construction. You might find the voltage regulator so hot that it melts the solder and slides right off the board if it doesn't do a thermal shut down first.

    Did you see the current requirement? I doubt half a Watt is going to start melting solder but I think half a Watt can produce more heat than we want.
    Sapphire wrote: »
    With such low output current, consider the 78L05. It accepts up to 35v input, and could provide 100mA output at your differential. Comes in a variety of packages: SOIC and TO-92. Pretty low noise.

    Thank you for the suggestion, I'll pass on the information. That part is what I was looking for as far as a linear regulator to use with a 24V supply. I'm betting we'll use Phil's option but I appreciate knowing there's a linear regulator option.

  • Duane DegnDuane Degn Posts: 10,588
    edited 2015-08-17 21:10
    Here's a bit more detail about what we're presently using as a 5V regulator. The current driver XTR111 has a regulated voltage output. Apparently when this is combined with some external circuitry it can be used as a voltage regulator. I see limits of 5mA and 25mA mentioned on page 18 of the datasheet.

    http://www.mouser.com/ds/2/405/xtr111-450892.pdf

    I'm pretty sure I've been interpreting page 18 incorrectly since I don't think there's a way the XT111 chip is outputting anywhere close to 25mA through the 100K resistor.

    As I study the datasheet a bit more, I think it's pretty clear the vast majority of the 5V supply is coming from the 24V source through the transistor in this circuit.

    Xtr111_DatasheetFig.JPG

    I'm trying to figure out what the limit of this circuit is. Is the limit determined by the transistor?

    I tapped into this 5V source in order to add a new sensor to the device I'm working on but I don't think using this 5V source is a long term solution and I think the next board revision should include a proper 5V regulator.

    Am I correct in thinking the regulator circuit above should be limited to currents below 25mA? If I'm wrong about this, we may not need an additional regulator at all.
    404 x 347 - 22K
  • jmgjmg Posts: 15,173
    Duane Degn wrote: »
    Am I correct in thinking the regulator circuit above should be limited to currents below 25mA? If I'm wrong about this, we may not need an additional regulator at all.

    You should be able to re-spec that design to 50mA
    There is no implicit current sense, but they have a 220 r series R in the collector, which works as a very simple brief short circuit limiter.
    No Trx part code is given, but change to a 'decent' part in a thermally capable SOT223 or DPAK with Hfe > 200 and you should be good for 50mA, with some thermal care.

    If you know the load current, you can also spread that heat around with multiple resistors and Power devices. (ie do not get a single part to cope with all the heat )

    If the power budget gets too high, there are also encapsulated 5V switchers that fit into TO220 footprints.
    A couple of examples :
    http://www.digikey.com/product-detail/en/R-78E5.0-0.5/945-1648-5-ND/
    http://www.digikey.com/product-detail/en/TR05S05/1470-2254-5-ND/

  • That seems like a very complicated way to get 5V and anyway the usual way is to switch down to 5V and use a linear reg for 3.3V logic as this removes any switching ripple in the process. The 78L05 mentioned would light up like a match with almost 1W of heat (19V*50ma).

    There are nice and cheap switching reg modules in 3 pin 7805 compatible packages for under $3. I mostly use an MCP1700 in an SOT89 for most of my 3.3V designs as they only require 1uf ceramics for stability.

  • LoopyBytelooseLoopyByteloose Posts: 12,537
    edited 2015-08-18 06:46
    The 78L05 mentioned would light up like a match with almost 1W of heat (19V*50ma).

    Agreed, it is all about how much heat the linear voltage regulator can dump, not the maximum voltage specification.

    Think about this... almost nothing in transistors goes below the maximum volt limit of 35VDC unless it is a special design. That is just the nature of the silicon and BJT construction. That doesn't mean that number is useful for operating input voltage. The smaller the package, the harder it is to dump heat -- the thermal limits of the package take priority.

    What that number indicates in how much voltage the linear regulator will tolerate in spikes. In automotive environment on might prefer a higher number as spikes from staring the motor, honking the horn, or having the alternator kick on and off can be significantly higher. Often a zener diode at about 25 volts is used to protect the regulator from higher than 35 volt spikes.

    In other words, about the highest I go with a 5 volt linear regulator is 14.2VDC input, and then it requires a formal heat sink. Generally 7.5 volts is ideal.

    Why 14.2 VDC? Considering that autos 12VDC system might really run at 14.2VDC when charging the battery, there are some good linear regulators designed specifically for automotive use.
  • Peter JakackiPeter Jakacki Posts: 10,193
    edited 2015-08-18 06:59
    The 78L05 mentioned would light up like a match with almost 1W of heat (19V*50ma).

    Agreed, it is all about how much heat the linear voltage regulator can dump, not the maximum voltage specification.

    Think about this... almost nothing in transistors goes below the maximum volt limit of 35VDC unless it is a special design. That is just the nature of the silicon and BJT construction. That doesn't mean that number is useful for operating input voltage. The smaller the package, the harder it is to dump heat -- the thermal limits of the package take priority.

    What that number indicates in how much voltage the linear regulator will tolerate in spikes. In automotive environment on might prefer a higher number as spikes from staring the motor, honking the horn, or having the alternator kick on and off can be significantly higher. Often a zener diode at about 25 volts is used to protect the regulator from higher than 35 volt spikes.

    In other words, about the highest I go with a 5 volt linear regulator is 14.2VDC input, and then it requires a formal heat sink. Generally 7.5 volts is ideal.

    Why 14.2 VDC? Considering that autos 12VDC system might really run at 14.2VDC when charging the battery, there are some good linear regulators designed specifically for automotive use.

    Don't know about the zener as that would just end up burning out but transient protection diodes might be used which sound similar to zeners but are different animals.

    There are cases where you could use a linear regulator with a higher voltage, even if the load current was quite high. It has to do with duty cycle and heat as well as chip hot spots. You could use a 78L05 from 24V if you are perhaps needing only a few milliamps mostly and sometimes short bursts of say 50ma, that woud work nicely enough and be simple and cheap, which may indeed be a requirement for a design. Remember, absolutes are only relative :)


  • Duane DegnDuane Degn Posts: 10,588
    edited 2015-08-18 22:07
    Thanks for all the suggestions guys. It looks like we'll use a switching regulator down to 5V and use a linear for the 3.3V stuff.

    @ Peter Jakacki, I like the MCP1700 too but I think it's a bit underpowered for this application. It's good up to 250mA which is cutting it a bit close for our 3.3V peak power needs.
  • Peter JakackiPeter Jakacki Posts: 10,193
    edited 2015-08-19 01:04
    Duane Degn wrote: »
    Thanks for all the suggestions guys. It looks like we'll use a switching regulator down to 5V and use a linear for the 3.3V stuff.

    @ Peter Jakacki, I like the MCP1700 too but I think it's a bit underpowered for this application. It's good up to 250mA which is cutting it a bit close for our 3.3V peak power needs.

    As you should know the 250ma limit is actually the average limit as the regulator has a peak limit of 550ma.
    6.5 Pulsed Load Applications
    For some applications, there are pulsed load current
    events that may exceed the specified 250 mA
    maximum specification of the MCP1700. The internal
    current limit of the MCP1700 will prevent high peak
    load demands from causing non-recoverable damage.
    The 250 mA rating is a maximum average continuous
    rating. As long as the average current does not exceed
    250 mA, pulsed higher load currents can be applied to
    the MCP1700
    .
    The typical current limit for the
    MCP1700 is 550 mA (T(a) +25°C)
    Besides, I frequently use more than one just to keep the processing section supply clean, maybe you can do the same?

  • Cluso99Cluso99 Posts: 18,069
    I use this switcher
    R-78E5.0-0.5 5V 28V 0.5A
    Digikey 945-1648-5-ND
Sign In or Register to comment.