I am having trouble with voltage regulation

Thread Starter

wallaby

Joined Jul 26, 2011
34
I have new LM317t regulators to work with, and still can't seem to get them to work. R1 is 280 ohm, and I used a trimming cermet as R2 to mock up the circuit. I can rotate the generator shaft at a given RPM and adjust the trimmer to get the desired output voltage... but if I raise the generator RPM, the output voltage also goes up.

If I plug an LED into the circuit, the voltage drops considerable. I can run 1 LED and my meter shows a drop from 5v to 1v output. How is it possible that the LED can light up at such low volts? If I add another LED in parallel, the second LED never lights up. 2 LEDs in series, they both work fine.

I am starting to think that current regulation might be a better option, but given my poor luck with voltage regulation... Lots of trial & error going on here.
Does anyone know of a pre-made LED driver (or plans for one) that can tame my variable input into a stable 3v output? I never imagined it would be so difficult to light up some LEDs with a generator.
 

Alec_t

Joined Sep 17, 2013
14,280
I think there must be a wiring error, or poor breadboard connection, if none of the 317's give the required output voltage.
If I add another LED in parallel, the second LED never lights up.
Not surprising. Unless the LEDs have closely matched forward voltage (Vf) characteristics, one will hog all the current. That's why it's generally bad practice to connect LEDs in parallel.
Why do you now want 3V out? Isn't your LED module rated 4-6V?
 
Last edited:

Thread Starter

wallaby

Joined Jul 26, 2011
34
I first thought parallel would be best in case one light failed. LEDs don't like this, I guess.
I am thinking of ditching the fancy blinking light controller. At a minimum, I would be happy with 2 LEDs that come on when the propeller turns at a reasonable speed.
The problem is that I need RED and GREEN lights... and the red always requires less power than the green.
I managed to find a pair of high-flux LEDs where the red and green colors have identical ratings. But they have 4 pins. They give a diagram showing the anode and cathode, but are there two of those, or the other two pins are dummy.
These LEDs (both red and green) take 3.2v and 20ma.
My generator makes 2.8v at 28 rpm
4.8v at 48 rpm, 7.7v @ 77rpm, etc . So I have enough to light up the LEDs when it's spinning at the speed of a record turntable. As I add regulators, it takes nearly double the speed to get the lights to come on.
I am just trying to find the least power-robbing way to keep the LEDs happy.
 

Alec_t

Joined Sep 17, 2013
14,280
Your generator may be able to create lots of volts open circuit, but have you tested how much current it can provide? Maybe the loading of even one LED is pulling the generator output voltage down and causing the 317 to drop out of regulation?
 

Thread Starter

wallaby

Joined Jul 26, 2011
34
I'm novice. I can put a meter on the output without load, and it shows rising voltage with increase in rpm.
I'm unsure how to test for current. Would that be done with load or without.
 

tracecom

Joined Apr 16, 2010
3,944
With. The meter should be set to the current mode and placed in series with the load. Read the manual for the meter; it's easy to blow the internal fuse if you connect it wrong.
 

Thread Starter

wallaby

Joined Jul 26, 2011
34
Thank you.
I looked at the link you gave for the controller, and it says voltage output is 80v. If my current is regulated, do I care what the voltage is?
I'd like to try some of these... here come the rookie questions: If I run 2 LEDs (maybe more) that need 20mA @ 3.2V, do I want a 20mA driver?
 

ronv

Joined Nov 12, 2008
3,770
Thank you.
I looked at the link you gave for the controller, and it says voltage output is 80v. If my current is regulated, do I care what the voltage is?
I'd like to try some of these... here come the rookie questions: If I run 2 LEDs (maybe more) that need 20mA @ 3.2V, do I want a 20mA driver?
It can take up to 80 volts, but it works down to the LED drop (say 3 volts) plus 2 volts. So in your case 5 volts minimum. Then it holds the current steady at 20 ma even if the voltage increases. That's why it is called constant current.
If you have enough voltage say 8 volts you can put 2 LEDs in series and each will still get 20 ma. I would start with 2 of them. One for each LED, just in case your voltage falls when the load (LED) is on.
This is the 20 ma one:
http://www.mouser.com/ProductDetail...OxUJ5NqHQkayNHMuRVa10rPcEzbntF/DuRk6oWfkjXw==
 

Thread Starter

wallaby

Joined Jul 26, 2011
34
Wow. Finally getting some results. I did the output test with LEDs hooked up and running. There was a big difference in the results. (vs no-load readings) I am turning my generator on a metal lathe.
30 RPM puts out 3.1v and 2.5mA. The lights are just "thinking" of coming on at this speed.
47 RPM puts out 4.7v and 8.8mA.
55 RPM, puts out 5.4v and 34.5ma
After that the voltage regulator kicks in and keeps the voltage steady at 5.7v, but the current continues to climb as speed goes up. I am putting out 60mA with 60RPM.
I think I am on to something when these current regulators arrive. :)
 
There's some very misleading advice in this thread. Good to see that despite that you've made some progress. Here's some tips to keep you going:
  • Your two pole brushed PM generator will deliver a pulsing DC voltage that is roughly proportional to speed. If the generator includes a capacitor across the terminals than the pulsing might be very small, otherwise adding your own will prevent periodic voltage dips as the brushes commutate. At such slow RPMs and only two poles the dips could be significant.
  • Given that PM generators deliver voltage roughly proportional to speed, your numbers (30~3.1, 47~4.7, 55~5.4) suggest the regulator is doing nothing.
  • The LM317, with appropriate feedback, will regulate a fixed output voltage (within 1% or so) provided:
    • The input voltage is <40V.
    • The input voltage is at least 3V higher than the fixed output voltage.
    • The load is drawing at least 10mA. Yes, that's right, a regulator does not regulate until there's a load!
  • The regulator dissipates a bit more than (Vin-Vout)*Iout watts. Provided you're drawing less than 100mA, you shouldn't have to worry.
  • To set a fixed output voltage, connect a resistor R1 between Vout and ADJ, and another R2 between ADJ and ground (not Vout and ground!). The output voltage will be close to 1.25*(1+R2/R1). So if you want 6V output, use 270Ω for R1and 1KΩ for R2.
  • Unless you're confident you don't need them, add an input and an output capacitor. Say 0.1uF input and 1uF output.
  • Put all that together, add a nominal resistive load of say 150Ω, and confirm the following behaviour:
    • Output voltage is roughly proportional to speed up to about 8V (80RPM).
    • Beyond that, output voltage falls and stays constant at 6V with increasing speed.
    • Do not exceed 40V input (400RPM).
Only then should you introduce the complexity of LEDs. LEDs are difficult to drive with a constant voltage source because they have a negative temperature coefficient. In other words, as they begin to conduct current, they heat up and begin to draw more current at the same voltage. This makes them hotter, drawing more current and the cycle may continue until the smoke comes out. If you're very careful with your voltage control, and don't exceed the thermal runaway threshold, it is still possible to drive them this way.

But you may still want to consider current regulation - if you do, consider adding it after the voltage regulator anyway, to give you a stable input to the current regulator. Note that the impact on power consumption of an additional regulator is not large, since the largest determining factor is the voltage of the input (controlled by how fast you spin the generator) times the current drawn by the load, which is dominated by the LED current.
 

Plamen

Joined Mar 29, 2015
101
I am working with a wind-powered generator that only needs to light up some LEDs when the wind blows.
I rewound my generator to make enough voltage at low RPM, but as expected, it provides too much as the wind speed picks up.
I figured a voltage regulator would cure the problem, but I can't seem to get it to work. I get it dialed in at a certain shaft speed, but then as the speed rises, so does the output. I'm using a simple LM317 adjustable regulator.
Isn't the regulator supposed to supply a constant output regardless of input? What might I be overlooking?

Car alternators regulate their output voltage by varying current in so called excitation coil.
Apparently you are using permanent magnet generator whose output voltage rises proportionately with speed.
The linear regulator you have started with operates as step down regulator absorbing the difference between input and output voltage.
This implies very poor efficiency at high speeds i.e. overheating.
Your best option is to use switching mode regulator. If the required output voltage is always below the input voltage - you need a Buck (step down) regulator. If at low speed the input voltage could be below the desired output voltage - you need Buck-Boost (or SEPIC) regulator.
Go to www.linear.com and study LTC3115 and LT8705. The first one is fully integrated regulator; the second a controller with external MOSFETs. LTC3115 can provide up to 2A from a voltage up to 45V. LT8705 can provide up to 60V from a voltage up to 80V.
These are both Buck-Boost switching mode devices. On the Buck side study LTC3639 (low power wide input range Buck).
LT web site provides selection tables, so you can fill your limitations and see a table with suitable parts. LT also offers LTSpice models (the SW is downloadable there for free). If you need specific assistance send me pm.


Mods Edit:
Please don't post your E-Mail address, it could bring the Spammer to the forum and you, thank you.
 
This implies very poor efficiency at high speeds i.e. overheating.
It really doesn't. wallaby is powering LEDs. The input is the wind. At maximum input the loss is (40-6)*0.03 = 1W. That's 25°C above ambient. No big deal. More importantly, what were you going to do with the extra power you save? Spin the generator faster? Save on wind?

Your best option is to use switching mode regulator.
wallaby, get it working with the linear regulator first. A switch mode is far more complex and unlikely to gain you anything.
 

Plamen

Joined Mar 29, 2015
101
Wind power is typically generated to be used (immediately or later).
One common use is to charge batteries i.e. for later use.
Another option is to store as heat.
If your goal is to just drive LEDs - fine. At light load linear regulators are fine. (although 150V, 50 mA Buck LTC3639 is not bad option either)
Most likely you can get away with smaller turbine.
You can set up your linear regulator not as voltage source, but as current source - LEDs like constant current.
To achieve this, connect your LED in the input wire in series with the regulator and load the regulator output with resistor R = Vref / I led,
Where I led is the required LED current, and Vref is your regulator internal reference. The LED could be also in the RTN wire to the generator.
 

Thread Starter

wallaby

Joined Jul 26, 2011
34
I am about ready to mock up this circuit on a breadboard. My LEDs are rated at 3.2v, 20mA. I am starting with just 2 of these LEDs.
If I wire them in series, what should I set my voltage regulator at? and what setting for current regulator?
 

Plamen

Joined Mar 29, 2015
101
I first thought parallel would be best in case one light failed. LEDs don't like this, I guess.
I am thinking of ditching the fancy blinking light controller. At a minimum, I would be happy with 2 LEDs that come on when the propeller turns at a reasonable speed.
The problem is that I need RED and GREEN lights... and the red always requires less power than the green.
I managed to find a pair of high-flux LEDs where the red and green colors have identical ratings. But they have 4 pins. They give a diagram showing the anode and cathode, but are there two of those, or the other two pins are dummy.
These LEDs (both red and green) take 3.2v and 20ma.
My generator makes 2.8v at 28 rpm
4.8v at 48 rpm, 7.7v @ 77rpm, etc . So I have enough to light up the LEDs when it's spinning at the speed of a record turntable. As I add regulators, it takes nearly double the speed to get the lights to come on.
I am just trying to find the least power-robbing way to keep the LEDs happy.
Petkan:
Using a linear regulator in this circumstances is questionable. Your input voltage could be below the required and your regulator is not even capable to start at so low voltage. Consider Buck Boost regulator like LTC3115 from LT (see www.linear.com).
Your regulator is only step down (cannot make more that the input voltage -3V). Buck Boost regulators can smoothly transition from stepping up (when input voltage is low) to stepping down (when it rises). Others have already pointed that LEDs, zeners, diodes poorly share current when connected in parallel. When connected in series they will require higher voltage. However Buck Boost regulator has no problem producing an output higher than its input. Also LEDs prefer constant current mode (to maintain constant brightness and colour).
This implies that your regulator (no matter linear on switching mode) could be made "current", not "voltage" source i.e. maintaining. constant output current, not voltage.

When a fixed voltage regulator is loaded with fixed resistor, the current through the said resistor is constant. The input current of a linear regulator is almost matching this constant current as the regulator has very low consumption. This implies the load (LEDs) are in the input of the linear regulator, not its output (ADJ wired to OUT and loaded by resistor R=Vref/Idesired).
With switching regulators to get a current mode we have to make the overall negative feedback current not voltage. This is achieved by a small resistor in series with the LEDs again selected as R= Vref/Idesired, where Vref is the internal voltage reference of the regulator and Idesired is the desired current.
 
I am about ready to mock up this circuit on a breadboard. My LEDs are rated at 3.2v, 20mA. I am starting with just 2 of these LEDs.
If I wire them in series, what should I set my voltage regulator at? and what setting for current regulator?
Assuming firstly there's no current regulator, use a potentiometer in your voltage regulator feedback circuit so you can adjust the output voltage. Start at about 4V to be safe, and increase from there while monitoring current. Do not exceed 20mA. In fact, don't go above about 15mA to allow for variation over temperature, component tolerance and voltage sharing. Will probably be about 6V, but beware that it will be far from linear - a small voltage change can cause a big current change. To improve voltage sharing between the LEDs, you could add a parallel resistor of say 1K to each one but it's a balancing act.

Now assuming you've got your current regulator hooked up, set the voltage regulator high enough to keep the current regulator happy (probably at least 7V) and set your current regulator at 20mA.
 

Plamen

Joined Mar 29, 2015
101
Assuming firstly there's no current regulator, use a potentiometer in your voltage regulator feedback circuit so you can adjust the output voltage. Start at about 4V to be safe, and increase from there while monitoring current. Do not exceed 20mA. In fact, don't go above about 15mA to allow for variation over temperature, component tolerance and voltage sharing. Will probably be about 6V, but beware that it will be far from linear - a small voltage change can cause a big current change. To improve voltage sharing between the LEDs, you could add a parallel resistor of say 1K to each one but it's a balancing act.

Now assuming you've got your current regulator hooked up, set the voltage regulator high enough to keep the current regulator happy (probably at least 7V) and set your current regulator at 20mA.
Petkan:
If LEDs (or any diodes zeners etc) are be parallelled, current sharing would be subject to huge part to part variation. Adding individual series resistors can help reducing it. However series connection makes the current in each LED the same (preferred for LEDs).
Operating from regulated fixed voltage requires ballast resistors - otherwise currents will dramatically change with temperature.
Voltage regulators can easily be set as current regulators. There is no need for separate voltage and current regulatots
 
Top