25-45 DC in, I need 12V DC out.

Question

I need stable 12v for the LED turn signals on my motorcycle but the bike doesn't have a battery (from the factory) so everything is in AC. I used a bridge rectifier (KBPC5010) which outputs anywhere between 25-45 volts DC (maybe even more since I didn't rev it too high and it depends on engine RPM) but my flasher relay works on 12V. I made a circuit containing a 1000uF capacitor, 7812 stabilizer and 2 smaller ceramic capacitors, the components were connected in that order. It worked, the output I got was a stable 11.9V (according to my butchered multimeter). I put all of that in a plastic shroud from a glowplug relay and hot glued everything in place. It worked for a little bit until it stopped working, the capacitors are fine but the 7812 stabilizer shows no voltage on the output pin. My question is, did I fry it with too much voltage coming out of the rectifier, or was my mistake inadequate cooling? What would you guys suggest I build to make my turn signals work? Thanks in advance!

michaelkellett · Accepted Answer

@ WarrenW - I don't think your values will work very well::

Suppose the load current is 1A, the TIP3055 isn't accurately specified for current gain at Ic = 1A but interpolating from the ST and On Semi data sheets we'll assume a gain of 40. So we need a base current of 25mA. To regulate reasonably the 1W zener current should never be less than 5mA and never more than 80mA so the resistor value should be about 475R, use 510 which gives a resistor current of about 75mA at a supply of 50V. 25mA will go into the base of the transistor and the rest into the zener. The circuit will start to lose regulation when the resistor current drops to 30mA, at which point the input voltage will be 12 + (0.03 * 510) = 27.3V. This may be just about acceptable.

It's not a nice design because the resistor needs to be rated at 5W (and could have to handle 3.3W continuously - the problem is that the TIP3055 doesn't have enough gain.

Lets try a BDX33C, cheaper than the TIP3055, gain minimum 750 at 3A, rated at 100V and 10A max, use a 1N4743A 13V zener.

The zener has a test current of 19mA so we can expect the voltage to be below 13V for lower currents. The darlington transistor will only need a base current of 1.3mA so, if we aim for lowest working input voltage of 20V, and 5mA zener current the resistor will need to be (20 - 13)/(.005 + .0013) = 1076R, lets use 1k. And at max input voltage of 50V the resistor current will be 37mA, and it will dissipate 1.7W.

So my advice is use R = 1k, 3W, zener 1N4743A, transistor BDX33c, all parts available from Farnell.

The output capacitor and zener capacitors can be small - 47uF at 25V will be fine, the input capacitor depends rather on the frequency of the AC, which we don't know, but I would go for 1000uF (will drop 5V in 5mS at 1A) to start with. It ought to be a decent part and rated at 60V or more.

MK

WarrenW · Answer

Firstly, msimon is correct the input voltages of a 7812 will not withstand prolonged high voltage. There is a high voltage equivalent but it may be better to consider a transistor series regulator.

I have attached a simple circuit for reference. Select a transistor that has enough voltage Vce rating and current.

Filter capacitors do not need to be huge. Input filter is the most critical. Make it perhaps 470Uf 63V. Zener diode is 12 volt.

WarrenW · Answer

What was the resistor value you were using? The Zener sounds like it was working overtime and dissipating a bit of heat! I would really suggest you look at the components I indicated and the values. The calculations are all based on Ohms law for power and voltages etc. P=V x I , V=IxR, I=V/R, R=V/I. Have a look at this for a better explanation Ohms Law Tutorial and Power in Electrical Circuits With a TIP142 transistor the power dissipations go right down due to the gain of the device. I would recommend you build the circuit. Sounds like the zener may now be fried. Replace it to be safe. Cheers Warren.

clem57 · Answer

OMG! The bridge rectifier is rated at 40-50 amps while the 7812 stabilizer is only rated for 1 amp. This is a big mismatch on current. Yes it is fried.

WarrenW · Answer

It would depend upon the final current draw of the indicators - I would suggest perhaps a TIP3055 on a heatsink and a 2k2 resistor (1 watt). Make the zener a 1 watt as well (softer regulation and more robust). Output filter capacitor can be 100UF 25V the one across the zener can be the same.

Very reliable circuit and extremely robust - but you should always have a fuse in circuit for protection (input side).

maroka · Answer

A little update, I found this: DC/DC Понижаващ модул с LM2596HV, LM2576HV | Захранване Ком
It's pretty cheap and the store is in my country. It's based on the LM2576HV and I've seen some reviews on the internet and people seem happy with it. On the page there's a "drawbacks" section. It says that when the difference between the input and the output voltage is greater 15-20V, the board can't reach its output of 2 amps without overheating, in order to do it, it needs a different inductor. I don't think a bunch of LEDs can draw 2 amps (though as I said I'm no expert so I could be entirely wrong) what do you guys think?

dougw · Answer

Technical Article: How a Motorcycle Charging System Works - ElectroSport Note the solutions suggested in the comments above will need significant heat sinks. You can buy add-on motorcycle regulators - they have fins to help keep the temperature down. Or you can build your own: How to make a reliable motorcycle voltage regulator The LM2576HV you found will be much more efficient (less heat) and probably work for low current loads. You can get similar switching modules with higher current capacity if you need higher current. The CPT power module you found also looks like it should do the job for a low current load.

WarrenW · Answer

Hi, What you want to do is make sure that the zener is regulating - as suggested remove the transistor. Your 18V transformer with a bridge should be fine for confirming the zener is indeed regulating. This simple circuit is effectively known as an emitter follower series regulator. The base of the transistor is held at the zener voltage and the output (emitter) will be approx. 0.6 volts lower than the zener voltage (with a small load on it - less than 0.5 amp). The regulation will be pretty stable as the current into the load increases. If you are using a BDX33 this is a darlington arrangement (effectively 2 transistors in one package), it requires less drive current into the base as you have 2 transistors to 'amplify' the current. With 2 transistors you have 3 base emitter junctions each at 0.6v so the output voltage will be 1.2volts lower than the zener voltage, but would be VERY stable under even small loads. Assuming the Hfe (current gain) is 100 and you want 1 amp of output. 1 amp divided by the gain would be the required input current to the base. Or in this case 10mA. For good regulation the zener current should be approx. 5 times the base current so let us assume we will want a total of 60mA (10 to the transistor and 50 to the zener). The zener voltage is 12 volts and when running at 50mA (P=I*V) the power dissipated would be 600mW so a 1 Watt zener is the smallest you would go for. The real problem you have is the very wide range of input voltage. If we say you have 45 volts of input, the voltage dropped across the resistor is going to be 45 -12 = 33 volts. We have to supply the zener / transistor with 60mA so that makes the resistor value 33/.06 = 550 ohms (closet is 560) and the wattage will be 33*.06 or 2 watts. As the input voltage decreases the amount of voltage across the resistor drops and therefore the current into the zener /transistor drops as well. At 24 Volts input we have only 12 volts across a 560 ohm resistor = 20mA if your load is 1 amp and the transistor is 100 gain then you have 10mA going into the base and 10mA into the zener. It will still regulate but is now dependant upon the characteristics of the zener. Best option is to find a higher gain transistor as this will allow you to drop the current through the zener / transistor combination. A TIP142 may be a better option with a VCE of 50V allowing you input voltage to get close to 60V. It also has a current gain of 1000. This means for a 1A load you only need to supply 1mA into the base! you can run with 5 -10mA zener current and this means your zener can go down to a 400mW version if needed for better regulation. The resistor value would now be 3300 ohms to give you a 10mA current flow into the zener /transistor at the 45V input voltage. When the input voltage drops to 24V the current in the zener / transistor would then be approx. 4mA. All good! The resistor wattage also reduces to a 1/2 watt version. Hopefully this long drawn out response will be of some assistance. Warren.

WarrenW · Answer

If you use a TIP142 you will only need a 3k3 resistor. The issue with that is that it will limit the current through the zener to smaller values. A larger wattage zener has a slightly softer regulation curve than a smaller one. Given for example you have a 45volt input to the regulator and a 12 volt zener. The difference in voltages is therefore 45v-12v = 33v. Using a 3k3 resistor results in a current flow in the zener of 33v/3300 ohms = 10mA. the wattage dissipated by the zener will be 12v x 10mA = 120mW so a 400mW zener rating would be fine. Oh and I just had a thought, you are running the generator voltage through a rectifier right? If not that could be the cause of the issues you have had with the diode smoking. A 470uF capacitor should be ok. As a general rule - allow 1ooouF per amp per volt of loss of regulation at 50Hz. your motor will be running at I guess at least 1200 rpm at idle (20Hz) and maybe 7500 at full rpm (125 Hz) so there is a bit of guesswork there. If you do NOT have a rectifier from the output of the generator to the input of the regulator you can add a diode into the feed +ve line (use a 5Amp rated device) or better still connect the output of then generator to a bridge rectifier and then to the regulator. I hope this helps. Cheers Warren.

WarrenW · Answer

Awesome stuff. I am glad that we could be of assistance. Learning by remote can be a very hard experience. I would have to say that you have done extremely well and taken advice on board that ultimately has helped you get your project operational. I also ask for assistance when stumped. No-one can know everything. I pays to draw on others experience to broaden your own knowledge base. Well done! Cheers Warren.

msimon · Answer

Hi Vladimir,

The datasheet says max input voltage is 35V for Vo = 5 to 18V. You have the 25-45V input (it is vague) which means it may exceed the absolute maximum rating. Overheat may also cause the system failure. I am not a power designer but If I were you I looked for a transformer to reduce the voltage to an acceptable level and use a rectifier with a voltage regulator.

MB

michaelkellett · Answer

It's a good suggested circuit for Vladimir's application - simple to build and cheap - just needed a little tweak (although he'll be OK if his LED lights draw only 100mA rather than the 1A that I guessed).. The trouble with using a nice efficient switcher is that you either buy one of the dodgy Ebay made up boards or put a lot of effort into building your own. I've been having trouble with the gain of 2N3055s and their clones for 40 years - they were used in audio amps quite a bit in the 1970s - the Quad 303 power amp used them. But the low gain (very low at high currents) and low Vceomax were always problems. I was amazed at just how cheap the BDX33C was - got me thinking about designing an audio power amp using them !! MK

michaelkellett · Answer

Try removing the transistor and measuring the voltage across the zener for input = 10V ..... 30V in 5V steps. Check your caps are all the right way round. Check it is a 12V zener ! MK

michaelkellett · Answer

Its a pretty poor zener specced for a test current of 50mA. With 12V and 2k resistor your current would have been about 2.5mA and with the 30V in about 10mA. The spec of the zener doesn't say what it will do at such low (compared with the spec) currents but 8.6V is not impossible. The zener I suggested is better (test current 19mA) but it will still vary as the supply voltage varies. To get good results you need to keep the current in the zener closer to the test current but this is very difficult without using more parts. To check the zener you need a range of resistors or a variable power supply. Your home supply needs a capacitor across the output (1000 uF would do) if you are going to use it. So long as you measure the current and voltage through the zener for different resistor values you can still work out the characteristics of the zener. (It's just a bit easier if you have all the gear.) MK

maroka · Answer

Allright, just got done testing the zener with the transformer I got, it was reading around 13V across it while radiating a considerable amount of heat. I think I baked it back at the lab because there was some smoke coming out when my teacher tested the zener, I guess the PSU there supplied too many amps. Flipping it around gave me around 0.03V while attached to my transformer but it was buzzing a lot louder. Warren, just to confirm, you're suggesting I build a new circuit with the values you stated above? I once again apologize for my newbie questions, it's all a little complicated to me.

maroka · Answer

So I just retested the zener and it seems that there was some dirt in the solder or something else which was preventing the current from going through the zenner to close the circuit. Right now with just the zener and the 2K2 resistor attached to my transformer I'm getting 7.8V across the zener and no smoke. Last time I had just connected the diode to my transformer with no load on it, which I'll admit wasn't very smart but it didn't want to work with the resistor. I will see when I can go to the store to buy the hardware. Thank you for the help guys, I really appreciate it. I'll post an update when I mount it on the bike.

WarrenW · Answer

YES YOU HAVE THAT CORRECT.

maroka · Answer

I got it all wired up today and it worked! Revving the bike didn't change neither the frequency of the flashing nor the brightness, so I guess it's safe to say that it works! Huge thanks to everyone who participated and helped educate me on the subject. And once again I apologize for being a such a slow learner. Take care -Vladimir Edit: It's really difficult to select the correct answer, since they are all correct . I think that Michael's answer will be easier to understand by someone who didn't read the whole thread. Don't get me wrong, I went with Warren's solution which works great! I just think that Michael's solution is easier to follow by newcomers to the thread, since it's in a single message. I hope I don't get anyone mad.

michaelkellett · Answer

I have to go home now but I'll make some suggestions first thing tomorrow. Have you got a reasonable voltmeter ? MK

25-45 DC in, I need 12V DC out.

Top Replies