I don't think this would ever go as high as 52V loaded. I'm only seeing 15.2V after about 20 mins running.

I was thinking of using a 7812, but that probably won't work quite right.

Full speed is needed as it doesn't spin very fast to begin with but keeps the LED surprisingly cool.

Do you happen to have a schematic/part numbers for the LDO and how it would be wired up with the zener.

 

Here's a discrete-component LDO reg that should protect the fan. Q3 needs to be able to dissipate >1W @ 32V.

 

As usual, my schematics use what is in my library to show the concept. You don't need a 1% resistor or a surface mount zener, but this is the general idea.

ak

 

Nice and simple AK. I try to avoud SMD, my big fat sausage hands have trouble with those little things.

So I could just use a regular resistor (5-10%?) and zener?

For the power transistor, would a D1761 work? or a D1885?? instead of the TIP31C?


Alec, also good, but I want to use as few components as possible.

 

Update with added diode to compensate for Vbe. Vout = 12V, headroom is less than 1 V, 4 parts.

ak

 

2SD1761 - yes.

2SD1885 - not the best choice. Massive spec overkill is not a problem, but higher saturation voltage and higher cutoff current can be. Interesting that you have this critter lying around.

I vote for the 1761, but use what ya got.

ak

 

Will any 12V zener work as long as it can dissipate 1w?

 

2SD1761 - yes.

2SD1885 - not the best choice. Massive spec overkill is not a problem, but higher saturation voltage and higher cutoff current can be. Interesting that you have this critter lying around.

I vote for the 1761, but use what ya got.

ak

actually bought a 4 pack while getting the parts to make the LED dimmer circuit as mentioned in my previous LED thread.

They are beyond overkill lol. So I'll probably just use the 1761 as I have two of them.

 

Using the first schematic, Recalc...

In post #9 you said the unloaded current source output voltage rose to 52V. If that still is possible at LED failure, then lets try this again. An inherent problem with the basic zener clipper regulator is that zener current is high at high Vin-Vout differential voltages.

Under normal operation the source voltage might be as low as 12V or 13V, so for low dropout you need enough base current for the transistor to sit just above saturation. (It can't saturate because the source for base current is the collector.) Assuming a transistor gain of 100 and a 50 mA load, you need 500 uA base current. So, how much headroom can you tolerate above the minimum 0.7V of the transistor? Lets go with 0.5V, for a total differential voltage of 1.2V. In this case the zener bias resistor is 1K. 0.5V / 0.0005A = 1000

With a 52 V input and a 12V zener, the worst case power dissipation in the resistor is 1.6 W. Warm. Zener current is 40 mA and zener power is 0.48W. That's more warm for a little diode package.

None of this is a crisis, just stuff to be careful about. If you're sure that 52V never will happen, substitute a lower Vmax and recalc. As shown, it is possible to get dropout voltage down below that of an LM317 or 7812 if that is necessary. If not, and Vmax isn't extreme, then it's a basic make-vs-buy decision.

ak

 

Assuming a transistor gain of 100

Unfortunately the TIP31C has a minimum gain of 20 according to its datasheet, implying a 2.5mA base current rather than 500uA and consequently either increased zener dissipation or increased dropout voltage

 

I'm going to use a D1761, it looks like it has a gain (hFE) of 60 - 120, but the datasheet is in Chinese.

although to be honest I don't fully grasp what's being talked about here. We learned all of this in school but trying to remember it all a year later when not doing it daily is a bit difficult.

I picked up some 1W, 12V Zeners a few ins ago. 1N4742A.

So given I will be using a D1761, and the 1N4742A Zener, will t his work?

Or just show/tell me what I need to do the calculations.

At any rate, the fan will easily handle 15V although I imagine this will shorten it's lifespan but it really doesn't matter. so as far as overhead, I know the fan can go up to at least 15V while still pulling only 50mA

 

so these are good components to use then??

The gain goes down as the current goes up?

 

what happened? I know there was a reply from AK up there...?!

 

Maybe he changed his mind.

 

Alec - Unfortunately, the minimum gain is 10 (ON Semiconductor, 09/2005, page 2), but that's at a collector current of 3 amp.

Fortunately, the gain is significantly greater than 100 at a collector current of 50 mA, Vce 2V, 25C, and gets better as it warms up (datasheet page 5, figure 8). Also, at Ic of 50 mA, Vbe is actually 0.6V, not 0.7V (Figure 10), and it's tempco is negative (Figure 11) so it too gets better as the device warms.

For decreased zener dissipation and decreased dropout.

GRANPNDR - The 2SD1761 is a different animal, but most of the characteristics are similar enough for this job. It comes in three gain grades (A, B, C). Which part do you have? A is similar to the TIP31, B and C are better for your application.

Be careful about overdriving the fan. Everything about fan life is bearing heat, and small fans have much less tolerance for this than large 4"-5" guys.

Excellent zener, your parts are fine, and most of the math is in my previous post. Two questions: What can you tolerate as the minimum fan voltage under all conditions? and What is the maximum input voltage the regulator will see? Those two numbers drive all of the calculations.

ak

 

Not sure which 1761 I have. It says "D1761 F 149"

 

You could drive the fan through a series resistor. I do it all the time, on things like 14.5v battery chargers and 12v fans.

The fan is a constant stable load so just pick a resistor that gives you 12v @ 50mA when the LED voltage is at its highest (15v?).

 

I'd originally thought of doing that and almost didn't post this thread because of it, but using a resistor would leave the fan underpowered if/when the driver dropped to 12V.

A 300 Ohm resistor will give me 50mA @ 15V but at 12V it's only 40mA.

Is this significant? no, not likely, but ideally I don't want the fan slowing down at all, so I may just stick with AK's circuit for the regulation.

 

Can you guys check over my schematic and board layout?

I might take a whirl at etching this today, which would also be my first ever board etch

 

Since your fan transistor is a bit undefined, per post #29 I'd start with R1 at 2K, not 20K, and adjust from there.

ak