Zener diode power dissipation: trusting the datasheet?

Thread Starter

minisystem

Joined Dec 4, 2008
32
Hello. I don't have enough experience to make an informed judgement about this:

I'm using a 5W 6.8V Zener diode to shunt 600 mA. datasheet:http://61.222.192.61/mccsemi/up_pdf/SMBJ5338B-SMBJ5369B(SMB).pdf

It needs to do this at a frequency of about 2 Hz with a duty cycle of 50%. ie. it is shunting the current half the time and each pulse lasts about 0.25 seconds.

The datasheet says the maximum Zener current is 700 mA. If I understand correctly, the power at that current is 6.8V x 0.7A = 4.76W. I'm shunting 600 mA or 4.08W, so about 85% of the maximum rated Zener current.

Again, if I'm interpreting the datasheet correctly, shunting 600 mA at 50% duty cycle should be OK. The diode's ground lead is connected to the ground plane on the PCB, which is probably a little over 1" square.

My lack of experience is making me doubt my design and wonder if I'm pushing this diode too far. It gets very hot when shunting the current. Hotter than I would expect considering its only doing it half the time. Should I add a heat sink to the package? Is it bad design? An alternative way to shunt the current is going to be complicated.

Should I trust the datasheet and just go with it?
 

Adjuster

Joined Dec 26, 2010
2,148
The data sheet power rating assumes 25°C lead temperature at the leads, which you may not obtain.
Note: 2. Lead temperature at 25°C = TL at mounting plane. Derate linearly above 25°C to zero power at 150°C
Having more copper than needed for the minimum pad size will of course help. On the other hand, from fig.3 in the datasheet, it looks as if 2Hz is fast enough to get a bit of advantage from not being on all the time.

This also begs the question however that you are really sure about what current is passing in the first place.
Is 200mA a design estimate, or have you actually checked it?
 

Thread Starter

minisystem

Joined Dec 4, 2008
32
This also begs the question however that you are really sure about what current is passing in the first place.
Is 200mA a design estimate, or have you actually checked it?
The power source is current limited to 600 mA, so when the Zener starts conducting at 6.8 Volts the maximum current passing through it is 600 mA. Of that I'm sure.
 

Thread Starter

minisystem

Joined Dec 4, 2008
32
What is your application?

There may be an easier way than using such a high power Zener.
Well, I was hoping nobody would ask that. :D

It is a circuit for flashing LEDs using a bicycle dynamo hub as a power source. Here is the schematic (sorry to strain your eyes)



There' a bunch of stuff there for charging a supercapacitor to keep the LEDs on while stopped which you can ignore. The dynamo hub is current limited to about 600 mA so when the wheel is spinning no extra driver circuitry is required.

D2 is the Zener. Q3 disconnects the LEDs (connected through J2 and J3) with pulses from the 555 to make them flash. When disconnected from the load, the dynohub voltage spikes. I need to keep it in check to protect the standlight circuitry. When it hits 6.8V the Zener conducts. I figured this was the simplest (but most wasteful) way of dealing with the dynamo hubs no-load voltage in between flashes. I don't really know what else to do with while still keeping it simple.
 

Adjuster

Joined Dec 26, 2010
2,148
One obvious way, if you must stick with the shunt regulator, would be to use zener to activate a power transistor. If the limit voltage is critical, you would want to get something like a 6.2V device to allow for the extra Vbe drop.

Incidentally, you talk of a capacitor to hold up when the generator is stopped, but there seems to be no smoothing capacitor on the rectifier output. Doesn't that give you problems?
 
Last edited:

studiot

Joined Nov 9, 2007
4,998
Even if you use a transistor the shunted power will be the same and present the same dissipation problem.

I would expect any semiconductor component running in excess of a couple of watts to need a heatsink.

Far too many semiconductor designs in overlook this, considering we have been using them for half a century now.
 

Adjuster

Joined Dec 26, 2010
2,148
Yes, but something like a TO-220 transistor would be easier to tie down to a lump of metal.

Something tells me that this isn't the right way to go anyway, as the system seems to be trying to waste about half of the generator output anyway. Wouldn't it be more attractive to store charge during the gaps between the flashes, so that the light can be brighter
(or are your lamps already at full power)?
 

Thread Starter

minisystem

Joined Dec 4, 2008
32
One obvious way, if you must stick with the shunt regulator, would be to use zener to activate a power transistor. If the limit voltage is critical, you would want to get something like a 6.2V device to allow for the extra Vbe drop.
So the idea here is you shunt the power through a transistor that has a package that can handle the heat better than the SMB package? (EDIT: saw your answer in previous post)

Incidentally, you talk of a capacitor to hold up when the generator is stopped, but there seems to be no smoothing capacitor on the rectifier output. Doesn't that give you problems?
I'm using a MOSFET rectifier instead of a diode rectifier in order to avoid the loss from the voltage drop across rectifier diodes (it really does work - less flickering at low speeds). If I wanted a smoothing capacitor on the output I think I'd need a diode to prevent the capacitor from discharging back through the MOSFET rectifier and since the point was to minimize diode drops (there's already one used for power ORing), I wanted to avoid using one again. The LEDs don't care about having a smoothed output and the 3.0V regulator doesn't seem to mind either. C6 is a smoothing capacitor for the supercap. I think I put that there because I was worried about the ripple for the supercap (I don't think they can handle as much as regular capacitors). When I was prototyping the supercapcitor charging circuit the waveform going into the supercap was pretty scary looking and C6 made it look much nicer!
 

Thread Starter

minisystem

Joined Dec 4, 2008
32
Yes, but something like a TO-220 transistor would be easier to tie down to a lump of metal.

Something tells me that this isn't the right way to go anyway, as the system seems to be trying to waste about half of the generator output anyway. Wouldn't it be more attractive to store charge during the gaps between the flashes, so that the light can be brighter
(or are your lamps already at full power)?
That's exactly what I'd like to do. I guess that would be some kind of charge pump? I've looked into it but the design is beyond my experience and capabilities. The idea is that during the LED off cycle the dynamo output is dumped into a capacitor and during the LED on cycle the energy in that capacitor is added to the dynamo output to make a brighter flash? I would love to be able to do that. Any pointers would be most appreciated. :)
 

Adjuster

Joined Dec 26, 2010
2,148
I would try either to solve the bleed back problem in the synchronous FET rectifier, which may not be impossible, providing you are using depletion types, or else dump it and use a bridge of schottkys instead. Then I would fit a reservoir capacitor, big enough to fill in the gaps between the generator pulses. I would think that with your present circuit there could be problems with AC ripple at the input. Have you been testing on the dynamo, or with a bench generator?

It might not be practical to use an ordinary cap to store all the energy generated in the gaps between flashes at 2Hz though. If you wanted to store the charge generated at say 600mA for 250ms, for 2V rise, from C = IT/V, we get C = 0.075F or 75000μF: rather big!

Perhaps a more modest capacitor in parallel with a rechargeable battery would be better, but then you would need to watch the charge current.
 
Last edited:

thatoneguy

Joined Feb 19, 2009
6,359
Instead of using a Zener, you could stack two 5Watt LEDs (and a small power resistor) in series to lock the voltage down, and giving you two very bright blinky things (which will need a heat sink) while doing so.
 

Thread Starter

minisystem

Joined Dec 4, 2008
32
I would think that with your present circuit there could be problems with AC ripple at the input. Have you been testing on the dynamo, or with a bench generator?
I actually have a hub dynamo in a wheel that I run on a motorized testing jig up to 50 km/h. The lack of a smoothing capacitor doesn't seem to be causing any problems. Should I be worried?
 

Thread Starter

minisystem

Joined Dec 4, 2008
32
Instead of using a Zener, you could stack two 5Watt LEDs (and a small power resistor) in series to lock the voltage down, and giving you two very bright blinky things (which will need a heat sink) while doing so.
I did think about this, and it's still an option. The problem is I'm retrofitting the LEDs into a vintage lamp housing, so I'm not sure where these extra LEDs would go.
 

Adjuster

Joined Dec 26, 2010
2,148
I actually have a hub dynamo in a wheel that I run on a motorized testing jig up to 50 km/h. The lack of a smoothing capacitor doesn't seem to be causing any problems. Should I be worried?
If there are problems, they would most likely show up at lower speeds. The flashing of the LEDs might for instance get quite irregular.

I suppose that it is just about imaginable that the switcher might come to harm in some way.

The outputs of rectifiers are customarily smoothed. Perhaps it is not necessary, or even undesirable in your application, but it is very orthodox.
 

Thread Starter

minisystem

Joined Dec 4, 2008
32
If there are problems, they would most likely show up at lower speeds. The flashing of the LEDs might for instance get quite irregular.

I suppose that it is just about imaginable that the switcher might come to harm in some way.

The outputs of rectifiers are customarily smoothed. Perhaps it is not necessary, or even undesirable in your application, but it is very orthodox.
Actually, now that you mention it, the 555 output was very erratic. Not shown in the schematic is a 10μF input capacitor for the 555's Vcc that I added. That sorted out the 555's output over the full range of speeds.
 

Thread Starter

minisystem

Joined Dec 4, 2008
32
I would try either to solve the bleed back problem in the synchronous FET rectifier, which may not be impossible, providing you are using depletion types, or else dump it and use a bridge of schottkys instead. Then I would fit a reservoir capacitor, big enough to fill in the gaps between the generator pulses. I would think that with your present circuit there could be problems with AC ripple at the input. Have you been testing on the dynamo, or with a bench generator?

It might not be practical to use an ordinary cap to store all the energy generated in the gaps between flashes at 2Hz though. If you wanted to store the charge generated at say 600mA for 250ms, for 2V rise, from C = IT/V, we get C = 0.075F or 75000μF: rather big!
Let me see if I can figure this out:

Let's say I go with a conventional diode rectifier and then put a large reservoir capacitor on the output - 75mF as you suggest. With the LEDs connected the dynamo voltage would be Vf of the LEDs (about 5V in my case), so the capacitor would charge to 5V, right? Then, during the off cycle the capacitor starts charging above 5V until it is clamped by the Zener at 6.8V. The size of the capacitor (or the timing of the pulse) is calculated so that it doesn't have time to charge to the clamping voltage of 6.8V, thus preventing the Zener from actually shunting any (or very little) current. Then during the next on pulse, the LEDs are on, dropping the capacitor voltage to 5V. That extra reservoir of energy in the capacitor is realized as extra current for the LEDs? Then during the next off cycle the capacitor is back at 5V and has 'room' to absorb more energy to prevent the Zener from clamping?

I must be missing something there. I don't quite fully understand.

I will need to do something: I put some thermal grease and a thermocouple on the 5W Zener diode and it only takes a minute or two for it to hit 100°C+ I turned it off before I let it get any higher but it was still climbing! :eek:
 

thatoneguy

Joined Feb 19, 2009
6,359
I must be missing something there. I don't quite fully understand.

I will need to do something: I put some thermal grease and a thermocouple on the 5W Zener diode and it only takes a minute or two for it to hit 100°C+ I turned it off before I let it get any higher but it was still climbing! :eek:
Retrofitting them in an old timey housing shouldn't be much of an issue, the LEDs as Zener could add a "flicker" to look a bit like a gas lamp instead of LED bulbs. If the outer casing is metal, mounting a tab to the housing would sink the heat, especially with the low duty cycle proposed above. Don't forget: a moving bicycle provides a ton of cooling if you leave vents.
 
Top