How to prevent transients on bench supply?

LEDBE
milkyspit

milkyspit

Flashlight Enthusiast
Joined
Sep 21, 2002
Messages
4,910
Location
New Jersey
Can one of our resident electronic gurus please tell me how I can prevent transients on a bench PSU? I've killed at least one Luxeon when testing it using my variable current, variable voltage bench PSU. I'd touch the wires to the emitter and poof! Some sort of initial buildup of energy, or transient, blasted the poor emitter. Is there a way to prevent this sort of thing from happening? /ubbthreads/images/graemlins/confused.gif
 
Hi there Scott,

If you are using a voltage supply to test an LED you should
*never* connect the LED directly to the output. You should
always use a resistor in series with the LED to limit
current. If you set the output to 5v with a meter then
connect a 10 ohm resistor in series with the LED you'll
get enought current flowing to test the LED (LS type)with.
For smaller LED's use a 100 ohm resistor.

You may wish to monitor the current through the LED too.

You'll have to be careful if you decide to turn up the
voltage also, because you might get too much current
flowing.

If you adjust the current setting without using a resistor
it may not be enough as the output capacitor(s) might
hold enough energy to blow out the LED regardless of
setting--use a series resistor at all times to be safe--
dont depend on the supplies setting to limit current.


Take care and good luck,
Al
 
[ QUOTE ]
MrAl said:
You should always use a resistor in series with the LED to limit current. If you set the output to 5v with a meter then connect a 10 ohm resistor in series with the LED you'll get enought current flowing to test the LED (LS type)with. For smaller LED's use a 100 ohm resistor.


[/ QUOTE ]

Good advice; I second that. I'd hate to see another Luxeon turn into an NED
(noise-emitting diode). The problem with NEDs is that they only work once. Briefly. /ubbthreads/images/graemlins/grin.gif
 
The other thing you can do is set your CC limit by shorting the output and adjusting the current control to the desired value. Then set the voltage control to some value below Vf, and connect the output to the LED. Turn up the voltage control until it just goes into CC mode, this happens at the Vf of the emitter plus the voltage drop through the leads. If you disconnect the emitter the voltage should not float more than a hundred millivolts above Vf, and reconnecting the emitter should be safe. If you're really worried, just turn down the voltage to below Vf before reconnecting an emitter.

MrAl is right, it is most likely the output capacitor. I have a similar power supply and, although I haven't blown up any LEDs, the sudden discharge is noticeable.
 
[ QUOTE ]
milkyspit said:
Can one of our resident electronic gurus please tell me how I can prevent transients on a bench PSU? I've killed at least one Luxeon when testing it using my variable current, variable voltage bench PSU. I'd touch the wires to the emitter and poof! Some sort of initial buildup of energy, or transient, blasted the poor emitter. Is there a way to prevent this sort of thing from happening? /ubbthreads/images/graemlins/confused.gif


[/ QUOTE ]

I've written this before. If the output voltage of a powersupply is more than a few volts about the Vf of the Luxeon you are testing - by the time the current limit takes action and the voltage drops the Luxeon will be damaged. Some are damaged totally - some are damaged so that at low current they don't light.

I have a Luxeon star that I now use to test my converters with a 4.7V Zener soldered across it. If I forget to lower the voltage on the supply or if the converter is powered up prior to connecting the load, the Zener will clamp that initial surge and protect the Luxeon. I just used a 1W zener that I had in my parts bin.

Even a 20uF capacitor charged to 10V will damage a 1W Luxeon. I've destroyed several bad bin ones proving to myself that this is indeed the case.

george.
 
Ditto MrAl etc

Most power supplies will have a capacitor, at least, across the output for stability reasons. Some have sluggish transient response. It's easier to just use a low value resistor in series with the LED than it is to modify to improve transient response or stability of the PS. A 1 to 10 ohm is ok, I'd prefer less than 1 ohm, but then you'll have to set the voltage just a small amount above the expected Vf ie 0.5 ohm, PS set to 4v for a 1/3W, 7V for a 5W.

Chris
 
In addition to all the other suggestions, I like the idea of using a dummy load for as much development as possible. Also I would leave the dummy load connected, add in the real luxeon, and then remove the dummy load.

George
 
I have a 20w 12v halogen light running off my variable voltage bench supply. I usually have it in parallel with any new load. Even tho it doesn't light at low voltage to a test Led, it makes a pretty good surge absorber as long as I don't try to increase voltage too much, too fast. And it lights up the bench area quite nicely at 10-15v. 90% of the time I'm using this supply to power/test 12v equipment. I really need to get the wiring to that light fixed, tho........
 
I use the same method mentioned earlier. Set the voltage to say, 2volts. Short the supply and adjust the current limit desired. Remove short and hook up test dingus. Ramp up voltage slowly until it gives desired results.
Always have a small (1 ohm) resistor in series is a good idea as well. it can soften the 'blow', and give a good sample point for current measurments. Measure voltage past the resistor to be accurate.

yours, drs the crazed, barely back from con..
 
Okay guys, this all sounds like some good advice, but I'd like someone to confirm some things I think I'm reading here...

1. It's okay to short the bench supply to set the current? I've never done that, as I thought it would be bad for the supply.

2. If I've set the voltage and current to specific levels, then put a 1 ohm resistor in series with the emitter, will I still get the corrent I specified, or will the resistor cause the level of current flow to decrease relative to what I had set?
 
Tried to Google a manual on how to setup a bench supply. Apparently everyone assumes this is common knowledge! Some fancy supplies have a switch setting for this. Under no-load set the voltage all the way up. Note: if current is set all the way down (a good idea) you may not see any readings. Now short at the point you are going to connect your load (including all wiring). Set your current. If you supply does not have a built in meter, or it is not accurate enough use a good meter in current mode instead of shorting the leads. Next step is to open the short (and set the external meter, if used, to voltage and re-connect) and bring your voltage down to the point desired. In DC circuits all components wired in series will have the same current flowing through them. All componets wired in parallel will have the same voltage. If you are using the resistor the voltage reading at the supply is not the drop across the Luxeon. Measure with a separate meter, which is a good idea anyway to eliminate wiring resistance. Note that your voltage will have to be enough for the the resistor and the Luxeon or higher. Which is were the current limit kicks in.
 
Hello again,

I just want to caution everyone that no matter how you
decide to test your LED's the best bet is to ALSO
connect a series resistor. There is more then one
reason for doing this...
1. It protects the LED during initial setup.
2. In case you turn the knob up too high it wont fry the LED
3. Sometimes when turning the knob up it seems like it's
only turning a little but the resulting output voltage
change is enough to blow a LED without a series resistor.
4. Older supplies sometimes develop an output noise voltage
caused by the aging of the pot that increases voltage.
As the pot is turned it is possible that the output voltage
shoots up somewhat until the pot arm comes to rest. This
in itself could cause high current pulses through the LED
and blow it out too if it doesnt have a series resistor.

Take care,
Al
 
OK. I took a 1 ohm 10w and attached it to one of the unused terminals. That's now my attachment point for LEDs or jumpers going to them.
 
Hi,

Could some kind soul explain why "shorting" the power supply would not destroy something or other in the power supply?

Jerry
 
[ QUOTE ]
GrayFox said:
Hi,

Could some kind soul explain why "shorting" the power supply would not destroy something or other in the power supply?

Jerry

[/ QUOTE ]

Sorry about that. In the case of a normal power supply, you are correct. (Like a wall socket or battery.)

The power supply referred to here is a "Lab" or "Bench" supply. These are designed to allow for both voltage and current adjustments. Used on the test bench for all sorts of nefarious purposes.. /ubbthreads/images/graemlins/grin.gif

You can adjust the maximum current you want by shorting the leads and adjusting the current limit control(s) to the level desired. Open the leads and adjust the voltage you want.

In a perfect world, these supplies would work perfectly for LED testing with no risk. Unfortunately, the design of this type of supply often has a caveat. To keep the output power stable and quiet, there's a filter capacitor at the output. Think of it as a tiny battery. It helps smooth and 'stiffen' the output. But it does hold a charge like a tiny Nicad. If you set your current to 300ma, and your voltage to, say, 10v, and connect your 4v 300ma LED to it, the LED may expire, because that little 'battery' is charged to 10v, and immediately 'dumps' it's power into the LED, letting the smoke out. /ubbthreads/images/graemlins/grin.gif If the LED survives that blast, the voltage will drop to 4v, and the current will hold at 300ma. The capacitor acts as a flywheel, storing and smoothing the power. If you've ever grabbed a spinning wheel and tried to stop it suddenly, you see the same effect.

Adding a smallish resistor to the circuit will slow down the 'dump' ("Resist" it.🙂 hopefully saving the LED from stress. Turning up the voltage slowly on my supply is what I use, so that the capacitor's voltage is never above the safe level for the LED. Of course, if you disconnect and reconnect the LED suddenly by accident, it can pop. That's why the resistor's a good idea.

I'd better get back to fixin'. Hope this helped.
Don't get me wrong, by the way, bench supplies are wonderful! Use 'em all the time..

DRS the crazed.
 
That's a great description, especially the "flywheel" part. /ubbthreads/images/graemlins/thumbsup.gif
 
[ QUOTE ]
milkyspit said:
Okay guys, this all sounds like some good advice, but I'd like someone to confirm some things I think I'm reading here...

1. It's okay to short the bench supply to set the current? I've never done that, as I thought it would be bad for the supply.


[/ QUOTE ]

Good question. It does go against the grain, doesn't it? /ubbthreads/images/graemlins/grin.gif
That is the way to set the current correctly. Turn the voltage up a volt or two, and start with the current limit turned down. Short leads and adjust for desired current. Open and adjust for desired voltage, or turn up voltage slowly when connected to load.

[ QUOTE ]


2. If I've set the voltage and current to specific levels, then put a 1 ohm resistor in series with the emitter, will I still get the corrent I specified, or will the resistor cause the level of current flow to decrease relative to what I had set?


[/ QUOTE ]

The resistor will drop some of the voltage you are applying. This will reduce current as well, unless you add voltage to compensate. Turning up the power supply will increase it's output voltage to the point where the set current is reached, and then it'll hold there.

Say you've got a 4V 1A LED. You set the P/S to 1A shorted and 4V open. Now you hook up the LED and a 1 ohm at least 2 watt resistor in series to the P/S. The Law says that 1A thru a 1ohm resistor would be a 1V drop. So you would need to increase your voltage to 5V to get the 4V to your LED.

Always remember to measure your test voltages after the resistor, across the LED, to not accidentally include the resistor drops in your numbers. The nice thing about using a 1ohm resistor is that you can measure the voltage across it with a cheap digital multimeter set to the 2V scale, and be reading current directly in A.XXX giving easy reading down to 1MA, and to .1Ma on the 200mv scale. Much better than most bench P/S current meters. (Use 1% resistor or accurately measure it for fudge factor)

By the way, a caveat that's very important for those of you working with any kind of switching regulators. Adding a resistor in series with one of these can royally foul up your measurements!! In some cases, even running the higher power ones directly from a small lab supply may result in odd readings. This is because the 'switcher' draws it's current in bursts, usually short in duration and of hgh current. This can confuse the P/S, and cause it to act 'funny', fail or even toast the device under test.
You may have noticed that a bench supply was flashing it's current limit light or 'chattering', even tho it's meters read reasonable or even rather low values. That's the switcher drawing an 'average' of 1A at 2V to make 400ma at 4v for your LED by punching 3A pulses of 1/3rd on time thru an inductor to boost the voltage up to LED levels.
A scope shows this quite well, but meters just average it out. So your 2A bench supply, turned all the way up will be showing an average meter reading of, say .75A, your LED is not full brightness, your P/S current limit light is flashing or on, and it doesn't make sense!
Take a large electrolytic capacitor (2000+ ufd at 15V or so) turn off the P/S, hook up the cap to the output, and slowly turn the P/S voltage back up and all of a sudden, everything works. That's the flywheel effect again. Too small of a cap and this happens, too big of a cap and you pop a lot of parts if you aren't real careful.
The cap can also be hooked up after the 1ohm, but remember that you have to watch it's storage now. Use a cap or direct feed for converters, don't use it and use a resistor for LEDs.

Eeep.. I'd better get off this before I get shot.

BTW, this stuff all from my experiences and knowledge, all derived from the school of hard knocks and much knowledge extracted from many other helpful folk. Been a component-level technician for 30+ years. (double Eeep! /ubbthreads/images/graemlins/grin.gif)

Some simplification and assumptions made herein. Switch-mode power systems can be very tricky and the above is by no means all-inclusive or applicable to any specific combination of DUT (device under test) P/S and wetware. (That's the guy/gal at the knobs. /ubbthreads/images/graemlins/grin.gif)

Meant to be helpful. Hopefully it is. Please also consult others and tread carefully. And one of my favorite roolz: Buy Spares!! Use Craig's cat urine green Luxeon before you drop in that U-binned snow-white wunderLED!

We have some amazingly talented (and patient!) people here on CPF that offer help and insights. And build neat stuff that I can't even SEE anymore out of parts I have accidentally thrown away, because I didn't know they were there. (Teeny black specks on black foam in a black box. Oops!) I like to play with battry-powered tube and early transistor radios for fun, as well as computerbits. So there's a bit of spread. But I can see the parts at least.. /ubbthreads/images/graemlins/grin.gif

I'm off! yours, drs the crazed
 
You must log in or register to reply here.

Latest posts

Top