Equipment setup for LED vf measurement

[clemence]

Hi guys,

I've been playing with LED for a while and now I'm getting deeper into LED measurements. I'm new to electronics and still learning. I'm still a newbie when it comes to find good equipment especially for electronic testing.
I need help to measure LED vf, with the following requirements:

- 10ms pulsed input or less. Non spike, flat regulated constant current 15A min. DC input.
- the measurement must has at least 3 decimal digits. For example: 3,215 V
- Voltage measurement range: 0,001 - 40,000 V
- logging/record feature.

What equipment/brand to start with?


Thanks,

Clemence

[danadak]

Do you want the measurement of Vfwd during the pulse, instantaneous, or
an average of several pulses ?

A PSOC with DelSig coupled with S/H should do the job for either situation.

The attached partial solution. DAC controls current thru LED. Mux measures driven current
accurately, and then Vfwd. Could use DMA to automate process, including averaging a
number of samples. And of course display to show results. A MOSFET could ground the
LED until measure time so that the current source control loop is fully settled. MOSFET
timing driven by on chip timer..... PSOC 5LP family. A/D could run continuously or triggered
mode.

All of this onchip except Q_1 and R_1.


Regards, Dana.

[clemence]

Hi Dana,

Thanks for detailed response/diagram.
Depends on the readings. Most cheaper power supply unit will give you spike or rather high curve reading in the beginning of whatever ms pulsed current. The square graph is ideal. I see that entry level DC load PSU like Rigol can provide me with such requirements.

The project I'm working in is to find a repeatable Tj measurements of various LED on various MCPCB under various current load under relatively constant controlled MCPCB temperature (regulated by peltier device/controller). We ran many tests to find LED peak output with success. And that were usually 3-4 times higher than manufacturer's absolute limit. What I want to get is the "real" data based from junction and phosphor temperature. Since LED vf is a very consistent measures of junction temperature, this should be the way to go.

TBH, I don't really get 60% of the information you wrote since I'm totally new in this field, still needs to learn a lot from you guys/girls. Correct me if I'm wrong, DAC = Data Acquisition Controller? DMA = Digital Multimeter A...?

- Clemence

[Giaime]

Have you considered using the ESD protection diode, that is usually incorporated into high-power LEDs in anti-parallel w.r.t. the LED junction, to measure LED Tj?

I've done this in the past... you don't need high current to measure ESD diode Tj. You just need to rapidly turn off the LED to switch a small current source on the ESD diode with opposite polarity. There is some highly specialized test equipment that I've seen that is doing exactly this.

Otherwise, to measure LED Vf, I suggest a good SMU, but those get expensive quickly at those current levels: I use a lot the Keithley 2425, but it's limited to a couple of amps.

[clemence]

Many of the board that will be tested don't have ESD or ESD devices footprint. Don't you think the junction temp location of ESD device would be too far away from the LED location?

15A is the maximum momentarily peak output for some "extreme-hard-to-kill" high power LEDs such as Luminus or Cree. Most of the useful LED tested are usually peaked in the 7-10A range. But it's always good to know the max death current. We knew the max. current for max. output. What we don't know is the max. safe reliable current. Majority of manufacturers use very conservative max. current using average MCPCB for their absolute max. current ratings to perhaps, still get the 50.000 hours LM80. While in practical portable flashlight uses, momentary boost output is desired.

The cheapest way to do the measurement is just to manually feed the LED with whatever preset current we want, while at the same time recording the vf using separate DMM. 10ms or less feeding time is easy to get manually or mechanically. The problem with such manual setup is current spike and minimum DMM response. I know with good and usually expensive DC load PSU this is an easy task. But I'm sure there must be cheaper option to get the same readings.

I was thinking about Dana's suggestion. From my newbie POV, to get the flat stabilized input to the LED from whatever PSU, constant feeding is one solution. Let me try to describe what I got from her idea:
1. constant current load from a DC source with load slightly higher than what would be when the LED connected. This load circuit bypassed the LED.
2. When abrupt 10ms stop applied to the main circuit, the load instantaneously fed to the LED without the common spike experienced from budget DC PSU.
3. The whole process (no. 1 and 2) is constantly recorded with resolution under 10ms.

Am I close to the idea?

- Clemence

[danadak]

DAC = D/A converter.

The reverse parasitic diode T measurement is useful in some situations.

You can eliminate offsets using the dual current measurement technique.


http://www.cypress.com/documentation/application-notes/an60590-psoc-3-psoc-4-and-psoc-5lp-temperature-measurement-diode



See attached.

[clemence]

Thanks Dana! That was very helpful to better understand the concept. I guess the best solution is to use the diode method since it's the diode junction temperature that I'm going after. The other solutions are indirect

- Clemence