You stated that You measured the Voltage across several of the LEDs, and all measured ~8.4-Volts,
This has to be a mistake, You are probably measuring Noise-Peaks from the Current-Limiter Switching.
You need to measure the Current, with the Noise filtered-out.
Probably the best way to do that is to measure the Input-Current to the whole Light while it's turned-on.

Do You know for a fact that the LEDs are wired in parallel ?,
or did You make an assumption based on the fact that You
measured the same Voltage across several LEDs ?

Have You actually visually followed the traces on the Board and
determined that, in-fact, all of the LEDs are physically connected in parallel ?

How many total LEDs are there in the Light ?

Having the LEDs all in Parallel has a big potential problem.
They won't share Current evenly.
If You could re-wire them so that You have 2-sets of 4 or 5 in Series, it would be much "safer".
.
.
.

 

I tested continuity across all the groups and most of the individual LEDs to confirm they are all in parallel. I drove them with my adjustable power supply, at about 8.4V, and got roughly the same wattage as when I drove them with the included circuit board at any rated voltage (12-24V rated). When I test them for Vf with my meter, they register sightly under 3V (about 2.8-2.9V), but don't have any meaningful brightness until about 7.5V. I tried to follow the traces, but i still have little experience so I don't trust myself in that regard yet, but I do trust the continuity measurements on my meter.

I agree these numbers don't make sense, but for some reason that's what my testing has shown, and this testing is using multiple pieces of equipment which all seem to be accurate based on other testing.

I don't have the light in front of me but I think it was 18 LEDs altogether. I have noticed, especially at lower voltages, that some are significantly brighter than others.

 

~3-Volts is something I can believe.

I think that Your "measurement-techniques" are highly questionable.

You're measuring Diodes in a Circuit, and possibly Resistors along with them.

Measuring with a Meter is not the same as careful
visual inspection of the wires or traces from start to finish.


You need to re-Wire the LEDs into series-connected-groups of 4-LEDs,
and then parallel all 4 groups together.
If this arrangement can not be accomplished, you're on your own.
.
.
.

 

Having the LEDs all in Parallel has a big potential problem.
They won't share Current evenly.

My cheap old flashlight had 24 ordinary LEDs all in parallel and their brightness was matched perfectly.
Its on-off switch failed so I saw then bought a new flashlight in a Dollar Store that has 60 surface mount LEDs that are extremely bright, even when powered from cheap 3 AAA Super Heavy Duty batteries. I did not take it apart yet to see if the LEDs are all in parallel but they probably are.

Amazon sells the flashlight for $13.55 but I bought it for only $3.99.

 

These are High-Power-LEDs, and he wants run them on the edge of reliability.

I personally don't believe that they are all in parallel.
.
.
.

 

If he put 8 to 10 V on them without destroying them, they are likely wired 3 in series.

Bob

 

BobTPH and LowQCab what do you think of this?

 

I think the guy doesn't know what he's doing.

There are many other Components on that Circuit-Board,
THEY ARE ALL, SOMEHOW, PART OF THE CIRCUIT.

There are not ~18-bare LEDs connected in parallel on that Board.
There are no "Special-High-Voltage-LEDs" being used here.

If You put ~9-volts across any LED, it will make 1-VERY BRIGHT FLASH, and never work again.
It's very likely that every LED on that Board is being powered through
some kind of a Current-Limiting-Resistor, or other Electronic-Device.

Testing Components or "Continuity" in a Circuit is
almost always going to create measurement anomalies or false readings.
.
.
.

 

I think the guy doesn't know what he's doing.

There are many other Components on that Circuit-Board,
THEY ARE ALL, SOMEHOW, PART OF THE CIRCUIT.

There are not ~18-bare LEDs connected in parallel on that Board.
There are no "Special-High-Voltage-LEDs" being used here.

If You put ~9-volts across any LED, it will make 1-VERY BRIGHT FLASH, and never work again.
It's very likely that every LED on that Board is being powered through
some kind of a Current-Limiting-Resistor, or other Electronic-Device.

Testing Components or "Continuity" in a Circuit is
almost always going to create measurement anomalies or false readings.
.
.
.

I'm not a micro electronics guy, but I am an electrician and I know how continuity works. In this case, for how it's set up, it's exactly as I said, and your understanding is incorrect. Yes there's other components, but unless they're parallel with the LEDs, they are disconnected. If you happen to live anywhere near Salem, Oregon, I'll drive these to your house to show you. I have not given any false information, I doubt I've given any incorrect information. I can't make sense of it, but this is what's happening.

 

The fact that You can't make sense of it indicates that there are factors in play that You do not understand.
You do not understand how a "Continuity-Check" may not work in an Electronic-Circuit.
You do not understand what factors must be controlled to make an LED operate, and have a long-life.
You do not understand that all those "extra" Devices soldered to the Board are there for very good reasons.

And finally, it is sometimes very confusing, figuring-out what the Design-Engineer was thinking when
he came up with some very unusual, and cheaper, method of making a better profit.
.
.
.

 

The fact that You can't make sense of it indicates that there are factors in play that You do not understand.
You do not understand how a "Continuity-Check" may not work in an Electronic-Circuit.
You do not understand what factors must be controlled to make an LED operate, and have a long-life.
You do not understand that all those "extra" Devices soldered to the Board are there for very good reasons.

And finally, it is sometimes very confusing, figuring-out what the Design-Engineer was thinking when
he came up with some very unusual, and cheaper, method of making a better profit.
.
.
.

You can't make any more sense of it than I can, without assuming I'm doing something wrong (an assumption you have no direct evidence for). And I am aware of how other components can confuse continuity, and I believe I've effectively accounts for that. I'm familiar enough that I can guarantee that all the LEDs are in parallel, and the circuit outputs around 8.4V. There's nothing wrong with saying, "I can't make sense of those numbers based on my understanding." You don't have to accuse me of not knowing what I'm doing, etc. I came to this forum for help, not to be insulted. If you don't know how to help me, you can just say that.

EDIT: I broke every component off that circuit board, the only change is that my meter no longer reads a Vf. There's now no components besides the LEDs, the wires, and the board, and everything else is still the same. The LEDs behave as though they have a Vf of about 6.8, and the are all in parallel.

 

There is no such thing as a single LED that has a Vf of 6.8V. It is likely that what you have are not single LEDs, but chips with 3 in series.


Bob

 

That's 2 or 3 LEDs on a single Chip.

And, I didn't insult you, I simply stated facts that You don't want to accept.

I already provided an excellent Circuit design for You,
now I have to talk You out of your preconceived notions so that You can hopefully apply it.

Everybody only learns from their mistakes, if You don't make mistakes, You are not learning.
.
.
.

 

That's 2 or 3 LEDs on a single Chip.

And, I didn't insult you, I simply stated facts that You don't want to accept.

I already provided an excellent Circuit design for You,
now I have to talk You out of your preconceived notions so that You can hopefully apply it.

Everybody only learns from their mistakes, if You don't make mistakes, You are not learning.
.
.
.

Now that you've suggested that, if I look really closely in the right light, it looks like 3 LEDs inside each "chip." Nothing I've said here has been wrong, except *maybe* you could argue the classification of the LEDs since they appear to be a collection of 3 smaller LEDs, not one single LED. Still doesn't change that all my measurements and testing have been correct, except that there was apparently something parallel to the LEDs as I can no longer measure that Vf of about 3V that I had previously been measuring. You've accused me of not knowing what I was doing, despite having a correct methodology and having been right, except on one apparently unrelated detail (that Vf measurement). You seem to be the one who can't accept the fact that my testing was correct, and all of these LEDs are in parallel. You were right that a single LED shouldn't ever have a Vf of 7~8 volts. But you did insult me by saying I didn't know what I was doing, and you argued with me unnecessarily.

I appreciate the schematic you gave earlier. Though for the effort and cost of building it, I'll probably go a different route.

 

There is no such thing as a single LED that has a Vf of 6.8V. It is likely that what you have are not single LEDs, but chips with 3 in series.


Bob

Now I understand, each individual component is 3 LEDs in series. Upon close visual inspection, this appears to be the case.

 

I never said that You didn't know what You were doing,
unless maybe You are the guy in the YouTube Video.
Was that You ?

No matter what solution You decide on, it's extremely important that You
Regulate the Maximum-Current, and let the Voltage go wherever it wants to.
Pay no attention to the Voltage, it will vary with the temperature of the Chips.
Then,
if You should decide to push-your-luck on increasing the Current,
and since there is no Data-Sheet available for the Chips,
gradually increase the Current in steps,
( for testing purposes only, you can manually vary the Voltage to regulate the Current ),
then make sure that nothing around the LED-Chips is getting over approximately ~75C, (~167F ).
The best way to check this is with an Infa-Red-Thermometer.
This is what is going to determine your Maximum-Current, and/or, Life-Expectancy.
The temperature that the LEDs are running at is directly related to
how many hours, (or minutes), (or seconds), of service that You will get out of them.
It would be a good idea to include a Current-Balancing-Resistor for every Chip,
this Resistor should not be used to set the final Current to the Chip,
it would be for extra insurance against uneven distribution of Current amongst the Chips,
which is very likely going to cause issues when using a Parallel-Wiring-Scheme.
Each Resistor should probably be around ~1-Ohm, and rated for ~1-Watt.

Another option is to arrange the Chips in pairs, ( 2-Chips in series ), ( which is really 6-LEDs in series ),
and then put all the "series-pairs" in parallel,
this is assuming a ~15 to ~18-Volt Supply,
each "series-pair" should also get a 1-Ohm Current-Balancing-Resistor.
This arrangement is "safer" for the Chips if something should go wrong with the Current-Regulator.
The Circuit that I provided will work for this configuration as well,
with no changes necessary after the Current-Sensing-Resistor is calculated.
Each pair should receive less than approximately ~500ma,
So, if You have 8-pairs, the total-Current would be 0.50 X 8 = 4-Total-Amps from the Current-Regulator.
.
.
.

 

I never said that You didn't know what You were doing,
unless maybe You are the guy in the YouTube Video.
Was that You ?

No matter what solution You decide on, it's extremely important that You
Regulate the Maximum-Current, and let the Voltage go wherever it wants to.
Pay no attention to the Voltage, it will vary with the temperature of the Chips.
Then,
if You should decide to push-your-luck on increasing the Current,
and since there is no Data-Sheet available for the Chips,
gradually increase the Current in steps,
( for testing purposes only, you can manually vary the Voltage to regulate the Current ),
then make sure that nothing around the LED-Chips is getting over approximately ~75C, (~167F ).
The best way to check this is with an Infa-Red-Thermometer.
This is what is going to determine your Maximum-Current, and/or, Life-Expectancy.
The temperature that the LEDs are running at is directly related to
how many hours, (or minutes), (or seconds), of service that You will get out of them.
It would be a good idea to include a Current-Balancing-Resistor for every Chip,
this Resistor should not be used to set the final Current to the Chip,
it would be for extra insurance against uneven distribution of Current amongst the Chips,
which is very likely going to cause issues when using a Parallel-Wiring-Scheme.
Each Resistor should probably be around ~1-Ohm, and rated for ~1-Watt.

Another option is to arrange the Chips in pairs, ( 2-Chips in series ), ( which is really 6-LEDs in series ),
and then put all the "series-pairs" in parallel,
this is assuming a ~15 to ~18-Volt Supply,
each "series-pair" should also get a 1-Ohm Current-Balancing-Resistor.
This arrangement is "safer" for the Chips if something should go wrong with the Current-Regulator.
The Circuit that I provided will work for this configuration as well,
with no changes necessary after the Current-Sensing-Resistor is calculated.
Each pair should receive less than approximately ~500ma,
So, if You have 8-pairs, the total-Current would be 0.50 X 8 = 4-Total-Amps from the Current-Regulator.
.
.
.

I am the guy in the video. Thanks for the info.

 

I did some more reading about LED drivers and constant current. It seems that constant current is the gold standard. However, since I have LEDs in parallel, using only constant current could result in a cascade failure if one LED burns out, so my best option would be be both current and voltage limited driver, to ensure maximum safety.

 

That's what the 1-Ohm Resistors are for,
they are not 100% protection against a shorted individual LED,
but they will keep things more stable so that a LED failure is less likely in the first place.

Don't be too concerned about regulating the Voltage, you'll just waste more Battery-Power.
( the Resistors waste a small amount of Power too, but help to keep things balanced )

You have no Spec-Sheet, so keep a close eye on the Temperatures.
.
.
.

 

That's what the 1-Ohm Resistors are for,
they are not 100% protection against a shorted individual LED,
but they will keep things more stable so that a LED failure is less likely in the first place.

Don't be too concerned about regulating the Voltage, you'll just waste more Battery-Power.
( the Resistors waste a small amount of Power too, but help to keep things balanced )

You have no Spec-Sheet, so keep a close eye on the Temperatures.
.
.
.

Oh gotcha, the 1-ohm resistor would help with that, yeah. I think I've found a board with voltage and current control built in.

If I have a constant current driver, can I use two in parallel? I know normally not to use buck/boost converters in parallel because a little voltage difference can cause a dramatic difference in current usage, but would using constant current sources fix that problem?