Help me customizing this Mean-Well driver

[anuragwap]

Recently I purchased the CEN-100-20 mean well driver (datasheet here) This driver is versatile since current and voltage can be varied (3.12-4.8A, 17-22V).



My circuit needs 20.7V and at 4.1A peak (average 2A or less). There are two screws for voltage/current adjustment. But how do I set them? Do I just connect multimeter probes to the outputs and turn the screws? Or do I need a proper load and measure the voltage across it? I realize this driver has CC-CV mode which makes things complicated.

EDIT: I guess the voltage and the max current values are factory set at the default values (i.e. 20V and 4.8A). So maybe I need to tweak the voltage screw only.

 

[SemiMan]

Recently I purchased the CEN-100-20 mean well driver (datasheet here) This driver is versatile since current and voltage can be varied (3.12-4.8A, 17-22V).



My circuit needs 20.7V and at 4.1A peak (average 2A or less). There are two screws for voltage/current adjustment. But how do I set them? Do I just connect multimeter probes to the outputs and turn the screws? Or do I need a proper load and measure the voltage across it? I realize this driver has CC-CV mode which makes things complicated.

EDIT: I guess the voltage and the max current values are factory set at the default values (i.e. 20V and 4.8A). So maybe I need to tweak the voltage screw only.

How do you know you need 20.7V?

 

[Steve K]

if your circuit needs a regulated current, then I'd suggest adjusting the CC pot.
If it needs a regulated voltage, then adjust the CV pot.

...or... if you really want to live life on the edge, check the manual or ask Meanwell.

 

[anuragwap]

How do you know you need 20.7V?

Well, I want it to operate in CV mode, with its output fed to CAT4101 based drivers from StevesLED with 0.5V overhead. The CAT 4101 will then drive 6 InGaN rebels (or 8 AlInGaP ones) in series at 700 or 1000mA. There are multiple strings of diferent colors, but only 4 of them will remain on at a time. Blue LEDs driven at 1000mA have the largest voltage drop. Adding 0.4V safety headroom to that, I got the figure. Won't go into more details but I'm confident about the voltage/peak current requirements.


I just connected 10 deep red rebels (not 8, for safety) and tried twisting the voltage screw in both ways. I was fooled. The current didnt change at all from 0.182A, the screw was fully tight initially and came out after twisting! Do I have to open the case to access the trimpots? Or maybe my deep red rebels don't allow more than 182mA at 22(max voltage)/10=2.2V? (pretty unlikely)

 

[Illum]

Considering the driver is outdoor rated, I would look for a trimpot slot underneath the screws. Very unlikely comes the need for a commercial driver like this to require trimming after installation, so likely the trimpot is protected underneath something that was field accessible during installation.

Note that an LEDs forward voltage increases with forward current, and that each LED [including those from the same batch] may have variations. IIRC I once bought a reel of 20 CREE XPGs, only about 11 measured "~3.3V" at a forward current of 1000mA, the rest varied from 3V to 3.5V.

As I understand it you are essentially feeding a linear LED driver using a switching LED driver with PFC... something tells me the MW is not giving you what you want because it can't figure out what you are doing to its load side.

 

[anuragwap]

Sorry, forgot to mention, those 10 red rebels were direct driven. It makes sense now: 20/10=2V should be the average drop for my deep red rebels at 182mA. (I could've verified that by putting the DMM across the LEDs, but it was connected in series)

Note that an LEDs forward voltage increases with forward current, and that each LED [including those from the same batch] may have variations. IIRC I once bought a reel of 20 CREE XPGs, only about 11 measured "~3.3V" at a forward current of 1000mA, the rest varied from 3V to 3.5V.

Thanks, I'll test the V_f of the strings before the final installation. My experience with rebel V_f spread at 700mA(from same reel) so far has been satisfactory, so I thought 0.4V headroom was enough. I know I should be careful as the V_f is quite high at startup, when the junction is cold.

SIDENOTE: This driver has a delayed start (perhaps to limit inrush), so it doesn't give out any nasty spark like my Xitanium drivers.

EDIT: Put my keychain light into the 'V_o adj' hole and finally saw a super-small blue trimpot deep down. Adjusted trimpot the wrong way and the current went down to 125mA! Tried to open the case, but resisted the temptation, as there are beautiful waterproof washers/ rubber threads everywhere.

 

[SemiMan]

Sorry, forgot to mention, those 10 red rebels were direct driven. And there is nothing but darkness under the screws. So I must open it.
It makes sense now: 20/10=2V should be the average drop for my deep red rebels at 182mA.



Thanks, I'll test the V_f of the strings before the final installation. My experience with rebel V_f spread at 700mA(from same reel) so far has been satisfactory, so I thought 0.4V headroom was enough. I know I should be careful as the V_f is quite high at startup, when the junction is cold.


SIDENOTE: This driver has a delayed start (perhaps to limit inrush), so it doesn't give out any nasty spark like my Xitanium drivers.

Keep in mind the Vf not only varies from LED to LED, but varies over time with a burn in period of 10s of hours for InGaN.

Planning for an overhead of 0.5V is really really tight on a 20+ volt string.

 

[Illum]

Keep in mind the Vf not only varies from LED to LED, but varies over time with a burn in period of 10s of hours for InGaN.


Planning for an overhead of 0.5V is really really tight on a 20+ volt string.

He doesn't have much of a choice, CAT4101 is a linear CC sink, had he applied an overhead of 1V would have resulted in an additional watt of heat output for every amp. That's not just power loss, but power loss that he has to deal with when designing cooling solutions.

On an interesting note, forward voltage is proportional to junction temperature too. Figuring I would get more efficacy out of an XR-E I had the heatsink/LED assembly stored in the freezer over night and powered it on next morning. At 350mA forward Vf was supposed to be ~3.3V, but it registered around 6V when chilled to ~30F

 

[SemiMan]

He doesn't have much of a choice, CAT4101 is a linear CC sink, had he applied an overhead of 1V would have resulted in an additional watt of heat output for every amp. That's not just power loss, but power loss that he has to deal with when designing cooling solutions.

That may be true, but if the LEDS have more than 0.5V deviation then it will not work.

Semiman

 

[Illum]

That may be true, but if the LEDS have more than 0.5V deviation then it will not work.

Semiman

well, it will still work, just not at the current he specified, below the forward voltage of an LED and it will start acting like a resistor. Nothing short of a current injector will allow the LEDs to operate on spec. I've actually used the CAT4101 with a purposely "undervolt" CV supply that matches the forward voltage of the string. I figured I could take the thermal properties of the LED to my advantage to roll back the current as the heatsinks they stand on gets hotter. The plan was sacked due to LED batch inconsistencies.