Analog signal for LED drivers

[brimarq]

I am using TRP LED drivers that do not except the pulse width modulation signal that the Meanwells require. Is there any way for me to control the TRP drivers with the Reef Angel?

[rimai]

There is another user that is using Thomas Research drivers.
I don't know how far into the project he's got, but it's very simple to do. All you need is one capacitor and one resistor.

[brimarq]

Do you by any chance know what values he's using?

Also I was curious if I add another relay box if I would be able to utilize two more additional PWM outlets? Would I be able to do the moon cycles and thunderstorms without the PWM expansion module using my 2-PWM outlets on my current relay box?

[rimai]

No, the extra relay box won't give you additional channels.
You still can do the moonlight and thunderstorm with just 2, but I'm not sure is it would look the same.
One channel for moonlight and another channel for all whites and blues, which would flash all at the same time.
Let me ask him to post his values.

[brimarq]

How much is the expansion module and is it available now?

[rimai]

Yes, on special order basis for $89

[brimarq]

Hmm, I need another relay block, I want the wifi, and I don't need this expansion module but it would be way cool for my system. I also just spent $600+ on LED's. Man, this is an expensive hobby! I clearly did not think this whole saltwater thing out enough before I jumped in.

[javisaman]

Hi I'm using the TR drivers and the components I'm using for the smoothing are 22kO resistor and 10uF capacitor. It's not perfect. The ramping is kinda slow and not linear. As in if you go from 0-100% with 10% increments the V goes from 0 to 1 to 2 to 2.9, 3.8, 4.7, 5.5, 6.4, 7.3, 8.2, 9.1 etc.

[brimarq]

Hi I'm using the TR drivers and the components I'm using for the smoothing are 22kO resistor and 10uF capacitor. It's not perfect. The ramping is kinda slow and not linear. As in if you go from 0-100% with 10% increments the V goes from 0 to 1 to 2 to 2.9, 3.8, 4.7, 5.5, 6.4, 7.3, 8.2, 9.1 etc.

Hey thanks! I have been doing a bit of research on this subject as well and was told to try using a potentiometer for the resistor so I could fine tune it and then use a multimeter to measure when I found the sweet spot. Not sure if this will smooth out the signal to make it more linear or not. I will start with a 10uF cap and a 50k pot and see what happens, I will let you know what becomes of my experimentation. Did you use a formula to figure this out or have you just experimented with it yourself? Is there any chance you know the frequency of the PWM that the Reef Angel sends, or the impedance of the TRP drivers (I'm using all 40w drivers). I know there is a formula to figure this out using the published specs for the drivers but I am a remodeling contractor by trade and this will require more homework on my end to figure out.


Brian

[rimai]

Here is some articles for your research:
http://www.arduino.cc/playground/Main/T ... Cheatsheet
http://provideyourown.com/2011/analogwr ... o-voltage/
http://sim.okawa-denshi.jp/en/PWMtool.php

[jonegross]

I bought a reef angel 2 weeks ago. Everything has been really easy to set up thus far, but the dimming. I knew it would be a bit interesting as I'm using an inventronics driver which I understand is the same as the TRC drivers. I have a 25 watt 700mA driver. Interested in what you guys come up with, I am doing research on how this PWM to analog thing works. I have managed to really confuse my understanding of electronics in the last week, lol. I will be following this thread. Hope to learn from you guys. The more info on this subject the better, as I am sure others are interested as well.

[rimai]

What I can tell about the slow response is that it is indeed slow with the values 220K Ohms and 10uF.
If you look at the graph produced by the calculator, it takes about 4 secs to reach about 90% and about 12s to stabilize.
Decreasing the resistor to about 22k or 10K still gives you a good riple to response ratio and gives you much better results.
Plug in the numbers and check out the graphs. The f_pwm is 490Hz.
Check out the graphs:
http://sim.okawa-denshi.jp/en/PWMtool.php

[rimai]

Scratch what I just said. I just re-read the post and it is 22k... Sorry about the confusion

[brimarq]

I think the cap may be too large, it takes time to charge up hence the delay (maybe). I'm thinking of starting with a 1uF and a 100K POT to experiment with. When you dial in the sweet spot you can measure with a multimeter and know exactly what size resistor to buy. Also it sure sounds like there is no way to eliminate the non-linear response just using a low bypass filter, I've read that the only way to convert a PWM signal to a linear analog signal requires a fairly sophisticated circuit that I personally am not capable of figuring out. Personally I think the Mean Wells may be just a lot easier to use with Reef Angel.

[rimai]

I don't think the human eye nor corals can really feel any difference on a linear vs non-linear (very small deviation, in our case) change of intensity on the LED string. Just my thought.

[jonegross]

I have played with .1uF-10uF caps and 10-300K resistors. Maybe I am wiring it wrong but I have not seen any difference between using the low pass and just plugging the Inventronics driver right into the PWM port. Is it bad for the driver? I seen no shuttering or shaking, lights and dimms the same as with RC. Maybe the dimming circuit has a cap in it? and just filters out the pwm anyway? Is there a reason to try to filter the signal that I am missing?

[rimai]

Can you post a link of the datasheet of the driver you are using?

[jonegross]

http://www.inventronics-co.com/uploads/ ... 10%20G.pdf
Its the 700mA version EUC-025S070DS(6)

[rimai]

Nothing on datasheet. I sent an email to manufacturer and asked

[javisaman]

you won't be able to tell the difference between the digital and analog signals with a regular multimeter. You need an oscilliscope. The frequency of the PWM is about 500Hz. I was told by an employee at nanotuners (where I bought my drivers from) that the thomas research drivers do not support digital pwm. I never tried directly plugging it into the reefangel because i was worried it might ruin either the driver or the controller.

[jonegross]

I have been running my inventronics/TRC driver for about 2 weeks without an RC. It seems to be working fine, to preface this my electronics skills are on the level were I am able to break most things (:.... I need another driver to split the white and blue channels, but am waiting to see if this unit holds up. I would much rather just buy one more of the same type driver, as it will make my project box simple. But really it isn't much price difference from just buying 2 meanwell pwm 30 watt drivers. Not sure what to do at this point. Is it possible for the RA controller to be damaged by running a 0-10v driver without an RC? Doesn't seem like it to me. My thought was the lights would flicker from the pwm, but I can't see any flickering. Maybe burn out the dimming circuit? Any word from the manufacture?

[rimai]

No response from the manufacturer.
The RA would not get damaged by your drivers.

[jonegross]

Thought I would post a quick update: It's been over a month with my Inventronics driver plugged directly into the pwm port, no issues as of yet. The more I look at designs of dimming circuits for these things, I don't think it is possible for them to distinguish between PWM and 0-10 analogue as every design I have seen a layout for has a cap in the dimming circuit running in parallel with the main circuit. Seems to me that would just filter out the pwm anyway... But what do I know, will update in a month or if the driver dies.

[08trdoffroad]

So bottom line, will the 48D's work with the new PWM to analog cables?

[rimai]

Yes.

[08trdoffroad]

Will they work correctly? Or will there be choppy dimming, etc?

[binder]

Will they work correctly? Or will there be choppy dimming, etc?

You most likely won't be able to dim 0-100%. There may be a drop off less than 10%. I have a 24D driver and it drops off at 6% for me (or maybe 7%, I forget). Other than that, it's a smooth slope from the min to 100% for me and back down.

[08trdoffroad]

Thanks for the reply binder! Just the information I was looking for! I am not worried if they drop off at 6%, even on a potentiometer they will only dim down to 5% before cut off.

[stevenhman]

Great information!

[TanksNStuff]

New RA owner here, trying to hook up 2 LED fixtures with a total of 4 Meanwell ELN-60-48D drivers (2 drivers per fixture, 1 for whites, 1 for blues). This is going to be long, so please bare with me.

I got a PWM expansion module, but have not hooked it up yet. I did all my testing using the 2 PWM channels that come with the main relay box. I also received a pair of PWM-to-Analog Lead wires with my controller. Once I get the PWM expansion module configured, I'll try to test again with that to see if I get different results than I did with the 2 stock channels.

First thing I did was test my lights using a 9v battery as a "dimming signal". The battery was at 9.4V as per my multimeter. All lights were very bright and I tested whites and blues for each fixture.

Next, using the Demo code on the controller, I went into PWM settings and made 1 channel 100%, the other channel 50%. Then, to verify the signal strength, I took multimeter readings right at the pins on the relay box. The 100% = 9.77V, the 50%= 4.94V.

Next step, I connected the PWM-to-Analog leads to the pins and took a reading on the open wire ends. The 100% = 3.17V, 50% = 1.86V. That's a pretty big dropoff between the pins and the other end of the converter! It seems it lost about 68% voltage on the 100% channel, and about 62% voltage on the 50% channel.

*LED FIXTURE INFO* - Dimming control is via an ethernet port on the fixture. I had to use a CAT5 patch cable to plug into the port, then cut/strip the other end to get bare wires. This is how I tested with the battery and verified the +/- wires for each channel.

Next step, I connected one driver's dimming wires to each converter. Lights worked for both channels, however, they appeared much dimmer than when I tested with the battery, even the 100% channel.

I then removed the ethernet plugs from the fixture ports, and tested the voltage at those pins. Voltage readings were the same as what I got at the end of the converter... so the patch cables were good.

Next test was to determine when the lights would shut off. I ramped the controllers PWM signals down in increments of 5 and then pinpointed the actual shutoff value was 21% intensity. At 22%, lights were all fine, at 21%, each channel (blue or white) shutoff. So a single ELN-60-48D driver on a single channel cutoff at 21% intensity when using the converter.

Next, I combined both drivers onto 1 channel. I twisted both blues and whites onto 1 converter. The results were basically identical... shutting off both at 21% intensity.

Next, I brought out fixture #2 and made another patch cable. I jumped right into adding all 4 drivers (2 from each fixture) onto 1 converter. This time, the results got worse. I started at 50% (thinking I'd work my way down) but none of the lights even came on. So I jumped up to 75% and worked down to 59%, where the blues went out on 1 fixture. At 58%, all lights shut off.

The last test was sort of irrelevant since I plan to only have 1 driver per channel... but I did it for the data and to satisfy my curiosity anyway.

So, being upset that I wasn't going to be able to utilize my new LEDs to the fullest extent, I remembered reading this thread and decided to see if my lights would work with a straight PWM signal, instead of using the converter.

I dug out my old box of computer parts and found a fan that had a 2-pin connector. (See, I knew I was going to need that someday!) I cut the wires off of it and I now had my lead wires without a PWM-to-Analog converter. Just a black and red wire.

OK, first test was setting the controller back to 100% channel and 50% channel. Took readings at the pins and got the same results as first time. 9.77V and 4.94V respectively.

Then I connected to the 100% pins and got 9.74V at the bare wires. Very minor dropoff, but I can live with that. I moved the connector to the 50% and got 4.91V at the bare wires. Again, acceptable IMO.

Next step, I wired 1 set of wires from the patch cable to these wires. I took readings at the ethernet jack pins and got the same as at the wired connection. Just double-checking that everything is still good.

Now, I was a bit scared of damaging either the fixture or the controller, so I did this set of test starting with 0% intensity and worked my way up to find out when the lights would turn on.

With a single driver on a single channel, lights came on at 10% intensity (PWM signal set on controller).

Seeing that was an improvement, I then added the other driver to the same channel. Again, starting at 0%, both whites and blues came on at 10% intensity. Voltage readings at the twisted wires was almost identical, maybe a few points lower.

After that, I got brave and added all 4 drivers to the same channel. Guess what, 10% was still the turnon point for all 4 drivers!

Next, I kept all 4 on the same channel, and ramped them up in 5% increments. I could barely notice the brightness increasing with each step with my eye, so can't comment on the scaling aspect. But, all lights ran continuously, no flickering or anything. And, at 100%, they were way brighter than when I had the in-line converter wired in. In fact, at around 50% they looked brighter to me than 100% with the converter!

I got all the way to 100% with 4 drivers, 1 PWM channel, and then took a reading at the wired connection to the new (non-converted) lead wires. I got 9.68V @ 100% intensity. A slightly higher dropoff than with a single driver one a single channel, but certainly one that most people could live with.

So, seeing that these seem to operate perfectly fine, my only other worry is whether the PWM signal will damage the driver. I put the lights away and didn't run them this way any longer. Roberto says he talked to people at Meanwell and it won't hurt the drivers to do this, so that's a glimmer of hope.

But, the main reason I posted this was to give data showing the PWM-to-Analog lead wires are not working as intended (or at least not to my expectations.) They do allow an analog driver to work, but at a very minimum capacity compared to their optimum performance.

I think my 3 x 3w moonlights, running on a constant current 1x3w driver was brighter than the 33 x 3W royal blue / UV mix running at around 2W each actinics.

Roberto, I'm the one that linked you this post in an email this morning. (Well, I'm about to do that right after I hit submit. Do you think you can take a look at the converters and test the voltage on the wires to verify my findings? Maybe the wrong resistor/capacitors were used to make these converters?