Optocoupler Necessary?

Hi All, I'm going to use my controller to provide a pwm signal for dimming 4x led drivers. The drivers themselves produce a 10v >10ma signal that is grounded via a potentiometer to vary the driver output, open circuit = max brightness and short circuit = max dimming. I'm going to replace the manual pot with this circuit.... |281x500 I've had it suggested to use an optocoupler between the controller and each dimming circuit and I was wondering if that is really necessary?

IMO, opto isolation would not be necessary in this application. I would recommend using a low Rds(on) logic MOS FET, but the transistor may meet you needs. I would not use the input capacitor as it will slow the rise time of the PWM.

You need to tell us what your LED drivers actually are. We cannot guess which one's you have.

LarryD: IMO, opto isolation would not be necessary in this application. I would recommend using a low Rds(on) logic MOS FET, but the transistor may meet you needs. I would not use the input capacitor as it will slow the rise time of the PWM.

I forgot to add initially the dimmer is analog not pwm, I'll look into the mos fet in place of the transistor

MarkT: You need to tell us what your LED drivers actually are. We cannot guess which one's you have.

I've no idea what they are, they come from a generic 165w chinese black box with no labels so your guess is as good as mine. I've tried to provide as much information as I can, if you'd like more specific details on the circuit let me know what you need.

The question is whether the dimmer will respond to PWM. Some are designed specifically to do so, others are not. That is why we ask.

Paul__B: The question is whether the dimmer will respond to PWM. Some are designed specifically to do so, others are not. That is why we ask.

As far as I can tell the driver works by grounding out the 10vdc signal produced by the vdm+ to the vdm- pin. The amount of voltage allowed to reach the vdm- pin controls the brightness. It is an analogue signal and the driver appears to have no facility to receive pwm or a traditional 0-10v analogue signal. So I'm trying to use a similar circuit to what would normally be used to produce a 0-10v analogue signal but as the driver produces it's own voltage no additional psu is needed.

Okay had a chance to give this more thought and have a play. An opto coupler will keep the drivers isolated from each other, I've no idea whether they would actually interfere with one one another and at this time I've only prototyped a single circuit so far. I still need to get my hands on the opto couplers but thoughts on this?|281x500

Your drawing.. I think that it's just wrong.

Yes, the RC would come after the transistor but not quite the way you have it drawn. A pot would not normally carry much current. You could likely have better results removing the transistor (coupler) entirely and imagine generating the voltage needed with an OP AMP.

Something to consider. Feed the op amp with the output of the RC filter from the pin. The OP amp should be able to create the 0-10V signal.

(No, I've not thought this through... just tossing out fodder for consideration.)

Your circuit will be isolated as long as there's no GND connection from the led driver to the Arduino. If your Arduino or MCU has 5V logic, then:

The current through the FOD817A's IRLED is (5-1.2)/200 = 19mA. The current transfer ratio (from graph) is about 132% = 25mA transistor current. The RC time constant of your filter is 15.5ms.

The way you have the filter connected, when the PWM signal goes HIGH, there would be instant response to full brightness. When the PWM signal goes LOW, it will get dimmer at a rate determined by the RC time constant.

To get near equal rise/fall response, try this push-pull opto-driver:

dlloyd: Your circuit will be isolated as long as there's no GND connection from the led driver to the Arduino. If your Arduino or MCU has 5V logic, then:

The current through the FOD817A's IRLED is (5-1.2)/200 = 19mA. The current transfer ratio (from graph) is about 132% = 25mA transistor current. The RC time constant of your filter is 15.5ms.

The way you have the filter connected, when the PWM signal goes HIGH, there would be instant response to full brightness. When the PWM signal goes LOW, it will get dimmer at a rate determined by the RC time constant.

To get near equal rise/fall response, try this push-pull opto-driver:

I'm feeling a tad out of my depth here so please bear with me..... In order to replace a manual potentiometer with a digital device you need some kind of digital to analog conversion. A pwm signal is a square wave on/off signal An NPN transistor can be used as an on/off switch which can be switched by the pwm. By using an RC filter to smooth the transistor output you can provide a consistent output voltage that can be varied by changing the pwm duty cycle, replacing said manual pot.

I'm not quite sure what your circuit achieves? If the pwm signal is low the second coupler flows and if the pwm signal is high the first coupler flows, would this not allow constant flow from vdm+ to vdm-? What is the purpose of the 680ohm resistors? I assume there should also be a 200ohm resistor between the 5v and the second coupler?

I trialed a similar circuit to my second one above earlier but in place of the opto coupler used a bd139 transistor and it seemed to operate without any flickering but I forgot to take my laptop with me today so dimming was running over about 30 seconds. I'm going to try map the driver output voltage vs the pwm duty cycle so I can make a lookup table for a log dimming curve so I'll have good check to see how the duty cycle effects the led output.

Just looking at your threads on this led controller ( inc RC), couple of things I would ask.

Any link or photo of your 165w driver ( 4 channel ?)

Also have you actually drawn out that little pots/components pcb , if we can see exactly how the it does it, it might be easier to suggest how to control it.

I'm not quite sure what your circuit achieves? If the pwm signal is low the second coupler flows and if the pwm signal is high the first coupler flows, would this not allow constant flow from vdm+ to vdm-? What is the purpose of the 680ohm resistors? I assume there should also be a 200ohm resistor between the 5v and the second coupler?

Your circuit would create a signal across the VDM ± would be similar to this:

The purpose of the suggested circuit is to emulate a basic RC filter to achieve an analog response.

For example, 50% PWM duty = 50% voltage across the VDM ± terminals.

The 680Ω resistors, when switched in parallel will charge the capacitor at about the same rate as the discharge rate when the lower 330Ω is switched in. No 200Ω resistor needed between the 5v and the second coupler. The 200Ω is shared as only one coupler is on at any time.

If your circuit works without flicker, then it looks like this is all you need.

Thanks for the explanation. Looking at this http://sim.okawa-denshi.jp/en/CRtool.php using 490hz it suggests using a 1uF cap with the 330ohm resistor, I'm only using a 47uF as that's what I has laying about. But is the calculation on the above site applicable to pwm?

@ricky101 I'll post a pic of the setup so far at some point today. I have mapped the little dimming board and from what I can tell the extra resistors etc are to modify the pots output curve.

|281x500 This is the set up so far, it's 4 separate drivers that I want to be able to control independently.

Hi,

Would still like to see a diagram of the dimmer control board, should only take a few minutes to draw it out, looks like a transistor and diode is used as well as res and cap.

Not sure whats on the top side ? - just the pot and connector ?

Seeing the actual circuit should help in what and where to apply a control signal.

ricky101: Hi,

Would still like to see a diagram of the dimmer control board, should only take a few minutes to draw it out, looks like a transistor and diode is used as well as res and cap.

Not sure whats on the top side ? - just the pot and connector ?

Seeing the actual circuit should help in what and where to apply a control signal.

Here is a (slightly) better pic of the dimmer board. |375x500 I'll draw a map tomorrow. The only things on the other side of the board are a pot/switch combo and the connectors, center pin of the pot is the wiper.

I set this circuit up today and it seems to work well with no flickering noticed I stepped through the pwm 10 steps at a time. I'm hoping to get a full 256 step - driver output map done tomorrow to give me an idea of the curve. I should be able to directly replace the BD139 with the opto coupler. |500x281