0-10V controlled via I2C

[permal]

Hi,

My kitchen LED lights are currently driven by one of these DC Maxi Jolly, (English text starts at page 5). The current controller is in turn controlled via RF signal and very unreliable. To remedy that I want to build a new controller.

• Output: 0-10V, 35mA
• Interface: I²C - I'm going to use an ESP32 as the MCU.
• Powered by a single 12V supply (the AUX power from the Dolly driver)

What I need guidance on is with the 0-10V output circuitry:
- Would a DAC101C085 (0-5V DAC, I²C) with a single-supply OpAmp (LM2904?) (non-inverting with a gain of 2; negative voltages not needed) attached to the output do the trick?
- How do I design the output to deliver, say 100mA to have some margin on the required 35mA?

Thanks in advance!

[frozenfrogz]

How about using a DAC like this one and an op-amp with x2 gain setting?

[permal]

How about using a DAC like this one and an op-amp with x2 gain setting?

You're suggesting the same kind of setup as I did; so my thinking wasn't all wrong then. The MCP4725 and DAC101C085 seems pretty equal except that the former one is half the price.

So what about the sourcing ability of this circuit? For example, LM2904 can only source 30mA (Vcc at 15V), how do I increase that to the required >35mA? Are there any Single-supply OpAmps that are capable of sourcing ~100mA or more?

[HB9EVI]

Controlling led brightness over the voltage is nonsense; what you need is a voltage controlled current source; so use a fet with a current sense resistor to build a variable current source

[permal]

Controlling led brightness over the voltage is nonsense; what you need is a voltage controlled current source; so use a fet with a current sense resistor to build a variable current source

I already have the LED driver (linked above). It uses a 0-10V control input signal to determine what current it should drive the LEDs with. I don't see how a variable current source would help in controlling it?

[tooki]

I think they misunderstood, and thought the 0-10V was to drive the LEDs directly. Must not be familiar with the 0-10V control standard.

[frozenfrogz]

You're suggesting the same kind of setup as I did; so my thinking wasn't all wrong then. The MCP4725 and DAC101C085 seems pretty equal except that the former one is half the price.

Sorry, my bad. I read your post, then had it sit in my mind for some time and later hit reply without reading the post again :/

One other thing that I thought about right now: A lot of the 0-10V signal inputs on dimmer circuits are built as constant current sources, so you can usually just connect a potentiometer and you’re good to go. Maybe try that as you can spare all the expense for a DAC + op-amp (feels like a bit of over-engineering - right?).
If it works, you can simply use a digital potentiometer with I2C.

Edit: The datasheet (p5 bottom left column) you posted confirms this, you simply need a 100kOhm pot.
Here is a selection of digital pots.

[Zero999]

There's no way you need 35mA for that. The LM358/LM2904 will be fine on its own. If you really want more current, two channels can be connected in parallel, using current sharing resistors.

See the application note linked below:
https://www.intersil.com/content/dam/Intersil/documents/an11/an1111.pdf

[permal]

There's no way you need 35mA for that. The LM358/LM2904 will be fine on its own. If you really want more current, two channels can be connected in parallel, using current sharing resistors.

See the application note linked below:
https://www.intersil.com/content/dam/Intersil/documents/an11/an1111.pdf

The datasheet states "Light regulation 0-100% by means of PUSH function, 1...10V interface (I=0.35mA) or 100 Kohm potentiometer.", that's where I got the number from. Thanks for linking the AN, I learned something new.

The datasheet (p5 bottom left column) you posted confirms this, you simply need a 100kOhm pot.
Here is a selection of digital pots.

It does mention 100k pot - but not where to actually connect it, or have I missed it? If you're right it just goes on the input. I suppose I can simply give it a try.  This'd be a much simpler solution if it works.

[Kleinstein]

When using a digital pot, keep in mind the voltage at the pot must be inside the supply of the chip. So it would need a digital pot for at least 10 V - the normal ones are 5.x V max.

[permal]

When using a digital pot, keep in mind the voltage at the pot must be inside the supply of the chip. So it would need a digital pot for at least 10 V - the normal ones are 5.x V max.

Yeah, I noticed that when reading a datasheet for one of the available ones.

I connected a regular 100k pot, the result wasn't what I expected. Yes, the lights can be dimmed, but it all happens within the lower ~20% of the scale. Why specify 100k when the entire dimming happens within the lower ~20k?

[Zero999]

There's no way you need 35mA for that. The LM358/LM2904 will be fine on its own. If you really want more current, two channels can be connected in parallel, using current sharing resistors.

See the application note linked below:
https://www.intersil.com/content/dam/Intersil/documents/an11/an1111.pdf

The datasheet states "Light regulation 0-100% by means of PUSH function, 1...10V interface (I=0.35mA) or 100 Kohm potentiometer.", that's where I got the number from. Thanks for linking the AN, I learned something new.

You got the decimal point wrong. 35mA, is 100 times 0.35mA.

[HB9EVI]

I connected a regular 100k pot, the result wasn't what I expected. Yes, the lights can be dimmed, but it all happens within the lower ~20% of the scale. Why specify 100k when the entire dimming happens within the lower ~20k?

could have caught a logarithmic pot perhaps?

[james_s]

A DAC seems a bit extreme for this application. How about a PWM output from a microcontroller with a RC low pass filter? To get the voltage up to 10V you could either use an op-amp to amplify the resulting DC level, or buffer the PWM signal prior to the low pass filter.