ET-AL01 - 5 channel Analog LED controller

Hi,

I want to share our 5 Channel Analog LED controller for use with ESPHOME and Home Assistant with you guys :slight_smile:

Please also check out our blog article about the ET-AL01 on our website: 5 channel Analog LED controller - ET-AL01 - espthings.io

PCB: espthings.io - ET-AL01 - Share Project - PCBWay

Soldering guide: Soldering - espthings.io

Our favorite tools and equipment: Tools and equipment - espthings.io

4 Likes

40 reads, 0 response. Damn, that hurts :wink:

1 Like

I´ve ordered 10 boards. PCBWay quoted 10 at the same price of 5. I’m building a new house, it seems i’ll install more LEDs than anticipated :wink:
As soon as all components arrive, I will asemble them and write my review…

Great! If you use the voltage regulator, place it closest to the terminal blocks as you need to bend to it to the right a little in order to fit in the enclosure.

Love to see photos of the finished product :slight_smile:

I will use the regulator, but I want to create a custom 3D printed box.

Will post photos once built.

Mind the “screw holes” are bigger in order to fit over the pilars of the enclore we used. We did not want to use screws, but rather lock the pcb within the closed enclosure.

1 Like

Thanks for the info, I’m planing to hold the board that way.

Since I’ll have PCBs in excess, I’m wondering if this driver could be suitable to replace the “dumb” drivers on a couple of LED lamps I have. The issue is that those lamps need a Constant Current driver and, as far as I understand, yours is a Constant Voltage driver, isn’t it?

It’s PWM dimming, so this does not apply. PWM dimming is a sort of Constant Voltage, it just switches on/off really fast :wink:

1 Like

First board built, a shame that I’m still waiting for the 470µF capacitors to arrive…
Next, build a 3Dprinted enclosure.

Nicely done!

Please share the 3d print enclosure design via thingiverse :slight_smile:

1 Like


I’ve finally got some time to create the box. This a first version. I’ve ordered some ET-AD01 and ET-DL01 boards from PCBWay and will wait until I receive them to make sure the screw holes fit them all.

The Espthings.io logo is engraved but with this filament is barely visible.

Regards!

Wow, great stuff! :ok_hand:

Thank you! I’ve been looking for something like this for quite some time

Hey,
thank you a ton for this cool thing! I was looking for hours for a similar device on Aliexpress, but no success or at least with a lot of invasive tinkering needed.

Before finding this I always wired everything up just manually on prototyping matrix boards. But that was so damn unefficient.
Now this here is awesome and I am about to order 10 pieces. And to be honest I think I will replace all my selfmade boards with that one here to have a consistent solution all over my projects.

Thank you!

Hello there. Where can i order the driver board?
my project is 24v analog rgbww esp32 controled for homeassistant

1 Like

Have a look at the very first post. The link to PCBway is just there.

I have a question regarding the “gate driver” 74HCT244:
as I understand the idea behind that it is just to ensure a proper gate voltage of 0V…5V with less tolerances than the ESP alone would have, right?
But when choosing a logic-level FET wisely (as you did with the IRLZ44N) you should usually not need that driver. In case of the IRLZ44N it needs at least 1,5V, which is far away from ESPs maximum low-level voltage of 0,33V) and in worst case 2V to be seen as high-level, which is also far enough below ESPs minimal high-level output voltage of 2,64V. So in that case one could just bridge the five channels of the 74HCT244, right?

And one more note regarding the IRLZ44N: that one is obsolete (EOL), see here because it is not lead-free. The RoHS conform model would be the IRLZ44NPBF. Even a little bit cheaper: IRLZ34NPBF or even the IRLZ14PBF if not much current is needed (high RDSon).

The answer to your question lies in this picture:

If you would drive the FET with just 3.0V, the max. current through the FET is (rounded off) 20A @ 12V VDS.
If you would drive the FET with 5.0V, you would be in between the curves of 4.0V and 6.0V, which show that the max. current is in the range of approx. 70-80A @ 12V VDS.
That means that the drain-source channel resistance is significantly lower when driving it at 5.0V vs 3.0V. This results in a lot less heat losses in the FET when you are driving 5 - 10 A on your output channel.

The same information is also “hiding” in the electrical characteristics table in the datasheet:


Look at the increase of the channel resistance if the VGS is dropped from 5.0V to 4.0V. The channel resistance will increase even more when the VGS is lower.

To answer your question if you can just “jump” / “bypass” the 74HCT244 chip; Yes, you can. However your output current capacity will be diminished, and most likely you will need to use a heatsink on the FETs to keep them cool.

I wrote a blog post about this specific design choice here . You can see what the influence can be of driving the gate with a (much) lower voltage.

@pzw Hey Paul,
thank you very much for your description! Of course you are totally right to point out to the relationship between V_gs and Rds_on.
But my question was really more to understand if there is another reason for the 74HCT244 than “just” gate driving. Of course one has to make sure that the individual usecase does not exceed any limits/thresholds :slight_smile: .

In my case: I will not draw more than 2,5A per channel. In most applications it will be even less, like around 1A or so. So assuming a very bad Rds_on of 0,05 Ohms and 2,5A would result in 0,3W losses. So really far away from pushing it to the limits :slight_smile: . But sure, this is just MY usecase.

This leads me to the conclusion: if I can get the gate driver IC, I will use it because it will make the MOSFETs feel more comfortable. If I can not get it (or run out of stock) I will not necessarily need is.

I did some math regarding the monetary costs of the additional losses when not using the gate driver compared to the costs of the gate driver itself. The result in my usecase was negligible - in MY case :slight_smile: .

Thanks again for sharing this project!