Design validation

Hi lads and gals,
this is the first time I use MOSFETs and the TLC5940 integrated circuit so, while I’ve been spending some time researching and I’ll start prototyping the thing in a few hours, still I wish to seek the community advice to validate my 5 x 5 x 5 LED cube circuit board.

I’ve already built a 3 x 3 x 3 cube using NPN BJTs and using directly Arduino pins to drive the LEDs, but because I wish to scale up to 8 x 8 x 8 I need a different approach and I believe an intermediate size might help.

A few considerations I already did:

  1. I don’t think I need a BJT to drive the MOSFET into saturation because of the part I picked, but I’m still not 100% sure, the idea is to keep the circuit as simple as possible
  2. I’m not sure I’ll be using the greyscale mode of the TLC5940, but I wanted to leave the possibility open
  3. I decided to optimize the data transfer and concentrate the padding at the transmission end rather than balance the power dissipation between the two TLC5940, but I’m not sure it’s the best choice

Please find the schematic attached and the datasheets links:

This will be my first time soldering SMD, I picked what I believe it’s a simple package to start with…

I think I might need an additional 10k pull up for each MOSFET gate to ensure the planes are turned off during bootstrap, but, on the other hand, the TLC should be turned off until latched…

I don't see any problem with that. 10K pullup on the gates to keep the mosfets off during reset would be good.

After some additional thinking I decided to reduce the gate resistor value down to 330 Ohm: still enough to limit the inrush current but should keep the MOSFETs cooler and allow for faster switching.

Not that I really need any high frequency switching having just to hit the POV frequency (aiming at 100Hz for a full cube refresh) , but I belive it will be a bit better for the circuit.

I’ve updated the schematics accordingly and included the pullup.

@CrossRoads thanks for your feedback, just a question: have you ever used that specific MOSFET?

I’m having second thoughts on this circuit related to the MOSFET Vgs: the datasheet reports 0.5 to 1.3 volts with a typical value of 0.9V.
I will be driving 25 LEDs in parallel (64 in the final version), but they are cheap eBay blue LEDs (no datasheet) which should have a voltage drop between 3.4 and 3.8 volts: it can get past the Vgs threshold!
Am I right assuming the consequence of not having a high enough Vgs the mosfet will stay in the linear region and quickly get hot trying to burn its way to the bin?

If I manage to measure the voltage drop of my LEDs what else should i take in account to evalute my capability to saturate the MOSFET? I will be powering the thing using a USB wall plug rated for 2A…

Why are you putting 330 ohm resistors in series with the FET gates? I don't see the purpose.

@charliesixpackTo limit the inrush current mainly, with the side effect of reducing the MOSFET oscillation… that’s what I’ve understood about MOSFETs.

MOSFETs are voltage driven and their gate impedance is very high, but they have a gate capacitance in the order of the nano farads (the one I’ve picked is reported as 700pF): when I will turn the MOSFET ON the gate capacitance will discharge (it’s a p channel) into my pin and without the resistor the current will have a spike. Very short in duration, but still there. Do it 800 times a second and the pins might die quite rapidly.

I’m having second thoughts on my circuit regarding the MOSFET Vgs, something I might have underestimated. The part I picked reports a Vgs between 0.5 and 1.3 volts with a typical value of 0.9V. I will be using it to source current to 25 LEDs in parallel (1A) with the goal to drive 64 of those (2.5A) but those are cheap eBay (no datasheet) blue LEDs!
Normally blue LEDs have a voltage drop between 3.4 and 3.8 volts which means I might be unable to get past the Vgs threshold!

Am I correct saying if that will be the case my MOSFETs will quickly become hot trying to burn their way to the bin?

If I manage to measure my LEDs voltage drop what else will play a role in the Vgs calculation?

I am planning to power the thing with a USB wall plug rated for 2A and a 3A one for the full scale version…

No. Just looked at its specs.
For cubes, I have built this card to drive a 9x9x9 cube, with thru hole N-channel MOSFETs to sink current from each layer.
cd74ac164 sources the current into each column.

I don't think you will blow anything up by omitting those gate resistors. Assuming 12V, 700 pf, and 1 MHz, I calculate 50 mW that will be dissipated in the driving device. That should be no problem.

@charliesixpack it's just 5V and the driving device is the Arduino pin...
In every search I've done so far the MOSFET gate resistor is almost always present and whenever it isn't there are comments suggesting to have one.

I'm not saying you are wrong, I've never played with MOSFETs before, but I will wait for additional comments before removing the gate resistor as per your suggestion.

Anybody willing to confirm I don't need a gate resistor?

No comments whatsoever on my concern regarding Vgs posted a couple of comments above?

Vgs is what turns the MOSFET off & on, and has nothing to do with Rds and the resulting Vds.
If the power supply is 5V, then 5V - 0.9 is what's needed to turn it on.

I'd add the gate resistor to keep the arduino output from overcurrenting into the gate capacitance of the MOSFET. It's just good design practice. If you were undergoing a design review by your peers at any company, you'd have a hard time justifying not having it in there. Spend the penny or two and enjoy a more durable & longer longevity project.