New boost driver with IMS PCB

[laserluke]

Here is my new high power and thermally efficient boost driver:
- insulated aluminum substrate (IMS) PCB for exceptional thermal efficiency
- current output exceed 2.5A
- very high precision current control and multi-tourn trimmer
- grounded load (only 1 wire to connect)
- open and short load protection
The PCB can be connected directly to the heatsink to obtain a perfect thermal contact.
The board is currently under tests and performs very well. :yh:
It is possible to make commercial boards if anyone interested.

 

[mattco2]

Looks nice price?? And it looks quite complicated

 

[laserluke]

Looks nice price?? And it looks quite complicated

The IMS PCB costs the same as a dual layer standard FR4 PCB. The technology limit is that an IMS PCB is a single layer PCB.
The circuit has an advanced current control to maintain an high efficiency.
As an example a driver based on TPS63020 with standard schematic uses a shunt resistor of value 0.5/Iload. For 2A output the resistor is 0.25ohm so the thermal power dissipated only by the shunt resistor is R*I^2=1W !!! reducing significantly the total efficiency of the driver and battery life.
The advanced current control reduces the dissipation of a factor of 25 and increases the control precision. The additional circuitry is almost ininfluent on total cost.

 

[Sigurthr]

My only question is how the designer compensated for the stray capacitance of the PCB, and how, if at all, it affects the circuit. It looks wonderfuly designed though, that is for sure.

 

[sinner]

Sounds pretty great, since we need 1.8A at most..

 

[laserluke]

my only question is how the designer compensated for the stray capacitance of the pcb, and how, if at all, it affects the circuit. It looks wonderfuly designed though, that is for sure.

I have observed no differences in circuit behaviour between standard FR4 PCB prototype (attached photo) and IMS PCB. The stray capacitance has negligible effects.
Many thanks for the compliments :yh:

 

[laserluke]

Sounds pretty great, since we need 1.8A at most..

I'm building a 445nm laser with this driver and an Ultrafire C8 host. I'll post pictures very soon

 

[lasersbee]

Nice job on the design...:gj:
What is the diameter of that PCB ??


Jerry

 

[laserluke]

Nice job on the design...:gj:
What is the diameter of that PCB ??


Jerry

Many thanks :yh:
The diameter is 1 inch (2.54cm). Fits perfectly in a cree torch.
You can mount the driver on top of the cree led pill as in this post (the driver needs only a single 18650 battery):
http://laserpointerforums.com/f52/2-05-watt-445nm-laser-cree-sacredfire-nf-009-mohrenberg-heatsink-tutorial-59251.html#post837213

 

[foulmist]

nice build! somewhat big but still nice

What's the input/output voltage range?

 

[TheReddish]

I would like to try a boost driver one day, for a handheld, I can make one for a labby, but for handheld I have used nothing but linear so far because of availability as I have been doing this a little less than a year, so I am going to lurk this thread to see what happens, but you have my interest.:beer:

 

[laserluke]

nice build! somewhat big but still nice

What's the input/output voltage range?

The input voltage range is 2.4V to 25V.
The output voltage absolute limit is the maximum switch voltage of 42V. The current control circuit limits the output voltage to 12V to protect output capacitors but this value can be easily changed replacing a component and using higher voltage output capacitors. The maximum switch current is 6A.
The driver can supply more than one laser diode connected in series.

 

[BShanahan14rulz]

You know, folks in the flashlight arena need a nice, high-current boost driver too. The current generation of emitters is generally powered up at about 3A, dropping ~3.3V across each LED, and wired in series. It would be bomb to have a sleeper light that blasts out ~3000lm off of one 18650 sized cell ;-) What sort of input voltage and current requirements would need to be met to get this to push 3A at 10.5V?

 

[laserluke]

You know, folks in the flashlight arena need a nice, high-current boost driver too. The current generation of emitters is generally powered up at about 3A, dropping ~3.3V across each LED, and wired in series. It would be bomb to have a sleeper light that blasts out ~3000lm off of one 18650 sized cell ;-) What sort of input voltage and current requirements would need to be met to get this to push 3A at 10.5V?

The "magic word" is efficiency. It's not trivial to exact calculate the efficiency of a boost converter.
Here you can find an article with problem analysis.
Anyway let's try to do a simple estimation. The power required for your "thermonuclear torch " is 3*10.5=31.5W
Suppose to have a converter efficiency of 80%... The input power required will be 31.5/0.8=39.375W... with a single 3.7V cell the required current is 39.375/3.7=10.642A !!! too much for the cell and for the driver.
If you try to recalculate with 2 cells (7.4V) and the same efficiency the current will be 5.32A. This value is very close to the switch limit of 6A and probably beyond the cell's possibilities...
If you try to recalculate with 3 cells (11.1V) the voltage is greater than that required by your leds (10.5V) so you can use a buck (better efficiency) or linear regulator (worst efficiency)...
...or you can use two boost drivers each with half load... the limit is fantasy:yh:

 

[BShanahan14rulz]

Oh, I see, this is why you mentioned the 6A switch current limit earlier! Thanks for that explanation :wave:

 

[DTR]

Wow that look awesome. Great work.:beer:

Would you have the ability and desire to post a test video for us like the one below. The main thing I would like to see is if the current stays constant @ 1.8A as the voltage decreases. Here is one I tested that did not.:yh: