hyperbuck preview

[georges80]

Well, after the success of the hyperboost driver I decided to see if I could leverage the same switcher core design as a buck converter. Received new boards a few days ago and have finished doing some preliminary testing of the new hyperbuck driver.

So far things look good. I've tested the driver at 60V input (the driver is capable of up to 80V input). The worst case for a buck is high input voltage and low output voltage. My initial tests are with a single 3W LED with 60V input and things are looking good.

Should make a good driver for ebikes etc.

Hyperbuck specs;

8V to 80V input.
Buck converter (obviously).
Reverse polarity protection.
Fully current regulated with on board multiturn trimpot to set output current (like hyperboost)
Up to 1500mA output.
Supports external Pot for dimming.
Open circuit protected, the driver will limit output voltage to 49V (to protect the output caps).
1.4" x 1.2" rectangular. Components on top side only, to ease mounting to a heatsink via thermal pad material.

Here's a picture:

Next I'll have to do some efficiency measurements at a few operating points.

cheers,
george.

 

[trout]

As usual it will be a great product George and no doubt folks will be plotting ways to use it to the Max .

If I may add a suggestion any chance of a couple of holes for mounting screws in these single sided boards

 

[marcopolo]

You don't ever stop for breath George!!!!!

I'm curious since you've designed the Hyperboost and now hyperbuck in quick succession are you going to flex them? Has there been enough demand for a flexed version yet? You mentioned issues with size for the extra componentry before I think.

regards,

Marco.

 

[georges80]

Mounting holes:

Well, it requires more real estate on the board and folk seem to have problems with the size of these boards already. So, I figure if folk want mounting holes then they would have a larger board anyway, so they could just put 2 holes in their heatsink to line up with the 'edge' of the PCB (say one to the left of C6 and one to the right of POT+/POT-). Then the "lip" of a pan head screw would catch the left and right edges of the PCB and hold it down to a piece of thermal pad material and the heatsink.

This would allow the folk that want to mount with screws a system to do it without impacting the board size for the folk that want to mount using some other scheme - like double sided 10mil thermal tape from a company like 3M.

Flex version:

Well, that will take some time. Obviously the board size would have to grow to accommodate the extra circuitry (regulator, ucontroller, programming pads, various resistors & capacitors etc). It would also require inventory of even more boards.

So, the key would be whether the Flex features are needed even though it would grow the hyperboost and hyperbuck board sizes by maybe 15 - 20% board area. I'd still want to keep the boards single sided to provide easier heatsinking.

Possibly for these two boards larger diameter/area wouldn't be too much of an issue since they are capable of driving so many LEDs and run off fairly large battery packs etc.

Anyhow, not many requests for flex versions yet...

cheers,
george.

 

[georges80]

More testing and I've run it up to 72V, basically the limit of my power supplies.

I've performed some efficiency tests and following is a table of some of the results, more info on my website.

Typical of switchers, the closer the input voltage is to the output voltage the higher the efficiency.

The power loss column shows the heat loss in the driver at various input/output combos.

cheers,
george.