OPEN SOURCE: "CC-Boost" - 2.4 Amp boost driver - RC1

[benmwv]

Hopefully when you test it on the newer scope the output is a bit cleaner! It would be very good news if that was the only problem.

About the efficiency, 65% is low, but that might just be how it is. Current regulators are always less efficient than voltage regs because of the current sensing components. You will never achieve the data sheet efficiency in current regulation.

That said, I would still expect it to be 75-80%. At least that's ballpark efficiency for most driver I've made/used. I would try reducing inductance, see what effect that has.

 

[rhd]

Hopefully when you test it on the newer scope the output is a bit cleaner! It would be very good news if that was the only problem.

About the efficiency, 65% is low, but that might just be how it is. Current regulators are always less efficient than voltage regs because of the current sensing components. You will never achieve the data sheet efficiency in current regulation.

That said, I would still expect it to be 75-80%. At least that's ballpark efficiency for most driver I've made/used. I would try reducing inductance, see what effect that has.

Frankly, I would just move on to a higher efficiency chip. This is a 500 khz IC. If a higher frequency is used, the inductance can be lowered, while keeping the output clean. That means you can potentially also improve the resistance of the inductor, giving similar size constraints, which means less power lost to heat in the inductor. I would also go fixed, and adjust the RSENSE on a per driver basis so that you're not dropping any more than the ~50mV (say 60mV for some margin) that the ZXCT actually needs across the resistor it's measuring. With an adjustable driver, at the high end of the range, you're dropping nearly 400mW by my off-the-top-of-my-head math, where you really only need to drop about 125mW if the driver was fixed and using an RSENSE chosen for a specific target current like 2.3A. That's an extra 275mW of inefficiency generated right there, just for the luxury of having a pot and no other reason.

 

[benmwv]

Very true ^

I like pots (hehe) but really they are a bit impractical. It's not like you need to adjust current to your diode on the fly. Might as well just calculate the value you need and set it fixed. Less cost, less board space, more efficient, more reliable.

Thing is lower frequency stuff is usually more efficient at the cost of bigger components vs high freq so I'm not sure why his numbers come in so low.

 

[jayhawker08]

Ran some more tests. I realized I didn't have the bandwidth limiter enabled on the scope. This cleaned up the signal a bit since most of the noise was >20MHz. Overall, same observations as before, namely:

Under a certain threshold current (~1.8-1.9A) the driver oscillates about 1V pk-pk

(yellow is output voltage at 1.5A, green is the feedback voltage into the IC)

(inductor voltage at 1.5A)


As the current is increased, the oscillation frequency slows

(output voltage at 1.9A)

Eventually it will reach the threshold current and stop oscillating, at which point it outputs a nice clean signal (94mV/4.57V = ±1% regulation)

(output voltage at 2.0A)

(startup at 2.0A)


Under the threshold, the driver heats up very quickly. Above it, they can run with no heatsink for a minute or so.

Both drivers I've assembled behave almost identically. Has anyone else tried running the RC1 version much under 2.0A?

 

[The Lightning Stalker]

IWIRE has. Is it possible that the ZXCT has
overheated?

 

[Hiemal]

Hey guys;

I've been reading through this thread a bit and it's definitely been a bit of help in regards to my own driver. I certainly appreciate everyone's hard work here and hopefully i'll be able to fool around with the E-drive again and see if I can't get it up and going again, in small quantities, while ironing out all of its various issues.

RHD, fantastic job on this driver. Hope you get everything figured out for it too.

 

[Alaskan]

I've saved the whole thread in a folder, have some of the bare boards Adam from down under sold, now to populate them

If anyone has one or more completed and working boards to help me put these together, I'd be happy to trade some cash

 

[crazyspaz]

just download eagle and open the files, you can right click on components and see what they are.

 

[Alaskan]

Thanks, I need to do that. Was hoping to also find a completed unit to see before I start buying parts for the bare boards I have. I have seen photo's and can do without a finished one, but having a working device in hand would be very nice.

 

[dubious speculation]

I know that I've probably missed it within the 31 pages here, but is the original CC-boost V08 10x10mm driver still useable?
The current RC1 13x12mm driver is obviously the newer and more stable version but for my applications only the 10x10 driver will fit.

 

[dubious speculation]

Perhaps I shouldn't have ended my post with a statement ^^^.

Is the original V08 10x10 boost driver still functional?

 

[The Lightning Stalker]

Personally, I could never get them to work,
but they are reported to be working. You
are welcome to try.