Buck and Boost Drivers, Testing, Modding, and Discussion (Pic Heavy)

297 posts / 0 new
Last post
JasonWW
JasonWW's picture
Offline
Last seen: 22 min 5 sec ago
Joined: 10/22/2016 - 11:41
Posts: 11527
Location: Houston Texas

Nightsword wrote:

My application is a bike light, specifically a IQ-XE light. The unique efficient design has the best cutoff for no glare onto oncoming traffic without losing lumen.

That looks like a neat design, but it’s dynamo powered? I guess it runs off the turning bike wheel? Seems odd.

Have you seen Lumintops new bike light, the B01? It’s a similar lens design, but cheaper and uses a 21700 battery. Plus it’s USB rechargeable so you should be able to plug in a power bank to greatly extend run time. Your spending a small fortune on modding that IQ-XE. It might be time to cut your losses.

Texas Ace Lumen Tube calibrated with Maukka lights

Click this to go to signature links.

Nightsword
Offline
Last seen: 5 days 9 hours ago
Joined: 06/30/2017 - 13:22
Posts: 20

The IQ-XE is entirely different league when it comes to a most lux at distance with distinct cutoff line for least lux up onto oncoming. The best comparison is the Lupine SL with broader and more even coverage, which many prefer over IQ-XE, but not me. We have very bad roads here and I need to see those potholes well ahead of time and the IQ-XE is the most efficient at putting lux right at the cutoff line directly ahead. The IQ-X is the dynamo version and IQ-XE is Ebike version. Had both. Same except for internal drivers. Dynamo version is cheaper, so it’s better for driver modding. When modding for higher output, also best to bond the LED assembly to the housing for better cooling of the LED. BTW, IQ-XE is already more than enough for just about anyone, but I can never turn down opportunity for improvement, and this mod does wonders.

I’m building the battery pack with regulated DC input for expansion and charging, and 5v DC output for phone charging while on the go. I wouldn’t venture OT but this is your thread, so I’m glad to expand.

JasonWW
JasonWW's picture
Offline
Last seen: 22 min 5 sec ago
Joined: 10/22/2016 - 11:41
Posts: 11527
Location: Houston Texas

I’ve messed with a lot of projector retrofits with hid when they started to be more common 15 years ago and recently started studying led versions of projector and even reflector housings. I love projectors for their sweet cutoff lines and very usable light on the ground.

These 90° led/reflector designs are still pretty new to me. I see them popping up here and there. Do you think their beam shape beats a projector housing?

I have seen these 90° led/reflector designs (do they have a specific name?) in aftermarket fog lights. I wonder if the design will be adapted into a OEM car headlight design soon.

Texas Ace Lumen Tube calibrated with Maukka lights

Click this to go to signature links.

Nightsword
Offline
Last seen: 5 days 9 hours ago
Joined: 06/30/2017 - 13:22
Posts: 20

Automotive is easy because efficiency isn’t detrimental. So simply blocking of the top half of the light works, and has inherent high beam built right into the same unit. European restrictions on bicycle lights has driven the market to create new ways to illuminate only the lower half without losing efficiency under small battery operation, and without need for high beam.

For what it accomplishes, I think the 90° design is genious. But because the automotive projector design is so simple while having inherent high beam, I don’t think it’s going to be replaced by the more efficient 90° designs soon, but it’s possible. Saving a few watts in automotive probably makes less difference than having a bug on the windshield.

However, you may find the Lupine SL interesting, as it’s an implementation of the 90° design in a projector configuration. That’s what makes it a much more pleasing beam pattern than the IQ-XE, but also what makes it have less peak lux.

JasonWW
JasonWW's picture
Offline
Last seen: 22 min 5 sec ago
Joined: 10/22/2016 - 11:41
Posts: 11527
Location: Houston Texas

Doh! Facepalm I forgot these new 90° designs can’t change to create a high beam. Still, a lot of cars use dual bulbs per side, one projector for low and typically a reflector for high beam. So still possible.

Maybe it’s just cost. I know projectors can have the cutoff line easily changed for left or right hand drive countries as well as special shapes. You just change the stamped metal shield. These new 90° designs might require more extensive modification (more Cash to adjust the cutoff line.

Automotive OEM’s have not really embraced led headlights anyway. There are a few ultra expensive led designs out there (mainly prototype designs), but I bet projectors are just way cheaper.

Texas Ace Lumen Tube calibrated with Maukka lights

Click this to go to signature links.

Agro
Agro's picture
Offline
Last seen: 1 day 1 hour ago
Joined: 05/14/2017 - 11:16
Posts: 4537
Location: Ślōnsk

Those 90° optics are new to me. I’m not sure I understand what exactly is going on in them. Has anyone seen some writup or a schematic on them?

Nightsword
Offline
Last seen: 5 days 9 hours ago
Joined: 06/30/2017 - 13:22
Posts: 20

Many years ago I was following the first experimenter of this design on mtbr.com (who doesn’t get the deserved credit imo). He was using portions of a large reflector to piece together the concept, somewhat crude but ingenious and it worked. Soon after, Edelux came out with the first commercial light with a reflector created specifically for the design, followed by many others. The first Edelux angled the LED at 45-degrees. Over the years there were many improved generations of these lights with the current IQ-X using 90-degrees, in which a convex combined with a reflector utilizes the light more efficiently.

Edelux Design

IQ-X / IQ-XE (Google-Translated Version)
The convex lens and double “series” integrated Oslon LED seen in last photo of post. The OSLON Black Flat KW H2L531.TE

Agro
Agro's picture
Offline
Last seen: 1 day 1 hour ago
Joined: 05/14/2017 - 11:16
Posts: 4537
Location: Ślōnsk

Thank you, that’s interesting Smile . I’ve seen Endermann calling that off-axis parabolic reflector.
It clearly works well with headlamp LEDs which are often sold pre-mounted as several-in-a-row configurations. That seems to be what IQ-X is using.
Any more info of Lupine SL A projection optics?

ADDED: Though I still don’t understand that IQ-X convex lens.
Is it used together with the reflector?

Nightsword
Offline
Last seen: 5 days 9 hours ago
Joined: 06/30/2017 - 13:22
Posts: 20

See the Lupine SL here
The side view is clear it’s a convex lens that refracts all of the output from ceiling-mounted LEDs.

Yes, The IQ-X convex lens only refracts the forward-emitted output, and rearward-emitted output is reflected off the reflector.

The Philips Saferide was the first to use multiple LEDs early on. Here’s some good history.

Edit: Note that the IQ-XE consumes less than half the watts of the Lupine SL, but produces more lux at the cutoff line. I do very long night rides on bad roads, so in my case, the IQ-XE is much more proficient. The Lupine SL puts a lot of light to the sides, which is especially nice for off road. But my friends and I realized we like light only on the roadway directly ahead and the least amount of light spill, because spill light kills the ambience of the ride. Just a preference.

JasonWW
JasonWW's picture
Offline
Last seen: 22 min 5 sec ago
Joined: 10/22/2016 - 11:41
Posts: 11527
Location: Houston Texas

With aftermarket LEDs that take the place of a filament bulb in a reflector based housing, they put 3 or 4 chips next to each other to simulate the size and shape of the filament. The reflector is specially designed to focus light from that grain of rice size light source.

With the new 90* reflector designs, they are creating a string of light for other reasons. They can probably use any shape they want, but the thin string probably helps give a wide view with sharp cutoff. They are starting with a clean design as opposed to adapting to an existing reflector.

We don’t have strict bicycle headlight laws in the US so I never even looked into the market. Here people just strap any old flashlight to their bike and blind everyone. Lol.

We might be getting a bit too off topic now.

Texas Ace Lumen Tube calibrated with Maukka lights

Click this to go to signature links.

Nightsword
Offline
Last seen: 5 days 9 hours ago
Joined: 06/30/2017 - 13:22
Posts: 20

Double LEDs create horizontal off axis, which spreads the beam pattern horizontally. Bringing this back around, you can see the need this creates for drivers like the H1-A.

JasonWW
JasonWW's picture
Offline
Last seen: 22 min 5 sec ago
Joined: 10/22/2016 - 11:41
Posts: 11527
Location: Houston Texas

Nightsword wrote:
Double LEDs create horizontal off axis, which spreads the beam pattern horizontally. Bringing this back around, you can see the need this creates for drivers like the H1-A.

They just need to wire the leds in parallel.

Texas Ace Lumen Tube calibrated with Maukka lights

Click this to go to signature links.

slopegatri70
slopegatri70's picture
Offline
Last seen: 2 days 12 hours ago
Joined: 06/01/2018 - 14:21
Posts: 37

Not sure if it’s right thread to ask. but is there any way to know which resistor is responsible for output current in below driver -

this is a from a 13w motorcycle headlight, designed for osram black flat leds (looks like it). output current is 1.5amp. but i want to increase it to 2.5-3amp and replace led to whiteflat.

Nightsword
Offline
Last seen: 5 days 9 hours ago
Joined: 06/30/2017 - 13:22
Posts: 20

Sorry to skip over your question slopegatri70, but JasonWW the LEDs are series embedded into the Oslon package.

Schoki
Schoki's picture
Offline
Last seen: 2 days 6 hours ago
Joined: 03/27/2017 - 15:34
Posts: 215
Location: Germany

@slopegatri70:
LED_Low: R4 and R5
LED_High: R9 and R10

slopegatri70
slopegatri70's picture
Offline
Last seen: 2 days 12 hours ago
Joined: 06/01/2018 - 14:21
Posts: 37
Schoki wrote:
@slopegatri70: LED_Low: R4 and R5 LED_High: R9 and R10

Nice, but how did you figure that out ? and what should be replacing resistor value for 3amp output ?

thanks.

JasonWW
JasonWW's picture
Offline
Last seen: 22 min 5 sec ago
Joined: 10/22/2016 - 11:41
Posts: 11527
Location: Houston Texas

Nightsword wrote:
Sorry to skip over your question slopegatri70, but JasonWW the LEDs are series embedded into the Oslon package.

Yes, that’s why I said “they”. I was refering to Oslon. In the the way Cree released an all parallel wired 3 volt xhp50.2, Oslon needs to release a parallel wired version running at 3 volts.

Granted, it’s not a very helpful post because it’s completely out of our hands. Does Oslon make a parallel version of that led package? I have no idea what leds they make.

Texas Ace Lumen Tube calibrated with Maukka lights

Click this to go to signature links.

JasonWW
JasonWW's picture
Offline
Last seen: 22 min 5 sec ago
Joined: 10/22/2016 - 11:41
Posts: 11527
Location: Houston Texas
slopegatri70 wrote:
Schoki wrote:
@slopegatri70: LED_Low: R4 and R5 LED_High: R9 and R10

Nice, but how did you figure that out ? and what should be replacing resistor value for 3amp output ?

thanks.


I looked at the circuit and my guess were the same resistors. Look at the values. Those are the only ones with a really low value, 0.36 ohm. That is typical for a sense resistor. Plus the package size is about right.

The big ones labeled 2000 (200 ohm) are too high a value plus are probably related to the power input section (middle of the board). They are also right next to the main input wires. Since it’s from a motorcycle I assume it’s a Buck driver so you need the power input section to take the 11v-15v input and smooth it out, get rid of the spikes and ripple. The generator/alternator on a car or bike is creating AC current on 3 output wires (some small bikes might use 2 wires) which then goes to a rectifier bridge turning it into a chopped up DC voltage then to a voltage regulator. Plus voltage can vary with engine rpm. I’m guessing this led driver needs a cleaner, smoother input. I only see resistors and diodes so not much is happening here, just smoothing out as far as I can tell.

Then the input voltage gets sent to the driver side you need with High on one side and Low on the other. The 2 driver sides looks almost identical in design and components. Same inductors, same IC chips.

I guess input voltage still varies up and down and the drivers just compensate so they put out steady voltage and current to the leds.

This is my guess and I’m just learning these things. Hopefully Schoki or another engineer here can tell us for sure.

——

I don’t know if there is a way to calculate the results of a new sense resistor. You might just have to experiment. R9 and R10 look to be in parallel, same for R4 and R5. This means the pair are creating 0.180 ohms. If you replace one of the R360 for an R250 the total resistance goes down to 0.15 ohm. Try a small change like that with the new led and see what you get.

https://www.allaboutcircuits.com/tools/parallel-resistance-calculator/

Keep in mind that driver may not be designed to handle much more current than it does. It’s possible you could damage it. So don’t make big changes.

Better yet, listen to Shocki instead of me. Lol

Texas Ace Lumen Tube calibrated with Maukka lights

Click this to go to signature links.

Nightsword
Offline
Last seen: 5 days 9 hours ago
Joined: 06/30/2017 - 13:22
Posts: 20

AFAIK the Oslon was developed for “beam shaping”, I don’t think come in parallel. But the IQ-X embeds the Oslon so a parallel OSLON couldn’t be retrofitted anyhow. I looked into this long and hard because 3v drivers are so much more common. Also the Oslon is the highest lumen density of all LEDs, I do believe.

Microa
Offline
Last seen: 5 hours 17 min ago
Joined: 06/29/2011 - 21:20
Posts: 236

The manufacturer employed SK26A which is a 2A60V diode for this buck driver. I guess that the drive current is around 1A to 1.5A. You just replace the sense resisters to boost the current to 3A may not be possible. The ability of the inductor and the driving IC are also to take into consideration.

slopegatri70
slopegatri70's picture
Offline
Last seen: 2 days 12 hours ago
Joined: 06/01/2018 - 14:21
Posts: 37

driver ic (A6213) is capable upto 2.5/5a depending on ic version used ) no idea about inductor.

Buck Switch Current Limit Threshold ( min – typ – max )
A6213 – 3.0min – 4.0typ – 5.0 A max
A6213-1 – 1.9min – 2.2typ – 2.5 A max (i guess i should go for 2.5a max)

i have only R180 resistor to experiment right now. will see how it goes. btw i was wondering if output current goes up/down linearly if resistor value is halved/doubled respectively ? or only way to find out is by experimenting ?

anyway thanks for helping me.

JasonWW
JasonWW's picture
Offline
Last seen: 22 min 5 sec ago
Joined: 10/22/2016 - 11:41
Posts: 11527
Location: Houston Texas

slopegatri70 wrote:

BTW, I was wondering if output current goes up/down linearly if resistor value is halved/doubled respectively or only way to find out is by experimenting?

You definitely don’t want to take a big step like going half or double. I have only messed with sense resistors in linear and boost drivers, but not Buck. In just about all those instances slightly lowering the resistance saw a noticable increase in output.

This one driver had a 0.04 ohm sense total and put out 4.6A. Changing the value to 0.03 bumped up the current to 7A. So small resistor change equals big output change. It’s probably one of the reasons this driver uses two resistors in place of one bigger one. You can’t always get the exact value resistor that you need, so the manufacturer can use two different values combined to get just the right value.

Using one .360 and one .180 gets you .12 total. That’s a pretty big jump from .18
Maybe get a few .300 and .250 to go along with the .180 you have.

A .360 and .300 will get you .16
Two .300 will get you .15 (same as .360 and .250)
A .300 and .250 will get you .14
Two .250 will get you .13
A .360 and .180 will get you .12
A .300 and .180 will get you .11
A .250 and .180 will get you .10

Yeah, that should work. Buy some R300 and R250 in the correct package size and you can get whatever value you need.

Texas Ace Lumen Tube calibrated with Maukka lights

Click this to go to signature links.

slopegatri70
slopegatri70's picture
Offline
Last seen: 2 days 12 hours ago
Joined: 06/01/2018 - 14:21
Posts: 37

JasonWW wrote:
slopegatri70 wrote:

BTW, I was wondering if output current goes up/down linearly if resistor value is halved/doubled respectively or only way to find out is by experimenting?

You definitely don’t want to take a big step like going half or double. I have only messed with sense resistors in linear and boost drivers, but not Buck. In just about all those instances slightly lowering the resistance saw a noticable increase in output.

This one driver had a 0.04 ohm sense total and put out 4.6A. Changing the value to 0.03 bumped up the current to 7A. So small resistor change equals big output change. It’s probably one of the reasons this driver uses two resistors in place of one bigger one. You can’t always get the exact value resistor that you need, so the manufacturer can use two different values combined to get just the right value.

Using one .360 and one .180 gets you .12 total. That’s a pretty big jump from .18
Maybe get a few .300 and .250 to go along with the .180 you have.

A .360 and .300 will get you .16
Two .300 will get you .15 (same as .360 and .250)
A .300 and .250 will get you .14
Two .250 will get you .13
A .360 and .180 will get you .12
A .300 and .180 will get you .11
A .250 and .180 will get you .10

Yeah, that should work. Buy some R300 and R250 in the correct package size and you can get whatever value you need.

hmm, okay seems like doing that will not damage the driver either. will order some R300-R250s today. thanks.

Windforce
Offline
Last seen: 1 month 2 weeks ago
Joined: 03/09/2019 - 12:15
Posts: 13
Location: Poland

Have you already seen new Convoy’s boost driver? Looks like well designed driver.
https://www.aliexpress.com/item/4000046041955.html

  • 22mm diameter
  • 2.3A max. current
  • MP3431 boost IC
  • big-ass 1.5mH inductor
  • large MLCC bank on the ouptut

MP3431 is very powerful boost controller:

  • up to 16V output
  • up to 40W to the load
  • up to 13A input current (programmable)
  • in theory should provide 5A output current for 6V XHP50.2

Link to IC:
https://www.monolithicpower.com/en/mp3431.html

Looks like promise to even higher perfrmance than H2-A or H1-C. I’ve ordered some quantity and it will land on my test bench soon.

Barkuti
Barkuti's picture
Offline
Last seen: 36 min 23 sec ago
Joined: 02/19/2014 - 14:46
Posts: 4142
Location: Alhama de Murcia, Spain

slopegatri70 wrote:

hmm, okay seems like doing that will not damage the driver either. will order some R300-R250s today. thanks.

Considering the driver comes stock with R360s, swapping them with R300s nets you a 20% current increase. Using R250s means 44% increase. Its an inversely proportional thing. With R220s the increase is 63.63%, and +80% with R200s. You may want to also swap the smaller schottky diodes in such a case, and even add some extra output capacitance. Ask the experts LoL.

Suffice to say, at your own risk. Innocent 

JasonWW
JasonWW's picture
Offline
Last seen: 22 min 5 sec ago
Joined: 10/22/2016 - 11:41
Posts: 11527
Location: Houston Texas

Barkuti wrote:

Considering the driver comes stock with R360s, swapping them with R300s nets you a 20% current increase. Using R250s means 44% increase. Its an inversely proportional thing. With R220s the increase is 63.63%, and +80% with R200s.


So if he wants to go from 1.5A to 2.5A he needs about 65% increase. When you say “with R220s” you mean two in parallel which equals 0.11?
A .300 and a .180 would also work.

Hmm, just replacing one of the .360 with a .180 is pretty close at .12. That might work. Hopefully it won’t burn up. Lol

Texas Ace Lumen Tube calibrated with Maukka lights

Click this to go to signature links.

slopegatri70
slopegatri70's picture
Offline
Last seen: 2 days 12 hours ago
Joined: 06/01/2018 - 14:21
Posts: 37

well, it does seem like it’s inversely proportional to that resistor. i found this pdf about the driver ic which is used. and it indicates same thing i guess same should apply for my driver too ? does this apply for all buck drivers ?

link to pdf


apart from those schottky diodes, anything else i need to change ? along with resistor ofc. hope it doesn’t burn, headlight costs 98$ lol

JasonWW
JasonWW's picture
Offline
Last seen: 22 min 5 sec ago
Joined: 10/22/2016 - 11:41
Posts: 11527
Location: Houston Texas

On my motorcycle, I just swapped in an aftermarket led bulb that gives me a stock cutoff pattern, but noticably brighter. Easy peasy. Silly

Texas Ace Lumen Tube calibrated with Maukka lights

Click this to go to signature links.

Nightsword
Offline
Last seen: 5 days 9 hours ago
Joined: 06/30/2017 - 13:22
Posts: 20

Windforce wrote:
Have you already seen new Convoy’s boost driver? Looks like well designed driver.
https://www.aliexpress.com/item/4000046041955.html
  • 22mm diameter
  • 2.3A max. current
  • MP3431 boost IC
  • big-ass 1.5mH inductor
  • large MLCC bank on the ouptut

MP3431 is very powerful boost controller:

  • up to 16V output
  • up to 40W to the load
  • up to 13A input current (programmable)
  • in theory should provide 5A output current for 6V XHP50.2

Link to IC:
https://www.monolithicpower.com/en/mp3431.html

Looks like promise to even higher perfrmance than H2-A or H1-C. I’ve ordered some quantity and it will land on my test bench soon.

Wow that MP3431 datasheet shows 97% efficiency @1A which is around the current I need. Just need a driver with more evenly-spaced or configurable modes, and a ~3V cutoff voltage.

EDIT: Appears to be PWM not constant current Facepalm

JasonWW
JasonWW's picture
Offline
Last seen: 22 min 5 sec ago
Joined: 10/22/2016 - 11:41
Posts: 11527
Location: Houston Texas

Nightsword wrote:

Wow that MP3431 datasheet shows 97% efficiency @1A which is around the current I need. Just need a driver with more evenly-spaced or configurable modes, and a ~3V cutoff voltage.

EDIT: Appears to be PWM not constant current Facepalm


I thought that driver was 12v out only, not 6v.

Have you seen Loneoceans boost driver? He does not sell them, he just designed it and shows others how to assemble the parts.

Texas Ace Lumen Tube calibrated with Maukka lights

Click this to go to signature links.

Pages