33mm, 25A, 35V, adjustable LED driver - fits D-cell tubes! - thermally protected

[DIWdiver]

This is the sales thread for the 25A driver I've been building, which has been discussed here:

http://www.candlepowerforums.com/vb/showthread.php?393212-Feeler-thread-0-25A-adjustable-linear-driver

For those who haven't been following, this driver has high/low/off modes controlled not by a pushbutton or clicky, but by a toggle switch. By connecting the H terminal to C, you get high, and by connecting L to C you get low. Disconnect both and you have OFF. The driver draws about 12 mA in off mode. This could be eliminated by using a double-pole switch.

The high and low settings can be reduced by adding a series resistor in the appropriate control line. This resistor does not carry the LED current, so it can be very small, or even a potentiometer (I've found that 50K linear taper works best, and gives zero to full scale output control. For those who can solder small surface mount components, you can change the settings by changing on-board resistors.

The maximum output can be adjusted using an on-board potentiometer. This pot should not be used frequently for dimming, as it is designed for 100 operations.

The driver is available in a 'single-cell' version, with input voltage from 3.0-5.5V, and a 'multi-cell' version, with input voltage from 3.5-35V. The single-cell is intended for lights using a single LiIon cell, while the multi-cell version is for use with 2-8 LiIon cells or 4-16 NiMH cells. Either voltage is available with 0-10A, 0-16.7A, or 0-25A adjustable output.

PLEASE NOTE: This is a linear driver, not a switchmode driver. What that means is that if the input voltage is much higher than the LED voltage, it will generate a LOT of heat. In most applications it will need to be attached to a large heatsink to prevent overheating. Most handheld lights will overheat if driven at high currents for more than short periods. The amount of power dissipated in the driver can be calculated using the following equation:

P = (Vin-Vled) * Iled.

The efficiency of this driver depends entirely on the choice of battery and LED, and the wiring configuration. In the basic configuration, the efficiency can be calculated using the following equation:

Eff = Vled/Vbat

The price is $35.00 each plus postage, prepaid. I use USPS flat-rate box at $5.80 within the US. Most international shipments are a flat $20.00 fee. PM me before ordering if you want it shipped outside the US. You may be liable for import duty or other fees, which can be substantial. Please contact your customs office if you are concerned.

I recommend 4-C shielded w/drain cable for remote location of mode switches, pots, etc. Since it's hard to buy in small quantities, I am stocking that too. Price is 1-9 feet, $0.50/ft. 10-49 feet, $0.40/ft. 50+ feet, $0.30/ft.

Payment is paypal (dlburdette 'at' juno 'dot' com). If you want to pay some other way, let me know. Please put your username in the Paypal notes so I can match up posts with payments. Thanks!

Since there are six different versions, these are usually made to order. I usually ship within 2-3 days of receiving your payment. Please make arrangements with me prior to ordering if you need a specific delivery schedule.

I currently have in stock:

Single-cell:
​

10A - qty 1​

16.7A - qty 1
25A - qty 1​

Multi-cell:​

10A - qty 1
16.7A - qty 1​

25A - qty 1

​

I can currently make at least 5 of any version in a few days.

 

[gofastman]

Any chance the 25A version could be made on a ≤26mm board for c cell Maglites?

 

[DIWdiver]

No.

Can you even get 25A from 26xxx cells?

 

[Epsilon]

Yes this is possible for IMR cells.

I have Sony 26650VT cells which can do 25A. The problem is the voltage at which they do, this might be a problem for single cell and single white LED use.

But, if you want to power a high power Red LED (Luminus 120 series) with low Vf of 2.2v, this might work. However I do think a single 26650 will drop below 3.0v quick at 25A.

Edit:
http://lygte-info.dk/review/batteries2012/Sony SE US26650VT 2600mAh (Green) UK.html

The Sony cells can handle 20A load at above 3.0v to pretty much the full capacity and 3.2v at 75%. So that might work .

 

[Hoop]

Compelling. You released it after all. I will have to come up with something to utilize this... oh yes, gutting the SR-90.

In the D mag it is possible to run the 26650's in parallel so 25A would be no big deal in that case.

 

[DIWdiver]

Yes this is possible for IMR cells.

I have Sony 26650VT cells which can do 25A. The problem is the voltage at which they do, this might be a problem for single cell and single white LED use.

But, if you want to power a high power Red LED (Luminus 120 series) with low Vf of 2.2v, this might work. However I do think a single 26650 will drop below 3.0v quick at 25A.

Edit:
http://lygte-info.dk/review/batteries2012/Sony SE US26650VT 2600mAh (Green) UK.html

The Sony cells can handle 20A load at above 3.0v to pretty much the full capacity and 3.2v at 75%. So that might work .

With white LEDs, you'd not have enough voltage from IMR cells to run to 25A, or probably even 20. But for red, you wouldn't have any problem running 25A until the cell is 3.0V or less.

 

[DIWdiver]

Compelling. You released it after all. I will have to come up with something to utilize this... oh yes, gutting the SR-90.

In the D mag it is possible to run the 26650's in parallel so 25A would be no big deal in that case.

Yeah, I figure I only need to sell about 20 to break even. And it gives me something to keep busy with in the evenings. Worst case I had some fun and I'm out a few bucks.

I wasn't seeing much interest after I opened the sales thread, so I started looking at building a light to see if I could spark some activity. Boy was I surprised when I started pricing the SBT-140! The driver is cheap in comparison.

 

[Hoop]

Yes they are not inexpensive. The Luminus chips with the round die seem to be discontinued/obselete products now on Mouser.

So is this driver capable of 875 watts continuous?

How many watts of heat can the regulator dissipate if heatsinked well?

 

[gofastman]

No.

Can you even get 25A from 26xxx cells?

That's what I guessed. Not that it matters but I was planning on 4 C NiMh.

 

[DIWdiver]

Okay, the first two drivers are packed and ready to ship to Epsilon in the morning.

I have two of each model built up and tested, in stock, and it only takes a few days to build more. Anyone interested?

 

[DIWdiver]

Yes they are not inexpensive. The Luminus chips with the round die seem to be discontinued/obselete products now on Mouser.

So is this driver capable of 875 watts continuous?

How many watts of heat can the regulator dissipate if heatsinked well?

Sorry Hoop, missed your questions somehow.

Yes, 35V at 25A would be 875 watts input, and that is the absolute max. Reaching that would take some doing, and the output would be less than that in any case.

"Heatsinked well" is pretty open to interpretation, so I'll assume some reasonable values and discuss two cases.

Case 1: Dive light, water cooled, or you've got some extreme cooler:
Figure the temperature of the heatsink at the interface to the FET is about 35C. You've got some high-tech thermal interface material that gives 0.15 C/W case-to-heatsink. A bit of cooling on the face of the driver (cool ambient air), so the thermal protection won't cut in until the FET die temp reaches 150C, which is well with the FET's capability.
Thermal resistance, junction to case (Rjc): 0.44 C/W
Thermal resistance, case to heatsink (Rch): 0.15 C/W
Temperature differential, junction to heatsink (dT): 115C
Power handling = dT/(Rjc+Rch) = 115/(0.44 + 0.15) = 195W

Case 2: some big air-cooled heatsink with fan:
Guess that the heatsink thermal resistance to ambient is around 1.5 C/W, ambient at 25C.
Use the thermal interface material shipped with the driver, that gives about 1.0 C/W case-to-heatsink.
Little cooling of the driver face, so the thermal protection may cut in around 100C die temp.
Thermal resistance, junction to case (Rjc): 0.44 C/W
Thermal resistance, case to heatsink (Rch): 1.0 C/W
Thermal resistance, heatsink to ambient (Rha): 1.5 C/W
Temperature differential, junction to heatsink (dT): 75C
Power handling = dT/(Rjc+Rch + Rha) = 75/(0.44 + 1.0 + 1.5) = 39W

As you can see, there are several things that need to be known or assumed before the calculations can be made, and the results can vary widely. That's why I try to be very careful to make sure people understand the implications of using a linear driver at high power levels.

The Luminus CBT-140 seems to be available from numerous suppliers, and has a round, flat emitter area (no dome). Unfortunately, it's around $100 each. Not really for the faint of heart.

D

 

[Hoop]

Luxury does have its price. Thank you for laying out the thermal scenarios.

 

[TRDmatrix]

Hello DIWdiver, I am planning on building some offroad lights using cut down D cell maglites as housing. The lights will be powered off of vehicle power at approximately 14 volts running and 11 volts although the light probably wouldn't be used with the engine off. I am looking for a driver to run 9 XP-Ls per housing. Just at a quick glance, this driver seems to be the perfect choice. My question is would this be the better choice over your dedicated automotive driver for what I am planning?

 

[DIWdiver]

If you are planning on mounting the driver in the light head then the automotive driver is a no-go, and you don't really have a choice.

But if you mount the driver in the light head, you are likely to have serious heat issues. By the time you consider optics and front plate (flat lens), you are probably looking at no more than 30% radiometric efficiency, meaning that of the 30W you put into the LEDs, only 9W makes it out of the head as light. What you are putting into the head is 3A at 14V, or 42W. Of that only 9W escapes as light and the rest is converted to heat. Thus you'll have about 33W to dissipate as heat. You need big fins and airflow to do that and keep the LEDs cool.

If you take the driver out of the light head, 12W of heat goes with it, reducing the heat in the head to 21W, which makes cooling somewhat less challenging. Also, you could then run all the lights on only one driver and use the extra $64-70 you didn't spend on drivers to get a nice heatsink for the driver. If you do this I would recommend putting a thermistor in one of the light heads to enable separate thermal protection for the LEDs and the driver. Try to pick the head that would have the worst cooling to provide the best protection.

In general, I think the dedicated automotive driver is better in automotive applications, not only because it is cheaper, but also because it can handle a few really rare automotive-only events that this one can't. In the unlikely scenario you experience one of these events, the automotive driver would completely protect you while this one could easily be destroyed, likely taking out your LEDs with it.

This one is only better if you need over 15A, have less than 8V input, want undervoltage cutout instead of overvoltage cutout, or need to fit it in a D-cell tube.

 

[Epsilon]

And received!

That FET is huge btw

 

[DIWdiver]

Shameless BUMP! Epsilon, if you're still there, did you get your lights built?

I just got an inquiry for a driver. Boy have sales been slow!

 

[las3r]

Do u have any other divers for sale still ? Pm me please

 

[Epsilon]

Shameless BUMP! Epsilon, if you're still there, did you get your lights built?

I just got an inquiry for a driver. Boy have sales been slow!

Not been here for years actually, and I am going to clean out some parts of project that I will not finish.
Was quickly browsing this part of the forums, and saw this topic.

Your driver, CBT140, maglite, massive heatsink and fivemega head are parts that I will keep to finish the project actually.
Still doubting about the way to power it. Probably I will make a 4x NiMH Sub-C pack.

Not sure when it will be finished though, priorities lie somewhere else.

 

[id30209]

PP sent as per PM

 

[DIWdiver]

PP received, thanks!

I will try to get it ready tomorrow, to ship Friday morning.