Sofirn IF25A 4000k increased driver consumption in off mode

Slava20 · November 19, 2025, 5:45am

Hello, my name is Slava and I’m from Kharkov, Ukraine. I’ve encountered a problem. One of my two IF25A 4000k flashlights (I bought it used) drains its battery very quickly in standby mode. About a month. It has a new Sofirn 5000mAh battery, and its capacity is even higher than stated. I checked the firmware versions of my two IF25A flashlights, they are the same - 0614. I reset both flashlights to factory settings. And I measured the readings of both flashlights in idle mode with a multimeter: 1) With the light off in the button 2) With the minimum light in the button 3) With the maximum light in the button. The results are as follows. On a working IF25A flashlight: 6500k: 0.01mA, 0.26mA, 4.04mA. The faulty IF25A has 4000kA, 4.55mA, 4.78mA, and 8.66mA.

For example, I’ll also provide measurements from other Sofirn flashlights: SR20 Mini 0.05
SP10pro 0.06
SC31Pro (1) 0.13, light in button at minimum
SC31Pro (2) 0.14, light in button at minimum

I emailed Sofirn on Ali, but got no response.

I also emailed Sofirn on the official website. They told me the model is old, long out of production, and there are no spare parts, such as a driver, and they won’t be available. Buy other, newer models.

Can anyone help me solve my problem? What part of the driver should I replace to reduce the flashlight’s power consumption when idle? Or does anyone have a spare driver, perhaps one left over from customizing an IF25A? And this person could have sent me this unnecessary driver.

Below are photos of my measurements and a photo of the IF25A driver found online.

As I understand it, the first versions with the first Anduril may have had slightly different details. In the same advertisement for the IF25A with a green button, Sofirn claimed 40 days of battery life in moonlight mode. A user measured the battery life with a 4000mAh battery and got 83 days, assuming the user was right. Later, these flashlights were added a power bank function, and their consumption may have increased. They also added a blinking blue LED. For example, my working IF25A with the minimum button light in moonlight mode consumes 5.62mA. That means it will last 29 days with a 4000mAh battery. If you use a 5000mAh battery, it will last 37 days at best.

Quadrupel · November 19, 2025, 10:29am

Welcome.

Who knows? MAybe some capacitor short circuited .

raccoon · November 19, 2025, 11:55am

Have a nice time here, Slava20!

Henk4U2 · November 19, 2025, 4:29pm

My m.o. is: when not in use (on the move or in my hand) I always give my lights a hard lockout. If possible, of course.
A soft(ware simulated) lockout is ALWAYS accompanied by the usage of energy. Sometimes more, sometimes less.

Slava20 · November 19, 2025, 8:22pm

That’s true, but the same SC31 Pro will last about 2.5 years with the button light on and a 3000mAh battery. And with a 4000mAh battery, it’ll last about 3.5 years. In theory, if the IF25A had the same power consumption as the SC31Pro, it would have discharged from a 6000mAh battery for about 5 years with minimal button light.

Slava20 · November 21, 2025, 9:35pm

I measured the quiescent current on all my Sofirn flashlights. On some, it’s negligible, on others, it’s high. So, I wondered if this is within normal limits for the Sofirn? Here’s the list:

SP33 v2 0.01mA 5500mAh will discharge in 62 years
LT1 0.29mA min light in button 4.5 years
SP40 Type-C 1.7mA 3000mAh 73 days
C8F 2.05mA 4000mAh 81 days
C8G 0.03mA 4000mAh 15 years
SP70 0.74mA 2x5500mAh 1.5 years
SD003 0.05mA 3000mAh 6.5 years
SD006 0.03mA 5000mAh 19 years
SR20 Mini 0.05mA 1100mAh 2.5 years
SP10pro 0.06mA 900mAh 1.7 years
SC31Pro (1) 0.13mA min light in button 3000mAh 2.6 years
SC31Pro (2) 0.14mA min light in button 3000mAh 2.4 years
IF25A 6500k 0.26mA min light in button 4000mAh 1.7 years
IF25A 4000k 4.78mA min light in button 4000mAh 34 days

I was surprised by how much power even diving flashlights with magnetic buttons consume. I was also surprised by some models, such as the C8F (2.05mA) and the SP40 (1.7mA). However, the SP40 has a total runtime of about two hours on low and medium modes. And that’s the power consumption out of the box.

TIFisher · November 22, 2025, 10:45am

This is excellent data. Thank you for taking the time to collate and share your observations. With as many Sofirn lights in circulation as there are, this information may be a great baseline reference for Sofirn collectors. I have many of the models you list and find this type of use case info of immense value. The IF25A (V1) is my favorite Sofirn torch, and would agree your high PD sample is defective. Hope you find a resolution.

Welcome, and thanks, again!

INeedMoreLumens · November 22, 2025, 7:20pm

There’s many things that could cause parasitic drain.
If the driver is a FET+1 as I suspect it is.
I would remove the 7135 chip and check it again.
If the parasitic drain is greatly reduced I would replace the 7135 chip.

Slava20 · November 22, 2025, 8:36pm

I removed the driver today and didn’t see any burn marks. I read on a forum about testing a 7135 chip: "That leak could be that the internal FET failed short-circuit in the drain-source path => always “on”.

Measure the resistance from pin 1 to pin 2 to verify; normally it should be open circuit (no continuity), but failure may read a few dozen ohms to 1 ohm or less."
I didn’t desolder the chip, but I think I measured 282 ohms with a multimeter. I measured the resistance on the left and middle pins. As far as I can tell, the left pin is output, and the middle pin is ground. So, is my chip faulty? I’ll email a specialist tomorrow who was recommended to me and try to resolve the issue with him. I’ve read that these chips are often defective. They are also prone to failure if overheated during soldering. I don’t have a hair dryer or a soldering iron with a thin tip. And the only measuring instrument I have is a multimeter. If I can resolve this issue with the repairman, I’ll definitely let you know.

INeedMoreLumens · November 22, 2025, 9:06pm

I’ve never tried the test one with a multimeter so I’m not sure what the measurements should be.
I ordered 150 7135 chips a while back so if I’m in doubt, I swap it out.

INeedMoreLumens · November 22, 2025, 9:13pm

I use a SK Tip to remove these style chips:

Rox · November 23, 2025, 9:39am

If you’re measuring the resistance between output and ground whilst in-circuit, you’re measuring the resistance of the LED and any current limiting resistors or any other components in that side of the circuit so it’s probably not really a measurement that’s entirely useful.

Slava20 · November 23, 2025, 4:08pm

I contacted the repairman. Let’s consider this a polite refusal. He said: “the failure was unusual and required a complete driver replacement. And anyway, this isn’t a failure at all, just normal. And you should always remove the battery from the flashlight. If you do undertake such a repair, you need to consider the flashlight’s cost, which will start at $35. And such chips are currently unavailable because they’re in short supply.”

For example, I bought a new IF25A 6500k with a battery and shipping for $33 a few months ago. I bought my broken IF25A 4000k used and without a battery a few months ago for $23. You can now purchase the 4000k version without a battery and shipping on our websites for $44-$48. The 6500k version with a battery and shipping costs $35. 7135 chips cost between $0.20 and $0.60. Shipping is about $1.20.

So I’m thinking about playing with microchips, or while I can, buying myself another new IF25A 6500k and simply re-soldering the diodes from my broken 4000k version.
And then installing the diodes from the 6500k in the broken one. Then it’ll act as a backup, like a backup flashlight, but with bugs.

And there’s another thing: if you send a flashlight for repair, anything can happen. Because there’s military action going on in the country, it often arrives through the post office. And the repairman himself is here today, and he’s moved tomorrow. Or, like I recently did, I contacted the seller about buying some aquarium fish, and he told me he’d been drafted into the army. So you could end up without a flashlight driver at all.

Slava20 · November 24, 2025, 10:54am

I remembered my wife has a soldering iron with a thin tip; she uses it for working with fabric. Experience has shown that my old soldering iron with a wide tip desolders the chip better, precisely because the wide tip fits all three contacts on the chip.But perhaps a soldering iron with a thin tip would be more convenient when I solder the microcircuit. I measured the resistance of the desoldered chip. Between pins 1 and 2, it was 321 ohms. Between pins 2 and 3, there was no reading. Between pins 1 and 3, there was no reading either. So, to summarize, is my chip faulty? Should I order a new one?

Quadrupel · November 24, 2025, 1:08pm

Removed 7135? OK, now measure off mode power consumption . Or maybe charging circuit consumes ? Usually when 7135 leaking you get constant moonlight in off mode.

Slava20 · November 24, 2025, 1:25pm

No, the moonlight wasn’t on. That means the chip was working. I measured the bottom of the board with the chip unsoldered, the consumption is 4.4 mA. In the photo, the chip is not yet unsoldered.

Slava20 · November 24, 2025, 1:59pm

Here’s another photo

Quadrupel · November 24, 2025, 2:16pm

Disconnect charging board, if no, check caps… a lot of fun.. Or maybe some Convoy driver will fit in?

Slava20 · November 24, 2025, 4:36pm

I actually measured it without the top part of the board, where the USB port is. Could the SMD S4 diode cause this kind of leakage?

INeedMoreLumens · November 24, 2025, 4:55pm

I don’t think so, that should just be for Reverse Polarity Protection for the MCU(T1616)