Convoy SST40 drivers now coming with R020 sense resistors?

Hello people,

Just recently I messaged Simon because I noticed the (∅17mm and ∅20mm afaik) Convoy SST40 drivers are now coming with different sense resistors: R020 instead of R010.

If this is so, the maximum current output is either halved or the current sense voltage is now twice the value, downgrading the driver, with a smaller regulation window and twice the amount of heat being dissipated in the sense resistor, which by the way is now coming with a smaller size, which means the power specification of the sense resistor is surpassed. :!!!:

I hope it is not the latter case, as in my lightful opinion sense voltage must be kept as small as possible.

In this ∅20mm 4-mode ad you can see this picture of a customer:

I have an unused 12-group driver of this type (bought months ago) which came with an R020, although I added an R050 on top of it. But it still is not tested.

Now that this stuff is happening I am going to make the build, and feed it with my regulated power supply so I know what is going on.

In the meantime, if some technically knowledgeable you want to provide some feedback (people who knows how to ensure a proper current measurement), it could be of help.

Regards.

Man, I really wish I could understand WTF you’re talking about. These kinds of post hammer home just how little I know about flashlight internals. *sigh

Which size is it ? 0805 or 1206 ? 1W 1206 are common, some 0805 are rated for 750mW but it’s rarer.

Why the change ?
Maybe the op amp was changed for a less precise one, requiring a higher sense voltage.
Maybe they increased the resolution for a lower minimin (10 bit ?) requiring again a higher Vsense if the opamp precision is the limiting factor.

If you find the voltage divider that set Vsense you can try changing it back to 60mV Vsense and 10mΩ Rsense.

Current clamp, or wire the driver outside of the flashlight and put a voltmeter across the sense resistor.

It's a 1206.

Nice information. However, if Vsense is higher @#$% off, I already changed the sense resistor in a driver and I simply do not want to spend more of my time with shenanigans like this. :facepalm:

Yes this is something I know, although when I wrote this I was thinking in people who had a more or less recent flashlight with one of these drivers and a current clamp, to get some feedback.

Now, as I said above I can install the modified driver in the pill and feed it with my power supply with increasing voltage. If the current continues increasing above what it should be according to my math for 60mV sense, it will mean Vsense was changed. :facepalm:

It is not the first time I see things like this with products manufactured in China, they make a good first version and after it is reviewed by the public, they downgrade the stuff in some way just to save a veery small amount of money. But the product is made worse, and the customer is in some way deceived, which is bad. @#$% off with this, honestly.

My rencent drivers from Convoy were using R020 (4 modes nor 12 groups). I thought that was their newer productions.

I still don’t know what other parts might change as it can produce above the rated current; 5.65A max (17mm), and 6.4 max (20mm & 22mm). All tested with 6500K Sst40 P2 bin.

Recently I mod an S2+ with 3v Xhp50.2 K2-2A 17mm 4 modes driver (swapped R020 with R010) and it only produced 10.35A max without cooling system.

Also an S2+ with Sft40 and 17mm 4 modes driver with stacked R050 on top of the R020 produced 7.47 max.

Battery for the test: Vtc5a.

Let's make some math: Vsense = I × Rsense

  • ∅17mm Vsense = 5.65A × 0.02Ω = 0.113V = 113mV
  • ∅20/22mm Vsense = 6.4A × 0.02Ω = 0.128V = 128mV

Previous Vsense values were 50mV (∅17mm) and 60mV (∅20/22mm). Now they seem to be above twice the old figures.

I do not mean to be judgmental, although I can't help but see this in a disappointing way. I often modify my drivers, and now I see that one of my latest builds (which I gave to a brother) ended up without a reasonable current limit, as I swapped the sense resistor to an R010. Not really a problem because that S2+ uses an LH351D and a low current cell, but… :(

Now, this is an extract of a conversation I had with Simon on Jan 23th:

The only answer I received from him was a :-) emoji. Of course, I should have inquired him further instead of believing that somehow he felt I was going to use in the same way as my faulty unit (old 12-groups firmware with memory bug) as I relate up there, customizing it for me… :facepalm: (I wanted such driver at 2.5A maximum for a tiny flashlight).

You can't expect the current flow to always max out with this kind of driver, the driver is just an element in a team; you need sufficient input voltage for this to happen. In a flashlight there are many elements which need to be considered, as they drop voltage too: switch, springs, contacts, cell. And the larger the current flow is, the more voltage drops in these elements. The load or led emitter also drops more voltage as the current flow into it increases. All in all, in my opinion getting 10.35A into an XHP50.2 is pretty fine.

These linear drivers behave like regulated variable loads, their purpose is to limit the output current to the load: the led. They sense the current flow in their sense resistors, and use their onboard MOSFETs as variable resistors to limit the current… if needed. If the driver senses the current does not reach the limit, the MOSFETs are kept unthrottled. The voltage drop at the sense resistor is how the drivers sense the current: the larger the voltage drop, the larger the current flow. Using a larger sense resistor and larger sense voltage just means they “cheaped” on the thing. The worst part about this, in my opinion, is still showing advertisement pictures of a product which is not like it used to be, namely for modders.

In some way I feel deceived with all of this. :(

I raised this with Simon 4 weeks ago as all my C8+ NM1 stock is with a R020. I measured the output current to be 5.7A which is too high

His response was that it’s designed for the SST40 and using different LEDs will change the current depending on the forward voltage.

That didn’t answer my question at all, but it’s sometimes hard to get anywhere. When they’re advertised as 5A, and also used with LEDs with a max allowable current of 5A, it’s extremely important that it’s not over driving the LED.

Simon said it’s not an issue and that he’s tried the Osram NM1 with his 6A drivers.

By the way, using the older version of the ∅17mm 4-mode driver, with 50mV sense, would have allowed you to attain a slightly larger maximum current. To attain this you would have needed to swap the stock R010 sense resistor with an R003 (for a max of 50mV / 3mΩ = 16.6̅A, or set in parallel an R005 over the stock R010 for a max of 15A. The main reason for this is the lower sense voltage of the old driver, making the sense stage more efficient and dropping less voltage in the driver. Remember the calculated 113mV in my above post, and substract 50mV in the old driver: ≈63mV. These 63mV would then have been dropped in the other parts of the flashlight, allowing a slightly higher current to the led emitter.

Thank you so much for the knowledges and idea, Barkuti! I might do that tonight and post the results here.

That’s too high for me too, but my NM1 in a shorty S2+ setup with “newer 4 mode driver” and 8x3mm brass for spring replacement only pulls 4.46A with 18350 YDL 1000mah. I will try with the Vtc5a what it pulls.

I asked another member on here who sells them and he measured only 4.8A tail, 3.8A at the LED, so maybe it’s not an issue.

I got 5.7A with my clamp meter and 5.4A with my dmm

I understand. The older version, at 5A, was perfect for the NM1. Now it can… slightly overdrive the emitter. At least the CSLNM1.TG, with 3030 footprint. Probably not a problem in practice, but the leds may go absolutely pedal to the metal.

That's incorrect. Besides the max current they are set for, which is selected for a kind of emitter, drivers are not really designed for a specific kind of emitter, this is wrong. Drivers can drive all sorts of emitters (by adjusting their maximum driving current, that is). The only difference between different emitters is their Vf (or Vf to current curves, to be more precise). But this does not matter because drivers are regulated, and so if the led emitter has very low Vf these drivers will throttle the current for it to stick to the required value (“5A” -cough- for the highest mode, and and percentage of this for the lower modes). Conversely, if the led has a high Vf the driver will probably sense the current flow does not reach the limit (because of insuficcient input to output voltage difference) and correspondingly keep the MOSFETs unthrottled. All in all, this is what regulation is for in these drivers: to limit the maximum current to the load (led or leds in parallel).

Of course the NM1 is not going to blow up with 6A or slightly more (I think this is what he means when he says “not an issue”, that it doesn't blows up LoL), but such level of current is certainly beyond ideal.

Concerning Simon, this smells of self-sabotage to me.

Not that fast, PeyoX. The sense resistor values I calculated before were for the older version driver, with 50mV sense voltage. Now, with ≈113mV in the newer driver you hit 10.35A max. With the driver modification you did (an R010) the driver will start throttling at ≈11.3A (twice your previously measured current with the stock R020). But you only hit 10.35A, which is under the current ≈11.3A limit. This means you will not go any higher just with less sense resistance in the driver.

These figures are dubious, wrong. Using a linear driver there's no way you can measure 4.8A at the tail, but only 3.8A to the led. Nonsense. Both tailcap and led current must be equal. Only with switching drivers (buck, boost and boost-buck) the average current at the tail is different that the current at the led emitter.

Barkuti: I forgot to tell you that I will try with the old version driver, as it may produce more Amps :smiley:

Added: the old version 4 modes driver with R005 stacked on top of the R010 pulls 12.38A on my dmm. Another giggle with Convoy “ramping” 3 modes driver by stacking another R005 pulls 15.24A on my dmm. They were tested with the same Xhp50.2 K2-2A.

Mine only pulls 4.72A at the tail with Vtc5a.

Hmm, very strange because I checked mine a few times and same result. I’m not going to worry with the stock I have, I sell them with a 12 month warranty anyway.

Come to think of it this driver is from older stock, so perhaps something has been changed on yours

Update: all new stock is still 5.5A or above

Okay, tested my L21B with the latest 5A driver and it’s only hitting 5.20A with dual springs and a Samsung 30T.

Obviously something has been altered. I made a demo video of the Osram C8+ vs Osram L21B and you can see that the old driver in the C8+ is turning the LED blue.

Glad I tested that as I was confused how the L21B looked much better at the same current.

Anyone have a link for the correct R010 so I use use the 15pcs drivers I have please

Now it makes sense why the C8+ was getting 820 lumens and the L21B only 760

Hope there will be no problem at all. Keep in mind that mine using 26Awg wire, not the stock 22Awg(?).

I took a look at the driver with one of the pics on the store :


(Not sure if C3 is wired like that but it should be)

So if you want to change Vsense back to 60mV for a 6A driver (10mΩ Rsense) you would change R2, top resistance of the voltage divider is R2+R1, bottom is R3, R1 and R3 shouldn’t be changed because it will affect the 1st and 2nd RC filter. (Convoy likes to use 2 RC filters, even though it’s not really necessary).

R2 = ((Vreg - Vsense) x R3 / Vsense) - R1

Vreg being the LDO voltage = PWM voltage (measured at 100% output).

Rsense was 1210 size, which was overkill anyway.

2 Thanks