Have I gotten this all wrong or did I just get a bad batch of AMC7135 drivers?!

[Korxi]

Hi there!

I've just bought a set of 10 AMC7135 1400mA drivers from DX, and tried making a sandwich of two, driving 5 SSC p4's i parallel of two 18650's also in parallel.. After hooking up everything I only got 1000mA through the circuit.. I double checked the wiring comparing to other threads in here and couldn't see that I did anything wrong.
Now I've tried testing 5 of the 10 boards and they all come out from 500-700mA - so now I'm thinking if I'm the idiot or I just got a faulty batch of drivers!? Btw. when I hooked up a single P4 to one of them, it only drew 100mA - I do not get it...

I attached pics of the wiring I did when testing the indivual drivers.
I put in jumper wires instead of the diodes and hooked up the battery on the back. I tested the current through the black and red wire on the front.

Hope you can help!

Christian

 

[Justin Case]

Make sure that all of the AMC7135 chips are connected in parallel. Use a DMM to check continuity between all of the Vdd legs. Some of these boards require that the end user put in some solder bridges to enable all of the chips.

 

[VidPro]

you did not indicate your input voltage?
because the chip adds in Some resistance, and only CLIPS when the voltage input is ABOVE the LED voltage.
so what your seeing sounds correct, depending on the voltage of the battery, and various resistances to get to your readings.

because the chip does not do any actual boosting of any sort, it is only "regulating" when the voltage is too high. the rest of the time it drops back to Direct drive with the slight resistance of the chip item some .12 ohms or something.

easier to check this stuff with a voltage controllable power supply as your initial source (instead of battery). that way you can observe the effects of various voltages applied to the input. (but dont go to high in voltage).
then you "simulate" various battery charge levels with the power supply, and have a good idea of what is going on, where your losses would be, and how the item works.

to stay "regulated" with this type of chip item, you feed in more voltage than the LED needs and the chip "current controls" by resistive heat burn off, when that voltage is to high FOR that current. Clipping my word for it. because it is resistive heat burnoff on the part, the input voltage should not go TOO high, or the part wont be able to toss out all your power and the part will overheat.

This actually works out rather well, as the battery does not get pulverised for more current as its voltage output is diminishing, instead the output declines , and then the battery lasts longer and longer. The only ramification is when the battery is low, the output is low, the advantage is you almost literally never run out of power. The other advantage is the battery does not get depleated to its very low point, and the "efficency" light output vrses power consumption is pretty high, as there is minimal losses untill you raise the input voltage, and then just Burn it back off again because the voltage is to high. (depending of course on the voltage of the input battery)


this type of clipping is good when the voltage of the batttery would go to high and burn out the led itself, but after that short time (with 4x high power dies and single li-ion) your quickly into DD with slight resistance instead of being "regulated. so some people dont like it because its not Full Bright ALL the time, i love it because if i want Full brightness then i have the battery fully charged. When you go out, your eyes being to slowly adapats to the darkness (what little there is with a Quad die) and you dont even notice the diminishing output, but you do notice that 10 HOURS later it is still running, and your battery is being treated nicely.

also there is 8X versions of that chip so you dont have to haump them together.

 

[Justin Case]

As I wrote before, first check that all of the AMC7135 chips are actually connected in parallel. You need to check two locations -- the brass-colored pad near the labels Q3 and D2 and the brass colored pad on the other end nearest Q1 and D1. Make sure the Vdd leg of the AMC7135 at Q3 is electrically connected to D2 (or what is now your wire that replaces diode D2). Similarly, make sure that the Vdd leg of the AMC7135 at Q1 is electrically connected to what had been D1. Use a multimeter to check for continuity.

Once you establish that you really do have 4xAMC7135 functioning, then you need to make sure that you satisfy the criteria to run in regulation. First, Vbatt must be greater than Vf + AMC7135 drop (=0.12V) for the board to run in regulation. At about 560mA per P4 (you have five P4s in parallel, driven with a total of 2.8A nominal, or 560mA per LED), the Vf for a P4 is probably quite low, let's say about 3.2V (I've measured several P4s at less than 3.4V@1000mA), so Vbatt > ~3.3V. That should be achievable using two 18650s in parallel.

Now you also have a fairly low Vf setup, so the diodes on the AMC board might matter. With the AMC7135 Vdd pin fed through a diode, the second requirement for regulation is that Vf also has to be greater than 2.7V (Vdd min for an AMC7135) + diode drop (~0.6V), or Vf > 3.3V to stay in regulation. If you remove the diodes, then Vf > 2.7V.

When you tested the 5 out of the 10 boards and got 500mA-700mA output, this suggests at a minimum that not all of the AMC chips are connected in parallel. Each chip provides about 350mA drive current. If either one of the areas I described above is not jumpered/bridged, then the board will probably deliver only half the advertised output current, or 700mA.

When you hooked up one of these boards to a single P4, did you remove the diodes? If not, then you also probably did not satisfy the Vf>3.3V requirement since it is quite possible that your P4 has a Vf less than 3.3V when driven at 500mA-700mA. When the AMC starts to fall out of regulation, it sends decreasing amount of current out, which could explain why you measured only about 100mA out.

The last thing is that you don't show a photo of how you hooked up your two boards in parallel to try to get 2.8A output. If that hookup was incorrect, that might also help explain why you got only 1000mA out.

 

[Korxi]

Thanks to both of you for your replies!

I'll do the check for conductivity when I get home from the weekend! I hope that's where the answer lies, but as far as I remember all the legs looked very soldered on to that board

The battery pack was charged to 4.0V prior to hooking it all up so I don't think that should be the problem. Furthermore, when I did the tests of the idividual boards I did it without the LED and measured right across the Vdd and OUT legs (the red and black wires on the pic) - is that "ok"? or would this method give me a false reading?? Anyway, the chips all got hot instantly so I guess that means they burned off the voltage as heat?

Btw. should I connect the LEDs through both brass plates (Q1 and Q3) or is one enough? I only connected to one, so that might be where my problem is? Edit: stupid question there they are ofcourse connected through the jumper wire

Since this is going to be a relatively high current circuit I guess I have to keep resistance in wires etc as low as possible, thicker wires should do the trick here right?

Thanks for helping out the newb

Christian

Edit: Oh and another quedstion, when I charged the two 18650's the cells got a bit hot and the charge was "only" at 4.0V. The charger didn't cut off - I guess because it didn't reach 4.1-2. Should I allways stop the charge when the cells start to warm up, or is it ok to let them be?

 

[Justin Case]

Your method of measuring the output current is incorrect, and the results therefore misleading. You need to measure driver output current in series with the load (the LED). Or measure the voltage across a resistor connected in series with the load and then compute the current from Ohm's law.

Use the CPF search to find out how to connect one of these boards to an LED. There is lots of previous discussion on it.

 

[Korxi]

Ok, I'll try that thanks I'm on the clear with the workings of LEDs resistors etc. and I managed to connect the five P4's correctly the first time - I think I'll also change the battery in my DMM, it started showing the low bat. sign while I tested last time.

Christian

 

[Gazoo]

Your method of measuring the output current is incorrect, and the results therefore misleading. You need to measure driver output current in series with the load (the LED). Or measure the voltage across a resistor connected in series with the load and then compute the current from Ohm's law.

Use the CPF search to find out how to connect one of these boards to an LED. There is lots of previous discussion on it.

This is a linear driver. Therefore current in will be exactly the same as the current to the load.

 

[Gazoo]

The battery pack was charged to 4.0V prior to hooking it all up so I don't think that should be the problem.

After you have everything hooked up again, measure the voltage of your battery with the LED on. It's possible your voltage from the battery is sagging. I have lots of experience with the 350ma version of these drivers. I know for a while DX had a bad batch of that driver, not sure about the higher current ones. These drivers should cause zero problems and are very easy to use. However if your voltage supply is sagging you will not get what you are expecting. Also if the driver becomes too hot the current limiting will kick in. But this should not happen if you are not applying more than 4.5-5.0 volts.

BTW...shorting the diode has very little to no effect. I have tested this myself and I know someone else recently did a test that graphed the differences. So really its best to leave the diode alone.

I found the graphs..hope this helps

http://www.candlepowerforums.com/vb/showpost.php?p=2412507&postcount=30

 

[Justin Case]

That's a good point and I forgot about that. For these AMC7135 LDO regulator boards, you can measure the current at the tail and that is also what is sent to the LED. However, you still need an LED load connected when making your measurement.

 

[Justin Case]

After you have everything hooked up again, measure the voltage of your battery with the LED on. It's possible your voltage from the battery is sagging. I have lots of experience with the 350ma version of these drivers. I know for a while DX had a bad batch of that driver, not sure about the higher current ones. These drivers should cause zero problems and are very easy to use. However if your voltage supply is sagging you will not get what you are expecting. Also if the driver becomes too hot the current limiting will kick in. But this should not happen if you are not applying more than 4.5-5.0 volts.

If voltage sag is the problem, then I would suspect the 18650 cells before the driver board. At an estimated 560mA drive current for five P4s in parallel (2.8A divided by 5 LEDs in parallel), Vf is probably around 3.2V, maybe 3.3V. To run in regulation, Vbatt has to be greater than Vf + AMC7135 drop, or about 3.3V-3.4V. The 18650s have to deliver ~2.8A of current, but that is only about 1C discharge rate. A decent 18650 should be able to stay above 3.6V at that discharge rate. So, unless the cells are bad, 18560s should be able to handle the load.

BTW...shorting the diode has very little to no effect. I have tested this myself and I know someone else recently did a test that graphed the differences. So really its best to leave the diode alone.

I found the graphs..hope this helps

http://www.candlepowerforums.com/vb/showpost.php?p=2412507&postcount=30

That post does not prove that diode removal has no effect.

In the first small table of that post, the test data shows that the SSC P4 used has a Vf~3.4V@~1A.

First, with the diodes installed, to run in regulation you need Vf > Vdd min of an AMC7135 + Vdrop of a typical diode ~ 2.7V + 0.6V = 3.3V (I've measured ~0.55V drop for the diodes on my AMC7135 boards). The SSC P4 in the test you linked already meets this requirement and removing the diodes does nothing for you (the P4's Vf is about 3.4V > 3.3V). This is proven by data in the 2nd and 3rd tables. Removing the diodes would matter only if you had a low Vf LED, i.e., less than 3.3V@1A.

Since you also need Vbatt > Vf + AMC7135 drop (=0.12V) for the board to run in regulation, the reason the drive current was less than 1050ma nominal at 3.4V in is because Vin (i.e., "Vbatt") is less than Vf + 0.12V (or about 3.5V).

The test that could have been definitive, but isn't, is the test in Post #41. Unfortunately, the test did not go low enough for Vin. It stopped at 3.4V. Vf for the SSC P4@350mA nominal is between 3.0V and 3.2V (see the first, small table in Post #30). Let's just say it is 3.1V. Then Vbatt > 3.1V + 0.12V ~ 3.2V for regulation. To test for the effect of diode removal, the test should have gone down to at least 3.2V, and preferably less than that to establish the trend.

 

[Korxi]

Phew, there's a lot of things to be checked here

Regarding the battery sagging and the current draw, the battery pack is 2 parallel 18650's at a total capacity around 4000mAh, so the draw will only be around 0,7C right?

Christian

I'll do the tests later today hope I get something cleared up - I guess there's still the possibility that half the 7135's on each board are defective? rather unlikely though I guess, unless they could get damaged by the heat when they are soldered on the board

 

[Korxi]

Ok, the results are in !

Started by testing one of the driver boards - all 7135's were connected in parallel - so I guess that rules out "or did I just get a bad batch of drivers"

Connected the board (1400mA) to the battery pack and two P4's in parallel and it only kicked out around 500mA..

Tested the battery voltage again 3.99V without load, but when I tested the total drop across the driver board and LEDs the output was only around 3.4V! So I guess I have sagging batteries..? (can anyone clarify what a sagging battery is? )

Then, I took apart the battery pack since I thought the problem might be in the construction.. I tried direct driving a single P4, the DMM measured a drop of 3.89V across the LED, but only measured 800mA (I guess the DMM needs replacing too then..)

As a last test, I connected the single battery directly to the driver and measured the output to two P4's in parallel, at this point I also got around 800mA..

Any suggestions??

Currently I'm thinking that driving p4's in parallel on a parallel pack of 18650's might not be a good solution for me.. The build I'm doing is a dive light and I need high output for around 2.5 hours.. The plan was to build a massive 10-15000mAh pack of parallel 18650's but I guess I would get a more stabile light if I did a series/parallel setup with a bit of surplus voltage that would then just get burned of in the 7135s.. However, that would mean that I'd have to get one of those RC chargers to handle a series/parallel 18650 pack..
Does that sound about right??

Christian

 

[Justin Case]

Ok, the results are in !

Started by testing one of the driver boards - all 7135's were connected in parallel - so I guess that rules out "or did I just get a bad batch of drivers"

Connected the board (1400mA) to the battery pack and two P4's in parallel and it only kicked out around 500mA..

Tested the battery voltage again 3.99V without load, but when I tested the total drop across the driver board and LEDs the output was only around 3.4V! So I guess I have sagging batteries..? (can anyone clarify what a sagging battery is? )

Then, I took apart the battery pack since I thought the problem might be in the construction.. I tried direct driving a single P4, the DMM measured a drop of 3.89V across the LED, but only measured 800mA (I guess the DMM needs replacing too then..)

As a last test, I connected the single battery directly to the driver and measured the output to two P4's in parallel, at this point I also got around 800mA..

Any suggestions??

Currently I'm thinking that driving p4's in parallel on a parallel pack of 18650's might not be a good solution for me.. The build I'm doing is a dive light and I need high output for around 2.5 hours.. The plan was to build a massive 10-15000mAh pack of parallel 18650's but I guess I would get a more stabile light if I did a series/parallel setup with a bit of surplus voltage that would then just get burned of in the 7135s.. However, that would mean that I'd have to get one of those RC chargers to handle a series/parallel 18650 pack..
Does that sound about right??

Christian

3.4V under a moderate load is not a good sign. Internal resistance of the cells is overly high.

In your direct drive test, you probably got a 800mA reading at 3.89V in because the DMM in the circuit presents its own voltage drop.

I would have done the simplest testing of the driver first -- one driver, one P4.

With the two P4s in parallel, you were measuring the current draw from the batteries and got 800mA?

Can you show the wiring you used on your direct drive and last test, and where you connected the DMM in the circuit?

 

[Korxi]

First of all thank you so much for taking the time to help me! I greatly appreciate it!

When I measured the current draw in the direct drive test, I measured 800mA in series with both LEDs yes.

I did the simple test of one driver one P4

I attached two pics, one showing the thick wiring I used to connect the battery to the driver, and the other showing where I inserted the driver in the circuit. I measured current in series with the LED through the wires marked with red arrows. I measured voltage drop across LED on + and - on the LED itself, battery input voltage I measured on the back of the driver where the two wires are connected. Here's what I got:

Draw from battery 3.83V

Drop across LED 3.74V

Measured current 530mA

With the battery input voltage at 3.83V shouldn't it provide enough for driving all four 7135's and the LED?

Christian

 

[Greg G]

You need to build a 2 series/X parallel battery pack so the driver has a bit of *headroom*.

Also, 1400ma is a lot of current for a Seoul P4. I realize that since it's a dive light it will enjoy excellent cooling, but will you be getting a good return on that extra current? It might be better to drop the current and add additional LED's.

 

[Greg G]

Holy Cow....I looked at the efficiency specs for a 7135 and at increased Vin it is terrible. When Vin and Vout are close it's pretty good though.

I'd consider a different driver. You don't want to get down there and have this thing fail. A safe solution is two circuits. 3 LED's per, and a good boost driver like a Blue Shark for each circuit, driven of a 7.4 volt nominal battery pack.

You'd have ~10.5 Vf and ~8.4 volts from the battery pack on a fresh charge.

 

[Korxi]

Hi Greg,

The initial idea was to run 5 parallel P4's on a total of 2800mA, which would keep the LEDs cool

I actually just tried putting the two batteries in series and hooking it all up again (just to one P4 still) and I think it blew out the 7135's - at least it didn't light up at all, not even on only one cell after that treatment

This stuff is much more difficult than I had imagined

Christian

 

[xenonk]

I actually just tried putting the two batteries in series and hooking it all up again

:green:

Maximum rated voltage for the AMC7135 is 5.5V IIRC, and it strongly protests at even that.

 

[Korxi]

Well that explains it - rookie mistake..

Good thing I got 10 then (unless ofcourse I'm not going to use them..)

Christian