AMC7135 Specs Inside **UPDATE**

[Drewfus2101]

1050mA version:


1400mA version:


Pretty good boards, as long as your in they're small voltage range. DX recommends that you stay within 3.6V to 4.5V. I read that is because between 4.6V to 6.0V (which it can take up to 6.0V) it gets very hot. In my testing I used a small fan to keep the board cool so this did not affect me.

These are a great alternative to resistors, even though resistors are cheaper. The best thing to do here is order 20 of the 1400mA version and you can remove each of the black AMC boards and each one you remove will cut down the output current by 350mA. I tested this and it does work.

Where these boards beat resistors is that these give you a semi-regulated output. If you use resistors and alkalines (for example) the flashlight will never be brighter than when you first fire it up. After that, the batteries will start to dim. If you chose your resistor based on the fresh battery voltage, as the batteries start to drain, your output will drop. With these, as long as your voltage stays above about 4.0V you get the full output. The 700mA version should need a lower voltage and the 350mA an even lower voltage. So that semi-regulation is worth the extra money to me. I can't tell you how much time and headache direct drive / resistors / alkalines have caused me.

I definitely love a good boost/buck convertor more than these, but these things are not bad little boards. I thought others could benefit from these numbers.

[TorchBoy]

Nice results. I'd be more interested in graphs of output current for input voltage, though. (I don't quite see the point of efficiency graphs for linear regulators.)

On the subject of them being linear regulators, how did you get more current out that current in?

[Drewfus2101]

I'd say that the current in and current out are probably equal its just that my accuracy in measuring them is slightly off.

Charts as requested:

[TorchBoy]

Wow, awesome! That shows just when the thing drops out of regulation. Now, why wouldn't it be completely flat when in regulation? Temperature effects in the AMC7135s?

[Drewfus2101]

Originally Posted by TorchBoy

Wow, awesome! That shows just when the thing drops out of regulation. Now, why wouldn't it be completely flat when in regulation? Temperature effects in the AMC7135s?

Not this time, at least I doubt it anyways. I had a small 12V fan running off of a separate power source an inch from the board running full out the whole time. It was moving more than enough air to keep the thing cool and the board was very cool to the touch when I picked it up as soon as I turned off the power. But I wouldn't run the same test without the fan, at least not up to 6V and I probably won't use these boards up to 6V without some testing first.

[ktronik]

Great info...

I did a test by mistake

I had a 3d m@g with some AMC's in... by mistake I put 3/ 3 AA carriers, way over voltage spec...

The light worked fine, until the AMC's heated up enough to melt the solder holding them on the board & slid off...

After re-solding them back on they still seem to work ok on normal voltage.

I would say they are pretty tough.

with in the right voltage range, they a great cheap circuit to work with...

Great for 4 nimh's, as by the time they are drained (3.2v min) the AMC's are draining very little, a great low batt warning.


K

[TorchBoy]

I knew they were tough, but that's amazing. You're talking 180-190 °C. It sounds like a very easy way to get them off the board.

I thought the thermal protection cut in at 150 °C.

On a multiple AMC7135 board and too many volts they'll presumeably starting blinking, but they'll all be blinking out of sync with each other. Would it even be noticeable?

[ktronik]

I think you may be right... they started blinking a bit, so I tapped the light head...thats when the AMC's slid off the board...

K

[mike2g]

I've been thinking about using these boards in a single li-ion powered light but am concerned with the loss drop in output current after the voltage drops below 4 volts. The reason I like these boards is that the efficiency of the boards is above 90% for most of the life of a single li-ion. However, the output doesn't seem to be very well regulated. According to this, for the case of the 1400ma version, the output current can be expected to drop substantially, 1.28A -> 0.85A, after the battery capacity has dropped to 80% of its' capacity. Basically, a 1/3 drop in out put after a 1/5 drop in battery capacity.

It strikes me that these drivers would be much improved if there were a way to shift the graph over so that the current output could be maintained at a lower battery voltage. Is there a way to do this?

Also, I'm wondering how the graph could be expected to change if the emitter had a lower Vf. For example it has been rumored that the Q5s have relatively high Vf, so would using a lower Vf emitter such as a SSC P4 shift the graph and increase the range where the output current could be regulated?

Thanks for putting this together, it's been very helpful!

[SnowplowTortoise]

Great post! Thanks for passing on the info. 60% efficiency at 6V is pretty scary.

Craig

[CampingLED]

Great review. Read a post from "Download" once stating that the diodes on the boards should be shorted out to improve output. This makes sense after checking the data sheets of the 7135s. Will appreciate it if you could do a simple test when the diodes are shorted out.

[Mr Happy]

Originally Posted by mike2g

I've been thinking about using these boards in a single li-ion powered light but am concerned with the loss drop in output current after the voltage drops below 4 volts.

This is unfortunately to be expected. The board is apparently a buck regulator, which means it can reduce the battery voltage but cannot increase it. Regulators of this type also need an extra margin of input voltage over output voltage to operate properly. In this case that margin is only about 0.3 V, so I think it's doing pretty well.

The reason I like these boards is that the efficiency of the boards is above 90% for most of the life of a single li-ion. However, the output doesn't seem to be very well regulated. According to this, for the case of the 1400ma version, the output current can be expected to drop substantially, 1.28A -> 0.85A, after the battery capacity has dropped to 80% of its' capacity. Basically, a 1/3 drop in out put after a 1/5 drop in battery capacity.

When the voltage drops below 4 volts the unit has dropped out of regulation due to insufficient input voltage. To obtain good results, the battery used should provide more than 4 volts until fully discharged. [Edit: For the particular LED tested in this case.]

It strikes me that these drivers would be much improved if there were a way to shift the graph over so that the current output could be maintained at a lower battery voltage. Is there a way to do this?

Not inexpensively, no. If you look at the voltage in vs. voltage out tables, you can see how the unit would have to boost the voltage to achieve what you wish, and this would be extra circuitry.

Also, I'm wondering how the graph could be expected to change if the emitter had a lower Vf. For example it has been rumored that the Q5s have relatively high Vf, so would using a lower Vf emitter such as a SSC P4 shift the graph and increase the range where the output current could be regulated?

I think it would help, yes. Anything that lowers the required output voltage would extend the range of regulation.

[Marduke]

Does anyone know if these can be ran two in parallel to provide the 2.8A required for a P7?

[Drewfus2101]

Originally Posted by CampingLED

Great review. Read a post from "Download" once stating that the diodes on the boards should be shorted out to improve output. This makes sense after checking the data sheets of the 7135s. Will appreciate it if you could do a simple test when the diodes are shorted out.

Wow, I almost forgot about that trick. I bookmarked some info about that, but forgot to try that when I was running this test. I should have a night this week to run more tests and I'll add this to my list.

[Drewfus2101]

Originally Posted by Marduke

Does anyone know if these can be ran two in parallel to provide the 2.8A required for a P7?

I'm pretty sure they can not. They are "buck" type boards so they can only reduce the voltage in. Notice how the current in and current out are practically identical (not counting multimeter accuracy error)? It would be pretty much the same as direct driving. The only benefit of using two of these (assuming that worked) was to buck the voltage down. But you can't go higher than 6.0V with these and DX recommends no higher than 4.5V due to heat.

I'm sure someone with a P7 will try it and it might just work, but I can't figure out how and it won't be efficient.

[Drewfus2101]

Originally Posted by Mr Happy

I think it would help, yes. Anything that lowers the required output voltage would extend the range of regulation.

I'll try and repeat the test with an SSC Z-power. Its the only other emitter I have that is star mounted and will work with my setup. I'll double check the emitter first and make sure it has a lower Vf too.

[TorchBoy]

Originally Posted by mike2g

It strikes me that these drivers would be much improved if there were a way to shift the graph over so that the current output could be maintained at a lower battery voltage. Is there a way to do this?

Use a lower voltage (Vf) LED.

Originally Posted by mike2g

Also, I'm wondering how the graph could be expected to change if the emitter had a lower Vf.

In the way that you want. The AMC7135 needs at least 2.7V and white LEDs have a Vf above that, so you'll be fine.

Originally Posted by CampingLED

Read a post from "Download" once stating that the diodes on the boards should be shorted out to improve output. This makes sense after checking the data sheets of the 7135s.

Let's see... The voltage drop across one of those diodes will be ~0.6V so if your input voltage is less than 3.3V it might make a difference, since the AMC7135 needs 2.7V to stay in regulation, but would your LED Vf be that low? Quite possibly - I've got one Cree with a Vf of 3.3V at 1A. Drewfus, it wouldn't have made any difference to your test, because your operating voltages were higher than that.

Originally Posted by Mr Happy

The board is apparently a buck regulator, which means it can reduce the battery voltage but cannot increase it.

No - it's a linear regulator - and yes respectively.

Originally Posted by Mr Happy

When the voltage drops below 4.0 volts the unit has dropped out of regulation due to insufficient input voltage. To obtain good results, the battery used should provide more than 4.0 volts until fully discharged.

That will depend on what the LED wants. For the particular LED tested, with a reasonably high Vf of 3.7V @ 1A, you do need at least 4V.

Originally Posted by Marduke

Does anyone know if these can be ran two in parallel to provide the 2.8A required for a P7?

Yes, not a problem. Someone recently even figured out how to connect a multimode board to a single mode multi-AMC7135 board to get lots of current and multiple modes. The efficiency, Drewfus, comes from running these things with a Vin a little higher than the Vf of the LED. Whether that's 3.3V for a low Vf Cree P4 or 4.2V for a SSC P7 doesn't matter.

[TorchBoy]

Here we go - multimode driver for P7. It has also has a diagram of how to connect two 4xAMC7135 boards in parallel. Easy.

[Supernam]

Darn, I found this post a little too late. I wanted to use this for 3D NiMH's, but I they provide probably a max total of 4v rested off the charger. I'm thinking the best would be 4D NiMH's or 3D Alkalines. Alkalines shouldn't sag much at only 1A, so I'm thinking I'd get some ridiculously long runtime right?

I guess I can't complain since they're less than $2 each!

[TorchBoy]

Supernam, if you're going with alkalines I'd definitely recommend 4 instead of 3. Their voltage really doesn't hold up and most of their life is with a lower voltage than would do the job if you only had 3.

With NiMH, it would depend what the Vf of your LED is and when you wanted the thing to drop out of regulation. Drewfus' test above showed his LED needed 4V for the board to stay in regulation, so 4 NiMH cells would probably work better in that situation, and drop out of regulation at a loaded 1V per cell. Note though that the board might get quite hot when the NiMH cells are freshly charged.

[Mr Happy]

Originally Posted by TorchBoy

No - it's a linear regulator - and yes respectively.

Very well, I bow to common parlance. Though I don't see why the term "buck" has to be confined only to step-down regulators that use switching technology

[TorchBoy]

Yeah, I don't know why that is. It certainly caused a bit of confusion for me before I figured out what people meant by it.

[Supernam]

Originally Posted by TorchBoy

Supernam, if you're going with alkalines I'd definitely recommend 4 instead of 3. Their voltage really doesn't hold up and most of their life is with a lower voltage than would do the job if you only had 3.

With NiMH, it would depend what the Vf of your LED is and when you wanted the thing to drop out of regulation. Drewfus' test above showed his LED needed 4V for the board to stay in regulation, so 4 NiMH cells would probably work better in that situation, and drop out of regulation at a loaded 1V per cell. Note though that the board might get quite hot when the NiMH cells are freshly charged.

The more I look at the data, the more I wonder what WAS this board designed for? One LiIon would mean that there would be no regulation for the most part. 2 CR123's is too much. 3 Nimh's would present the same problem as 1 LiIon. It seems as though the only way to go with this board is to use 4 NiMH's. I think 3 D sized Alkalines would work fairly well, at least with the 1050mAh board. 4 would bring you to at least 6v which will run the board pretty darn hot when you're using new batteries.

I was planning to use this with a SSC P4 with a Vf around 3.3 in 3D Nimh host, but I'm going to just kill my wallet with a buck/boost GD1000 instead (about $25 shipped for 1!). Perhaps this board will be better with a 4C or 4 SubC mag.

[Drewfus2101]

Originally Posted by Supernam

The more I look at the data, the more I wonder what WAS this board designed for? One LiIon would mean that there would be no regulation for the most part. 2 CR123's is too much. 3 Nimh's would present the same problem as 1 LiIon. It seems as though the only way to go with this board is to use 4 NiMH's. I think 3 D sized Alkalines would work fairly well, at least with the 1050mAh board. 4 would bring you to at least 6v which will run the board pretty darn hot when you're using new batteries.

I was planning to use this with a SSC P4 with a Vf around 3.3 in 3D Nimh host, but I'm going to just kill my wallet with a buck/boost GD1000 instead (about $25 shipped for 1!). Perhaps this board will be better with a 4C or 4 SubC mag.

I bought these specifically for modding Mags with alkalines and P4s. Like you said, 3 alkalines are pretty perfect. The 700mA setting is slightly lower than you really want, but still plenty bright and works great for 3 C batteries. And the 1050mA setting works well for 3 D batteries that can handle that current draw better than the C size batteries.

[mike2g]

Originally Posted by Supernam

The more I look at the data, the more I wonder what WAS this board designed for? One LiIon would mean that there would be no regulation for the most part. 2 CR123's is too much. 3 Nimh's would present the same problem as 1 LiIon. It seems as though the only way to go with this board is to use 4 NiMH's. I think 3 D sized Alkalines would work fairly well, at least with the 1050mAh board. 4 would bring you to at least 6v which will run the board pretty darn hot when you're using new batteries.

I was planning to use this with a SSC P4 with a Vf around 3.3 in 3D Nimh host, but I'm going to just kill my wallet with a buck/boost GD1000 instead (about $25 shipped for 1!). Perhaps this board will be better with a 4C or 4 SubC mag.

I completely agree. Efficiency and regulation occur in two separate parts of the voltage curve, with only a small overlapping area. If you want efficiency greater than 90%, then you have to stay below 4.2V. If you want, regulation you need to stay above 4V. This leaves a 0.2V window to get good regulation and efficiency.

If you are willing to have an efficiency down to 80% the voltage must only stay above 4.6V/4.8V for the 1050ma/1400ma version respectively. Regulation still requires 4V so the window for regulation and efficiency is now 0.6V/0.8V (1050ma/1400ma).

From the battery perspective however, this is difficult to achieve. Li-ions will start at 4.2V and quickly drop below 4V and out of regulation. Regarding alkaline D batteries, it seems that the voltage drop is roughly linear from 1.4V to 0.8V, resulting in a 0.6V change. However because there are multiple cells the change is now 1.8V for 3 cells, and 2.4V for 4 cells. Both of these are much larger than the 0.8V window allowable for 80% efficiency.

Nimh may be the best solution however for it's ability to maintain voltage up to the bitter end. At the 1A draw rate for AA, the discharge is primarily between about 1.35V to 1.15V giving a change of 0.2V. This change is 0.6V and 0.8V for 3 and 4 cells respectively. This range seems reasonable for the 80% efficiency minimum and the best compromise from the available battery choices.

The only optimism I have left for this board is in the coming SSC P4 test with the lower Vf. Hopefully the regulated range will increase. Thanks Drewfus2101 for doing the meticulous testing!

[freedom2000]

Originally Posted by Drewfus2101

1050mA version:

I definitely love a good boost/buck convertor more than these, but these things are not bad little boards. I thought others could benefit from these numbers.

Hi Drewfus2101,

Many thanks for this nice technical review. (I have just bought such a driver, and you already helped me a lot).

You mentionned however that you prefer boost drivers...
Would you have a recommendation for me to find a good (and cheap) driver to drive a Q5 Cree led at 1A with one or two Nimh ?

My intention is to mod this flashlight, which has curerntly a rather inefficient driver see here

Again, thanks for your help
JP

[TorchBoy]

Originally Posted by Drewfus2101

And the 1050mA setting works well for 3 D batteries...

Why?



(That Energizer Max doesn't do too well compared to the others.)

Within a handful of minutes the cells are down to 1.4V, and by the time they hit 1.2V there's, what? 75-80% of the life left. And even 3x1.2 V would be enough only if the Vf of the LED is <=3.4V. Yeah, they exist in quite a few P4s, but I don't think many Q5s. From 1.2V onward, even a 3.4 Vf LED would be out of regulation.

I still say three alkaline cells are not suited for this driver (unless you want to just use the top 20% of each cell's capacity).

mike2g, your point of voltage difference is very good. I'll still say 4 alkaline cells will work quite nicely, though, with a qualifier or two. Because of the way alkaline cells discharge, the board when used with the LED tested above would be in the higher efficiency area (median would be 84.5%) for four times longer than it would be in the low efficiency area. So for efficiency, it would work nicely. If regulation is your goal, however, you'll have to use a low Vf LED to have it stay in regulation until 0.8 V per cell, and that would push the median efficiency down (to ~75%), which I'd still be happy with for efficiency as well. I don't deny you might have trouble finding the right LED though.

Originally Posted by mike2g

Efficiency and regulation occur in two separate parts of the voltage curve, with only a small overlapping area.

That's a contradiction. I suppose the efficiency area isn't quite a subset of the regulated area, so you could say the areas are in different, but not separate, places.

Originally Posted by mike2g

Thanks Drewfus2101 for doing the meticulous testing!

It's a great test, and both very interesting and very useful, but meticulous probably means something other than what you intend. You can't have 1.24 A in and 1.28 A out with a linear regulator - no disrespect intended to you Drewfus.

(Actually, one definition for meticulous I found is "ridiculously obsessed with extreme minutiae" - which would be me posting this. )

Originally Posted by freedom2000

Would you have a recommendation for me to find a good (and cheap) driver to drive a Q5 Cree led at 1A with one or two Nimh ?

freedom2000, I don't think such a driver exists, if $20 (for a GD1000) doesn't come into your idea of cheap. Have you seen the driver list?

[Drewfus2101]

Originally Posted by freedom2000

Would you have a recommendation for me to find a good (and cheap) driver to drive a Q5 Cree led at 1A with one or two Nimh ?

I wish I did. I've bought the only two single mode boost drivers that DX sells but have only had time to test one. I'm trying to do the same thing.

This is the one that seems better, but I havn't had a chance to test it. I'll post the results when I get a chance:
http://www.dealextreme.com/details.dx/sku.4735

Here is the other one. You can see my post at the bottom of the page. These boards are useless as boost boards. Maybe they are just mislabeled.
http://www.dealextreme.com/details.dx/sku.4382

[Drewfus2101]

Originally Posted by TorchBoy

Why?



(That Energizer Max doesn't do too well compared to the others.)

Within a handful of minutes the cells are down to 1.4V, and by the time they hit 1.2V there's, what? 75-80% of the life left. And even 3x1.2 V would be enough only if the Vf of the LED is <=3.4V. Yeah, they exist in quite a few P4s, but I don't think many Q5s. From 1.2V onward, even a 3.4 Vf LED would be out of regulation.

I still say three alkaline cells are not suited for this driver (unless you want to just use the top 20% of each cell's capacity).

mike2g, your point of voltage difference is very good. I'll still say 4 alkaline cells will work quite nicely, though, with a qualifier or two. Because of the way alkaline cells discharge, the board when used with the LED tested above would be in the higher efficiency area (median would be 84.5%) for four times longer than it would be in the low efficiency area. So for efficiency, it would work nicely. If regulation is your goal, however, you'll have to use a low Vf LED to have it stay in regulation until 0.8 V per cell, and that would push the median efficiency down (to ~75%), which I'd still be happy with for efficiency as well. I don't deny you might have trouble finding the right LED though.


That's a contradiction. I suppose the efficiency area isn't quite a subset of the regulated area, so you could say the areas are in different, but not separate, places.


It's a great test, and both very interesting and very useful, but meticulous probably means something other than what you intend. You can't have 1.24 A in and 1.28 A out with a linear regulator - no disrespect intended to you Drewfus.

(Actually, one definition for meticulous I found is "ridiculously obsessed with extreme minutiae" - which would be me posting this. )



freedom2000, I don't think such a driver exists, if $20 (for a GD1000) doesn't come into your idea of cheap. Have you seen the driver list?

To your first question about using the 1050mA with 3 D alkalines, I should have said that was my choice OVER direct driver or resistors. As I said, DD with resistors was always a headache. These are better, but definitely not perfect.

For the Mag that I made for a friend, I used a SSC P4 and the 700mA version with 3 D alkalines. It was that or a 1 ohm resistor.

[Drewfus2101]

**UPDATE**

Following the suggestions, I tried a few different things. First I found a SSC Z-Power P4 emitter and tested it:



So it has a more lower Vf than a Cree Q5 does.

Then I hooked it to the 1050mA version of the AMC7135 and re-ran the normal test the exact same way as before. Next I took out the diodes and soldered in a leg from a resistor in place and re-ran the same test. Both tables are below followed by the graphs.





Removing the diodes didn't seem to make much of any difference at all in the efficiency: