The SF M6-R - Part 2

[js]

This is a continuation of The SF M6-R:A regulated rechargeable SF M6 (HOLA) which is near enough to the 200 post limit that I am starting Part 2 now.

As it turns out, this is timely because I can make an updated and better summary of the SF M6-R by editing this first post, which I will do shortly.

 

[tvodrd]

Hope I'm still in line! /ubbthreads/images/graemlins/smile.gif /ubbthreads/images/graemlins/thumbsup.gif

Larry

 

[js]

Larry,

I'm still aware of your interest, and you're on the notify list, but the actual line won't start until I open the B/S/T signup thread. /ubbthreads/images/graemlins/grin.gif

OK. pyro and Catdaddy have gotten in touch with me and do not want to forfeit their positions as field testers. I am working on their kits as we speak and should be mailing them out soon.

AND, Prolepsis should be getting his M6-R any day now.

 

[js]

It's high time for an update. Sorry I've been remiss; I've been working on the field tester M6-R packs and chargers. Details on that subject are posted in the field tester thread, but the short story is that I am behind schedule and have only finished Catdaddy's charger at the moment. His pack will be completed tonight, and pyro's pack and charger will be done in a couple days, assuming that everything goes smoothly.

A few technical subjects: Willie Hunt's LVR3's have a 20K resistor across the power FET gate and source which lowers the gate after power is removed. This is so that the gate doesn't stay halfway open when power is removed from the LVR circuit, because if that happened, the FET would burn out almost instantly because it is meant to be either open or closed and is not able to dissipate the amount of power which would be dropped across it if it were in a half-open state.

Now, in the M6-R application, the LVR is always powered up because it is integrated and hardwired into the battery pack. So the only time R3 would come into play is if someone left their M6-R turned on. In that case, when the voltage of the pack dropped past 5 volts, the LVR would not have enough voltage to power itself, and would possibly go haywire, and would normally shut down, and then the voltage of the pack would rise, and the LVR would turn back on and try to draw more current from the pack, and so on until the pack were destroyed.

So, on Prolepsis's, Catdaddy's, and pyro's pack, I have left off R3 entirely because I would so much rather change out the $5 IRL 3803 power FET, than build an entirely new battery pack--so much more labor and cost involved in that. What I'm saying is that if the LVR ever gets to the point where it doesn't have enough voltage to function, I really want the FET to blow, as it will save the pack.

Plus, leaving off R3 means that there are only wires to and from the FET on the back, exposed, side of the board, and I like this.

On another technical topic, I have been "plating/coating" the copper, rounded triangular end-caps of the packs with a 96 percent tin and 4 percent silver lead-free solder. This is electrical grade solder, which means it conducts well, and it is nicely resistant to corrosion, and it is the hardest of the soft-solders.

However, it is not nearly as hard as I would like it to be, and I am always looking for ways to improve my designs, so I have been investigating the so-called hard solders, which are brazed on with a torch, instead of soldered on with an iron. I found some really good stuff, which conducts better than copper, is specifically used for making contacts out of pure copper pieces, and which is still low enough temp that it will not change the conductivity of the base copper. This means I get to play with an oxy-actelene torch. hehe /ubbthreads/images/graemlins/evilgrin07.gif Should be fun. Anyway, I will be using this on pyro's pack, and if all goes well, it will be the standard for the production packs.

Don't get me wrong, the tin/silver solder works fine, but it does get dented by the springs, and for some reason that bothers me a little, so I hope this hard solder will work fine. I don't see any reason why it won't.

I haven't edited the first post in this thread with an updated summary yet for precisely this reason: I'm waiting to see if this new contact plating will work out. Plus, I'd like it to have pictures of the charger with a glass epoxy printed circuit board instead of a phenolic proto-board.

And that's the weekly update. Thanks everyone.

 

[js]

I tried the silver solder/braze, and WOW! does it work well. It's a nice gold color and it is very hard. It looks exactly like the plating on the SF M6 MB20 battery holder contacts, actually. And it wasn't hard to do at all, so I think that as long as my resistance tests turn out well, this will be the standard plating for the build-run. I'll post pictures at some point.

Meanwhile, I am anxiously awaiting Prolepsis's initial report in the field testing thread.

 

[js]

The silver braze I am using is Harris Safety Silv 50N brazing filler metal, and it is about 50 percent silver, 20 percent copper, 28 percent Zinc, and 2 percent nickel. It melts at 1300 F and has a conductivity of 49 % I.A.C.S., which means that it is about half as conductive as copper. "I.A.C.S" stands for "international annealed copper standard" and is used to compare conductors against a copper wire standard. This means that a wire made of this silver brazing compound 1 meter in length with a cross sectional area of about 1 sq mm would have a resistance of .86 ohms. This may sound like a lot, but it is not. In our application, this material will add negligible resistance. I couldn't measure any difference at all. Let me put it this way, 6061 Al has a % I.A.C.S of 43, and how many flashlights use the entire length of the Al body as the conduction path? I'm just using a fraction of a mm of this silver braze at the contact plates, and it is hard and corrosion resistance and--as it turns out--looks pretty, too. /ubbthreads/images/graemlins/smile.gif

I am making new plates up for Catdaddy's pack and I am still working on his charger, but I am doing some things in parallel, so pyro's M6-R is coming along as well. I hope to have Catdaddy's pack (new and improved) and charger completed tonight.

 

[js]

The problem with Catdaddy's charger . . . it was the stupid chip itself. Of course I didn't find this out until after I had built an entirely new board and wired it up. I then pulled the chip from the old board and inserted it into the new one, and began charging a pack. Which charged for only 40 or 50 minutes and then pre-maturely stopped charging. I was like "AHHH MAN!" and then I was like "Wait a minute. What if it's the chip!? D'OH!" Yes. Yes. I should have tried swapping out the chip BEFORE I built a new board, but truth be told, I was worried about the solid-core wiring I had used, and which I was stripping with a razor-blade. This is a bad idea, because you can almost nearly cut through the wire, which then breaks when you bend and solder it in place, leaving an intermittent connection. Or it can break with vibration. Plus, the wiring layout and the solder joints weren't exactly *ahem* pretty.

So, I'm actually glad I built a new board with stranded core, teflon coated wiring, and with neat looking wire layout and solder joints. It's something I'm proud to have my name attached to. Plus, I increased the current to the green LED and now it appears to be about the same brightness as the red LED.

And yes, it's working prefectly. So, Catdaddy, I still need to put your pack together with the new silver-brazed coated/plated contact plates, but I should have no problem doing that tonight, and then mailing out your stuff tomorrow. I'm really sorry for the delay.

 

[js]

More delay.

It was actually NOT the chip. It is the pass transistor overheating. And I've decided to can the whole thing and go in with bwaites on buying Hitec CG-340 chargers and 12V 13A powersupplies. With the quantity discount we'll be getting, I can offer these two things for about the same $75 price-tag as my custom, dedicated charger was going to cost. And the 340 will charge any battery pack and not just the M6-R pack, and it will charge NiMH batteries at rates of up to 1.5 amps, which means it will charge the M6-R pack in about 45 minutes. PLUS, it will work in the car off the 12V adaptor power. I'll just need to make up adaptors as a special option.

Plus, if anyone is going to be buying BOTH a USL and an M6-R, the very same setup, right down to the charging cable, will work with both.

And, if anyone is interested in upgrading to a Triton charger (which I highly recommend, BTW) Bill and I are looking into how much we could get them in quantity.

That's the short of it. Here is some more for entertainment value:

So, I was installing the new silver braze contact plates on Catdaddy's pack, and had just finished discharging his pack, and was periodically watching it being charged by my new and perfected and carefully constructed custom charger. And I was feeling very pleased about it.

And then I look over and the red fast-charge LED was out, and I felt the pack and it was still cold and I was like [insert bad word here] [bad word] [bad word]!!!!

After doing more examination of things and some thinking, I realized that while I have the cover off, and am watching and checking the circuit, everything works fine. It is only when I install the cover and try doing a full charge that fast-charging terminates pre-maturely. And it is the exact same cicuit and components as Brightnorm's charger, which works just fine. The only difference is in the heat-sinking and size of the box.

I had plenty of people laugh at me for using the heat-sink I did, saying that it was over-kill. Well, yes, in free-air it will handle 8 watts, but inside a box, who knows? I de-rated it to 4 watts, thinking that that was conservative given that I was using a metal, perforated box.

Apparently not.

So I could mount a different style heat-sink to the outside, or use a MOSFET and an indcutor and go to a switch-mode charger design, and/or use a lower voltage power supply and just use two different ones for USA/Canada and overseas, but then I thought:

screw this!

And I called Bill right away and threw in the towell for the night.

Anyway, given the choice bewteen a Hitec and p.s., and the charger I could build, other things being equal, I'd say the Hitec is the clear winner. So this is a good thing.

pyro and Catdaddy, you need to decide if you want a refund for the charger part of your M6-R kits (and you'll find your own stand-alone charger and power supply) or if want to take advantage of the excellent pricing Bill and I are getting on the Hitec CG-340 and power supply. I can send out these to you for the same price, but I don't know how fast we will be getting them. In either case, you will still need a $10 charging cable to connect the M6-R pack to the charger.

I'll send you PM's.

 

[paulr]

JS, you're clearly doing the right thing dropping the custom charger. There's no sense spending your time making chargers that duplicate the efforts of mass produced chargers, when you could spend that time instead doing unique mods.

What's the diff between a Hitec and a Triton? My wallet isn't ready for an M6R yet, but I'm on the USL list, and if both chargers are offered I'm not sure which one to prefer.

 

[js]

paulr,

Yes. My point exactly. I was going to do it at first, because there was a significant price difference. Actually, at first I was going to do a slow-charger using the LM317T, but then I found that the MAX712 circuit wasn't that much more complicated. So I built that. But whatever.

The Triton is a much better charger than the CG-340, and is not only a charger, but also a discharger and cycler. You can add a thermal probe for another $10, which gives such peace of mind. You know that if for some reason the peak-detect doesn't work, then the thermal probe will cut off the charging.

Also, you can change the charge rate during charging, can charge at up to 5 amps, can set maximum mAh input, can set a time-on-charger maximum, can do a "top-off" charge after the end of fast-charge in order to equalize the cells, and most importantly, the Triton uses a detection scheme whereby the charging is stopped temporarily and the voltage of the pack is measured, whereas the Hitec measures the applied voltage necessary to keep a constant current and how that changes over time.

The Triton brings NiMH batteries to peak with the least amount of heating up possible. It slows down charging right at the end in order to gently bring the pack to full charge (this is different than the top-off feature, which is optional and user-selectable).

SilverFox is doing a comparison of different chargers in a thread in the Batteries forum. Check it out.

In any case the Hitec CG-340 is a nice charger, too, but it's just not as sophisticated.

 

[js]

HOLD EVERYTHING! (again).

As I was driving home today it dawned on me that there was one other difference between brightnorm's charger and the ones I am building for Catdaddy and pyro. Well actually there are two, but the main one is that on BN's, I tried a 15VDC wallwart at one point and changed out R1 from 1.2K to 1.0K ohm in order to supply 10 mA to the V+ node of the MAX712. When I decided that the 15V wallwart wasn't going to work (it wasn't even 15V under load) I left R1 at 1.0K, knowing that a couple extra mA into V+ was fine.

And then I realized that the new diffuse red LED I am using (BN's has a clear) has a Vf of 1.6 not 1.7. Plus, I also couldn't remember ever calculating the current through that branch. I had 5 mA in my head, but when I sat down to do the calculation, I found that it is actually pulling a little more than 7 mA from the V+ node. So with 10 mA comming in, it means only 3 mA for the chip itself, and the 712 needs at least 5 mA's to power it's internal logic.

So I replaced R1 with an 811 ohm resistor, which will put 16 mA into the V+ node, which will leave more than enough for the 712. And I am discharging Catdaddy's pack right now so I can try out the modified charger to see if that clears up the problem.

I have that EUREKA feeling, so I am very sure that that was the problem and I remember that the very first fault was a stop of fast-charge and then a re-start. And this with the cover off. So that could not possibly be the pass transistor.

Ah. LOL. Given my playing my cards out on the table for everyone to see method of conducting my modding work on CPF, everyone gets to see all of the screw-ups and false starts and stupidity that goes along with doing something like this. I hope everyone finds it amusing and not annoying. Sorry if it's the latter.

So, OK, assuming that the charger now works fine, this means that I can get Catdaddy's and pyro's M6-R kits out sooner rather than later and field testing can get underway in earnest.

But this still leaves open the question of what will happen with the build-run and I still am of the opinion that given the same costs that the Hitec CG-340 and 12V 13A powersupply are the better set-up.

Should I post a poll thread to get people's opinions? Let's see. The only advantage I can think of with the custom charger that I have designed is that the circuitry is a zero-delta-V detection scheme, which means that it has any number of chances to terminate charging after the cells have peaked, and that it goes to a trickle charge after fast-charge, which allows for equalizing and topping off the cells. These two things aren't that big a deal, and as I mentioned the advantages of the 340 are that it will charge in 45 minutes, will work in the car, and will charge any type of pack containing between 4 and 16 cells.

What do people think? Should I start a poll thread, or just make an executive decision?

I'll be PM'ing pyro and Catdaddy again.

I'm a freak-show, I am. /ubbthreads/images/graemlins/crackup.gif /ubbthreads/images/graemlins/grin.gif /ubbthreads/images/graemlins/grin.gif /ubbthreads/images/graemlins/grin.gif

 

[Starlight]

Here's your poll, spend your time making our battery packs, which no one else makes, and buy chargers that several other companies make.

 

[bwaites]

I love AH HA moments!!

Bill

 

[js]

Starlight,

Yes. You're right. But I'm stubborn. And when you've invested considerable time and no small amount of money into something, it's hard to abandon it.

But, alas, that is what is called for here. The low current flowing into V+ was not the problem. I fixed that, although I had to put R1 back to 1.2K and decrease the current flow through the red LED, because the chip itself was getting too hot for my taste. Yet even with 7 mA flowing into the chip, it still terminated fast-charging too early. It must be the pass transistor overheating and dropping its beta to the point where the chip's DRV pin can no longer keep it passing enough current. Or maybe it's something else. Who knows. Brightnorms charger works fine.

But, time to cut my loses. Yes. No poll needed. At this point I will be offering Hitec CG-340 chargers with 12V 13A powersupplies and a charging cable. I'm pretty sure I can offer these for the same cost as my custom charger was going to be: $75, and this is a really good deal, if you do some looking around the web for prices on these two items.

There is also a chance that bwaites and I will be able to go in together on a group buy of Tritons and that we will be able to get a good price on them. So that may be an option for those who want one of the best chargers made to go along with one of the best flashlights made and with the pack to make it rechargeable.

 

[brightnorm]

[ QUOTE ]
js said:
...Brightnorms charger works fine...

[/ QUOTE ]

Perfectly in fact

BN

 

[js]

brightnorm,

Yeah. I know. It must be the smaller box and different heat sink and the pass transistor junction temperature.

But, whatever. The 340 will be a better charger, all things considered.

For now, for the field testers, what I may do is send the 10 hour slow charger circuits which I use to break-in the battery packs. This will hold Catdaddy over until I get the 340's and power supplies. pyro actually has a stand alone charger that should work, but I need to look into that to make sure.

Both packs will be completed today and I'll run some initial tests to make sure they're working fine, and then I'll send them out on Monday. Well, at least Catdaddy's. Still waiting to hear what pyro wants to do. The Triton charger upgrade option that bwaites and I are looking into is still up in the air and we don't know how long it will be before we get an answer on pricing/quantity options.

 

[PaulW]

[ QUOTE ]
js said:

. . . Ah. LOL. Given my playing my cards out on the table for everyone to see method of conducting my modding work on CPF, everyone gets to see all of the screw-ups and false starts and stupidity that goes along with doing something like this. I hope everyone finds it amusing and not annoying. Sorry if it's the latter.

I'm a freak-show, I am. /ubbthreads/images/graemlins/crackup.gif /ubbthreads/images/graemlins/grin.gif /ubbthreads/images/graemlins/grin.gif /ubbthreads/images/graemlins/grin.gif . . .

[/ QUOTE ]

Jim,

Better than amusing, this is a wonderful show. I appreciate your sharing your work. I find it educational (and dare I say, endearing). I am reminded of the awkward and surprising left turns some of my own personal projects have taken. But, not as brave as you, I tend to delay posting on CPF until after I've achieved some measure of success and can edit my findings.

I think comprehensive reporting about progress of a project that many are interested in is refreshing. It give me a sense of engagement in what I feel is important work. And . . . this way we get to cheer you on. /ubbthreads/images/graemlins/grin.gif

Paul

 

[SilverFox]

Hello Jim,

While you are "in the kitchen," why don't you cook up an adapter cable to charge the TigerLight with the Triton... /ubbthreads/images/graemlins/smile.gif

At any rate, put me down for a Triton and a temperature probe when you get that going.

Tom

 

[JimH]

Jim,

I can fully understand all the mistakes and false starts. As an experiment, I have been working on building a minimum cost McGyver light using a small pinapple juice can and the reflector from a busted lantern. I've been working on this off and on for a month. I never realized all that goes into making something like this from scratch.

Your project is several orders of magnitude more complex than mine. It boggles the mind how much planning, designing, protoptyping, procurement, and work that goes into a project like yours.

Keep up the good work, and thank you for continually looking for better solutions, even if it may be at the expense of a lot of your own sweat and labor.

 

[js]

WOW! Thanks for the encouragement, guys. I really appreciate it. The charger thing was a big downer, but I've moved on.

So, I just finished testing out Catdaddy's pack with the new contact plates with the harder silver braze, and just as I suspected, this stuff is several times harder than the tin/silver lead free solder I was using. The LOTC spring doesn't dent it at all, but only makes the contact area directly underneath it more shiny and rubbed. This is the ticket.

Plus, brazing is pretty cool. I like it. It's actually even easier than soldering, really. Plus, using an oxy-acet torch is a bonus. The silver braze is quite expensive, though. I'm glad I don't need a lot of the stuff. One 1/16" diameter 12" long rod costs $9. Yikes!

I've done one run test on Catdaddy's pack. Findings posted in the field testing thread.