NickReiser's Biped

Let's Make Robots Projects Showcase
81

Nick & Anyone else who can help

Looking at the pic of your bot standing up, can you tell me how wide the AHC-01 is between the hip servos?

I’m wondering whether I could mount the top half of the body (from the waist rotate up) onto the top of a Lynxmotion Biped Scout!

…just thinking…if I turn the biped scout round so that the knees are now forward facing, then mount a torso in the middle, the nose of the scout now becomes a sort of backpack with battery and processor and I have a humanoid with 6DOF legs!!

82

Hey, I take no credit; give all the kudos to Jim.

:smiley:

Arg!!!
Pitch… roll… these airplane terms hurt my head.
It sounds like you’ve got it right, though.
In essence, a humanoid has the same setup as the Biped Scout, except there’s no hip rotation servos.

That’s a good question about the topmost servos on the Scout…
I never really used them when I was building the Scout’s gaits.
I’d like to hear if anyone has found a good use for them.

How does a humanoid turn?
Well, to turn in place, they take one step foreward, and drag their foot back to where it started.
The friction of the move takes care of the turning.
The distance of the first step directly correlates with the degree of the turn.

Turning while walking is more complicated, but it’s accomplished in the same fashion.

Before each step is taken, the forewards foot is dragged back slightly.
This creates a gradual curve.

I love the 5645’s.
Economical, powerful, and, most importantly, reliable.
Their whine isn’t loud (although I’ve never used other servos, so I have no comparison).
And, they seem to have “magic smoke” protection.

I wouldn’t dream of building a biped with anything less (although I’d think of using 5995’s, if I could afford it).

In my opinion, there’s no place on the biped that can be sacrificed, except for the head.
Thus, I skipped the head servo entirely, for now.
:smiley:

The arms are needed for punching, lifting, getting up once it falls over, and the list goes on…
They’ll need to be very powerful, because they’ll be moving in a larger range of motion than the legs will, and usually maintaining difficultly extended positions.

I agree.
The PC is an amazing processor that I have at my disposal to use.
However, it’s important not to underestimate lesser technologies.
In the case of the onboard micro, I’m using it to do a bit of simple math that the FSR balancing will require.
I doubt that I’d get responses that were fast enough to save it, if I were to stream that round trip through an 11 MB/s (at absolute best) semi-reliable connection.

Keeping the entire balancing mechanism onboard provides an extra check and balance in the case of a transmission failure.

The section that you’d mount onto is 1" wide (also the size of the hub that it sits on).

I can’t see why you couldn’t mount it directly onto the Lexxan.

It should require a bit of drilling, but you’ve got the space (where you attach the nose).

I had actually debated doing so, myself, but I was concerned that the slight wobbliness of the original Scout would turn into a not-so-slight wobbliness if I added more weight and height.

83

Nick

Thanks for the update and I think I caught up with you in another thread. My intention is to build the Biped scout with HS5645 servos and get used to these (oh and try to build a walking gait with PCT) and then buy the brackets adn do something similar to yourself - the only bit from the Biped scout I intend to keep is the waist assembly giving the humanoid 6DOF legs.

I think I’ll stick with HS5645 in the arms rather than HS5475 assuming I can get enough cash together, and I already have 2 HS645s for the head

I didn’t mean to put down other technologies other than PCs and agree that keeping the balance on an onboard micro is the way to go - I use Atmel Mega Micros for onboard processing as they are easy to use and program (www.dontronics.com) but if you have the WiPort then there is an incredible capacity to offload a lot of processing to the PC - and the PC could also do things like map building from the rangefinder data!

keep us all updated as it’s looking good!

84

You just had to go and say that, eh?

:laughing:

Turns out, things are starting to look bad.

I’ve been having real battery issues.
I broke up my three BAT-08’s and soldered a square bolt from the Meccano Erector Set onto the + end of each.
This way, I had the male protrusion that would properly fit each AA battery holder that I’m using.

I placed two cells behind each shoulder and one cell in front of each shoulder.
I used household electrical wire (a slightly larger gauge than the regular battery connector wire) to connect the relocated cells in series.

At first, I was very conservative with the solder, using as little as I could, and being sure not to heat the batteries more than absolutely necessary.
However, after I realized that it wasn’t working (more on that in a bit), I went back and used liberal amounts of solder at each connection.
I then checked the resistance of each connection with my meter and kept adding even more solder until they were all 0.1 ohms (which seems to be 0 ohms, on my meter).

I went so far as to even solder the batteries directly onto their battery holders.
I also added an extra cell, and I’ve been recharging and swapping the batteries frequently, to rule that out.
I’m using a seperate 9V for the logic.

Here’s what’s wrong:

I measure the voltage of the batteries when it’s not connected, and it’s 7.48V.
I connect it, the servos start holding a position (the biped’s laying on the table, so there’s little current being drawn) and it’s 7.44V at the terminal.
I tell him to stand up, he holds the position, but the voltage drops to 6.09V, at the power terminal.
I take a leg and try to push it out of position, and the voltage quickly drops until it’s well below 3V, and the servos fail to maintain a position.

Here’s what I believe is happening:
There’s something wrong with the batteries, or there’s some weak connection that I’m not finding.
I’m thinking that a weak connection would be like a kink in a hose, which would allow less current through.
The servos would demand more current than they are being supplied, and the meager amount of current flowing through the hose would be sucked dry as soon as it’s produced.
This would explain the voltage drops that I’m seeing, right?
Of course, my meter’s amp measurement doesn’t seem to work, so I can’t be sure.

I do know that it didn’t fail to maintain position so easily before (and I was only using 5 cells, back then).
I can’t remember if the voltage dropped significantly, back when I was using the complete packs, though.

I’m going to go try my old-school meter and see if I can get an idea on the amps.

I’ll post back in a bit with those results.

Let me know if you guys see a “DOH” that I’ve missed.

:stuck_out_tongue:

85

If there’s a store that builds battery packs nearby they’ll probably be willing to spot weld tabs onto your cells for a nominal charge.

86

What’s a BAT-08?

Still cringing from the description of what you did. :open_mouth: It is very difficult to solder directly to batteries that do not have solder tabs. It can be done, but you usually have to scrape any plating off the metal ends first, then tin the ends, then solder the wires using reflow. While the voltage is 3vdc on the terminals, measure the voltage across each connection, but do it on the battery end, not on the wire. When you find your missing 3+ volts you will find your bad solder connection. I recommend you start over. The way it is you are likely to fix one connection just to have another flake out. Adding more solder to a bad connection will not help. Remove all the old solder. Get a small file and get down to bare metal, then solder everything back up…

87

Sorry, I meant BAT-03.

I was a bit fried, yesterday.
(no pun intended :stuck_out_tongue:)

Thanks for the replies, guys.
Just to clarify, the voltage at the terminals should be 7.2V when I’ve got everything connected properly, right?
Even when I put the servos under a bit of strain?

Thanks for the checking method, Jim.
I’ve been going berserk trying to figure out a good way to debug this.

You said that I should file down to bare metal… but I’ve taken off all of the shrinkwrap, so everything is bare metal.
I removed the spot-welded tabs that connected them in the pack by the simple expedient of pulling hard.

Do you mean that I should score the metal so that the solder has something to grip, or am I already down to the “bare metal” that you were talking about?

88

If you use a small file on the ends of the batteries it will change from a bright chrome finish to a copper colored metal. This will accept solder, but the chrome (plated) metal will not. Funny that you pulled off the tabs. They usually accept solder well. :open_mouth:

89

Well, I got real frustrated with unsoldering everything, filing, and resoldering, so I picked up a couple packs of Energizer 2500mAh NiMh AA cells (1.2V).

I plugged them in, made sure everything was connected well, fired her up, and finally started programming gaits again.

After less than a minute, though, my bot collapsed.

I checked the batteries individual voltages after taking them out of the bot, and they were down to .89V (started at 1.24V).

So much for “high-capacity batteries”.
Has anyone else used these batteries?

I’m not thinking that this should happen, so quickly, even with store-bought batteries.
After all, my others are 1600mAh, and they hold their charge much longer.
Perhaps I have a short somewhere?

I’m thinking so, too.
When I measured the voltage across one of the chest cells, it was -.59V!
Yes, I had the red probe on the male battery head, and the black on the negative side.
I tripple-checked to make sure that I didn’t solder anything backwards, but it’s not like it’s complicated wiring.
They’re definitely connected in series.

Can anyone think of why a battery would be showing negative voltage?

I’m going to try peak charging them and give it another go.
Perhaps they weren’t charged out-of-the-box.
I can’t imagine why they’d be at 1.24V when they’re not charged, though.

Well, as I said, I’ve gotten tired of trying to resolder those batteries with a very crappy iron.

Is it possible to arc-weld Lynxmotion NiMh batteries?
The label on the Energizers says not to short circuit them, or they’ll blow up.

So, would it be fine to ground the positive while I weld a tab to the positive, and vice versa for the negative?
That way, the charge isn’t passing through the battery.

90

I’m not sure if arc welding will work with such thin metal. I’m thinking it would simply burn a hole on the electrodes. Are you talking about an industrial arc welder? Even the small arc welders that Home Depot sells I think would be to much for battery terminals.

Did you try using Flux when soldering on the filed ends of the battery? Also, because the battery has a lot of metal surface, it takes a while for the solder to heat up the terminal to the point where it will flow and stick. Flux can help a great deal, and a nice hot iron is important. Another problem is the type of soldering tip you are using. If the soldering tip had a small tip, it will not transfer the heat well at all, in fact, small tips can actually cool down on contact causing the solder to solidify! What you need is a nice fat chisel type of soldering tip to transfer heat better.

91

Nick,
Guessing that you are using NiMH, in answer to your question about how a cell can be displaying a negative voltage, this can happen when a cell is deeply overdischarged and wired in series with several other cells. Keep in mind that you can damage cells by excessive temperature, such as may have occurred when soldering wires to them. A short summary of the situation is the single cell has been completely discharged but current continues to flow (it is continuous, remember) through the string of cells and starts to charge the cell in reverse polarity. Here is a link to a manufacturers cells that we use in several commercial/industrial applications, and some of this information is around pages 15-16 or so.
hardingenergy.com/pdfs/NiMH.pdf
I don’t imagine they sell to end users much so hopefully there is no harm in my posting the link here on Lynxmotion’s forum. They do have a lot of very pertinent information about specifying, usage, charging, and handling of several battery chemistries.
:mrgreen:

92

Mike:
As long as it’s not going to blow up in my face, I’ll give it a try.
It’s a small mechanic’s arc-welder that doubles as a… either TIG or MIG, I can’t remember those acronyms (whichever is the weaker-bonding one).

By now, I’ve got quite a few dud batteries that’ll no longer hold a charge, so they’ll do as guiney pigs.
:stuck_out_tongue:

If I turn the amps and the spool down to the lowest setting, I can easily write my name on a car’s hood (don’t ask me how I know that :laughing: ).
So, the thinness of the battery shouldn’t be too much of a problem.
Although, I believe that the batteries are alluminum, so I’m not sure that it’ll accept stainless steel being injected into it, very well.

Well, if you don’t hear from me for a week or so, call in a team of bucketeers to collect my remains.

:stuck_out_tongue:

About the soldering:
I used pre-solder acid, as well as flux.
And, yep, I used a flat tip.

Eddie:
Hmm… good to know.
I’m still confused as to why one cell would be drastically below the others, when they all started at exactly 1.24V…
But, that’s probably just a freak occurance, or a faulty cell.

93

I think the MIG welder bonds weaker, but don’t you need the special wire to weld aluminum with a MIG welder? I’ve only ever welded aluminum with a TIG welder or a straight electric.

I’ve used those same energizer batteries in my submarine. They work fine for me, but are a bit on the heavy side.

When you soldered those batteries together, what type of solder did you use? I’ve never been able to get standard solder to stick to aluminum, I’ve always had to use special flux and solder. I think I got it from cheapbatterypacks.com.

94

Nick,
All the batteries in a pack are never going to be exactly the same. As you use a pack, especially if you use it hard and have a relatively inexpensive charger, the cells are going to get further apart. Batteries are chemistry based and temperature plays a large part in the equations of how they work so it is yet another example of where practice differs from theory. In theory NiMH packs should never develop a memory and never go bad. In practice, well they can a lot better than their predecessors NiCd, but they still have their own set of issues. Here is a thought, if on an individual basis you could discharge each cell in a pack to a known state and then recharge each cell according to the manufacturers directions for charge rates, termination, etc., then you could build a battery that may last a very long time, especially if you graded and matched the cells to begine with. The cost of the charging system and the time it would take to “charge” a pack would be prohibitive in all but a few very unique markets. So in practice things get lumped together, averages made, if you read how they design NiMH cells you will see they made compromises in the cell design to be overly certain certain user caused problems were difficult to occur, and what you wind up with is a ‘commercial’ product that is more difficult to destroy but has a lower watt density and a shorter product lifetime.
I doubt seriously that when you received your pack you had bad or faulty cells in it. Reading about your experiences attempting to reconfigure the pack and soldering to the cells, my guess would be one or more of the cells got overly warm and then when you tried to use them they just did not have the expected capacity remaining for either using or charging. I believe polatity reversal only occurs on discharge. Once one cell becomes damaged the rest of the cells become over stressed though and the whole dominoe thing soon follows. If your charger allows it you may want to consider charging each cell individually once your pack is built just to lessen the likelyhood of another cell being radically different than the rest and going out the same way as the one you discovered.

95

Another thought, if you have them in your area, is to go to your local Batteries Plus store. They stock all kinds of batteries, and I’ve had them build up packs for me in the past when I couldn’t find them at a local hobby shop or was too impatient to wait for mail-order. Some of the stores actually have a small shop with welding and soldering tools just for this kind of stuff.

96

It is very comon for a NiMH battery to show a negative value if it has been sitting for a while. This will fix itself when it has been charged. Don’t freak out…

Forget about the arc welder! Nothing good will happen. Jim…

97

I have had a bad experience with these shops. The batteries are usually of very poor quality, and anything they do as “custom” will be very expensive. You can use any premade battery pack, just cut the solder tabs in half, don’t pull them off. The “solder” tabs are easy to… um, well, solder to! :open_mouth: Jim

98

Sigh…

Right you guys are.

Arc welding doesn’t work, nor does MIG.
It’s just not hot enough to do aluminum.

I’ll just get me another couple of complete packs and split the tabs of one in half.
The other will be kept as a back-up in case it still doesn’t work.

Well, I suppose learning the hard way is always best, in the long run.
:stuck_out_tongue:

99

I thought there was a special sort of solder for use on aluminium.

100

I found it. It was this stuff:cheapbatterypacks.com/main.asp?sid=533180&pgid=misc Bottom of the page.