Friday, November 26, 2004

Echo

Halllloooooooooo to all you out there in blog land.

well looking at the last blog (what was it 2 months ago now?) some things have changed
and in the course of things they were photographed.

Anyway, we never really did a post fight teardown of Plan-B after the last event, it got put on a shelf and left to mature.... yeah thats it.
We have been working on a new bot, (no not plan-C, current name for this one is Plan-B Alloy) which was going to have many added funkynesses but it got left a bit late, so rather than turn up with a dodgy bot (TM, besides robort fills that spot nicley) we are giving plan-B a last minute going over. Plan-B Alloy still will happen, hopefully we will have it or a derivative of it ready for robowars in melbourne, though it may have a new and improved wepon over and above what we were building for it.

Currently Plan-B has another layer of triaxial glass, and a new internal layout so we should be able to fit 2 booster packs in, wether we do series or parallell is yet to be decided. (testing will show us the path ;->)


realised i didnt show all you people in blog land the outside of our bot after the last change so here ya go.




and here is a magical mystery photo.



for reference the diamiter is ~100mm


ooh yah and also I have been getting some 2.4 Ghz radio Lurve.
I have successfully got radio waves flowing through the ether, admitadly they are in the non funky direct mode of the laipac modules but soon my prettys soon ;->
btw if you are writing to laipac support state your question in the first line.
they dont bother reading more than one line of emails it seems.
went back and forth 3 times with this guy before he actually answered the question i asked in the first email. and each of the 3 times he told me that what i had actually done couldnt be done. (and then went on to reccomend i try doing the thing i had asked for help on in the first place. grr)

in the end i managed to get a sample config word out of him (note to google laipac TRF-2.4G )

A configuration word for shockburst mode
1 byte of address say 0xAA, 1 byte of data and 8 bit CRC, 1 channel receive,250Kbps,2410MHz,0dBm output power
120-bit Laipac RF config:
0x08, 0x08, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
0x00, 0x00, 0x00, 0xaa, 0x21, 0x4F, 0x15 for reciver/0x14 for transmitter

Note well young friends this is untested. I have NFI if that will work or not.

Sunday, September 12, 2004

miscelanious stuff

well making some progress on all fronts.

current theory is that our armor is too good.
it needs to be tough enough to stand up to everyday wear and tear (sorry and 90% of cobra) but theres no point worrying about trying to stop nick at full speed cos that just aint going to happen. As a result Plan-B is now nekked, no side armor, looks like the frame weighs a little under 2kg sans armor, (was 3.5 with)

we hope the new armor we are hoping will weigh about 1kg.
going with a foam core then some triaxial glass which should be really tough and we are going to put some kevlar over that.

CADing up 3 new disks for Plan-B, going to get it plasma cut from bisalloy 80 ($77 including material so ya cant really complain about that) shooting for 1.3kg total mass with the same energy as the current disk, i'll write me some software to do the KE calcs on it tommorow. heh we were driving to a plastic place (they also sell transmission parts...) to get some plastic for our washing machine (heh how many people do you know lathe up replacement parts for the suspension in their washing machines ;->) when i saw a large factory type place with the name "oxy cut" (106 long st smithfeild 9757 3833 MOQ $70 +GST) and being the mad roboteers we are we pulled in. They werent madly into robot wars but seemed keen enough to help (may have helped we had the bot with us). The guy at the plastics place on the other hand was pretty much useless when we asked about armor. FMS were good to us with the kevlar again, and we got the special scissors for it too which will help. Special thanks again to DIAB for the foam samples we are making our armor core out of. hint bringing your bot along when you speak to people can garner you some respect, possibly some glossy action shots too?

mmm been having some problems with Plan-C weapon but we think we have got it worked out now. When we went from the first prototype to the second we didnt think we changed anything and in truth we didnt change anything inside the wep, but we changed the supporting equipment we didnt realise that this would make a difference and we were coming up with all these weird ass theories on what was going wrong and generally getting frustrated, wasnt untill we realised we had made the change, (and a little bit broke one of the critical parts) that we did a quickie fix and got some 3/4 power shots out if it *booyah*!! till the quickie fix part broke.
but point is we found the source of the problems and so Plan-C is back on track, though we may need some more engenering work on the firing mechanism to get it up to speed.

ooooh this being secretive bit is tough innit lol.

oh almost forgot
the bit you have all been waiting for PHOTOS!



cross section of the armor
shamefully we wernt truly australian and the cardboard didnt come from a beer carton *hangs head in shame*



nekked Plan-B, its actually much smaller than it seems with the sides on ay



on the scales, mmm lightweight bot chasis, keep in mind this still has the bearing for the weapon in it (note pounds in black on the outside, grey inside is kg)



some Plan-C action, thats 3mm AL with the hole in it, backed with a roughly 6kg rock, at about half power, might be a little higher, still no tip, so that punched the hole with a 16mm flat shaft. heh tis going to be lethal at full power with a armor peircing head on it.

Thursday, September 02, 2004

post battle stripdown (of the bot you dirty minded people)

well we pulled plan-B apart today
somewhat aided by the total absence of electrical power making this the most constructive thing we could do.

anyway it looks like we actually got damaged *gasp*
worst of all it looks like sorry came the closest to doing anything dramatic....
*hides head in shame*

we took a hit to the front that fractured the belt guard and cut through the first kevlar the glass but was stopped by the second layer of kevlar.
even so it still managed to damage a belt, put a split in it about 5cm long down to the fiber backing in it.

there was also a black mark where it looks like the belt has been pressed into the glass



and here is where nick took a little chunk out of our wheel too


gratitious top view


did a weigh in of the various bits too today
ItemWeight (kg)
Shell3.45
Lid0.65
Kevlar Inserts0.22
Wep Motor + pully0.63
Left Drive assembly1.15
Right Drive assembly1.15
Wepon1.83
Drive Motors1.1
Misc foam bits0.05
Ni-MH pack1.45
Electrics0.5
Bolts and Brackets0.2
Wepon Booster0.15
Fan0.2
9v batteries0.1
Ni-Cd pack1.15


each individal wheel weighs 750 grams, we want to bring that down to 250 and take the fan out. might rework the front guard around the wep to protect the belt some more.

Saturday, August 28, 2004

marryong event - first blush

well we found the weakest point in our robot
9v battery clips need tape when the batteries themselves arent secured.

also in the lessons learnt column something i noticed for the builders of other bots, you have to "stress relieve" your connections when you are going up against spinners.
now i dont know the KE of nicks bar but when it hits a bot i wouldnt be suprised if for a very short time the innards of your bot wernt pulling 200G's
that 50 gram wire from your motor to speed controller whatever will now weigh 10Kg, this makes a mess of most plugs and isnt so crash hot for solder joints.

basically what you have to do is where you have your wires coming into a plug/join you have to mechanically hold the wire behind the plug and use that to take the load.

our ring terminals seemed to work pretty well in that regard especially as more than once when we took it out of the ring some of the nuts had actually come undone slightly (which is why ring terminals kick soo much ass in my book lol, ok so they are a pain in the ass to work with in this high change environment but they sure are rugged) but it would have been better if we could have stress relieved them (lesson for Plan-C lol)

Also for the last half of our fights it seemed that we had lost forwad drive speed, the IBC/servo travel needs some tweaking, we could spin at full speed but full forward stick did basically bugger all. kinda neutered our, "suck basilisk into hitting the wall / sombody elses bot then attack him" plan.

on a good note though, perfect radio all day, no jitters, no nothing smooth controll from start to finish, all on one set of 9'v batteries so i'm well pleased with that. still might take that LED off though, it pulls about 1/3rd of the remaining current draw of the IBC.

bits it seems people want
FET switches, esp for basilisk lol. a fet switch for that would be smaller and lighter than that hugeass thing you had in there lol.

Temperature monitors, i'm just "spitballing" here but some kind of light/series of lights you can put on your bot to tell you when your motor/batteries are between "well done" or "carbonising as you watch"
adjustable trigger points, 2 levels of warning per sensor "hot" and "overtemp" say? should be fairly cheap, 2 sensor setup ~ $20-$25 possibly less with a source of high brightness cheap LED's
anybody interested?

OptoIsolation of the IBC
with us it seemed to *really* help with radio noise problems, doubled radio range even with the switchmode PSU in the IBC. Basically you plug the servo leads from the IBC into it and then it plugs into your Rx unit (which must run off batteries,(note you need the BEC of the IBC *on* to power its half of the opto)) $25 should cover that pretty easily


ToDo on Plan-B, grind out the delaminations and re-glass the damaged sections, all pretty straight forwad there, little more difficult on the front top where basilisk had a bit of a gentle gnawing, (if we bother fixing that part back to full thickness that is, its not structural out there and its just a bit of a pain to fill the corner back up.)
mount the battery eliminator eliminator batteries more securly than jamming them in behind the motor.
mount the opto isolators perhaps.

Thursday, August 26, 2004

Plan-B weapon testing

well just to do things completley different we have a robot finished more than 8 hours before an event *gasp* yes friends a shocking part with tradition I know but all things must make way for progress.

anahoo

quickie update
heres Plan-B making sure our welding practice piece has plenty of holes in it to weld up. For the curious its largley 3mm alumium i think the tube bit is 6060 and the plate is 5000 series.
FET switch is fully opperational and seems to work quite well.
Check the forum topic in the marketplace if you are interested in getting one them.


pre-attack


attack 1


attack 2 impact 1


attack 2 impact 2

Tuesday, August 24, 2004

mmm optical

no progress on the 2.4 ghz radio units (havent touched them yet, need to get the electronics PC working)

however optical is showing some decent progress ;->
had a 20W fluro pointed directly at the reciever from 1M and it rx'd low speed data with no noticible effect, putting a hand 10cm in front of Tx and seperating rx by 4 meters had some impact but a 3x sampling would still read it out ok

no RFI in optical mmmmmmm
small hitch for the users of cheap digicams though, they are sensitive to IR so it shows up *really* well.

IBC boost

hooked up the new seperate power supply for the IBC
looks like most of the radio noise problem is solved.

now (with no connection between IBC and RX) there is no difference if the ibc is on or off, 26 meters range both ways (small top hat and tx antenna down)

driving a motor we get about 15 meters range (tx antenna up), so thats double what we had with the opto and (scary) steel box, and well about 15 times what we had with the standard setup.

Photo

I think i'll try optoing the rx from the IBC again and see what range we get but I doubt it'll improve things much there shouldnt be any spikes from the motor making it into the power supply. I'll just try the BEC off and running off RX battery again see if that helps things at all.

looks like some more RFI screening on the motors would give us a further range boost, anybody got some handy tips on suppression?
hmmm i wonder about a sheilded can around the motors mwaaa ;->

now to make the wepon driver (with inverter)

testing and screening IBC

yesterdays testing

setout to measure the affect of IBC noise on the radio setup and see what could be done to improve it.


we are hopefully going to use a "tophat" style antenna which should give us good range in a small package inital testing qualitivly proved the idea (ie it worked fine) now we are putting some numbers to it and seeing what can be done to improve it.



Rx batt voltage 5.4V, Tx bat voltage 17.4V



Testing was conducted using a servo and commanding it to traverse smoothly from stop to stop. Signal strength was rated out of 5

5 - perfect signal
4 - slight interuptions - mostly noticablle only in the sound of the servo drive
3 - intermittant drive - servo visibly twitching
2 - poor drive - servo still travels but with much inteference
1 - unusablle - servo movement eratic
0 - no servo movement


Results

Antenna Orientation IBC Range Signal Str Notes
Series : 1 Tuning antenna, no sheilding on IBC, Photos 1,2
90 Off 15m 3 See Photo 1 for detail
0 Off 27 5 there seemed to be some blockage near the gate signal strength dropps off quite quickly there see photo 3
0 On 15 3 signal degridation is abrupt, 5 - 0 over about 1.5 meters
90 On 3 2 possibly an optimistic result, the rx had better line of sight than normal
Series : 2 Alumium foil wrapping photo 4 sheild connected to -ve Vbat
0 off 31 3 baseline
90 on 9 3
0 on 25 3
Series : 3 Steel Shield (basic wrap) sheild connected to -ve Vbat Photo 5
0 off 37 3 baseline
90 off 18 3 baseline
0 on 9 3 the steel posed a large LOS obstruction which might inhibit it signifigantly
90 on 9 3 same signal strength with 0 and 90 deg antenna
Series : 4a Steel Shield (basic wrap) sheild connected to -ve Vbat, reciever is connected to IBC for the first time in the test series, note BEC is *off* rx is running on its own battery, driving a rs550 drill motor
0 off 1.5 2 some LOS issues here
90 off 1 2
Series : 4b Steel Shield (basic wrap) sheild connected to -ve Vbat, same as above using stock antenna photos 7
0 off 6 2
Series : 5 Complete steel sheild, IBC optically isolated from reciever, tophat antenna 7
0 On 15 3
90 On 12 3



the optically isolated complete steel housing gives us the best isolation from the RFI from the IBC.

note that with the transmitter *off* if the ibc was switched on next to the reciever the servo started jerking around all over the place, it was reduced with the steel sheild but not eliminated.

Sunday, August 22, 2004

bit more looking into the RF noise

looks like most of the noise is coming from the power converter.
i'm reading almost a volt of noise just holding the probe of my scope near the big cap there at around 250khz, the top of the cap is dancing around with 25V peaks (charge pump is working lol, only 14V batts)
The converters cap (and diode) also has a fairly pronounced ringing at about 8 mhz which is what I was seeing before.

looks like there is also some 12 mhz clock noise or something close leaking out along the power lines to the RX but its kinda small (~.3V) but still way more than it would expect to get from a battery.

reference note
beat freq of clk + ring is 4mhz

now to find the IF of a superhet RX unit
hmm looks like it will be around the 450khz range?
roughly double the freq of the switching freq, not a good thing.

IBC radio problems

well folks its that time again.
I have gotten off my ass and have some news for youse.
As some of you may know plan-B lost on the second day primarily because of radio trouble (nothing to do with our compeditors or nothin).

Anyway I have been moving on making a 2.4 Ghz 2 way digital radio setup so we can do our own plan of funkyness on it.
now whilst i have the hardware for that wired up there is still some bug fixing to do before it works.

Now for the interesting part...
because it dosent look like we will get the new digital radio going for marryong 7 we started looking at ways to get better signal with our existing radio. To that end we are putting in a base loaded top hat (capacity hat) antenna. Now we were just looking for a proof of concept so we wired it up and i did a range test inside our house (tx antenna down).
We got about 30 meters range... before i ran out of house.

This is how it was set up.


Note the extensive use of the TLAR principle (that looks about right) in the antenna design and coil construction. Point is it worked and more than well enough for robot wars.

now heres were it really gets interesting

we put the reciever and antenna in the bot (no changes) powered the bot up and promptly found our radio range (as judjed by IBC and servo) was around 2 meters. After much fiddling and tuning we got about a 10% signal at 4-6 meters with perfect tx and rx antenna (tx extended, and rx out of the bot and nicley presented)

now heres where it really really gets interesting

i had a bit of a brain wave, left everything where it was, radio, bot etc, just unplugged the ibc and plugged the rx battery back in. range went up to about 8 meters ( at 10%). then it got interesting. we switched the bot off. range went back up to the point we didnt have enough land to get the servo to twitch noticablly with the antenna down. Worst we got was a slight vibration when i had the antenna down the radio on the ground and pointing horisontally away from the reciever. Then we switched the bot on, servo went spastic for a second then died.
radio range was about 3 meters with tx down.
Now keep in mind that the radio and bot *were not connected* by moving the radio away from the bot radio range went up.

Now for some less interesting stuff, unless your an electronics nerd in which case its probbly hella frustrating. feel free to skip to the conclusions.

heres a bad photo of my scope

This is taken from ground to the *antenna* ie this should be basically only be seeing random low level noise, instead we have a .1V signal blasting across the spectrum *and* this is after encasing the IBC in a grounded foil sleve.

it seems to have 3 or more frequency components and 4 distinct stages
the 2 pulses are 4 usec apart
the high freq noise has a period of .125 usec and there is another signal at about .8 of a volt that i cant clearly determine the frequency of but it isnt much higher.

quick quackulation
the high frequency noise is at 8 mhz, this is pretty damn close to what is mixed with the input freq to make the IF(intermediate frequency) (in our case the radio has an 11 mhz xtal in it so some of those higher freqs may be almost directly on this mixing freq)
the AVR runs off a 12 mhz xtal so that shouldnt be it.

the pulses are spaced at .25 mhz and occur at about .1mhz so i dont think they will be doing anything much.

also we noticed that the input line from the rx sits high but gets pulled low at 1Hz

Conclusions
The IBC is spewing RF noise like a football team after a bad curry and 2 hours on a dirt road.

possible solutions.
a thick grounded AL box may help to attenuate the RF being spewed out.
run the reciever off a battery.
opto isolate the RX from the IBC and keep the IBC leads inside the box.

we are just about to test the IBC battery connections and motor leads for noise.
hopefully this is mainly confined to the logic side of the board.
(personally i'm pointing my finger at either the SEPIC converter (possibly) or perhaps more likley the HIP chip's charge pumps.)


P.S
just did some more testing
without motors connected and radio off we saw simmilar noise with the pulses and hf at the same freqs coming out the motor lines, there was a small <0.01V of that signal on the battery lines.
i'm thinking its probbly the charge pump in the HIP chips more now
i'll do some more playing later and see what i come up with