Thanks

I will be needing the source code for the IBC eventually. No rush though.

Where might I find out how to make the programmer?

the PIC i replaced it with was actually much slower lol, 5mhz instruction rate, but it has hardware PWM and hardware serial ports so it spends 99% of its time doing nothing.

The reason i did a PIC rather than AVR is i know PICs and to do AVR's would have meant learning a new thing. A serial version of the IBC's code would have loads more time for doing other stuff as you could make the PWM out use the timer and just preload it to expire when one of the PWM's needs changing.
_________________
Mechanical engineers build weapons, civil engineers build targets

You know you want to make that NSW controler Jake...
_________________
( •_•)

( •_•)>⌐■-■

(⌐■_■)

YEAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAH

looks like the irf1405 is on the out
all hail the new king IRFB3207
3.9 RDSon Vs 5.3

Its not free though
input capacitance is 7800pf Vs 5400

price is apparently about the same (from the post i read about it)
"we paid $1.07 recently for 500 qty of the 3207 vice 0.94 for the 1405."

there are some better ones but they are apparently $$$
_________________
Mechanical engineers build weapons, civil engineers build targets

What's the inherited issues with the higher input capacitance?
_________________
https://www.halfdonethings.com/

it can take longer to switch states and cause issues with inductance and ringing.
the fets do alot of heating when they are in the "half on" state so its good to get them out of that as fast as possible.
the added capacitance also means you get higher currents to fill the gate which can make inductance problems worse.

its not *drastically* higher though and the fet is rated to 70V so i think a drop in replacement for any OSMC based board (rs80, IBC etc), but probably best to do them all at once not just one at a time. It should make the TVS diodes (which are set at 50V) actually do a decent job protecting the fets, rather than the 50v diode on a 50v fet.
_________________
Mechanical engineers build weapons, civil engineers build targets

Higher input capacitance results in a longet time to turn on/off (switching losses.) The most effecient system is where your switching losses equalls half of the total losses, (the other half being running losses.)

If Rds is lower (running losses) and Capacitance is higher (switching losses) then you need to swich at a lower frequency to make these losses equal which will increase effeciency and therefore reduce heat.

***EDIT***
Sorry Valen, we must have been typing at the same time.

Last edited by Fish_in_a_Barrel on Wed Nov 15, 2006 3:17 pm; edited 1 time in total

or switch harder
_________________
Mechanical engineers build weapons, civil engineers build targets

Hey there guys,

Im having a bit of a brainstorm at the moment and Ive
come up with a weird concept and want to see what the tech-heads
think in the way of pro's and cons.

Im planning on designing a modular control system from scratch
that uses fibre-optic cable to control localised power boards with
pwm signals.
Eg central controller sending a pwm pulse train down a fibre-optic
link to a (FET) power-board attached as close as possible to the motor,
in order to shorten the leads carrying high-current switched loads. One link
and one power-board per motor. The power boards are basically a
fibre-optic reciever feeding a pic/picaxe which drives a Fet boosted h-bridge.

Bear in mind, my first bot is still only in CAD, and Im only just starting
to study Fet designs now.(OSMC project)

The advantages I see are as follows...
1) Central processor/reciever opto-isolated from current-bearing circuits.(damage prevention)
2) Blown power circuits replaced without full nervous-system replacement.(ease of maintenance/repair)
3) Easier sectional testing (ie test single motor without needing full system power-up).
4) Ability to completely (electrically) isolate receiver from motor power supply.
5) Upgrade power circuits without full nervous-system replacement.(say in case of motor upgrade)
6) Lower internally generated RF, due to fibre-optic signal wires ?!?(advice please)
7) Lower internally generated RF, due to shortened switched-power leads ?!?(advice please)
8 ) Lower power-lead losses, due to shortened switched-power leads ?!?(advice please)
9) Reduced weight, due to shortened switched-power leads and lightweight signal leads ?!?(advice please)

As for dis-advantages, I cant see any on paper but my ears are open.
And please dont count complexity of the design as a dis-advantage as
this project is (at this time) largely a learning process for me to come
to grips with fet power designs.

I'm mostly looking for advice on the feasability of the distributed
powerboard "concept" and fibre-optics as a weight-saver.

Thanks guys
_________________
"The future is not set. There is no fate but what WE make."
........CEO Cyberdyne Systems

one disadvantage I can think of after just reading this is that you'll have to buy fiber optic cable at the desired length because you can't cut fiber optic cable......this is because the edges formed from anything other then a perfectly square cut will lead to loss in transmission signal and may cause interference....the best way to get around this is to buy a flexible SPDIF cable at the desired length. SPDIF cables are used in audio setups where u don't want to lose any sound quality through the wires.

I will add more as I think about it more...

Would it be possible for someone to email me the source code for the IBC....got mine in the mail a few days ago and want to play around with it a bit and see what I can change maybe.

Thanks

andrew

I dont have time for a detailed reply now, but in combat-robot equipment, a key-concept is "Failure Points"

ie. The more things that can go wrong, the more that probably will.

Seperate PCB's and mounting/enclosures, Fibre Optic Connectors, Multiple Power supply points etc etc.

Also keep in mind the end-users. Most combat robot builders are not electronics guru's. Basic Electronics is usually within their grasp, but not 1 in 10 would know more about a FET than "its the bit that goes boom."

If you are building it for your own education fine, but if you are seriously intending it for use in combat, look at it the same way the army looks at a gun - If you cant give it to a grunt with an IQ less than their bicep size, throw it in the sand, drown it in the ocean, leave it in the jungle for a year or two and then pick it up and fire a bullet, then its not going to be reliable in combat.

More later
_________________
Great minds discuss ideas. Average minds discuss events. Small minds discuss people

Last edited by Spockie-Tech on Tue Nov 21, 2006 10:13 am; edited 1 time in total

we thought of the same thing.
We are doing an ESC now for serious brushless lol. The problem is you still need to get boost power to them all anyway. We have settled on opt-isolators so we can drive the high side (and will use the same on the low side to make it even) Then to protect the boost supplies we are going to diode things up the wazoo so if one bit fails we don't loose the power supply or the other fets (atm power supply is looking like a stack of rechargeable lithium coin cells)
_________________
Mechanical engineers build weapons, civil engineers build targets

Ok, replies...

Andrew12 - Ive been playing with fibreoptics for a few months
now and have been able to manufacture a reliable connector
using an Led at one end and a photo-transistor at the other.
By drilling a small hole in the led and transistor i can cut the
cable easily and just jam it in the drilled holes, secured externally
with blu-tak, cable cut with a hobby knife.
I have been getting reliable 20-30kbit per second transmit rates
around the house & workshop , so I'm sure it will manage a
5-10khz pwm signal fine.

SpockieTech - I agree with the failuree point concept, but shouldnt
it be balanced with the "dont keep all your eggs in one basket" concept?
Ive no practical experiance with IBC's or any other controllers, but
wouldnt it be common for lose function to the whole IBC if a fried
motor blows the fets or bridge? A distributed design allows the bot
to continue to hobble onwards, and the localised powerboard could
be replaced quickly if you have spares on hand.
Also, this is really for personal use only so I dont need to worry about
any end users besides myself.

Valen - Glad to see someone else warped enough to thinks into every
corner of the box. Im not planning for brushless at this stage so Im
unlikely to need boosters, just a main-power rail running the length of
the bot. Good luck with your design too, and rather than diodes, have
you thought about slo-blow fusable links or thermal circuit breakers
on the outputs instead?

Anyways, thanks for the advice all, and everything everyone says is
being taken into consideraton, no matter how badly I phrase my replies.
As this is still early stage in the design for me, so no rush on the replies
but please keep them coming.
Catch you'all later
_________________
"The future is not set. There is no fate but what WE make."
........CEO Cyberdyne Systems

Using optic fiber for reducing weight may be an advantage, and I know the optic fiber in machinery works very reliably.

Panasonic use this tech knowledge in there high speed surface mount loaders, glue dispenses, etc


the main problem is that you still need to supply power to all the boards.
That could mean that there is no weight saving.

what effect that would have on the RF I don't know.

As said, it would increase failure points, but is a very reliable technology.


PS. for the curious the pic is of a Panasonic MV2F high speed chip shooter.
weighs around 3 ton
It has 15 axis
has 12 heads that are connected to a turret
places a SMT chip every 0.1 sec
Can hold 300 8mm feeders. on 2 feeder platforms
fully loaded feeder platform weighs 300Kg and can move 25mm in <0.1 sec

and how do I know all this. I helped Installed and serviced the first one that arrived in Australia
_________________
It's all about the destruction.

Last edited by Ajax on Wed Nov 22, 2006 8:18 am; edited 1 time in total

that is cool