Heathkit Hero Brain Transplant

Why? Doesn't the Hero have like 8 8080 processors?

I don't know what is in the VEX, but I suspect it is something cheap and not very powerful.

Rich

Since the Vex uses some non-standard connectors, etc., I - personally - would go with something that lends itself better to parts scrounging. But then again, it's *your* robot --- that means you're in charge & can do whatever you feel like doing !

I wonder if anyone has done a Vex autopsy or tear-down site on the net yet ?

In the mean time, these may be of interest - found by Googling "heathkit hero upgrade" ...

Compared to an 8080?

: I have been looking at the Radio Shack discounted Vex robot and it : occurred to me that it might make a possible "brain transplant" for the : old Heathkit Hero I.

While it's been about 20 years since I built Hero in my High School lab -- I would say no.

Two PICs.

"Why? Doesn't the Hero have like 8 8080 processors? "

The Hero 1 (HERO ONE) used a 6800 series processor.

While I am not impressed with the Vex Robotics System for overall value for your money, I am impressed with the integration within the system.

The power of a robotics computational system lies in how much software there is that exists...hopefully the Vex system would have more available than the ancient Heathkit series. Also the sophistication of the development system is greater.

TMT

If you're looking to replace the CPU of a HERO, I think an OOPic, Basic-X, BASIC Stamp, PIC, AVR, HC11/12, etc. is the better choice. These are more generic processors and are designed for interfacing, and all have a plethora (as in pinatas) of example code and libraries.

Support is probably the greatest for the BASIC Stamp, but as a processor it's one of the weakest of this bunch. It is, after all, the oldest product. (Parallax has new processors coming out, such as the Propeller, that you might also consider, but because they are new there is less existing code.) BASCOM and GCC are common languages for the AVR, and example code and libraries exist for both. The Motorola HC11 and HC12 chips are often used in university-level robotics labs, with robotics books (e.g. Mobile Robots) and cirriculum written around it. And of course there are thousands (literally) of examples of using the PIC -- mostly in assembler -- for robotic tasks. The OOPic and Basic-X both have Yahoo user-groups where a lot of robot-centric code is shared. The OOPic has numerous built-in objects directly related to robotics (reading encoders, ultrasound sensors, PWM output, etc.), as its developer (Scott Savage) is a robotics experimenter.

-- Gordon

I agree with Gordon. You're probably better off working with a more open and expandable processor like the ones he points out. Your choice could depend on your experience with these controllers, what languages you can program in and what you are going to interface to. I'm not familiar with the Hero, so I don't know the number of I/O lines you'll need.

I personally prefer Bascom. It's one of the best basic compilers available... The author, Mark Alberts, provides a ton of sample code and you can find lots of examples on the web.

Eddy

If I remember correctly the 68HC11 has a M6800 core with extra instruction, I/O etc. added If you have the source code of the software it should be easy to port it to a HC11.

Bill K7NOM

: If you're looking to replace the CPU of a HERO, I think an OOPic, : Basic-X, BASIC Stamp, PIC, AVR, HC11/12, etc. is the better choice. : These are more generic processors and are designed for interfacing, and : all have a plethora (as in pinatas) of example code and libraries.

After thinking about good old Hero from a while, and seeing where this discussion went, let me expand on my earlier "no" with some more specifics. Forget about processing power, you need to start with what it physically takes to connect to the Hero. The Vex controller has 16 IO pins which can be digital or analog, and 8 PWM which I don't see readily from the info on the net if they are shared with the other 16.

The Hero-1 has a single drive motor plus at lest seven stepper motors for steering, head, and the arm. So first, you are going to figure out how to interface a stepper controller to the Vex, with at most 16 IO pins, and have enough left for the light, sound, sonar, and voice synth if you want it.

The specific answer to your question would probably come from looking at the schematics for Hero. I seem to recall only ever moving one arm motor at a time, because of limitations in it's built in language. Is there a single stepper driver multiplex to the motors, or a driver for each motor ? How many pins do the driver(s) use ? Do you want to replace all the motor drivers too ?

If you seriously want to retrofit any new controller into a Hero, you are going to want to seriously sit down with the schematics and see what it is needed to drive it. Sure you can use IO expanders, shift registers, etc. But 16 IO is not a lot at all for something the size and complexity of Hero.

If someone is going through all the trouble to replace the brains of a HERO, why bother with a zero or marginal performance increase? Slap a PC104 or Mini-ITX Pentium in there. USB cameras and vision processing.

802.11 wireless networking. Leave the PIC for the tabletop terrors, the HERO has room for much more. Single-board computers are plentiful on eBay and a surprising amount of performance can be had for well under $100.

I had a Hero Jr. and it had a 6803. I wrote a tiny program that allowed a device connected to the serial port to read/write the io subsystems. Then I connected a

Why stop there? Why not just start from scratch and build a new chassis. A vintage HERO in operating condition has some resale value, probably enough to fund a brand new robot from the ground up. No limits. Even in non-operating condition it's often possible to get a few hundred $$$ for it.

-- Gordon

I'd go with a Gumstix

It should work.

-- D. Jay Newman

Bah, those PC104 boards everyone talks about, but no one uses, because they pull so much current the robot can't haul the power line or big batteries around! :)

Check out Fred Eady's site

for PIC or AVR based WiFi solutions and the ATMega128 (or newer 256 when they come out) as a co-processor and you get tons of speed and power for small money and way smaller power needs. A solution like this has the advantage of getting a WiFi connection to your desktop or laptop machine for serious power to crunch numbers and you get the benefit of lower costs, better range (on the batteries) and a form of autonomous abilities as well.

There are so many possibilities these days the mind boggles!

IMO, DLC

cbm5 wrote:

Eh, I'm getting about 20 hours off a 400MHz SBC on a motorcycle-battery sized SLA. With Wifi.