Problem- deep thoughts into encoder logic....

[amontgom]

the answer is 42...... but what was the question???



for some time now i have been looking into a touchy problem..
i have tons of servo driver and amps.. but a mis-matched selection of motors..

i currently have a set of leadshine dcs810 servo drivers .. i like them very much .. however the pittman geared head servos i use with them have a single ended quadrature encoder. which is nice but the dcs810 is differential..
i have been sneaking around the back door a bit by simply not hooking the BAR A, BAR B phases.. and simply tying the a /b to the a/b inputs..

the rub is i get some oscillations .. and with the geared head setup that was eaten up in backlash..

now that i want to do some real work and not *****foot ( talk about touchy content parsers..) with the backlash crap ( encoder 500lpr = 2000ppr (quad) X 17.9:1 gear ratio x 20 tpi screw)
and just drive the screw at 2000ppr(quad) X20 tpi .. i give up resolution but save headaches in factoring out the backlash...


now on with the story..
as i said i have a mis match of motors.. and encoders and drivers..

the servos i am currently playing with are not equipped with encoders..
so i have some 1/4 in shafted dynapars.. single ended that they are..

i mounted them and tested them.. same oscillation...

so i am thinking why not open the cans up and install an inverter on each of the single ended lines and send the inversion back down the pipe..

hey,,, a 7404 should do the trick....

so i pop a can.. and trace the lines.. and glue a 7404n dead bug style to the other chip.. hook up the power and blue wire glue logic together..

( i did this for years at Intel) and viloet` the logic fails...

oh.. the chip is good.. and the signals are there .. sort of..
a rails at 5 volts .the new bar a rails at 3.2 v the same for b..

have i forgotten something over the 30 years i have been working in the field...
i am feeding a clean 5v signal directly into the input of an inverter ...

and getting a little over half of that out..
since i am using a 7404n.. should it be a 74ls04 instead?


< codgering>
now i am certain that there are a few people out ther that will say ... "just go out and buy the right servos stupid.."

that is not the point of my exercise.. first .. i'm broke.. no money .. 3 jobs.. and little time to play. much less sleep..

i buy food and pay my #@$%#@ taxes...

i have the basic parts to make due with; even if i have to modify things a bit... i am too old and set in my ways to casually go out and drop a grand per axis to get the toys i like to play with..

<insert senility>

if i can't buy it .. i'll build it.. if i can't build it .. well then i don't really need it... if i don't need .. i don't want it.. errr something like that...

damn the torpedoes and hang the bankers.. and screw the lawyers..( they had it coming anyway( they'd do it to you.. you just have to do it to them first..))

i had the brains once.. they're just a bit rusty..


whew' enough codgering back to business..
</codgering>
< recover a momentary lapse of reason>


the problem is a simple one .. vcc is ideal at 5.vdc ground is just that..
the issue is the chip set selection or should i throw out the inverter in lou of good old fashion DTL logic... there is about 11 ways to skin this cat..

just looking for some other ideas..
http://www.cnczone.com/forums/images...s/rolleyes.gif han opening up my dusty, hollow, thin wallet.

( sorry for the rusty brain bit .. AL- the man.... we have both seen better days. and i have been playing around in the salt water too long )

[vger]

I hate when that happens...
Some of your oscillation problem my be in the tuning of the servo amps / system. Typical output for an SN7404 (VOH) is 3.4 volts. Same for the LS and S verstions. I had to think to remember what the SN stood for..
Might try a dual op-amp that will work on 5 volts single ended supply to get your inversion to go to the supply rail.

Steve

[Al_The_Man]

You should not need an inverter in the first place, It might pay to look in the encoder and see what kind of driver, the only thing to watch is that some are open collector, but if the output switches now from 0to ~5v then there is most likely a line driver there already.
If you want to use a RS422 differential driver, you could use a DS9638 or 26LS31.
The encoder resolution does not have to be matched to the motor, as Steve said, it is more like tuning causing oscillation.

BTW, in spite of what you often see on sites like this, the correct terminology/definition is quadrature x4.
IOW, the basic pulse count IS the quadrature pulse.
Al.

[amontgom]

i knew that.. but was lazy..
looking at the spec sheets.. to drive them .. it requires 7 vdc in order to get 5 v out..
.. but there are ways to do just that.. my source is only capable of handling about 50 ma though... i must find a better power solution...

i think i've got one though..

it has been a while since i played around on the inside of these.. usually it is just a wire match-up thing.....

this is an old encoder.. ( personally i would rather be playing with a HEDS series encoder) those i have down to a science..

Originally Posted by Al_The_Man You should not need an inverter in the first place, It might pay to look in the encoder and see what kind of driver, the only thing to watch is that some are open collector, but if the output switches now from 0to ~5v then there is most likely a line driver there already. If you want to use a RS422 differential driver, you could use a DS9638 or 26LS31. The encoder resolution does not have to be matched to the motor, as Steve said, it is more like tuning causing oscillation. BTW, in spite of what you often see on sites like this, the correct terminology/definition is quadrature x4. IOW, the basic pulse count IS the quadrature pulse. Al.

[bharbour]

Originally Posted by amontgom i knew that.. but was lazy.. looking at the spec sheets.. to drive them .. it requires 7 vdc in order to get 5 v out..

I would be pretty surprised if 7V does not let the smoke out of a real TTL version of s 7404.
My ancient Motorola TTL data book says 5.5V is the maximum operating voltage for LS and I expect straight 7400 would be similar.

If you use a CMOS version, 74HC04 it should swing pretty close to 5V for a 5V power supply. I'm with Al on the 26C31 though. It's a real differential driver and they are cheap.

BobH

[amontgom]

Originally Posted by bharbour I would be pretty surprised if 7V does not let the smoke out of a real TTL version of s 7404. My ancient Motorola TTL data book says 5.5V is the maximum operating voltage for LS and I expect straight 7400 would be similar. If you use a CMOS version, 74HC04 it should swing pretty close to 5V for a 5V power supply. I'm with Al on the 26C31 though. It's a real differential driver and they are cheap. BobH

7 vdc would be the max as my fairchild part datasheet says..
i did semiconductors at Intel.. the rails can take a bit.. 10 dvc will let the smoke out for sure... maybe.. it depends on the wafer location..
and actual bus trace width ( it varies due to production characteristics) ..


i am also with al on this one.. and i am searching through my parts bin..