Re: [CAD_CAM_EDM_DRO] Re: Position counter stop updating....Still there....

[John Sheahan <jrsheahan-at-optushome.com.au>]

On Sat, Sep 14, 2002 at 08:14:21AM -0400, Babar wrote:
> 
> 
> >Do you ever disable the 26LS31?  (which is what I
> know the 
> > SN75174 as in this part of the world)
> > Better not to here.
> 
> John, We already have connected the Active High Enable
> signals of 75174
> (pins 4 & 12) to Vcc=+5V
> please tell us what do u mean by disabling the 26LS31?

Hi Babar
If the 75174 ever gets disabled, then it can be susceptible to noise 
when hi impedance.     You have it always enabled, so thats not the
problem here.
(in some applications, bidirectional or multidropped, enables are
 actively switched) 


> 
> >Is the pair terminated (ie ~100 ohms across it) at
> the load end
> >  well worth doing! 
> >  (the pair is the two wires connecting the the A and
> /A output pins
> >   through to the far end)
> 
> No we havent terminated the pair of each encoder
> outputs.
> And if we need to terminate it then the best place
> which i understand is at the inputs of receiver ic
> "26LS32" present on STG-II. meaning we
> will have to solder the resistors on STG-II??
> 

the receiver input is the best place.  You could consider trying 
maybe with just one channel to see if it helps. 
In the plug would work about as well, but near the receiver end of the
cable.



> >Is the pair run as a twisted pair?
> >  helps, but depends on your wiring.
> 
> The cable that is coming from encoders is a 13 core
> sheileded cable. out of which 8 wires are being used.
> the Sheild is connected to system earth + ground. this
> sheild is connected to ground only at single side.

thats the way to do it.

> Now this cable which has differential signals come to
> a board from where it goes to STG-II using simple
> ribbon cable (not shielded).
> This board also contains connections for DAC outputs
> of STG-II.

ribbon cables are pretty crummy - but a diff pair is remarkably noise 
tolerant.  Shielding the cable is unlikely to help - coupling here
tends to be between adjascent wires.   I ripped a ribbon in half
(along the cable, stop near the plugs) and fixed an odd one once - but
that was lower level.


for mine - I'd try terminating both pairs for a single channel of 
the encoder (I missed the start - this is a quadrature encoder).
I'd make sure both wires of the diff pair are connected end-end
(diff circuits work about the same as single ended when one wire is
 broken if not terminated.)
and if its still broken I'd start wondering about the software driver. 

sorry this seems not to be helping
John




> 
> >Do you use a 26LS32 or similar to receive the
> differential signal?
> >  essential
> 
> No we use 75174 ic at encoder side and since STG-II
> can accept differential signals therfore we didnt use
> any Receiver ic.
> 
> >do you have any test equpiment (eg oscilloscope) to
> measure 
> >noise levels?
> 
> 
> We have osciloscope. but We have rarely seen noise on
> osciloscope.
> although the stg2diag shows big errornous values
> following is the respone of stg2diag utility:
> 
> [root-at-localhost bin]# stg2diag
> Servo To Go Board IO address = 0x200
>          +1.000          +1.000          -2.000
>          +1.000          +1.000      -65538.000
>          +1.000          +1.000          -2.000
>          +1.000          +1.000      -65538.000
>          +1.000          +1.000          -2.000
>      +65537.000      +65537.000      -65538.000
>          +1.000          +1.000          -2.000
>      +65537.000      +65537.000      -65538.000
>          +1.000          +1.000          -2.000
>      +65537.000      +65537.000        -258.000
>          +1.000          +1.000          -2.000
>          +1.000          +1.000      -65538.000
>          +1.000          -3.000          -2.000
>          +1.000          -3.000      -65538.000
>          +1.000          -3.000          -2.000
>          +1.000          -3.000      -65538.000
>          +1.000          -3.000          -2.000
>      +65537.000     -131075.000      -65538.000
>          +1.000          -3.000          -2.000
>      +65537.000     -131075.000      -65538.000
>          +1.000          -3.000          -2.000
>      +65537.000     -131075.000        -258.000
>          +1.000          -3.000          -2.000
>          +1.000          -3.000      -65538.000
>                                                       
>                        
> Looking at this response we see big errornous values
> after which
> the next value is same as before the erroronous value.
> Isnt this shows that this problem is perhaps something
> to do with the software?
> I meant if the actual problem is with errornous counts
> originating 
> form encoder, then why these counter do not retain
> that errornous
> value? 
> 
> Les you wrote:
> >With computer ON, encoder supply ON, and servo amps
> OFF (at supply)
> >does the encoder count still smoothly change when the
> motors
> >are turned by hand with all axes?
> 
> >If so when the servo amp is turned on does the
> problem appear
> >with all axes? (of course it should have 0 input when
> turning it on)
> 
> >Does the problem disappear if the servo amps are on
> but not enabled?
> 
> >Is the behavior noticed with BOTH EMC and stgdiag
> programs?
> 
> We used external supply this time thinking that this
> might help us. isolating this encoder power suply from
> the machine. Our computer was
> ON servo amps and all the machine supplies were OFF.
> What we saw on STG2DIAG noise such as shwon in above
> response. We ran the EMC,
> run the motors with hands and didnt observed Counter
> stucking problem.
> 
> Here please note that when the drives are powered, pc
> is ON and EMC running, When we run any axis drives
> using Testing dialog box then for 
> upto 3 volts counters dont stuck (Although we see
> continous errornous big values in stg2diag)but for
> more than this 3 volts counters stop. 
> To do this check we run stg2diag in parallel with EMC.
> 
> When Stg2diag program is running it never stucks
> counters but gives
> errornous big values after some random period of time
> which is around
> 1 to 3 seconds.
> 
> >Can you download the DOS version of stgdiag and run
> it from a >temporary
> >DOS boot floppy? If so what are the results?
> 
> We'll try to do that but could you please explain,
> what is the 
> difference in running stg2diag in linux and running
> that utility using
> MSDOS bootable floppy??
> 
> >Is the 75174 IC enabled By +5 to pins 4 and 12 (if it
> is 16 pin dip
> >package)?
> 
> YES, We already have connected the Active High Enable
> signals of 75174
> (pins 4 & 12) to Vcc=+5V
> 
> >Now after the IC you should have 8 wires to each
> encoder. For me
> >there are 3 separate shielded  twisted pairs for the
> encoder outputs
> >and +5 and a separate ground. The shields for the
> twisted pair are
> >hooked to star ground near the computer but LEFT
> UNHOOKED at
> >the encoder end 2 meters or so away.
> 
> Yes We have 8 wires after 75174 but all the wires run
> in a single 13 core sheilded cable whose one end near
> pc side is connected to Earth+Ground.
> What is star ground?
> 
> >Can you get a scope to measure the signals? Channels
> A+, B+, and I+
> > should be high and A-,B-, and I- low or vice versa
> relative to >ground.
> 
> >I still use a 1 meter ribbon cable in the control box
> to go from a 
> >breakout
> >board to the card. It is a special ribbon cable that
> uses twisted 
> >pairs.
> >What do you use?
> 
> Yes we measured the signals using osciloscope and have
> found that differential signals are always opposite to
> each other.
> 
> We too have a ribbon cable which is one meter in
> length but is simple one (found mostly in computers to
> connect hard drives)
> 
> >What model of encoder are you using? What is the
> count per revolution?
> >How many rpm do you want to turn it?
> 
> >What is the base address of your stg card? Have you
> tried changing it?
> 
> >How are your other card dip switches set?
> 
> The X and Y axis Drives have 200 lines encoders. Index
> is working
> The Z axis has 250 lines encoders. (Z axis encoder was
> damage so we had fitted another encoder.) Index isnt
> working alignment is out. but we'll fix it.
> 
> We dont know how many  RPM should we run it at?
> 
> Base address of our STG is 0x200 and we didnt try to
> change it.
> Can changing this address may affect counter stucking
> problem?
> the DIP switch have default configuration:
> 
> Analog input range J2 = +/-10V 
> DAC latch on interrupt J3 = latch on interrupt setting
> Watchdog J5 = Unlatched
> Watchdog timeout slection = 333 msec
> Source for CLK2 J7 = OSC/2
> 
> 
> 
> Jon you wrote:
> >My servo amps have LC filters on the output, so that
> no more than 1 
> >Volt of the
> >100 KHz chopping waveform is allowed to escape the
> servo amp.  What 
> >brand of
> >servo amp is used in this machine?  I don't have to
> ask if the output 
> >is filtered, it
> >almost certainly is not.  Unfortunately, adding an
> output filter may 
> >not be easy,
> >at least without retuning the servo amp.  But, you
> would want to put 
> >inductors of
> >probably several hundred microHenries in series with
> each motor output 
> >lead
> >from the amplifier, and then a capacitor of perhaps
> 5 uF to the DC 
> >power supply
> >ground from each motor wire.  Without knowing the PWM
> chopping 
> >frequency and
> >the motor current, it is hard for me to be more
> specific about this.
> 
> 
> We are using servoamps are from WestAmp.
> Peak current : +/-22Amp (for 500 msec)
> peak voltage : +/-100VDC
> Continous currnt +/-10 Amp
> Continous voltage +/-92VDC
> Switch frequency (constant) : 5Khz
> Frequncy responce : 500 Hz minmium
> Signal Input voltage : +/-10 V(typ) +/-50V Max
> 
> And we are not sure about the presence of LC circuit
> in the drive.
> 
> 
> >One possibility is to change
> >the wiring location of the encoder signals.  If the
> cable shield is 
> >grounded at both ends, this
> >is incorrect practice.  The encoder cable should be
> grounded only at 
> >the STG card end.
> >Using an external power supply may be
> counterproductive, you may do 
> >better with the
> >internal power source from the STG card.
> 
> >Especially since the stg2diag ALSO shows the same
> problem when the 
> >servo amps are
> >on, EMC is not to blame.  And, in fact, nobody else's
> software will 
> >run, either.
> 
> >Jon
> 
> No the encoder cable sheild isnt grounded at both
> ends.The encoder cable is grounded only at our
> breakout board and from there it goes to STG-II via
> ribbon cable.
> 
> We were using STG-II supply to power encoders but
> since we used 75174 
> we observed the Voltage droped to 4.3V. so we now are
> using external 
> +5V PSU.
> 
> Please comment
> best regards
> 
> Babar
> 
> 
> 
> __________________________________________________
> Do you Yahoo!?
> Yahoo! News - Today's headlines
> http://news.yahoo.com