Re: [CAD_CAM_EDM_DRO] Stepper Wire ?
Posted by
Matt Shaver
on 2001-06-19 00:29:51 UTC
Hugh & Denise Currin wrote:
couple of other prior posts on this subject, but didn't take the time.
I've been doing a lot of thinking and research on cables to use for cnc
control applications. There're several different circuit types in a
typical control:
1. Drive Electronics -> Motor
Most modern motor drives, whether stepper or servo, use pulse width
modulation (PWM) to control the motor current. Basically, the transistor
switches that connect the power supply to the motor leads are turned on
and off at some fixed rate (usually in the 2kHz to 20kHz range depending
on the inductance of the motor windings). The current is increased or
decreased by increasing or decreasing the ratio of the "on" time to the
"off" time (this is a somewhat simplified explanation). Since the
transistor is either fully on or fully off, power dissipation, and thus
heat generation, in the transistors is minimized. As a consequence of
this switching, the motor cables carry high currents with a square
waveform. While the base frequency is in the low kilohertz range, square
waves are rich in harmonic content and if they're not carried in a well
shielded cable, will radiate signals in the RF range with significant
field strength throughout your shop (this is a bummer if you listen to
AM radio). So, the main consideration for motor cables is keeping the
radio frequency interference (RFI) controlled. Since bipolar stepper
drives require four conductors arranged as two twisted pairs, the cable
I like best for this application is:
Alpha 5620B1602 (5620B1802 is 18g if you can't do 16g...)
http://209.208.232.71/pages/322.CFM
or
Manhattan M9746020 (18g = M9748020...)
http://www.manhattancdt.com/catalog/Instrument121.htm
This cable consists of two 16g twisted pairs, each individually foil
shielded (with a drain wire) and then the whole cable has an overall
foil shield (also with its own drain wire). Since most servo
applications that I've seen carry the tachometer signal in the same
cable as the motor current, this cable would work for those applications
as well.
2. Encoder -> Computer
Encoder signals are very low power since the input impedance of most
quadrature decoders is at least 10k ohms and the signals only swing from
+5V to ground. Since these are all low power digital signals, crosstalk
isn't an issue, and most encoders I've dealt with have differential
outputs to maximize noise immunity. The main goal, therefore, of encoder
cables is to shield them from outside electrical interference radiated
from conductors that carry high currents. My picks are:
Alpha 2214C
http://209.208.232.71/pages/215.CFM
or
Manhattan M39132
http://www.manhattancdt.com/catalog/Control104.htm
or
Weico 7708
http://www.weicowire.com/specpage.asp?nGroupID=370
Alpha does make cable specifically for servo motors and encoders but the
servo cable costs about $7 per foot and the encoder cable costs about
$14 per foot so I don't think I'll be using any of this...
For the gory details see:
http://209.208.232.71/pages/106.CFM
and
http://209.208.232.71/pages/107.CFM
3. Limit, Home, and Other Switches -> Computer
I don't think shielded cable is necessary for these applications. In
the controls I build, I use a 100 ohm pull-up resistor from the TTL
input to +5V and the switch is wired from the TTL input to ground.
Since the circuit's impedance is so low (even if you used 1k pull-up
resistors), there is almost no chance of interference causing the bit to
toggle (unless you wrapped the cable around an arc welder...).
4. Parallel Port (or other TTL Outputs) -> Step, Direction, Other TTL
Logic Inputs
This is basically the same situation as the encoder cable. If everthing
is going to one small area, a cut off printer cable is good. If you need
to send bits all over, I like 2 conductor microphone cable (from Radio
Shack if you need it quick at any price...).
5. AC & DC Power Cables
Other than ensuring that they're big enough, and rated for the max
voltage, that's about it. If your motor drives are leaking RF energy
back into the AC or DC supply lines, you need to filter this out before
the cables exit the shilded enclosure to prevent them from radiating a
bunch of hash and wrecking the radio reception ;).
wires soldered to the connector shells at both ends. They usually have
all 25 wires, but it's straight multiconductor cable, not twisted pair,
since they don't know your application. Cheap printer cables usually
only have 18 wires since pins 18-25 are ground. An IEEE-1284 cable has a
couple of the signals in twisted pairs with a ground wire from the pin
18-25 range that's ground on a parallel port. I think you'll get along
fine with a regular old cheap cable.
Some control cable is only rated for 90V. I'd use a piece of a PC power
cord. Some of these are shielded if you need that (it says this on the
jacket), and you'll get a free ground wire too!
you need 20', I can just send that to you gratis. I'll look tomorrow and
see.
Matt
> Does anyone have a source (preferably on-line) for shielded wiring?Thanks for bringing up this subject again. I'd meant to reply to a
>
> Also, if anyone has thoughts on my scheme below please let me know. :-)
couple of other prior posts on this subject, but didn't take the time.
I've been doing a lot of thinking and research on cables to use for cnc
control applications. There're several different circuit types in a
typical control:
1. Drive Electronics -> Motor
Most modern motor drives, whether stepper or servo, use pulse width
modulation (PWM) to control the motor current. Basically, the transistor
switches that connect the power supply to the motor leads are turned on
and off at some fixed rate (usually in the 2kHz to 20kHz range depending
on the inductance of the motor windings). The current is increased or
decreased by increasing or decreasing the ratio of the "on" time to the
"off" time (this is a somewhat simplified explanation). Since the
transistor is either fully on or fully off, power dissipation, and thus
heat generation, in the transistors is minimized. As a consequence of
this switching, the motor cables carry high currents with a square
waveform. While the base frequency is in the low kilohertz range, square
waves are rich in harmonic content and if they're not carried in a well
shielded cable, will radiate signals in the RF range with significant
field strength throughout your shop (this is a bummer if you listen to
AM radio). So, the main consideration for motor cables is keeping the
radio frequency interference (RFI) controlled. Since bipolar stepper
drives require four conductors arranged as two twisted pairs, the cable
I like best for this application is:
Alpha 5620B1602 (5620B1802 is 18g if you can't do 16g...)
http://209.208.232.71/pages/322.CFM
or
Manhattan M9746020 (18g = M9748020...)
http://www.manhattancdt.com/catalog/Instrument121.htm
This cable consists of two 16g twisted pairs, each individually foil
shielded (with a drain wire) and then the whole cable has an overall
foil shield (also with its own drain wire). Since most servo
applications that I've seen carry the tachometer signal in the same
cable as the motor current, this cable would work for those applications
as well.
2. Encoder -> Computer
Encoder signals are very low power since the input impedance of most
quadrature decoders is at least 10k ohms and the signals only swing from
+5V to ground. Since these are all low power digital signals, crosstalk
isn't an issue, and most encoders I've dealt with have differential
outputs to maximize noise immunity. The main goal, therefore, of encoder
cables is to shield them from outside electrical interference radiated
from conductors that carry high currents. My picks are:
Alpha 2214C
http://209.208.232.71/pages/215.CFM
or
Manhattan M39132
http://www.manhattancdt.com/catalog/Control104.htm
or
Weico 7708
http://www.weicowire.com/specpage.asp?nGroupID=370
Alpha does make cable specifically for servo motors and encoders but the
servo cable costs about $7 per foot and the encoder cable costs about
$14 per foot so I don't think I'll be using any of this...
For the gory details see:
http://209.208.232.71/pages/106.CFM
and
http://209.208.232.71/pages/107.CFM
3. Limit, Home, and Other Switches -> Computer
I don't think shielded cable is necessary for these applications. In
the controls I build, I use a 100 ohm pull-up resistor from the TTL
input to +5V and the switch is wired from the TTL input to ground.
Since the circuit's impedance is so low (even if you used 1k pull-up
resistors), there is almost no chance of interference causing the bit to
toggle (unless you wrapped the cable around an arc welder...).
4. Parallel Port (or other TTL Outputs) -> Step, Direction, Other TTL
Logic Inputs
This is basically the same situation as the encoder cable. If everthing
is going to one small area, a cut off printer cable is good. If you need
to send bits all over, I like 2 conductor microphone cable (from Radio
Shack if you need it quick at any price...).
5. AC & DC Power Cables
Other than ensuring that they're big enough, and rated for the max
voltage, that's about it. If your motor drives are leaking RF energy
back into the AC or DC supply lines, you need to filter this out before
the cables exit the shilded enclosure to prevent them from radiating a
bunch of hash and wrecking the radio reception ;).
> I have a small opto-isolation circuit which is probably best right at theGeneral purpose extension cords are foil shielded overall with the drain
> computer parallel port? One side powered from the computer and the other
> from the 5V supply in the driver box. This isolation circuit is to be
> connected to the driver box via a D-Sub 25 pin extension cable. I think
> these have all 25 conductors through and are paired & twisted where some
> cables are not?
wires soldered to the connector shells at both ends. They usually have
all 25 wires, but it's straight multiconductor cable, not twisted pair,
since they don't know your application. Cheap printer cables usually
only have 18 wires since pins 18-25 are ground. An IEEE-1284 cable has a
couple of the signals in twisted pairs with a ground wire from the pin
18-25 range that's ground on a parallel port. I think you'll get along
fine with a regular old cheap cable.
> Then, inside a Sescom enclosure, use the 22 AWG twisted shielded pair cableMake sure the cable you're using for 120Vac is rated for that voltage.
> for 5V supply, step & direction signals. I can use some of the 18 gage
> double twisted pair to get motor winding power out of the box. But then
> should the 120V AC and higher DC power be in shielded cable? If so I could
> run them in a cable of one 18 gage twisted pair.
Some control cable is only rated for 90V. I'd use a piece of a PC power
cord. Some of these are shielded if you need that (it says this on the
jacket), and you'll get a free ground wire too!
> If this seems OK, then I can use close to 100' of the 18 gage dual twistedI think I have a bunch of 18g stranded, twisted, foil shilded cable. If
> pair, and 100' of 22 gage single twisted pair. However only 10' or 20' of
> single pair 18 gage. I suspect I can't buy less than 50' or 100'
> spools. Anyone have a spool and wants to sell 20'? Or maybe a couple of
> people to go in on part of a 100' spool?
you need 20', I can just send that to you gratis. I'll look tomorrow and
see.
Matt
Discussion Thread
Al Schoepp
2001-06-16 17:17:27 UTC
driver wiring questions
Jon Elson
2001-06-16 17:30:09 UTC
Re: [CAD_CAM_EDM_DRO] driver wiring questions
Tim Goldstein
2001-06-16 19:19:44 UTC
RE: [CAD_CAM_EDM_DRO] driver wiring questions
Hugh & Denise Currin
2001-06-18 17:11:00 UTC
Stepper Wire ?
Matt Shaver
2001-06-19 00:29:51 UTC
Re: [CAD_CAM_EDM_DRO] Stepper Wire ?
ptengin@a...
2001-06-19 04:02:42 UTC
Re: [CAD_CAM_EDM_DRO] Stepper Wire ?
Jon Elson
2001-06-19 17:10:46 UTC
Re: [CAD_CAM_EDM_DRO] Stepper Wire ?
zephyrus@r...
2001-07-10 14:36:13 UTC
Re: Stepper Wire ?
mariss92705@y...
2001-07-10 15:31:45 UTC
Re: Stepper Wire ?
Eric Keller
2001-07-10 17:14:53 UTC
Re: [CAD_CAM_EDM_DRO] Re: Stepper Wire ?
Jon Elson
2001-07-10 20:57:58 UTC
Re: [CAD_CAM_EDM_DRO] Re: Stepper Wire ?
van halcomb
2001-07-10 21:20:26 UTC
Re: [CAD_CAM_EDM_DRO] Re: Stepper Wire ?
Randy Gordon-Gilmore
2001-07-11 22:35:38 UTC
Re: [CAD_CAM_EDM_DRO] Re: Stepper Wire ?
mariss92705@y...
2001-07-12 07:51:08 UTC
Re: Stepper Wire ?