Re: [CAD_CAM_EDM_DRO] Servo Dynamics 1525 tuning?

shawncd@... wrote:

>Does anyone have experience tuning the Servo Dynamics
>SD1525 series servo drives? I'm having problems. This
>is the first time I have tried to tune analog servo
>drives. Any info or direction on where to go from here
>would be great.
>
>

Not specifically that model, but the SD amps are totally classic
analog servo amps. I have some amps that are made after the same
model. What problems are you having? Do you have a digital
storage oscilloscope? If not, are you using EMC (it essentially
has this built in)?

I've included the servo tuning info for my servo amp. Much of
it should apply.

Jon

----------

Setup and tuning procedure for Pico Systems PWM Servo Amplifier

Connect +12 Volts to card edge pins U and 17, -12 Volts to
pins S and 15, and ground to pins 21, 22, Y and Z. Check
current draw for any shorts, reversed chips, etc.

Step 1 is to set the frequency of the PWM oscillator. Use an
oscilloscope at pin 1 of U22(A) or the junction of R28 and R29.
You should see a trinagle wave with about a 4 V P-P amplitude.
Set it to 100 KHz. Note that the signal should be symmetric
about the 6V level, so it should go from about 4V to 8V. if it
is not symmetrically balanced around the 6V level, make sure the
+12V DC power is correct, and then check the 6V reference,
at U11B, pin 1. It should read exactly 1/2 of the 12V power
source.

Once this checks out, turn Tach and Command gain pots (R61 and
R92) to minimum (full counterclockwise until they click).
Turn Current Limit (R68) to minimum also. Set Current gain
(R39) to half way, by turning CCW until it clicks, then CW
10 full turns. Since no power is available to the full
bridge switch, there can be no current in the output sense
resistor R22. So, use this opportunity to zero the current
error amp. Note that in this condition, the current error
amp (U10A) is almost open loop, since its output is barely
fed back to the input. So, it won't null easily, just get
it close. Use a scope or DVM to read voltage at U10, pin 1,
while turning R105 (Current Offset). When the DVM swings from
rail to rail (-12 to +12) you just crossed the null.

Now, null the velocity error amp (U11A) by reading voltage
at pin 1 and set R69 until it crosses the null, same as
above.

Now, you have to hook it up to close the loops! So, you
need the DC tachometer connected to pins 18 and V. Connect
the DC motor to pins H, J, 7 and 8 for one side of the
armature, and D, E, 4 and 5 for the other. A variable
DC power supply for the power switch is a big plus at this
stage. It is best to not hook up the computer, and get the
current loop and velocity loop stable before adding the
positioning loop. So, enable the amp by driving pin 20 to
+12 Volts, and then start turning up the variable power
supply, with the + connected to pins A, B, 1 and 2. (Ground
the - output to 21, 22, Y and Z, the common ground point.)
Observe the current feedback signal at U8, pin 6. After
turning the power supply up to about 3-6 Volts, observe
if any current is indicated by a non-zero voltage from
U8, pin 6. If so, adjust the current offset, R105, and
see if the current indication changes. the motor may
also begin to turn. Set it for zero current.
If oscillations develop, turn R42 clockwise, to increase
loop compensation. Then, turn the current limit, R68,
about 2-3 turns CW from the zero where it was. If
current develops, readjust velocity offset, R69, to
null it out as best as you can. Then, start turning
the tach gain up with R61. At some point, you should
notice that the motor either resists efforts to turn it,
or starts to spin wildly. If it spins wildly, reverse
the tach leads (or motor leads). If the motor resists
turning by hand, then the tach is correctly phased, and
your attempt to turn the motor is sensed by the tach, and
a compensating current was delivered by the amp to
prevent it from turning. If oscillation develops, turn
R46, velocity compensation, CW. Once you get here,
you can start turning the variable power supply up and
look for instability. (Increasing the DC power supply's
voltage acts to increase system gain, because the same
pulse width delivers more voltage and current to the motor.)
When you get it stable at, say, 24 Volts, you can try
connecting the computer, and turning up the command gain
control (R92) to get desired servo performance.

Note: Don't try to run a motor using one +12 Volt
power supply both for the logic and the motor power.
It WON'T work! I already know that! The motor load
will affect the power supply's output, and the amp
will go haywire!

Finally, you have to accomplish servo tuning and
loop compensation including the CNC control, which
sets up a whole new ball of worms. I have found,
through many hours of frustration, that it is best
to tune the servo amp without the computer, and then
DON'T change any setting on the servo amp (with the
possible exception of command gain) while getting
the positioning loop to run right. Hopefully, your
CNC control gives enough control over its software
servo loop that you can damp out any vibrations
or instabilities. If you will be using the NIST
EMC software, I know how to tune this. If not,
you pretty much need a Digital Storage Oscilloscope.
This stuff just happens too slowly, and is non-
repetitive, for an analog scope. but, once you get
it moving the machine, you can do a good deal of
tuning with your ear. Any resonances or instability
are easily heard.

[Non-text portions of this message have been removed]