Compare 7I43 (LPT) and 5I20 (PCI)

More
18 Feb 2013 01:31 #30221 by aike
Hi!
I tested two cards. The same hardware, the same lathe. The same HAL (only change loaded software).
Software for all cards - SVST_2_7I47.
2 step/dir (X Z axes), 1 pwm (Spindle), 3 encoders (X, Z, spindle) 8 output pins, 19 input pins.
The axis processing speed:
25 mm/min Z axis, 2,5 mm/min X axis.
The processing form cone. All axis moved simultaneously.
100 pcs details were processed for each card.
The sizes were measured and made the table. According to math statistic formulas I found: mediana, sigma.
Results:
7i43 - sigma is 0,032 mm mediana 0,047
5I20 - sigma is 0,020 mm mediana 0,020

The conclusion.
For higher precession use PCI cards. LPT cards (7I43) make more errors in small sizes and low speed tuning.

Please Log in or Create an account to join the conversation.

More
18 Feb 2013 01:58 - 18 Feb 2013 02:02 #30223 by PCW
I think before you make any broad conclusions, you would need to tune both systems optimally
In a closed loop velocity mode stepgen setup, the 5I20 and 7I43 will differ slightly in tuning for two reasons:

1. Reference clock differences: the 7I43 uses a 50 MHz reference clock (+-50 PPM) for its stepgen but the 5I20 uses the PCI clock which may differ as much as 1% from nominal 33.33333 MHz. This means the optimum FF1 value of the feedback loop will be slightly different between the two systems

2. The 7I43 has slower access so the time between sampling the encoder position and applying the velocity correction will be longer, also requiring a slightly different FF1 value
Last edit: 18 Feb 2013 02:02 by PCW.

Please Log in or Create an account to join the conversation.

Moderators: PCWjmelson
Time to create page: 0.178 seconds
Powered by Kunena Forum