Advanced Search

Here is the 7I96/7I96S input circuit

Notes:

1. The quad optos are AC input type

2. The solid GND symbol is input common

 
 

I installed Linuxcnc on an 8GB Pi5 using this file since I want to use a MESA board, it looked like the right choice.

www.linuxcnc.org/iso/rpi-5-debian-bookwo...25-01-27-0140.img.xz

Everything booted up well(but there was no initial setup wizard like I have seen some people mention), and the install looks good, but I don't see wifi anywhere or network options of any kind. I tried a USB wifi dongle, and that didn't change anything and unfortunately, ethernet is not an option for me. Does anyone have any tips for getting wifi working? Should I have used the other .iso? Thanks!

tommylight wrote:

HEXFET, TrenchFET, Epitaxial FET, .... damn i am getting old and this time i am not blaming it on being 3:15AM right now, there are many more types of MOSFET's, and almost all have flyback diodes, you can spend days sifting through catalogs to find any that does not

 

I have been making servo amps since about 2004, and had some issues with them blowing up under certain conditions.  I used some beefy 56 A International Rectifier HEXFETs.  Using a digital scope, I was horrified to see that the body diodes in these transistors did not start to conduct until about 10 us after they were forward biased by up to 12 V!  So, I had to add in some ultra-fast diodes across the transistors to prevent this.  The issue was that when the high-side transistor was sourcing current and then shut off, motor and filter inductance would pull the source of that upper transistor down to large negative voltages, and cause the gate driver chip to turn both transistors on at the same time, leading to fireworks.  The  extra diodes solved this issue.
Jon

Any clue why adding fast diodes would cause them to blow up?

IT WILL NOT!
Unless you reverse polarity.
I wonder where are you getting such wrong info from and why are you so stuck at diodes ...
-
Are you aware that miss wiring the stepper motor will blow the drive?
The "fuses" you mention, are NOT fuses, they are resistors used for current limiting in the drives.
Also mentioned before, do NOT cut the drive power on the DC side, ever, meaning, if you wire the drive and motor and use any power supply that is already powered, then touch/wire the DC power to the drives = they are dead in most cases.

Your hal file shows a software encoder rather than the 7I73's encoder,
A software encoder sampled at the servo thread rate will limit you to less than
1000 counts per second. You would need to use the 7I73s encoder counter to go
faster. From the 7I73 manual:

"The 7I73 can track encoder inputs up to a 50 KHz maximum count rate (4X modes).
Maximum count rate is 12.5 KHz in 1X modes and 25 KHz in 2X modes. These maximum
count rates will be reduced by quadrature distortion (deviation from 90 degree phase
difference between A and B ) so practical maximum count rates are normally somewhat
lower.

Short answer:
Yes, you can use encoders, even 2 encoders per axis/joint, or an encoder and a glass scale, both wired to Mesa/LinuxCNC,
and
Yes, you can have one encoder (usually the one on the motor shaft) wired to drive and a glass scale wired to Mesa/LinuxCNC
and
Yes, you can have new drives that take step/dir pulses and have their own encoder and add glass scales wired back to Mesa/LinuxCNC
and
Within reason, pretty much however you like/need.
As for backlash, the best and easiest way is the last version from above.

unknown wrote:

The issue is that interface is setup for both dhcp and a static address.

Yeaaa man, sorry i missed that.

so to be clear, I need to make this folder. it is not one that should already exist?

I get the general principle of inputs in your MESA cards, I understood. You have made a very simple and easily understable manuals with you cards, btw.

But I am trying to solve simmilar inputs possibilities in one of my arduino wirings. I would like to have a little universal solution like you have on 7i96 cards, not just TTL as usual, but cant figure out what parts is best to use to achieve this.

What caused the original failures ?
Are the existing motors good or are they the cause of the initial failures ?
I've lost count of how many times a tech has replaced the failed component is a system, not investigated why it went only to destroy the custom ASIC they just replaced.

Goto the gecko drive home page, they have some great stuff about stepper motors and drivers, and it's broader than just their own products
After becoming wiser pick drivers that match your motors, don't guess.
Carefully set one axis up, don't take any shortcuts or do anything additional you think is right, this would seem how you got into this mess.
The drivers are going to come from china, probably the same factory, only thing that may change is the silk screen on the front cover.
I've bought from stepper online, sellers from eBay and from aliexpress, no issues as long as you take care as would with any other job.
Do one axis at a time, thereby only having to debug one axis, wire the rest up the way you do the first one once it's works and you'll be fine.
Spend time doing research, you'll be wiser for it, you'll do a better job, you'll stop wasting money.
And for crying out loud keep your soldering iron away from the drivers.
If there is an area you are not totally grounded in get an expert in and don't suggest to them how to do their job.

Typically you can only use mixed voltages with sinking inputs (input common grounded)
In this case, any positive input voltage from +5V to +36V will be detected by the 7I96
isolated inputs.
 

Hi PWC will any 5-axis bob work with this firmware? The Sainsmart boards are hard to get in the UK

LC-E series drives dont have an auto tune feature. Has anyone successfully tuned these drives even when only one motor is used per axis (not a gantry)? Considering if I buy a kit as well over Inovance or Leadshine

Read this page, not just the chapter in the link.
wiki.debian.org/NetworkConfiguration#Con...e_interface_manually