TB6560 removed optocouplers - runs perfect - but how to secure?

[atair]

after reading 20 pages of the "how i fixed my TB6560" -thread, and reading the info here:
http://reprap.org/wiki/SeeMeCNC-H1_B..._Mod.2FSkewfix
i removed the optocouplers as shown in the image. now the TB6560 went from "give me back my money" to supersmooth.. at least without any load - just the motors on the table.

now - is there any (simple) way to protect the pc? i read some things about diodes, resistors etc.
i am not into electronics too much - can anyone point me to a specific component i can place in there?

also the rerap text states:

Now typically you would never want to bypass/eliminate the functionality of an optocoupler but it was found that both side of the optocoupler share a common ground, almost defeating the use of optocouplers to begin with.

is this true? can i leave everything as it is in the picture and be more or less safe?

[Al_The_Man]

I only use Opto couplers to go from one system/supply to another, I have always used a common earth ground, if done properly with a central star point it there is seldom a problem.
As I have pointed out many times, your PC P.S. common is already at earth ground and so is your mains fed router, so they are already common.
The best thing you can do is set up a central Earth ground and take a bonding wire to it, this includes your incoming system ground.
Al.

[atair]

wow thanks - that was a fast reply!
does it mean if both components (PC and TB6560) go into the same power strip - is this already a "star ground"? or how do i achieve this?

[Al_The_Man]

This is a bit of a long document, but it may give you the gist or part of the picture.
http://www.automation.siemens.com/do.../emv_r.pdf?p=1
Al.

[MoneyPit*10]

The TB6560 optos are not protecting much as a whole and can be by-passed. The one exception is the opto on the RELAY line unless you re-wire it to work without the opto and that involves more than jumpers.

In the setup pictured they are by-passing the DIR and STEP optos but leave the ENABLE optos as is. The ENABLE lines WILL cause blown TB6560 chips (read smoked) and should also be removed. And your best bet is to wire ALL the ENABLE lines to the same, single, pin so they ALL get turned off/on at same time instead of cascading down the line. There is a delay when you switch the ENABLE lines via even MACH3 and this could be enough to blow a TB6560. In the more current design of the Chinese TB6560 type boards, the ENABLE is a single line, common to all, if you browse the setup information.

That being said. The sockets are bad. Using any type of non-soldered wire is bad news down the road. Oxidization. Soldering jumpers would be the best approche.

As to the optos, the ones for the limit switches and Estop are even more questionnable as they all take 5V power and ground from the computer power supply. There is nothing live from the outside world unless you go external 12V or 24V for limits/Estop power via relay and even then it is questionnable if they are required as the relay must be protected via a diode from BEMF.

MP*10

[atair]

this may be a stupid question, but here:
forum.zerspanungsbude.net • Thema anzeigen - China-Steuerung mit TB6560 - ein paar Hintergründe you can see in the first image how a guy bypassed them - what i dont get is why some are diagonal and some are straight.

[MoneyPit*10]

You must look at the thread on " How I repaired my TB6560 " for the schematics the poster graciously made availlable for us. It explains/shows the reasons.

Reading the specs of the PC817 also helps understand why you jumper in this manner.

Also keep in mind jumpered from the top (via the sockets) will need to be a mirrored image from the bottom of card, inverted connections. This can become confusing with out a diagram to follow along.

MP*10

[H500]

Even worse than the common ground is the common power supply on both sides of the opto. That essentially bypassed all protection from the optos. The design is unbelievably bad.