Problem with delay DIR-STEP

Mehmet · Posted at 2018-8-12 00:28:37

Dear ytliu and nikolay
First of all. Thank you for a lot efforts for find a good solution about a problem. Everyday I am following your posts.
I am learning a lot of things from your posts (systems, stepper drive, signal ,step-dır, step lose etc...)

I am a simple DDCNC 2.1 user , but i don't have a cnc. I need a lot of parts like stepper driver, screws, spinle ..For now, I have a DDCNC 2.1 controller.
I want to make cloosed loop cnc so , next week order a Lichuan closed stepper driver. If you are use closed loop step (with encoder) never losing step ?  Is it correct ?
I am not sure but Lichuan stepper driver's Logic input current : 10 mA at 60vDC @200kHz  Ia=(5V-0.7V)/(270+100)=11.6ma > 10mA. (formula from ytliu). is it ok. for controller dncnc?

I need your experiences, If i use this closed loop  driver , machine doesn't lose steps when fast changing direction. I it correct or not?
Should I use a pulley belt system for solve  torque , microstepping , step loses.

http://servo.xlichuan.com/prod_v ... 166&FId=t3:69:3

Soory for my english.I hose so you understand..

Nikolay81 · Posted at 2018-8-11 14:47:19

Last edited by Nikolay81 In 2018-8-11 15:03 Editor

ytliu Posted at 2018-8-11 06:24
Thank you very much for your analysis of this issue, I will use the first way you said to make cha ...

Thank you, for your desire to solve the problem, dear Ytliu.
I understand that solving this problem is quite difficult for you.
Making changes to hardware logic can cause many other glitches.
I certainly do not ask you to solve the problem, right now.
In such cases, you can not hurry.
It is necessary to think everything over.
You can solve this problem in six months, or a year.

The main thing is that now DDCSV, a little "crippled" or "problem" controller.
Since it does not meet the specifications of the leading manufacturer of drivers.

And the solution to this problem will immediately bring DDCSV to a whole new level of quality and reliability.
It will be possible to raise the price for it.
For example, now I would not risk, sell a lot of DDCSV, like your dealer.
Because, by reason of technical problems with the product, this business can "eat" me later.
To earn good money, i need to sell at least 40 controllers a month.
This is 500 controllers a year.
If even 10% of buyers have a hard-to-solve problem with the loss of steps, I will spare all my free time, I will spend on technical support of users.
But life is too short to waste it so stupidly :-)
I ask you to understand my position.
Have a nice weekend.

ytliu · Posted at 2018-8-11 06:24:11

Nikolay81 Posted at 2018-8-11 01:14
I wrote a new article for you.
I hope it will help you better understand the reasons for missing t ...

Thank you very much for your analysis of this issue, I will use the first way you said to make changes.

Nikolay81 · Posted at 2018-8-11 01:14:07

Last edited by Nikolay81 In 2018-8-13 03:41 Editor

ytliu Posted at 2018-8-10 07:22
The current motor drive signal generation logic is: in each signal processing period (t), the DIR si ...

I wrote a new article for you.
I hope it will help you better understand the reasons for missing the steps and find a good solution.
This problem is quite difficult to understand. Perhaps machine translation will distort the meaning of some phrases.
But I'd better do it until I can.
Please read the article and tell us your opinion about it.

https://drive.google.com/open?id ... 9uvzhUmf3IJdtMNvpR_

Nikolay81 · Posted at 2018-8-10 14:38:58

ytliu Posted at 2018-8-10 07:22
The current motor drive signal generation logic is: in each signal processing period (t), the DIR si ...

Please cancel the order DM556.
Do not waste your money. You can not buy the same driver as mine.
If you can not cancel the order, I'll pay you. Just tell me how much it cost.
The problem is not only in DM556. Leadshine produces a huge number of driver models.
In addition, now most of their products are counterfeits (say 70%).
It is necessary, that the controller correctly worked with any drivers, and not only with 556.

In addition to removing the resistors and removing the oscillograms, yesterday I went to a company that has been repairing CNC machines for many years.
I asked their specialists to consult me for money, but they did it for free, for which I am very grateful.

They read our correspondence with you and explained the problem. We were both mistaken.
Everything is simple enough. Briefly, it is necessary to increase the delay "DIR-STEP" to 7000-10000 nanoseconds.
Then any drivers will work correctly with DDCSV.
You can partially bypass the problem, individually in the case of each specific driver.
But, only the correct delay DIR-STEP will make it so that there will be no problem at all.
With any drivers.

If you can not do this programmatically, you can hardware shift the signal STEP relative to DIR.
The simplest option is to install a circuit from the resistor and the capacitor between the driver and the controller.
It's a little bit ugly in the form of the signal, but as I saw yesterday, the optocoupler, it's still very ugly.

Today I will write an article in which I will describe in detail what they explained to me.

Please do nothing until you read the article. You also need to understand what is going on to solve the problem.

ytliu · Posted at 2018-8-10 07:22:02

Last edited by ytliu In 2018-8-10 09:22 Editor

The current motor drive signal generation logic is: in each signal processing period (t), the DIR signal is first established, and the PLU signal rises at time #416 and falls at t/2.(see Attachment)

I will make the following changes:
The current signal processing period DIR signal does not change, the PLU signal rises at time 0, and falls at t/2;
Otherwise, proceed as before.

Since the motion of the system is planned, the speed at which the direction changes is generally low. At this time, t is relatively large, so the case of #416>t/2 does not occur.

This should be a relatively simple way to solve this problem.

I have already ordered LEADSHINE 556 and I will verify it.

Nikolay81 · Posted at 2018-8-10 05:49:44

ytliu Posted at 2018-8-9 23:20
I have encountered many customers who have lost steps due to the "Time interval between DIR & puls ...

I replaced the resistors with jumpers.
I had resistors 100 ohms.
The driver input circuit diagram is incorrect.
Everything is completely different there.
Perhaps this scheme is version 1.0. And I have version 2.2.

Before and after removing the resistors, I shot the oscillograms, between the different terminals of the optocoupler with a logic output of 6N137.
This is a very common optocoupler. Datasheet on it, you can easily find it in Google.

Conclusion - nothing has changed. Perhaps there was a little more interference. In general, the signal at the output of the optocoupler has a lot of noise.
Unfortunately, the theory of improving the performance of the optocoupler, after shunting the resistor, was not correct. You can not remove the resistor.

I uploaded a waveform to Google Drive. They can be opened by the program "Multi VirAnalyzer".
You can download it here: http://english.instrustar.com/upload/image//English%20Version(3_10_7_1).zip

In some oscillograms, I was wrong with the division of the probe (put X10). There, the voltage should be divided by 10.
I think everything will be clear.

In some oscillograms, a different frequency.
This has been the case before. Often on the same programs, there is a different frequency. I do not know why this happens.

I will write more in detail tomorrow. It's very late now.

Here is a link to oscillograms:
https://drive.google.com/open?id ... dpnRbFb1pjd4J3THV3r

ytliu · Posted at 2018-8-9 23:20:58

Nikolay81 Posted at 2018-8-9 14:37
Thank you. I also read the instructions to these drivers and consulted with familiar electronics en ...

I have encountered many customers who have lost steps due to the "Time interval between DIR & pulse". By changing the pulse level, this problem can be solved very well.

As for the problem you are currently experiencing, you must first replace the resistor, and subsequent analysis is meaningful. After replacing the resistor, if there is still a problem, I will purchase LEADSHINE 556 and verify it myself.

Since this update involves hardware logic, I have to be very cautious, please understand.

Nikolay81 · Posted at 2018-8-9 14:37:46

Last edited by Nikolay81 In 2018-8-9 14:42 Editor

ytliu Posted at 2018-8-9 06:48
Ia=(5V-0.7V)/(270+100)=11.6ma < 16ma

When operating at high frequencies, the logic signal current ...

Thank you. I also read the instructions to these drivers and consulted with familiar electronics engineers.
I will certainly remove the resistors today. Just for the experiment.
This takes time. In this situation, the dismantling of the controller is rather complicated.
The result "before" and "after", I will describe in detail.
But, even with the help of logic and mathematics, one can understand that this can improve the problem but not solve it.
Because increasing the delay DIR-STEP solves the problem completely. Without removing the resistors.In this, I was convinced yesterday on my experience.
In the Purelogic controllers, there were no resistors at the outputs, but they still skipped the steps with the Leadshine drivers.
While they did not increase the delay of DIR-STEP. Then the problem immediately decided.
Still, the question arises: why did they even add the "Time interval between DIR & pulse" setting, if its value can not be changed?
After all, if there is something, it should work. If it obviously does not work, it should be removed, so as not to embarrass the users.

Perhaps I do not understand something. Please tell me how difficult it is to fix the bug with the setting "Time interval between DIR & pulse"?
If it takes a very long time, or is impossible at all - I'll understand.
Then we will look for "ways to bypass" the problem.

ytliu · Posted at 2018-8-9 06:48:27

ytliu Posted at 2018-8-9 06:15
Have you replaced the resistor? The problem now is that the optocoupler does not work properly durin ...

Ia=(5V-0.7V)/(270+100)=11.6ma < 16ma

When operating at high frequencies, the logic signal current cannot meet the requirements, resulting in lost steps.

Problem with delay DIR-STEP

View forum before

TEL

0755-87654321