Bug in PicStep - bad FW

Hello,

I made PICstep driver with FW with torque compensation but this SW not work. FW without compensation worked but if I testing max rpm i working gut only one side rotation.

I write my own FW. I lost many months development, because bug is mainly in datasheet for LMD18245. Very difficult is timing algoritmus for generating outputs for bridge.

Unfortunately my driver is comercial project and I cant public source code for PIC16F628.

My driver is public on my web http://www.cncnet.info/view.php?naze...nku=2006050002

Parameters Bistep:
-------------------------------
microstepping: 1/2 1/4 1/8 1/16 for step
input frequence: 45kHz
optocouplers on inputs compatible directly with LPT
current settings: 0.5 - 3A step 0.5A
DC supply: 12-48V...driver has on board step-down converter for 5V logic
timeout: is not implemented

By testing this driver with bipolar motor 3V/1.5A rotate above 1500rpm with full strong.

Pinout PIC16F628A with my FW is compatible with original version Picstep.

RadekCX

Quote:

---

Originally Posted by RadekCX

Hello,

I made PICstep driver with FW with torque compensation but this SW not work. FW without compensation worked but if I testing max rpm i working gut only one side rotation.

I write my own FW. I lost many months development, because bug is mainly in datasheet for LMD18245. Very difficult is timing algoritmus for generating outputs for bridge.

---

K'eh? PICStep has been sucessfully used by hundreds of people around the world, what bug are you talking about? I'm guessing FW==firmware, what does SW equal? SW==software maybe?

If english isn't your strongest language please don't confuse your post with acronyms.

Quote:

---

Originally Posted by RadekCX

Unfortunately my driver is comercial project and I cant public source code for PIC16F628.

---

Seeing as PICStep is a GPL project I hope you haven't "based" any of your code on my firmware else you're in breach of the GPL.

Alan.

What language is the webpage? is there a compatable english translation page?

Quote:

---

Originally Posted by Garfield2

K'eh? PICStep has been sucessfully used by hundreds of people around the world, what bug are you talking about?

---

Yes? Maybe, but your FW not work on both side rotation identically. I testing this FW and I dont know, why max rpm is not identically - this is not identically even if change phase on motor. I'm testing progress current through phase with osciloscope and here also not current identically. With good motor is bug better evident. This bug verify two people not dependent on me.

Quote:

---

Originally Posted by Garfield2

I'm guessing FW==firmware, what does SW equal? SW==software maybe?

---

Sorry, this is my bug. Correct is only FW==firmware. FW is software for MCU.

Quote:

---

Originally Posted by Garfield2

Seeing as PICStep is a GPL project I hope you haven't "based" any of your code on my firmware else you're in breach of the GPL.

---

This is very simple and long history.
I need fom may CNC drivers and PICstep was very nice and I working with PIC MCU. I made my own version PCB and flash FW version with torque compensation. This version not worked. I flash version without compensation - this version worked ( seemingly). But I need microstepping 1/16. I rewrite your FW new. No with ASM but I use mikroPascal for PIC. This FW worked as your FW - have extra only 1/16 microstepping and not support full step. This version FW and PCB is on my website http://www.cncnet.info/view.php?naze...nku=2005120802 with all source code, HEX and all documentation.

But when I this version testing I find bug see above.
What to do!!! I have four PCB and bad working. I flash again your FW and when I thoroughly test have aqual bug as my FW.

I again and again rewrite FW and testing but driver not work well. But I finaly uncover why driver not working OK. This bug is already in datasheet LMD18245. This argument for equal pinout on MCU.

Then I made new PCB with DC-DC converter and this project is Bistep. PCB not equal with Picstep, FW is not equal with Picstep - only pinout is identicaly with Picstep. My new HEX have 7.5kB - almost 2x better as PICstep.

Is this violation GPL - IMHO NOT!!! Any driver with PIC16F628, LMD18245 is PICstep?

Sorry for may english - better I dont know.

RadekCX

Quote:

---

Originally Posted by RadekCX

Yes? Maybe, but your FW not work on both side rotation identically. I testing this FW and I dont know, why max rpm is not identically - this is not identically even if change phase on motor. I'm testing progress current through phase with osciloscope and here also not current identically. With good motor is bug better evident. This bug verify two people not dependent on me.

---

I've not seen this bug of which you speak, how are you coming to the conclusion that the speed difference is the FW and not the multitude of other variables between the IO pins of the PIC and the coils of the motor?

The coil currents are simply a look-up table output to the DAC of the LMD's. There really isn't much to go wrong here as far as my reading of the datasheet.

Quote:

---

Originally Posted by RadekCX

Is this violation GPL - IMHO NOT!!! Any driver with PIC16F628, LMD18245 is PICstep?

---

No any driver with an LMD and a PIC isn't PICStep, but any driver with an LMD and a PIC programmed with my firmware or a derivative of it is. If you've written your firmware from scratch with pascal and haven't simply ripped the headers off my asm then that's fine.

But at the same time, you cannot cry "WOLF WOLF" about my firmware without showing where this "bug" is. That's just not very nice to the community that has "helped" you produce your "commercial" firmware.

Alan.

Quote:

---

Originally Posted by Garfield2

I've not seen this bug of which you speak, how are you coming to the conclusion that the speed difference is the FW and not the multitude of other variables between the IO pins of the PIC and the coils of the motor?

---

This is simply - I know where is bug and bug is in DS LMD18245. Your FW is write exactly by this document.
I have two diferent version PCB and result is every equal - bad rotation (resonance, not max rpm) on one side not dependent on subtitution coil on motor (this eliminate bug in motor). When I write first version my FW was equal problem. When I flash new FW in both PCB then all problems lost.

Quote:

---

Originally Posted by Garfield2

The coil currents are simply a look-up table output to the DAC of the LMD's. There really isn't much to go wrong here as far as my reading of the datasheet.

---

Problems regulation stepper motors is not simply!!! But bug in DS is fundamental. :-(

Quote:

---

Originally Posted by Garfield2

But at the same time, you cannot cry "WOLF WOLF" about my firmware without showing where this "bug" is. That's just not very nice to the community that has "helped" you produce your "commercial" firmware.

---

Alan - I send to your mailbox (before I posted message in this forum) two e-mails - first 12/13/2005. Here I writed about my version PICstep and you can download my FW with torque compensation.
Before many weeks I write second email - here I write about the bug in our FW.
On my emails you do not replied. Maybe in this time was possibly cooperation.
Then I posted message in this forum. Alan don't cry now!

How community "helped" produce my version FW!!!?? Schematic PICstep is identicaly with recommended schematic in DS LMD18245. Is it not new. New my FW is absolutely different.
Driver Bistep now completed company not dependent on me and I cannot FW release.

I have not obligation to community!! Information about bug you know now. It's all.

RadekCX

Still note:

On my web is most projects free and people can download all documents free - I not need GPL - its simply free - for example weatherstation with web output, LCD with RS232 input and other.

I working as devoleper SW for PC and working with MCU is my hobby. When I get any money from donate or when I sell any project with MCU, then I put profit from this projects into SW or instruments for develop other free projects.

RadekCX

Quote:

---

Originally Posted by RadekCX

Alan - I send to your mailbox (before I posted message in this forum) two e-mails - first 12/13/2005. Here I writed about my version PICstep and you can download my FW with torque compensation.
Before many weeks I write second email - here I write about the bug in our FW.
On my emails you do not replied. Maybe in this time was possibly cooperation.
Then I posted message in this forum. Alan don't cry now!

---

I receive nearly 600 emails everyday, I do not reply to everyone because I wouldn't get anything else done. Also if the english is particularly hard to comprehend I'm even less likely to respond.

As for crying, I've done nothing of the sorts other than trying to decifer what you are talking about and trying to reduce your bad mouthing of _my_ many months of work.

I think you are misunderstanding my responses as aggressive when all I'm trying to do is figure out what it is you have found. Yes I understand you have produced a commercial driver, but using my driver and it's "buggy" firmware as a reason to purchase yours isn't very friendly.

Quote:

---

Originally Posted by RadekCX

How community "helped" produce my version FW!!!?? Schematic PICstep is identicaly with recommended schematic in DS LMD18245. Is it not new. New my FW is absolutely different.

---

The fact you even stated you got your ideas from my driver is a prime example. I think the language barrier here is proving to be the biggest bug in this dialog. I doubt I could understand the issue you have found even if you put it as simply as possible. I get it has something to do with the datasheet information for the LMD about coil current control but what the bug is I am still scratching to understand.

Quote:

---

Originally Posted by RadekCX

Driver Bistep now completed company not dependent on me and I cannot FW release.

I have not obligation to community!! Information about bug you know now. It's all.

---

Nice, good fortune to you too.

Alan.

from what i understand

1. there is a problem when u turn the motor on i direction the speed is not the same as the opposite direction (it cant reach its max rpm)

2. this is due to a mistake in the datesheet of the lmd

3. this project is commerical but u are posting all the documentation on your web site, i dont know how comes ?


so please send alan the page number and line number where is the mistake in the lmd datasheet and what should be corrected, that will be better that trying to decrypt your messages

also take it easy alan, your driver is great and everyone knows it
i am sure that the langage is causing problem and thats it.


max

Hi,

Somebody who knows Czech language read for me the page about Picstep. I mean the Picstep you built from original schematics. Unfortunately, it is not clear there if the FW you modified from original Picstep is corrected or not. It seems that you did not correct it just modified for full step and 1/16 step. Is this true? Also, you do not mention what jumper setup is needed for each step/microstep function.
For your own version of controller/driver it seems that the schematics is the same with picstep, just some opto added. Is this true?
Do you sell (the company) already programmed chips and if yes, how and at what price?

Thank you.

Zoltan

Quote:

---

Originally Posted by zoltan

Hi,
Somebody who knows Czech language read for me the page about Picstep. I mean the Picstep you built from original schematics. Unfortunately, it is not clear there if the FW you modified from original Picstep is corrected or not.

---

My first version is on this URL http://www.cncnet.info/view.php?naze...nku=2005120802 this version is on web with source code, HEX and other documents.

When I write this version FW I dont know about bug in DS LMD18245. In this FW is identicaly bug as in Alans FW. This FW have only microstepping 1/16 and have not full step.

Quote:

---

Originally Posted by zoltan

It seems that you did not correct it just modified for full step and 1/16 step. Is this true? Also, you do not mention what jumper setup is needed for each step/microstep function.

---

No, my all version (only for 20MHz clock) support only: 1/2 1/4 1/8 and 1/16 microstepping - 1/2 is with torque compensation

Quote:

---

Originally Posted by zoltan

For your own version of controller/driver it seems that the schematics is the same with picstep, just some opto added. Is this true?

---

First version is on link above with schematic and PCB - this version is identicaly with Alans PICstep - I add only jumpers for easy change currents through coils.

Comercial driver BiStep http://www.cncnet.info/view.php?naze...nku=2006050002 is partly identical (PIC and LMDs - so FW are compatible) but added 3x optocouplers (DIR, STEP, BRAKE) and DC-DC step-down converter. Linear stabilisator (like 7805) is not suitable for supply 48V.

Detail picture Bistep are here: http://www.cncnet.info/gallery.php?m...&galerie_id=11

Quote:

---

Originally Posted by zoltan

Do you sell (the company) already programmed chips and if yes, how and at what price?

---

Yes, its possible.

programmed PIC16F628A-20MHz: 15 euro (I cann send directly)
completed Bistep: 80 euro (company mail: zstloukal(at)volny.cz)

Next comercial question please on privat contact cncnet.info(at)gmail.com - no on this forum.

RadekCX

to all simple test:

Leave good motor (above 1.5A) freely on wood board and rotate on left and right side with 1/2 microstepping.
Listen sound and see resonance - wood board increase sound. One side is different (hard sound) and max rpm and torque on this side rotation is slight.

woow
is the for only driver !!!
80 euro for 1 axis ??
thats about 105$ , i might as well add 10$ and get a gecko
if it is for 1 axis this is over way to hi for it, alan driver will be much better in that case, and i dont mind if it has a small flow like that.

anyway good luck

Hi

Interestingly I have a similar problem.

I have been testing EAS electronics microstep controller with some FL57STH56-2008B motors. This board also uses the LMD18245 and I observer the same problem. The sound and feel of the rotation is definitely different and a current probe also shows the stepping is different.

I have checked the circuit and the source code against the latest LM18245 data sheet and I can not see any errors.

So RadekCX could you tell me what your solution is?

Stephen

Quote:

---

Originally Posted by stephenkench

So RadekCX could you tell me what your solution is?

---

I doubt he will want to tell you if his firmware costs 15 euro. It's probably something to do with the DAC values not being linear (or the same) at either side of the centre. The Cos-Sin curves are correct, but if the DAC is offset slightly you'd see more power in one half of the cycle and less in the other. I'd suppose this would be what is inducing the resonance at high frequencies.

I'll have a play around with it when I've got some time. It's not circuit related so it should be fairly trivial to figure out once you know where to look.

Alan.

Alan, Alan have you osciloscope!!! Sin curve is not identical. In attachment is picture this sinus. :-)

This bug i in DAC table on DS LMD18245. Its not identical - forward not equal as reverse. Static are identicaly but dynamic isnt equal.

Meditate!!!

HW forum in Czech republic this bug detect in 6 hours. :-)

When this bud detect (its simple repair in FW) have you hard work on your FW but your FW is write slowly and its not good - this manifestate equal as bug in DS LMD18245.

Timming, timming and timming.

RadekCX

Quote:

---

Originally Posted by RadekCX

Alan, Alan have you osciloscope!!! Sin curve is not identical. In attachment is picture this sinus. :-)

This bug i in DAC table on DS LMD18245. Its not identical - forward not equal as reverse. Static are identicaly but dynamic isnt equal.

Meditate!!!

HW forum in Czech republic this bug detect in 6 hours. :-)

When this bud detect (its simple repair in FW) have you hard work on your FW but your FW is write slowly and its not good - this manifestate equal as bug in DS LMD18245.

Timming, timming and timming.

RadekCX

---

First I'll look at the potato, banana banana. Looks like red smells green and the dish ran away with the spoon...

Thanks I'll figure it out faster than these posts I think. Viva the Czech Republic for being so smart!

I'm off to "Meditate!!!" on your scope picture.

Thanks,
Alan.

Can you write your message in basic english. I dont undestand.

You find this bug? :-)

RadekCX

Quote:

---

Originally Posted by max_imum2000

woow
is the for only driver !!!
80 euro for 1 axis ??
thats about 105$ , i might as well add 10$ and get a gecko
if it is for 1 axis this is over way to hi for it, alan driver will be much better in that case, and i dont mind if it has a small flow like that.

anyway good luck

---

Yes, unfortunately local dealer LMD18245 sell this IC per 21 USD!!! Its reality in Europe and Czech republic.

RadekCX

Quote:

---

Originally Posted by RadekCX

Yes, unfortunately local dealer LMD18245 sell this IC per 21 USD!!! Its reality in Europe and Czech republic.

RadekCX

---

Thats kind of cheap. This part in Norway, goes for 37$ or something...