Thread: A misadventure with StepperOnline stepper drivers....

Hi All,

Just received a new email from the team at Stepper Online, telling me that they've fixed the incomplete table for the dial switch in the CL86T v.4 manual. So you shouldn't have to download the CL57T manual anymore.

Yay for progress!
Hamish.

Hi Everyone,

I've been engaged in some rather in-depth emails with Stepper Online, and they've been particularly communicative and helpful. There's some major progress and news. I hope it helps others out there.

As you know, the documentation has already been improved since I complained about it (Thanks Bo).

However, that's just the start!

The big news is direct from the Stepper Online team. However, please note that it's "early days" and this is just for information purposes, and what you do with that information, is entirely up to you... as is your responsibility should you act upon it.. Don't blame me. I'm merely relaying information that I think will help others.

So in numbered dot point summary, from most important to "kinda interesting, and we'll see what happens":

1. Despite information to the contrary on the Stepper Online site, the V.4 of the CL86T driver DOES have a programming port.... At Stepper Online's support team's suggestion, I cracked open the case, and found it (as described) inside the unit. Please note, that it is not accessible from the outside of the case by default. (See image below) However, opening the case to access it, damages the "QC Passed" sticker which when peeled off says "VOID" (so it's a warranty void if removed sticker too). So you've been warned. (See left side of image below)
2. If you're wondering where it is, it's next to the 5V/24V stepper input voltage selector switch. (See right side of the image below).
3. At present, there's there's no "Complete programming cable solution" on their store for the V.4 though. (It does exist for the V.3 and below though). It is assumed you need to use a standard USB to RS232 cable (most modern PCs don't have serial ports after all), and then use the "debugging cable" to wire it up to the board. Hopefully that will become available soon.
4. I've been told that the next version of the CL86T (the V4.1) will have an externally-accessible port to "debug" it once again. In the mean time, for those considering which drivers to buy, I recommend looking into the other models listed and compared in the post above, since the price difference isn't huge.. and they have everything you need to program them on the Stepper Online store.

CL86T_v4_SummaryImages.jpg

Now, I've discussed... at length, the ideas I've had to improve the product, the product information on their store, and what I've found both good and bad about the whole situation. I'm hoping it will lead to real-world benefits for other existing V.4 owners... and owners of future generations of Stepper Online products. Some of the ideas include:

• Placing all DIP switches and sockets for driver configuration on the side facing the user. (CNC control boxes can be cramped.. and this makes adjusting them with side-mounted, difficult-to-reach switches and sockets more challenging). As someone who's strained to read the DIP switch settings on the side of a driver.... Instead of having the DIP switch settings labelled on the side of the driver, placing a QR code on the front would be a space saving, and easy to read/use improvement too.
• Since most progamming interfaces on drivers use serial interface based tech, why not standardise all models to use the same cable and sockets in future gen drivers. If you have many CNC retrofitted tools in a maker space or workshop, finding the right cable/software/computer with compatible OS... for various drives can be a real hassle.
• Consider providing a text based/serial/command line interface for people who don't use Windows. The app is literally just sending and receiving serial data in a GUI.. so accessing that with any old computer via the command line would be a huge benefit.
• List all default configuration settings for ALL drivers somewhere on the site. The CL86Y models I use have a huge 4000 step threshold before the alarm is triggered when it comes from the factory. Ok, it might not be huge if your microstepping down to 56000 pulses per revolution on a high precision mill... but for CNC routers, that's quite a delay and defeats much of the point of a closed-loop system.
• Make downloadable manuals far more detailed, in colour, and written by a native speaker of <insert language here>. Copying and pasting images into a document is easy, and limiting the manual because of printing costs (most people lose/toss their paper manuals anyway) is counterproductive.
• Standardizing the wiring colour codes. Why are the motor cables differently colour coded to extension cables?
• Improving the panel-mounting capability of motor/encoder cables. Longer threads on aviation plugs, and/or panel mount adaptors for the D-SUB15 encoder cables.
• I also outlined that it's unfair that earlier (and hopefully later) versions of the driver can tune/calibrate/adjust their models without potentially voiding the warranty... when the V.4 does indeed have the interface, and the user has paid for ALL of the unit's capabilities. I'm still waiting for a clear response on whether I still have a warranty.

If you have any other ideas? I'll pass them along!

I've volunteered to test their configuration cable/software when they send it to me. I'll keep you posted.

Take care, have fun, and make something awesome.
Hamish.

That is an interesting read and a lot more in depth than I've gone with my kit.

Originally Posted by BoneInspector

That is an interesting read and a lot more in depth than I've gone with my kit.

Hi Bone Inspector,

Are you a radiographer perchance? Just curious about the avatar/handle/nick name.

Thanks for the comment. I was wondering if there was any actual interest in my little tale. I freely admit it's not up to my usual humour standard, and I understand that there are many makes and models of stepper drivers, and of course, servo "amplifiers" too. So my focus on this particular post is quite specific... and consequently, less interesting and/or relevant to many. However, I do think it is a cautionary tale that might be useful to anyone new to the DIY CNC building community.

Also, what sort of kit/setup are you running?

Kind Regards,
Hamish.

Hi all,

I hope you're well. I was discussing this tale with a newbie local CNC builder, and I was asked:

"Since Stepper Online says that most of the time you don't need the calibration cable/tool, why do you care that the CL86T v.4 doesn't have an external debug/calibration port? Isn't open loop good enough? It works for me!"

Well... that's an interesting tale!

What started as a "minor glitch", had some potentially expensive consequences...

Closed loop steppers are a significant improvement over the open loops. However, just because the driver controlling a particular motor can notice a problem... it doesn't necessarily mean that it does so in a timely fashion. What is "timely" anyway? Well that depends on how much "real world movement" is allowed before the alarm is triggered, and whether that alarm circuit is wired into your control/break out board. A closed loop motor and driver can solve a variety of issues, but by themselves there's a lot they can't.

While closed loop steppers are incredibly versatile, and can be used in any number of ways... the default settings, while perfectly fine for some situations, are utterly hopeless for others. I was in the "Others" category. If you're building a wood carving CNC router... you're also likely in this category.

Honestly, I was having a great deal of trouble with my early CNC projects whenever movement lagged behind the controller. I was confused because I thought the driver should be handling all that. Unfortunately, I didn't have the cable to view the configuration and calibrate my drivers. It wasn't until the amazing guys on the LinuxCNC forum told me of the default (and unpublicized) settings on the CL86Y drivers, where and how some of my issues could be found and fixed.

Of course, you can definitely use drivers in their factory settings in many cases. Stepper Online even says "Hey, you don't really need this programming interface in most cases". I think this is probably true if you're running a DIY converted mill with high precision microstepping. However, if you have a 200 step/revolution motor configuration on a CNC router, and a 5mm pitched ball screw... each step is 0.025mm of real world travel. More than enough precision for my CNC router needs. But a default threshold of 4000 steps or 20 revolutions of error is 100mm of movement lag in my setup... and that's before the alarm is triggered.

A lot can happen in 100mm. As I'll explain!

Now, imagine I'm carving a decorative inlay into a large Jarrah slab. I do the simulation run, the "air cut" and even tested it on some scrap. All goes to plan. So I clamped up the slab, did my touch probe, homing/zeroing dance, loaded up the file, and hit "Go!"

As it starts carving, I loiter about, cleaning up a few odds and ends, a mere step or two from the machine when I get a call. I step a few metres away so I can be heard... and as I'm talking...

The Y axis jammed on the rail at a certain point. However, it was about 4cm short of the of the carving boundary, then it unjammed as it went back as the carve progressed... this happened three times at the same point on the Y axis in the 5 minutes I was chatting with my friend (who wanted me to do this project for him). I glance over after things "don't quite look right"... It got pretty awkward when I informed them of the problem, and it was a $500 piece of timber.

As far as the driver was concerned, the error threshold wasn't reached, so there weren't any problems. It had never happened before... I was unprepared.. so it certainly put the friendship to the test. I told them, "Let me think of a solution, and I'll get back to you". Fortunately, the carve was still quite shallow at this point, so I simply planed the slab down 3mm, and restarted after fixing the rail problem.

Fortunately, when he saw the successful second attempt, he said "I thought there was a problem? You really had me worried!"... He didn't even notice the thinner slab.

Regardless, when I found out that the threshold was so high after telling the LinuxCNC folks my tale, I decided I needed the calibration cable, and I was going to get it. However, even then, it's not always "easy" to use.

The calibration stuff too overwhelming?

There's a lot of stuff in the tuning/calibration software that I don't fully grasp. It's complicated because there's a number of inter-related parameters which impact torque vs. speed vs. precision and consistency (not the same thing) of position control vs. smoothness of movement vs. power consumption/over heating issues. Unfortunately, there's no "one best way" to configure these and they often need to be tuned and calibrated on a case by case basis... and tested thoroughly if you mess with them. Consequently, I mostly run off the defaults for that.

So how do I use the calibration software?

There's really two main settings that I'm concerned with, because I use the closed loop driver alarm to signal a problem to the CNC control card. If I don't, an axis motor driver might get jammed, the driver enters an alert state, and this simply turns the jammed motor off. Meanwhile, the control card is unaware that a motor has an issue, and continues to move all the other axes as best as it can, according to the G-code. This can lead to some really weird and sometimes tragic results.

So the two settings I care about (and are easy to understand) are:

1. The "positional error"/"lost step count" threshold that it takes to trigger the alarm.
2. The "alarm polarity" (also known as "normally open" or "normally closed") function of the alarm optocoupler. I have drivers wired to the alarm in series, so any driver alarm pauses the carve, and for that to work, I need "normally closed" on all of them. However, depending on how you've built/wired/configured the machine, perhaps yours needs to work in a normally open way. For newbies looking at buying their first stepper kit. I recommend using "normally closed" since the alarm signal wiring is constantly powered, and the power is turned off when there's a problem. This ensures that any wiring problems will trip the alert as well... rather than defeat a signal that is usually off, but suddenly gets turned on, and consequently ignored. It also means you can use a single digital input to manage the OK/Fault condition of all your drivers.. which can be handy if you're running out of inputs on your BOB/control card.

Anyway, I feel that if your closed loop stepper kit has the capacity to confirm "All ok/fault" status to your control card/cam software, then your CNC machine is much more likely to address a problem in a timely manner, and may even avoid further complications. If you pay an extra $20 per axis to get that functionality, it doesn't take long to pay off! I'm a little disappointed that the v.4 has made this more difficult than previous versions and older-gen models. So I'm glad they're planning to bring it back in the 4.1.

I'll keep you posted with any further updates, once I get the CL86T V4 cable.

Take care everyone!
Hamish.

Hello Hamish,

no not a radiologist, it’s a username I picked back when I was 15, a couple of decades ago and I thought it was cool. It’s based off a book character that was asked about their name. They claimed all the good names were taken so that was what was left over. Fits when you think of the internet.

I run an xyz carve, modified x carve sold from Australia. Any missed steps the unit stops and was about as good as plug and play as you can get for a kit. Belt driven too so the accuracy isn’t quite as good as yours, close enough for wood imo.