UC100 error lost connection

[Kevin Johnson]

Back in February I purchased a new UC100 (genuine -- I was very careful about this and purchased it from a distributor here in Florida. I also had been using a genuine one for about 5 years prior which I purchased for my K1212 when it was new). The event that precipitated purchasing a new UC100 was a failure of a computer (hard drive errors) so to be cautious I used that event to justify renewing associated hardware including a new tower.

I began having random but frequent lost connection errors which of course meant that the part being cut is scrap (Teflon, so expensive material).

I am and have been running Windows 7 Professional with Mach 3; I tried three different towers over the past nine months with the same results. I replaced the UPS with a new unit twice. The machine is well grounded and the power supply to the machine itself has a dedicated surge protector. The equipment surrounding the machine has not changed in 5 years (I was thinking -- possible noise issue, maybe?).

Finally I checked some forums and read about possible issues with the USB cable. I was using the brand new cable included with the UC100. I went to Endless Cables in Port Richey, FL and purchased a new cable with a ferrite core for $10.00. The gentleman there tested the new cable supplied with the UC100 and it did not show any issue.

Regardless, I installed the new locally purchased USB cable with ferrite core and the connection errors have ceased completely. The machine, in fact, is running now and has been all night.

FYI

[cncdrive]

The longer the USB cable the higher the impedance of the cable and so the more sensitive to noise. The same when for cable thickness, the thinner the cable wires the higher the impedance which means more noise sensitivity.
Remember that normally noise effects your computer's ground because the UC100 should connect to optoisolated breakout board so basicly the UC100 can get noise only via the USB cable. (Except if you have a ground loop which means that your breakout board is not isolating and it is grounded somewhere.)

And ofcourse the higher the noise level in your control system the worse it will be.

With an avarage system where the EMI noise is on a tolerable level the original cable supplied with the UC100 is fine.
With a system where the EMI noise level is higher than the avarage or in other words if it is higher than what is defined as maximum for home electronics equipments in the electrics standards then you might need a thicker and/or shorter and/or better shielded cable or you might need to reduce the noise at it's source which is usually the spindle motor, a plasma cutter and sometimes the stepper or servo motors and drives.

For this reason when running the controller in an industrial environment we recommend to use the UC400ETH and UC300ETH, because ethernet gives you isolated communication where the above described issue does not exist.

[Kevin Johnson]

The longer the USB cable the higher the impedance of the cable and so the more sensitive to noise. The same when for cable thickness, the thinner the cable wires the higher the impedance which means more noise sensitivity.
Remember that normally noise effects your computer's ground because the UC100 should connect to optoisolated breakout board so basicly the UC100 can get noise only via the USB cable. (Except if you have a ground loop which means that your breakout board is not isolating and it is grounded somewhere.)

Do not worry -- I saved everything (all the parts):

Five years ago, the cable supplied with the UC100 was white, 3.8mm in diameter and labeled "USB Shielded High Speed Cable 2.0 Revision 28AWG/1P+28AWG/2C AWM 2725 80DEGREES C 30V VW-1" it has a small tag "2.0M" (so, about 6 feet) and another small tag "Certified HI-SPEED USB 2.0".

The cable supplied with the UC100 purchased in February, 2018 is black, about 20cm shorter than the original, 3.8mm in diameter and labeled simply "USB 2.0 HIGH SPEED MINI CABLE"

The replacement cable that I purchased locally with the ferrite core is black, 10 feet long, 4.4mm in diameter and labeled "USB SHIELDED HIGH SPEED CABLE 2.0 REVISION 28AWG/1P+24AWG/2C MONOPRICE(TM) E119932-T (LOGO) AWM 2725 80 DEGREES C 30V VW-1 CSA LL84201-F4 AWM I/II A 80 DEGREES C 30V FT1 V10143 http://WWW.MONOPRICE.COM"

I have the S/N for both the UC100 modules if you would like them -- you can see that different molds were used for the case and the original also has a sticker "QC Passed" and the newer one does not.

On February 21st at 9:07am EST I wrote to your company asking whether the reseller was selling genuine products but then I located them on your website and placed my order at 10:01 am EST. I never did receive a response to my query, however.

I should mention that I also purchased an optoisolator at that time and it had no effect on the random errors (ABA testing protocol). I did not use an optoisolator in the prior five years and did not have these errors.

...
And ofcourse the higher the noise level in your control system the worse it will be.

With an avarage system where the EMI noise is on a tolerable level the original cable supplied with the UC100 is fine.
With a system where the EMI noise level is higher than the avarage or in other words if it is higher than what is defined as maximum for home electronics equipments in the electrics standards then you might need a thicker and/or shorter and/or better shielded cable or you might need to reduce the noise at it's source which is usually the spindle motor, a plasma cutter and sometimes the stepper or servo motors and drives.

For this reason when running the controller in an industrial environment we recommend to use the UC400ETH and UC300ETH, because ethernet gives you isolated communication where the above described issue does not exist.

It seems pretty straight forward that a lessor quality cable was supplied this time around and that the cable was indeed the problem. I have tried to supply as detailed information as possible so that you yourself can determine if in fact your merchandise was being supplied or pirated copies. As you can see, this type of thing does directly impact your customers.

[cncdrive]

Thank you for the detailed informations.

If you using your UC100 with our Mach3 or 4 plugin or with the UCCNC software then it is sure a genuine UC100, because the fake UC100 is not even a UC100, it is a chinese board with their own software, it does not work with any of our software codes, because it is not an UC100 in real.
The only common between the real UC100 and the chinese fake one is the name, because the chinese gave the device the name UC100 to misslead customers, because they see that the UC100 is popular on the market.
The fake "UC100" does not even have a serial number as far as I know.

We buying the USB cables at TME store in Poland, and you are right that sometimes we supply the UC100 with white cable and sometimes with black. It only depends on which one is in stock at the supplier.
But by datasheet both cables should be fine and we did not hear more or less issues with the 2 different cables, they have the same parameters.

By the way, if you had this kind of noise issue then next time you might try one of our ethernet motion controllers (UC400ETH or UC300ETH).
Ethernet gives isolated communication and so it tolerates high level of noise, the desribed issue which can happen with USB devices can't happen with ethernet.

[Kevin Johnson]

I am certainly willing to accept that this particular cable was out in the one or two standard deviations of the normal distribution curve, i.e. still functioning but barely or not meeting the USB 2.0 specification for shielding. That is based on empirical evidence.

I completely empathize with you with respect to product piracy. Quite a number of my product designs have been stolen over the years.

There is another building where I use a CNC plasma cutter and there are all sorts of heavy duty electrical motors running so I appreciate the issues with and need for shielding. Where the K1212 is running there are only light fixtures and the machine with computer.

Good luck with your business and a curse upon those who pirate your products.

[Kevin Johnson]

Update:

A suggestion was made that a grounding problem in the K1212 control box might be the real issue and that the ferrite core on the new cable is masking this.

I decided to re-install the original white cable from 5 years back on the next cutting job. It has been running continuously for more than 3.5 hours with no error message. It does not have a ferrite core so hopefully this test rejects that hypothesis.

[cncdrive]

The hypotesis is wrong for sure, because all cables we buy are from the EU and so they are CE certified and so they must comply the EU regulations on EMI noise sensitivity of home electronics.
This is why I wrote that if the noise level is not greater than what is defined as maximum in the standards then there is no problem.
A problem occurs if the noise is larger, if the environment is kind of industrial with lots of noise, then the system may fail to reject all the noise and not only the cable is the problem in this case, but the home computer with no industrial EMI filter etc.

[Kevin Johnson]

I think there may be some confusion about USB standards. I found the document addressing this for USB 2.0 cables; you can read it here: https://www.usb.org/sites/default/files/CabConn20.pdf

The document is very clear that manufacturer compliance with the standards is VOLUNTARY unless they desire to have their product CERTIFIED as being in compliance. As such there would be no CE violation if a maker simply labeled a cable as "USB 2.0 HIGH SPEED MINI CABLE". There is a moral and ethical duty but since compliance is voluntary there is not a CE violation.

This means that even in the EU, a manufacturer must be on guard to read carefully what is being claimed by the product labeling, ergo caveat emptor.

[cncdrive]

CE is a must in the EU. Products like electric goods can't be sold without CE certification, it is law. So, they can't sell it here without CE certification.

[Kevin Johnson]

CE is a must in the EU. Products like electric goods can't be sold without CE certification, it is law. So, they can't sell it here without CE certification.

Yes, but CE certification does not address that the product match voluntary specification requirements. The cable that you supplied initially had the USB Logo for a certified USB product. That means the manufacturer needed to supply test data from a laboratory. A product that does not meet the tight specs like the second cable might be fine, as you mention, for non-industrial use. However, a CNC router (a large one) is industrial per se.

Maybe a good analogy is resistor tolerance:

Resistor tolerance is a measure of the resistors variation from the specified resistive value and is a consequence of the manufacturing process and is expressed as a percentage of its “nominal” or preferred value.

Typical resistor tolerances for film resistors range from 1% to 10% while carbon resistors have tolerances up to 20%. Resistors with tolerances lower than 2% are called precision resistors with the or lower tolerance resistors being more expensive.

A non-certified USB cable might have a greater tolerance -- that is to say that the characteristics may deviate more greatly from the preferred specifications set forth in the USB documentation.

As a practical issue, just make sure the cables are USB certified from your supplier and this should not be an issue.