Reliability Issues with iGaging Digital Scales

Thursday, July 25, 2013

Many people are wary of digital readout setups that use capacitive scales because such setups tend to be sensitive to electrical interferences. A shop full of motors, fluorescent lights, transformers and other equipment is a noisy environment and when that noise gets into the DRO system it can cause havoc with the readout. The most common symptoms of such issues include random resets and floating position readout. While it appears that all “chinese” scales and calipers are somewhat prone to these issues, the majority of the complaints fall on IGaging Remote Digital Readout scales. Surprisingly the scales themselves are rarely the culprit. In my experience 9 out of 10 problems with the iGaging scales can be traced to the wiring and/or the power supply.

The Basics

Before going much further let’s talk about the wires. For convenience’s sake we treat them as perfect conductors but in reality they are not. Every piece of wire, no matter how short, has some resistance, impedance, capacitance and inductance. In other words a USB cable that connects the scales to the controller act as a resistor, capacitor, a weak transformer and an antenna.

To better visualize the idea you can set up a simple experiment: Take a 6-10 foot piece of thin wire and lay it alongside a power cord of a running motor. Now measure the AC voltage between one end of the wire and the ground. When I did that with a 10 foot piece of unshielded USB cable my Fluke multimeter registered over 3.5V of noise voltage. Additionally, the wires had about 0.6 Ohms of resistance (at 500 mA this resistance would cause close to 0.3V drop).

A piece of wire sitting next to a power cord picks up over 3.5V of noise voltage

Consider a basic DRO setup that uses three iGaging scales that are connected to the controller using the supplied 6-foot unshielded USB cables; the controller is powered from a wall wart or an old phone charger. At the very least this setup has about 24 feet of unshielded cables in close proximity to the powerful spindle motor, fluorescent lights, compressor(s) and other noisy shop equipment. All this interference is picked up by the wires and translated into voltage fluctuations that can exceed the 3.3V DC power supply voltage.

There is a good amount of resistance
between the frame and the ground lead

Additionally there might be one more, less obvious, noise hazard: long ground loop between the machine and the controller. More often than not the wall wart uses a two-prong plug. If it happens to be one of the newer transformerless power supplies, controller’s ground is directly connected to the neutral line of that outlet. Keep in mind that in the USA the neutral line is tied to the ground at the switch panel, so in effect there is a connection between the controller’s ground and the machines frame created by a long span of wire (from the neutral to switch panel and then back to the machine via ground line). At the same time, since iGaging scales’ frames are connected to the negative side of their power supply, there are three ground connections to the machine’s frame though the USB cables. When the motor is off, in my garage I can measure about 0.5V of potential difference between the neutral and ground lines in the outlet used by my X2 mill; when the motor starts I can often register spikes as large as 5V (the outlet is only about 20 feet away from the switch panel).

Random Resets

If the DRO has a ground loop as described above, every time the machine’s frame is at or above 3.3V potential in relation to the controller's ground, the power supply voltage the scales receive is effectively reversed. Although less likely, a large spike on the ground wire in the USB cable can create the same situation. When this happens the scale immediately looses power and are effectively reset.

The scope is seeing 0.6V pulses between ground and neutral lines

You might recall that iGaging Remote Digital Readout (and most of the other capacitive scales and calipers) are incremental encoders. Rather than measuring the position, they measure distance traveled since the the power was applied. To do so the number of clicks is stored in the scales memory for as long as the power is applied; remove the batteries and in a few minutes the scales “forget” their position. The reason iGaging scales are more susceptible to this problem is that the batteries are held in the display unit while the encoder is in the reading head. When the head loses power the click count is lost, so the readout “jumps” to 0.

Floating Position

When this occurs, the position is, in simple terms, jumping all over the place. To understand why this happens we need to look closer at the communication mechanism used by the scales. The “remote DRO” scales use a serial communication protocol, where the display unit (or the wireless DRO controller) provides a number of clock pulses and the scale sends the data in binary format by turning certan pin on and off as needed. In other words the controller toggles a pin on and off 21 times and on the “off” cycle reads a value of the data line coming from the scale. If one of the 21 bits is not read correctly the whole readout will be off. For example, consider these two binary numbers: 1011 1111 1111 1111 and 1111 1111 1111 1111 (49,151 and 65,535 respectively). As you can see a single misread pin state can throw the position off by around 25%.

There are a few different things that can cause such misreads. In the simplest case a short voltage spike in the data line caused by the interference can be misread as “high” by the controller; a negative spike can cause a false “low” reading.

Similarly, glitches on the clock line might cause the reading head to set the bits prematurely. I.e. in response to a false “high” the scale can send bit number 6 (for instance). By the time the controller sets the real clock #6 the scale is already thinks it’s sending bit 7 and so on.

Bottom capture shows what happens when the scale detects glitches on the clock line

Finally the wires can create enough impedance that the “high” voltage falls below the microcontroller’s threshold and is never interpreted as logical 1. For instance, Arduino running at 5Vcc requires above 2.6V input for a logical 1 and below 2.1V input for a logical 0. Since we are already pushing the envelope a bit with the 3.3V scale’s power supply, there is only 0.7V to spare.

Side Note

Although this is not very common, the issues with stability can be caused by poor connections between the scales and the DRO. The Mini-USB connector that plugs into the controller board is the weakest link in the system, since it doesn’t provide very solid mechanical interface. After some time it can vibrate itself loose and cause intermittent connection issues. Similarly, after a while a conductor can break in the cable, leading to the same issues.

Conclusion

If you’re having stability issues with the iGaging Remote DRO scales, I hope this post will help you better understand the underlying causes. In the next post I will describe a few easy remedies that can eliminate the majority of those issues.

4 comments :

  1. Great info here; thanks!

    Scot

    ReplyDelete
  2. Hi Yuri
    Really great work, I'm planning on buying a couple of iGaging scales but am afraid for bad read outs.
    Will the use of STP Ethernet cable be a good idea, I don't care about the connectors.
    Hop you can help.
    Cheers,
    Marco

    ReplyDelete
  3. My DROs kept resetting and turning on all by themselves. Put a Tek 465B between gnd and netural at another outlet on the same power line and saw 3Vp-p 60HZ. Turned on the Grizzly G8689 and what a mess of spikes peaking at 12V. No wonder there is a problem. I removed the ground from the mill's power cord and haven't seen the problem since.

    ReplyDelete
  4. I took the route of installing a new cable but hardwiring it to the scale board. This despite the small work space was infinitely easier than locating and modifying a USB Mini-Mini cord. I also shortened the cord to 6" which leaves no excess "antenna"

    ReplyDelete