Battery replacement for battery-only Igaging DROs?

[homebrewed]

I've got 5 Igaging remote-reading DROs on my mini machines (three on my mill and two on my lathe). They can only use CR2032 batteries, and they all seem to eat them quite rapidly. I have stacks of dead batteries waiting to be recycled. I got fed up with continually swapping them out so initially modified one of my DROs by opening up the controller and soldering wires to the internal power and ground connections, then connected them to a battery holder with two D cells. D cells have a MUCH higher mA-Hr rating than CR2032's. Unfortunately, the extra wires (and perhaps the slightly lower supply voltage of 3.0V vs. 3.3) caused the position readout to become extremely unstable due to noise pickup. I had traded one annoying problem for an impossible one. I also didn't like the necessity of modifying the DROs in order to hook up the external power.

Thinking about it some more, I realized that the USB connector brings out Vcc and Gnd (to power the sensor on the scale). So I could access power and ground by making a little board with a male USB connector on one end and a female on the other. The male would plug into the controller and the female would accept the cable from the sensor. An external (noisy) supply line would be conditioned by a pi topology RC filter on the board. Much cleaner, and a way to do it without modifying the controller. PCBs are so cheap now that the whole thing would pay for itself fairly quickly. Not to mention the reduced annoyance factor.

This idea seems so obvious (in retrospect) that I'm wondering if anyone else has tried it, and, if so, how it worked for them. Even if their solution involved just opening up the cable and soldering wires to Vcc and Gnd, I'd like to know.

 

[AGCB97]

I've been switching a lot of my things over to use Li-on batteries. The 18650 is about 4 volts when fully charged. There are all kinds of cheap voltage regulator modules in any voltage you want. Some are adjustable. So my stock pile of parts contains battery holders for 1-4 cells, regulators in most common voltages, and a good 2 cell charger with all the safety cutouts (over voltage, over current over temperature and maybe a few more OVERS) oh yes and a bunch of batteries. Start with one of each and if you like them then buy in quantity.
That's just my take.
Aaron

 

[markba633csi]

Ferrite snap-on filters might work there, or toroid cores from an old computer power supply. Florescent lighting could be one culprit
-Mark

 

[homebrewed]

Ferrite snap-on filters might work there, or toroid cores from an old computer power supply. Florescent lighting could be one culprit
-Mark

The main source of EMI is the mini mill's motor controller. I could see the DRO readout numbers' wild fluctuations settle down as soon as I turned the motor off. I did try shielding the motor cables -- only a slight improvement; and a snap-on filter on the mill's power cord -- no improvement. Shielding the DRO cables produced a noticeable improvement but did not completely cure the problem. Electrically Isolating the scales from the mill did nothing.

The best results came from a combination of inserting an RC filter in each sensor's Vcc connection, shielding the DRO cables, and hard-wiring the scale bodies to their respective DRO grounds. The last step seems counter-intuitive, because it introduces a ground loop; but it worked for all 3 DROs. Even with all this, on occasion a DRO will STILL go nuts. That's a sign that the batteries in that particular DRO are starting to go flat. I think low supply voltage decreases the circuit's noise immunity, so that's another reason to go with an external supply. I will design the supply to include battery backup in case of power failure.

I suspect the hardwired ground approach worked because the scale is made from aluminum, which readily forms an insulating oxide skin. The sensor (sender?) has a sliding contact to connect the scale to its internal ground (I ohmed it out to verify this) , but if it's highly resistive the scale can bounce around, electrically speaking. Since the sensor uses capacitive coupling to read out its position, this is a Bad Thing. One way around this would be to use the more expensive stainless steel rules -- no oxide skin to interfere with the sliding contact. If/when any of my current DROs fail, I will replace it with the stainless steel version.

 

[markba633csi]

So the EMI is coming from the motor controller directly through the air. That's a hard one to solve, you can't easily stop it at the source.
I didn't know about the oxide skin issue, kinda sucks
-M

 

[homebrewed]

So the EMI is coming from the motor controller directly through the air. That's a hard one to solve, you can't easily stop it at the source.
I didn't know about the oxide skin issue, kinda sucks
-M

Yes, I've found it a difficult nut to crack. The oxide skin problem is an hypothesis, but it appears to be consistent with the solution I found (pretty much in desperation ).

For additional confirmation of the oxide skin theory, I have two DROs installed on my mini lathe (it also uses a PWM style motor controller). I bought them after I solved the DRO problems on my mill, so they DO use the stainless steel scales. Initially I had some noise problems with them, too; but just implementing the inline supply filters took care of it. No additional shielding or external connections to the scale. They still eat batteries pretty fast so if my external supply idea works I'll install the external adapters for them, too.

BTW, implementing the inline filters did require some board surgery. I cut the Vcc trace, inserted a 1K resistor and installed a 1uF cap to ground on the other side of the resistor. The sensor draws at most 70uA -- that's if it draws the lion's share of current, which likely is not the case. So worst-case there's just a 70 millivolt drop across the 1K resistor. The RC time constant is 1 millisecond so it's not filtering out 60 Hz noise -- just the higher frequency stuff coming from the motor controller. But that appears to suffice.

 

[RJSakowski]

A different setup as I am using the iGaging scales with the Touch DRO. I had EMC problems with my lathe motor when the motor was switched off. I tried a number of solutions and finally end up soldering bypass capacitors across the power leads in the pickup heads. That solved my problem.

 

[addertooth]

A ten microfarad tantalum capacitor soldered at the DRO end of your run should fix your induced noise problem. Keep in mind that tantalum caps are polarized, so get the plus and minus connections right. Tantalum capacitors have VERY low impedance, and they snub stray signals quite well.

As someone said, and 18650 lithium with a low drop out (LDO) regulator can work, or a disposable 3.3 volt lithium battery can work too. The disposable 3.3. volt cell will not require a regulator. Even a 3.3 volt disposable which is half as long as a AA battery should give you a run time at least 30 times longer than your existing CR2032 button cell.

 

[homebrewed]

I finally got my power adapter boards in and built one up to see how it works. It didn't power-up the DRO at all, so I got out my DVM and did some measurements. I won't go into all the stuff I did, but I discovered that the particular DRO I chose to test my battery replacement scheme draws _much_ more current than the one I initially used to "vet" my design. I'm talking about 70 micro-amps for your average Igaging DRO, but THIS one draws 1.5mA!!! OK, it is probably defective -- but the darn thing works fine if it has a power source that is stiff enough to provide that current. Of course, that kind of current draw translates to short battery life -- but that's what I've been dealing with all along.

The bright side is that my USB jack/plug connections are OK. I built up a second board but bypassed the filter so it was hardwired to the 3.3V power supply, and that one worked just fine.

So version 2 will use a board-level power supply regulator to condition the voltage going into the DRO. At this point I can't return the deviant DRO so, since it is working OK other than the supply current problem, that's the most expedient solution. 10 PCBs are less than $10 in total so it also is the cheapest.

One issue I ran into with this project is that, as far as I could determine, all inexpensive USB plugs are meant to be used in a USB cable assembly, so they're not compatible with a PCB. I looked at the plug dimensions and figured out that, at least for the older-style mini-B connectors used on my Igaging DROs, I could mount a cable-style jack if I used a thinner PCB and added cutouts on the PCB to accommodate the housing body. At least that part of the project worked OK....

Fortunately, the PCBs were really inexpensive so I'm not out much, other than time. And for a hobbyist that's OK.

 

[AGCB97]

Were the PCBs you used 'off the shelf' or special made? If special, from where?
Thanks
Aaron