Cheap o-scope recommendations?

[garyhlucas]

It is important to recognize there are two completely different grounding methods for completely different purposes. The goal of Power grounding is to have as many ground loops as possible. This is so that when a fault occurs where 5000 amps may flow no two pieces of adjacent metal will have a potential difference that could be dangerous.

The goal of Signal grounding is to keep noise out and that to work you want no ground loops at all that might allow current to flow. So a single ground point with a separate ground wire from every device and every shield is way you make this happen.

The two methods can be tied together but only at one point.

[J Tiers]

Even with signal grounding you may not get good results from the typically recommended "star" grounding.

The "real" goal of all shielding ("grounding" is included in that) is to keep noise out of signal wiring.

You can do that by shielding the signal wiring entirely, so that noise cannot reach it. You can do that in some cases by eliminating any resistances/impedances that are common to the signal and the noise. That is the main goal of "star" grounding, to be sure that signal and a potential noise source do not share a wire, or a PCB trace.

However, various things matter, conditions of frequency, geometry of the cabling, etc affect the goodness of shielding. "Ground" in this discussion means anything that can be a reference point for the signals inside the shield, and which is connected to the source or receiver of the signals.

Frequency matters.
1) A shielded cable, such as RF coaxial cable, maybe RG-59, is a pretty good shield against RF. Currents on the outside do not penetrate to the inside, so RF currents on the outside are separated from those on the inside.

But make one of the signals , for instance the one "on the outside", a low frequency, and it will conduct throughout the cross-section of the shield, appearing on the inside as well as the outside. The shielding does not work against it.

Those signals can be separated, because of frequency, and a high frequency might not generally be affected by a low frequency. One may choose to ground the daylights out of such a cable, without bad effects, so long as it is done in a good way with regard to geometry..

However, the shield is not perfect, so large currents, even of high frequencies, on the outside, will probably have some conduction of current on the inside, where it can mix with the signal carried on the inside surface and the center conductor.

2) For low frequencies, in the lower audio range, and typical power frequencies, star grounding is very effective. At higher frequencies, star grounding systems may start to cause more problems than they solve, when the wiring to the "star point" begins to create an "antenna" that picks up noise. Then it is necessary to revert to a different grounding scheme.

That can even happen at low frequencies, when there is a strong magnetic field, for instance from a transformer or inductor carrying power current. Transformers can produce strong local fields, and often make trouble of that sort.

Geometry matters.
Routing of a cable affects how much current it can pick up on the outside. There is a "loop" formed by the shield vs the nearest grounded item. So if the cable is routed such that it is spaced up off the surface of a grounded item, such as a chassis, etc, then there is an opportunity for a signal to be induced, if the cable is grounded to it at points far apart. That risks pickup since the shield is not 100% impervious (even if it is solid, and not braid).

So you want to keep the cable down against the grounded surface, and minimize the loop area. In cases where you cannot do that, it may be best to NOT ground the far end of it. You still may pick up a signal, due to high "antenna currents" on the cable, but that is usually only in areas near high power transmitters. A VFD probably will not produce such a field.

Even with unshielded wires, keeping them close to a grounded surface will minimize both pickup and emission of signals. A VFD produces a strong "common mode" signal, between the chassis of the VFD and the 3 power wires. The capacitance of the motor wiring to the case completes that circuit, and can create a good "loop antenna" that broadcasts a strong noise signal. Connecting the VFD chassis to the machine chassis, routing the wires close to the surface of the machine, and along, not "over" any gaps, will lower the noise problem. In general, a star grounding scheme may be counterproductive in that situation, if it is even possible. The best answer may have essentially nothing to do with grounding.

You really need to figure out whether your problem is due to a "shared impedance" issue, regular "antenna type pickup", capacitive coupling, or a "loop induction" issue, etc. The solutions can be very different. I never like to see someone making a blanket recommendation of a particular ground strategy without considering the rest of the wiring, the nature of the noise vs the signal, the source of the noise, etc.

[Paul Alciatore]

Well, my post was long enough. Thanks for taking it a bit further. I can remember many times when I was looking at a digital signal in the tens of megaHertz range and wondering what it really looked like. Unfortunately I never had a scope that went into the gigaHertz range much less beyond. There you are looking at a pulse that does not look like either a square wave nor a sine. And you wonder just what it really is and what (something in the circuit or something in the scope or it's probes) is making it look the way that it does.

If I had my druthers, I would ask for a scope that went to at least 100 times the basic frequency in both response and sampling rate for a digital scope. And it is not only the frequency of the signal you are looking at, but also the frequency response of the circuit that will receive that signal. Some logic chips can react to very fast spikes that will never show up on any but the absolute best scopes. But, I dream.

That is really not true. It LOOKS true, but actually it is misleading.

The fundamental issue is misunderstanding frequency and frequency response. You alluded to that earlier in the post, but then seem to ignore it in the bit above. For this I am assuming the sampling is sufficiently above the frequency response, which is generally true.

The frequency response you need is NOT related to the pulse frequency. It is the highest frequency in the waveform. A square wave theoretically has an infinite frequency content. But because it is produced by some physical device that itself has limits, the upper frequency is limited. The difference shows up in the "sharpness" of the "corners", and the slope of the rising edge.

To get the correct "picture" of the square wave, you must have a 'scope that has a frequency response of at least that highest frequency. If your 'scope is not of at least that capability, you will see a picture that has blunter corners, rounded off instead of square, and also the rising and falling edges will be sloped instead of almost straight up and down.

Those sorts of distortion can affect needed measurements, producing errors in apparent timing, apparent rise time, etc, which may be specified factors that you need to know.

The "rule of thumb" is to have a frequency response 10x the pulse frequency. But that is just an estimate, something that is "usually true" for "typical measurements".

If you need accurate rise time, or relative timing measurements, you may need considerably wider frequency response than the 10x rule gives. You will need to base the frequency response on the rise times you need to measure, and not on the pulse frequencies.

[Paul Alciatore]

A lot of talk about grounding. In my experience there often were two distinct ground systems in use.

The first was the building's electrical ground as prescribed by the code. This did radiate out in a star fashion from the point where power entered the building. Most buildings will have a single, code compliant ground rod at or near that point. The problem with it was that the ground rod often was not a really good ground. It was simply installed per the code and nothing was done to verify it's effectiveness. And these power ground systems often contained loops that were created by interconnections, mechanical as well as electrical, between the various machines. Another characteristic of the power ground was the size of the conductors. They are sized to trip breakers, not to swallow electronic noise. They are simply too small for good noise suppression.

The second ground system was one that was deliberately installed to handle electronic noise, aka a signal ground. When I specified and installed these signal grounds I used terms like "2 inch wide copper strap" and specified wire sizes with multiple zeros ("000 gauge"). Another important point in specifying a signal ground is how the connections are made. I liked to overlap the 2" copper straps for a full four square inches of contact and then drill a bunch of holes in that overlap and use a torch and a lot of solder to ensure a good, low resistance connection. Just one such joint often took a half hour or more to complete. Connections between the copper strap and heavy gauge wires had the strap wrapped around the wire and were also soldered as were the oversized lugs that connected the ground wires to the individual machines. The point here is that a mechanical connection can develop corrosion over time and have a few Ohms of resistance. Even that, admittedly small resistance, can allow signals or noise to be picked up in the poorly connected branches.

I liked to keep all signal ground conductors as close to the actual, earth ground as possible. Even a properly grounded conductor that sticks up into the air will become an antenna that will pick up any signals or noise in the air at that point. In case you do not believe this, consider the fact that one type of AM radio transmitting tower (in AM the tower itself IS the antenna) is actually grounded at it's base and the AM radio signal is fed to the tower at a point above the ground. That radio signal is radiated, it's not just shunted to ground. Also remember that any antenna will work equally well as both a transmitting and a receiving antenna. A grounded conductor can and does make a good antenna. Keeping it close to the earth (real ground) is the only way to defeat this.

I liked to install a signal ground with it's own, central point. At that central point I liked to have several ground rods extending over 15 feet into the earth. Those ground rods would be spaced around 5 to 10 feet apart. And yes, holes were drilled through the concrete slab to install them.

Overkill? Perhaps, but with such a signal ground system, properly installed I almost never had ground or ground loop problems between machines or equipment. You may not need all of the above in a machine shop, but if you are having difficult problems it does not hurt to know these things.

[JRouche]

[SIZE=14px][FONT=Arial]I want to test the encoder on my ORAC lathe with Centroid Acorn control board. 8,000 to 40,000 encoder counts per revolution

Sorry, no help for the other thing. 40,000 steps? I would love to see that. 8000, that also.

Tech just breaks me down all the time. I dont mind. I cant keep up. JR

My encoders do not come close to those numbers.

[PStechPaul]

For calibration of high current circuit breaker test sets, we used a 2500 amp 50 mv shunt, mounted on bus bars which bolted onto the output plates. The millivolt signal was measured with a precision true RMS DMM or a proprietary device based on the circuitry of the Ortmaster, which I designed. It used a balanced differential input to an instrumentation amplifier. The high currents (up to 50,000 amps) involved created a strong magnetic field which would be induced into the multimeter leads, creating measurement error. It helped to twist the leads together, but the best results came with using a shielded twisted pair. It is very difficult to shield against magnetic fields - they are only attenuated by means of high permeability steel, such as mu metal, and may require multiple layers. But a twisted pair cancels the induced EMF, and full differential instrumentation with high common mode rejection is also effective.

When using a scope, it is important to connect the shield closely to a ground point of the circuit. You should try connecting the probe tip to the ground clip and hold it close to the circuit you are investigating. You will often see noise signals even with the probe grounded, showing that the loop is acting as an antenna. It may be necessary to use two channels of the scope in differential mode for some measurements, and you must watch out not to exceed the scope's common mode spec.

[J Tiers]

Most probes come with a little clip that goes on the end of the bare probe. It connects to the ground ring, and has a "tail" the same length as the center pin. On a ground-plane type PWB, you get the ground clip "tail" on the plane and the center pin on the signal point.

In most cases that is as good as it gets. In a few cases, more has to be done, but the "loop" created by that connection is very small indeed.

[PStechPaul]

Some words of wisdom from the Guru himself, Bob Pease:

https://www.electronicdesign.com/tec...e-stuff-anyhow

His book on troubleshooting:

Discussion of scope probes:

https://www.edn.com/are-you-measurin...r-scope-probe/

[JRouche]

My Job sucked. Some times, most often. My short timers calendar is 3+years long. I made the squares, 1152 of them. And marked them off for three years of war time hours in the Navy. .

I have my original but...

Same reason I do not post my machine or otherwise work here.

You folks are too judgely for a short timers sheet.

FYI Its a nice one. Never folded. JR

[J Tiers]

That's quite a good little article by RP (he wrote no bad ones). There are a number of books on the subject as well. If anyone is interested I can find the ones I have and give the titles etc.

The best article I saw I think was one of RPs, but it involved instructions to an engineering student who was working for a smart professor, managing the lab. The professor said that whatever was wrong with anything, that student had to fix it himself. Calibrations, repairs, whatever. The result, as you might expect, was a very good education in practical electronics, good for a student who had little practical experience.

That leads to the home shop..... same thing, if it's broke, you fix it, if you want a tool, you make it, etc (if it is in any way possible). Nothing like that to give a person the experience to tackle anything the shop is capable of.

JR.... go ahead, post stuff. I've posted things for years, sometimes I get yelled at for some perceived fault, sometimes not. But I'm stubborn at least 3 ways by family background, so it does not stop me.

Go for it, I'm not gonna punk you for it. We can both holler at anyone that does!

[MaxHeadRoom]

how much was that new? i cant find any price, is it discontinued?

Spoke to my contact and the Company changed hands and he is in the process of restoring stock, so they should be ready soon.
https://www.gsdelectronics.com/mfo-1/

Evidently the universities have been buying them like crazy!
Max.

[andywander]

...Incidentally, concerning the practice of using a quadrature (90°) spaced waveforms, with all of the common optical head unit, the detected signal starts out as a sine wave...

Really? How does an encoder that is looking at a series of ON/OFF transitions output a sine wave?

[J Tiers]

Seems nice enough, I wonder what their "affordable price" is.

[MaxHeadRoom]

Seems nice enough, I wonder what their "affordable price" is.

Not sure now, the original model I purchesed when they first came out was $140.00.

Really? How does an encoder that is looking at a series of ON/OFF transitions output a sine wave?

More observable on low res at slow RPM, the aperture does not expose the infra red LED all at once, the transition creates a gradual increase in brightness.
They usually use a Schmitt trigger to square up.

Max.

[Sparky_NY]

More observable on low res at slow RPM, the aperture does not expose the infra red LED all at once, the transition creates a gradual increase in brightness.
They usually use a Schmitt trigger to square up.

Max.

+1 Your average encoder does the "squaring up" internally with the encoder pcb.

Heidenhein has made a LOT of encoders and scales that output sine waves, they are somewhat famous for that. They also have "converter" boxes they make to convert the signal to TTL quadrature. A lot of people have gotten the surprise of discovering they were not TTL quadrature outputs when converting machines with heidenhein controls to mach3/linuxcnc etc. They were common on the bridgeport Boss cnc mills with Heidenhein controls.