AC voltage, Multimeters, ocilloscopes, etc.

Lots of info to be discussed but one thing you could do in buying a meter is to look for the words True RMS for the AC voltage reading. Look at Fluke for one brand that reads true RMS.

The 18 VAC is RMS.

That stands for Root Mean Square. If the AC voltage were displayed on a graph, the RMS is calculated by measuring a large number of values, squaring them, computing the mean, and taking the square root of the result. That's why the inexpensive meters just assume it's a sine wave, read the peak voltage, multiply by .707 and call it the RMS AC voltage. How the True RMS meters do it, I don't know...maybe I'll Google it...sounds interesting.

Note that when the voltage is controlled by chopping up the waveform, when the voltage is on all the way like 18 VAC, then the inexpensive meter will be fairly accurate since the waveform is almost a complete sine wave at max voltage.

Here's a description...

https://en.m.wikipedia.org/wiki/True_RMS_converter

Thanks.

So I did a little product research and ordered a Fluke 115.  Should be at the front door Saturday.

Although an oscilloscope is an interesting piece of kit to have around, it really isn't needed or suited to virtually anything you would do in a model RR environment from a diagnostic perspective.  Although one can invent things to measure with it (as I have done with mine, an HP 1743A analog "CRO"), it's only needed if you really want to see how clean the output waveform on a transformer is - and you will very seldom, if ever, really need to do that.  If you're into looking at the digital signals in TMCC systems, maybe - but then you need more than a basic oscilloscope.  But judging by the original post, my guess is that's a way off. 

GRJ,,gunrunner john somewhere has compared wave forms with various equipment using his scope. He breaks it down nicely with plenty of images.

JTrains posted:

Although an oscilloscope is an interesting piece of kit to have around, it really isn't needed or suited to virtually anything you would do in a model RR environment from a diagnostic perspective.  Although one can invent things to measure with it (as I have done with mine, an HP 1743A analog "CRO"), it's only needed if you really want to see how clean the output waveform on a transformer is - and you will very seldom, if ever, really need to do that.  If you're into looking at the digital signals in TMCC systems, maybe - but then you need more than a basic oscilloscope.  But judging by the original post, my guess is that's a way off. 

While it's true that the average model railroader probably won't have much use for a 'scope, anyone that's actually working on the electronics would likely find one beneficial.  I agree that the older analog models are much less useful, I have one gathering dust in the corner, haven't powered it on for several years.  OTOH, my digital 'scope gets a lot of bench time, and I'd be lost without it.  With the prices of digital 'scopes starting under $200, it's not that expensive to have a really useful tool at your disposal.  A lot of questions come up that there is continued speculation about, and I can lay them to rest in minutes by actual observation of circuit behavior.

Fluke is industrial, "the ISO inspector coming next week", "UL checking up on ya", "DCMA inspecting" stuff. Have Fluke and its calibrated and they look someplace else. You can drop kick it, knock if off a vibration system and have it fall to the concrete floor and it still works.

You can get acceptable stuff and Harbor Freight or tequipment.net, Hioki is one. Also Extech.

Lotta people like Rigol scopes. I don't like the idea of working at a defense contractor and having a scope with an internet connection that was designed and made in China. Keysight, formerly Agilent, formerly HP, has a new line of lower cost scopes that competes with Rigol and Instek (Taiwan). Doubt you care if the Chinese are eavesdropping on your scope when you are fixing three rail trains. Doubt you are gonna hook it up to your router either.

JTrains posted:

Although an oscilloscope is an interesting piece of kit to have around, it really isn't needed or suited to virtually anything you would do in a model RR environment from a diagnostic perspective.  Although one can invent things to measure with it (as I have done with mine, an HP 1743A analog "CRO"), it's only needed if you really want to see how clean the output waveform on a transformer is - and you will very seldom, if ever, really need to do that.  If you're into looking at the digital signals in TMCC systems, maybe - but then you need more than a basic oscilloscope.  But judging by the original post, my guess is that's a way off. 

Once early in my career, I was designing an RF amplifier and measured a dc voltage different from one end of a wire to the other end. I looked with the scope and saw a huge self oscillation signal on the wire. I often measured dc voltages with the scope after that...the scope when on when I got to the lab and off when I left for home.

illinoiscentral posted:

Lotta people like Rigol scopes. I don't like the idea of working at a defense contractor and having a scope with an internet connection that was designed and made in China. Keysight, formerly Agilent, formerly HP, has a new line of lower cost scopes that competes with Rigol and Instek (Taiwan). Doubt you care if the Chinese are eavesdropping on your scope when you are fixing three rail trains. Doubt you are gonna hook it up to your router either.

 

What makes you think that Keysight scopes are not made in the Far East?   They're also about twice the price of the Rigol 'scope with the same specifications.  I have an ATTEN 100mhz 'scope that's served me well for several years, and it was $250 a couple years ago, it has similar spec's to the popular Rigol models.

Since my 'scope isn't connected to the Internet, the Chinese are not privy to my valuable train secrets in any case.

I had reason to need a $225 Tektronix probe because they were not especially robust and one of mine broke at the junction of the compensator where it plugs into the scope. So I was at the Dayton Hamfest and asked the Rigol guys if they had any probes for sale. They said sure and sold me a pac of two X10/X1 probes with all the accessories for $30. I went back and bought a scope from them I was so impressed. BTW, their digital multimeter is a winner also.

I've used the Rigol 'scope, it's a very nice unit for the price.  When I bought the Atten, the were considerably more expensive or I'd have gone that way.  As it turns out, mine has been more than sufficient for what I do, I can't complain.  Beats the pants off the old 50mhz dual-trace analog one I was using for years.

gunrunnerjohn posted:

JTrains posted:

Although an oscilloscope is an interesting piece of kit to have around, it really isn't needed or suited to virtually anything you would do in a model RR environment from a diagnostic perspective. 

OTOH, my digital 'scope gets a lot of bench time, and I'd be lost without it.  A lot of questions come up that there is continued speculation about, and I can lay them to rest in minutes by actual observation of circuit behavior.

Certainly - someone like GRJ definitely needs a scope!    Out of curiosity: where do you find yourself using it the most, outside of your circuit design work?

I use it for checking stuff like transformer waveforms.   When a question about a circuit comes up, I can usually answer it far more effectively by hanging the 'scope on it than trying to track it down with a voltmeter.  For something as simple as trying to connect to the TMCC serial data, it's useful to see why something impacts the serial signal.  The 'scope enabled me to quickly see why the new ERR RailSounds Commander attenuates the serial data enough that the older ERR Cruise Commander M goes batty if they're both connected.  The addition of a simple three component buffer solved the problem, but finding the source of the problem was key.

GRJ,

I'd love to hear more about the TMCC serial data situation and the three component buffer you created.   

Your posts are great, and I love learning from them.

Thanks!

Here's the buffer, pretty simple.  It just removes the load of the RailSounds board from the serial data.  The optical encoder on the RS board loads down the serial data line enough that the Cruise Commander M doesn't see the serial signal.  I steal the 5V from pin 19 or 20 of the R2LC, serial data comes from pin-24 of the R2LC.  The serial data out goes to the RS board.  Since I shrunk this in heatshrink, the colors were just the wire ID so I knew what was what.

Edited to actually add the buffer drawing!

What does the RS input look like? A transistor base with a resistor to common? Or a data input on what chip? Just curious.

The serial data is driving an opto-isolator, so there's a 100 ohm resistor and the opto LED to ground.

This whole conversation started because I used my old Radio Shack meter to measure the smoke output voltage for an upgrade I am doing.  The voltage  at the resistor was measured at about 8.5v  and about 16v AC with the "boost" button.  I understand these values should be 12V and 14V respectively.  I was advised that I need to use a meter that was capable of measuring True RMS AC.

The Fluke 115, advertised to measure True RMS, arrived yesterday.  Much to my confusion, I got the same results - about 8.5v  and 16v AC.  I verified track voltage at a bit over 18v.  I measured the voltage across chassis ground and at the resistor.  I checked also at the smoke switch, and at the R2LC, all aginst chassis ground at various locations with same results.  I ressistance of chassis ground to rail and get 0.1 ohms

As part of the upgrade, I did suggested tweaks, including using a 20 ohm resistor.  The smoke output is satisfactory (though I expected and have seen better), but am I missing something here?

Any chance the new meter is NG?  (any test to verify proper operation?)

Thanks

I suspect that you're asking a bit too much from a meter that doesn't have four digits in the price.   I have noticed my true-RMS meters don't necessarily give my the computed values of non-sinusoidal waveforms either.  However, my digital 'scope is much better at the task.

So it seems that I just dropped a buck-thirty on the Fluke for no useful reason... 

That said (grunt), what is the "flavor" of the R2LC smoke output, in terms of wave form?  Is the voltage before the triac track voltage? Any chance of a one line for the smoke circuit?  Just trying to understand things better.

Then I'll know enough to be dangerous...

The R2LC smoke output in command mode is half-wave track power, only the positive half of the waveform.  The "boost" when you first start or hold down the 9 key is full-wave track power.  Conventional mode operation outputs full-wave track power.

Bear with me here... (I'm looking for a good book on this at the beginner level)

So, then why does the half wave of the 18v equal 12v?

Jim Harrington posted:

...

The voltage  at the resistor was measured at about 8.5v  and about 16v AC with the "boost" button.  I understand these values should be 12V and 14V respectively.  I was advised that I need to use a meter that was capable of measuring True RMS AC.

...

As part of the upgrade, I did suggested tweaks, including using a 20 ohm resistor.  The smoke output is satisfactory (though I expected and have seen better), but am I missing something here?

With a 20 ohm resistor, where do the numbers for the "correct" values for RMS voltage ,12V (or 7.2 Watts = V x V / R = 12V x 12V / 20 ohms) and 14V (or 9.8 Watts), come from?  That's quite a lot of power to only get "satisfactory" smoke output.  I'm thinking something like half of that (in Watts) should get you acceptable smoke.

If you're truly getting 16V AC RMS across a 20 ohm smoke heater, that's a whopping 12.8 Watts - and only "satisfactory" smoke?

Not familiar with what "boost" button is doing...I recall a recent thread where some transformers engage some kind of AC "boost" when the whistle button is pressed.  There are different methods to measure "True RMS".  Some methods (meters) do not account for any DC offsets and measure just the AC RMS.  Obviously the heater resistor is agnostic with respect to DC or AC.  I don't think that's what's going on here but something to be aware of if messing with conventional track voltage measurements when pressing the whistle or bell buttons.

Jim Harrington posted:

Bear with me here... (I'm looking for a good book on this at the beginner level)

So, then why does the half wave of the 18v equal 12v?

GRJ unveiled the man-behind-the-curtain.  If the voltage is half-wave, that is effectively a large DC offset.  If your meter only measures the AC RMS portion of a voltage signal...

Jim Harrington posted:

Bear with me here... (I'm looking for a good book on this at the beginner level)

So, then why does the half wave of the 18v equal 12v?

The integral shown on this site is solved to calculate the RMS value of any waveform. You can find the derivation worked out on many sites. But the table shows the results for many popular waveforms...

http://meettechniek.info/compe...erage-effective.html

7.2 watts is about right for really good smoke from the 20 ohm resistor.  Remember, it's a 3W resistor, so in free air it'll handle 3 watts without getting nearly hot enough to vaporize the smoke fluid.  As far as the exact power delivered, I stopped trying to figure that out, the 20 ohm minimum value was determined empirically by good smoke output and keeping the R2LC smoke triac in a safe operating temperature range.  I've seen the triac from 70C to 80C after extended running at 18 volts with the smoke on, so that's as much current as I feel safe in pushing the triac to handle.

Like Stan says, even my true-RMS meters don't really give me an accurate reading of the half-wave or the chopped wave transformer waveforms.  If I really want to know what's going on, I hook up my 'scope and look at the actual waveform.  I think Stan's comment that the meter is not measuring the DC component is probably why it's so far out.  I suspect they're assuming a symmetrical waveform with an average DC value of zero, even with the true-RMS meters.

The "boost" for TMCC mode is holding the smoke on key down.  That changes the triac operating mode to full wave output.  Obviously, you should not do that for very long, you'll have around 15 watts being dissipated by the smoke resistor.  Figure on about a 1 volt drop through the triac, you have about 17 volts RMS across 20 ohms.  That's OK for a few seconds, but not for any length of time.  If you don't smoke the triac by overheating it, you'll probably cook the smoke unit with extended running with that kind of power.

cjack posted:

Jim Harrington posted:

Bear with me here... (I'm looking for a good book on this at the beginner level)

So, then why does the half wave of the 18v equal 12v?

The integral shown on this site is solved to calculate the RMS value of any waveform. You can find the derivation worked out on many sites. But the table shows the results for many popular waveforms...

http://meettechniek.info/compe...erage-effective.html

That points out why a true-RMS meter can't possibly be accurate for all of them, no way they know exactly what waveform they're dealing with.  Only by sampling the waveform and computing the value would the have a shot at accuracy.

Here's something about that from Fluke. They measure RMS "very carefully" they say.

http://www.fluke.com/fluke/use...trical/true-rmsfacts

There they admit that most meters can't deal with the full calculation and presume that it's a symmetrical AC waveform with no DC component, just as we previously suspected.  I hadn't really dug into it enough to find those, good catch Chuck.   Nice to know the facts about the meters.

Clipped from Fluke;

"What can you do if you have only the ac-coupled rms measurement capability in your DMM and you want to measure the combined ac and dc energy? Bring on the calculator.

You can first measure the ac voltage and then the dc, recording both values. Next, you square both terms and add these squared values together. Finally, you take the square root of the sum and you have the ac+dc true-rms value of the signal."

Still doesn't add up, but it's closer...

√(8.5²+7²)≈11v

Having received the same readings with my old Radio Shack meter, I will be returning the Fluke; unless someone can convince me otherwise...

For 99% of the stuff you'll be using the meter for, I doubt the Fluke will be a huge improvement over an inexpensive DVM.  I have a bench model Fluke, but I bought a Mastertech MS8226T DMM when my old Fluke 77 finally ceased to make accurate measurements.  The MS8226T meter does about everything but mix drinks, and I've been using it for a couple of years daily.  For $42, it was a good deal, and has taken a lot of punishment, a few drops, etc.

"Having received the same readings with my old Radio Shack meter, I will be returning the Fluke; unless someone can convince me otherwise..."

Hard to say. I like an accurate meter...how does it compare for less significant digits? Like when you measure the same DC voltage and the RS reads say 1.566 and the Fluke reads 1.620. That's usually what separates the expensive from the not...whether it matters or not is the question.

Jim Harrington posted:

...

Still doesn't add up, but it's closer...

√(8.5²+7²)≈11v

Having received the same readings with my old Radio Shack meter, I will be returning the Fluke; unless someone can convince me otherwise...

I still want to know where your "12V and 14V" RMS readings come from.  After all, you are measuring 16V (not 14V) in the "boost" mode and that is for (apparently) a full-wave, no-DC, waveform.   In other words, if you had indeed made the separate DC and AC measurements, then did the math, and it had come out to be 12V (instead of 11V) would you then keep the meter? 

My MS8226T is very close to my 4.5 digit bench Fluke for all the measurements I've checked.  It is easily a 1% or better meter, and I have no need for anything more accurate.

This would be neat to watch a waveform as the engine or whatever goes around the layout...

https://www.crowdsupply.com/ae...ireless-oscilloscope

stan2004 posted:

if you had indeed made the separate DC and AC measurements, then did the math, and it had come out to be 12V (instead of 11V) would you then keep the meter? 

Being that the RS and Fluke both took the same readings (or close enough), I'll put the $130 back in my pocket to buy other stuff...

stan2004 posted:

Jim Harrington posted:

...

Still doesn't add up, but it's closer...

√(8.5²+7²)≈11v

Having received the same readings with my old Radio Shack meter, I will be returning the Fluke; unless someone can convince me otherwise...

I still want to know where your "12V and 14V" RMS readings come from.  After all, you are measuring 16V (not 14V) 

12v and 14v taken from an older post...

https://ogrforum.com/t...-unit-output-voltage

Jim Harrington posted:

stan2004 posted:

if you had indeed made the separate DC and AC measurements, then did the math, and it had come out to be 12V (instead of 11V) would you then keep the meter? 

Being that the RS and Fluke both took the same readings (or close enough), I'll put the $130 back in my pocket to buy other stuff...

 

Understood.  Just want to be clear that I don't think the meter itself is "defective".