Whistle Waveform from Model ZW Transformer

Kudos on your write up and analysis. Out of curiosity, does this ZW have the selenium original rectifier discs in it or some other type of rectifier?

That's a great question.  There is no indication that the rectifier has ever been replaced.  Plus, the ZW has been sitting in my basement for over 40 years, and I never replaced the rectifier. So, it probably has the original selenium stack.

But that raises another issue. I ran the simulation using a silicon diode, and I suspect the selenium rectifier has a different forward voltage. I need to check that.

As you may know, the [EDIT: selenium] rectifiers don't work as well, if at all, after that amount of time.  I still plan to replace them in my ZWs.

From here:

Selenium rectifier - Wikipedia

the forward voltage drop is 1 volt per plate. Each plate can withstand a reverse voltage of 25 Volts, but I don't know how many plates were used in the ZW. (I suspect either 1 or 2.)

I increased the forward voltage of my diode to 1 and then 2 volts in my simulation, but that didn't help. The only thing that does help is increasing the value of the 1.5 Ohm resistor. That low value was always suspicious. I need to measure them resistor in my ZW.

I believe it was a single plate in the ZW.

@vroberts  there was a recent discussion linked here about silicon diode vs modern replacement discs for the whistle circuit that may be interesting to you.

EDIT: My understanding of the 1.5 Ohm resistor is that when the whistle lever is advanced into the hold position, it drops the voltage from the starting position to keep the whistle motor running on PW tenders.  This is incorrect as explained by John below.

As far as the number of plates, the A and D outputs each has one.

@vroberts posted:

From here:

Selenium rectifier - Wikipedia

the forward voltage drop is 1 volt per plate. Each plate can withstand a reverse voltage of 25 Volts, but I don't know how many plates were used in the ZW. (I suspect either 1 or 2.)

I increased the forward voltage of my diode to 1 and then 2 volts in my simulation, but that didn't help. The only thing that does help is increasing the value of the 1.5 Ohm resistor. That low value was always suspicious. I need to measure them resistor in my ZW.

The rectifiers in the Lionel whistle controllers are copper oxide, not selenium. 6 or 7 years ago I measured the characteristic curves and posted them here; I seem to recall that they looked like a very small voltage drop, maybe 100 millivolts, in series with about 1/2 an ohm. It was not a sharp knee, there was significant current flow at very low voltages.

I'll see if I can find my old data.

@SteveH posted:

My understanding of the 1.5 Ohm resistor is that when the whistle lever is advanced into the hold position, it drops the voltage from the starting position to keep the whistle motor running on PW tenders.

Actually, the 1.5 ohm resistor is connected across the rectifier.  That reduces the DC offset, but also boosts the voltage a few volts for to compensate for the voltage drop due to the current for the whistle tender motor.

Thank you, @SteveH, @gunrunnerjohn and @PLCprof. This is all very useful information.  I measured the resistance between A and U when the whistle switch was activated and that should give me the value of the resistor. It was well over 100 Ohms.  (Or I may be misunderstanding the functioning of the Whistle/Direction switch.)   

So, I increased the value of the resistor to 100 Ohms in my simulation and that produced the expected offset sinewave.  Clearly very different than my measured waveforms.  So, I believe I have an issue with the ZW that I must resolve before I put it up for sale.

At a minimum I will read the thread on replacing the old rectifier with a modern silicon diode, and then make that replacement.

I've attached a copy of my whistle circuit simulation for the A circuit, if you are into such things. As I said, I may have misunderstood how the Whistle/Direction switch works.

Thanks again!

@gunrunnerjohn posted:

Actually, the 1.5 ohm resistor is connected across the rectifier.  That reduces the DC offset, but also boosts the voltage a few volts for to compensate for the voltage drop due to the current for the whistle tender motor.

John, thanks for providing the correct info about the resistor's function.  Looking at the schematic, what you're saying about how the current flows through the compensating winding et. al. makes sense to me now.  Thank you.

@PLCProf  Thank you for the Copper Oxide information.  I had apparently read some incorrect information elsewhere about the rectifier disc's composition.

I sure am glad you guys are here to keep us mortals straight.   Thank you.

Based on this additional info I put the resistor back to 1.5 Ohms, but also lowered the rectifier forward voltage drop to 0.1 Volts and the forward resistance to 0.01 Ohms. I also noticed that my load resistance was far too high. I had run the test with 22 Ohms, but had used 100 in the simulation, for reasons I cannot explain. I dropped it to 12 Ohms to get more load current. (We need enough current so that the voltage across the resistor is high enough to activate the rectifier.)

These changes did produce a small DC offset in the sinewave, but not much, less than 1 volt, which I do not think is enough to activate the whistle relay in a tender.

That brings up another question. What is the typical operating current of a Lionel engine?  Perhaps 12 Ohms is still too high. 

The Rev 2 circuit is attached.

@vroberts posted:

...That brings up another question. What is the typical operating current of a Lionel engine? ...

About 2 Amps is about what I figure on up to 5 Amps or so if including whistle motor, smoke, incandescent lighted passenger cars, etc.

@vroberts posted:

Based on this additional info I put the resistor back to 1.5 Ohms, but also lowered the rectifier forward voltage drop to 0.1 Volts and the forward resistance to 0.01 Ohms. I also noticed that my load resistance was far too high. I had run the test with 22 Ohms, but had used 100 in the simulation, for reasons I cannot explain. I dropped it to 12 Ohms to get more load current. (We need enough current so that the voltage across the resistor is high enough to activate the rectifier.)

These changes did produce a small DC offset in the sinewave, but not much, less than 1 volt, which I do not think is enough to activate the whistle relay in a tender.

That brings up another question. What is the typical operating current of a Lionel engine?  Perhaps 12 Ohms is still too high.

The Rev 2 circuit is attached.

You're missing the point.  The first detent on the whistle switch puts the rectifier in series with the boost winding to generate several volts of DC offset to pick the relay.  The second detent of the whistle switch parallels the 1.5 ohm resistor in parallel with the rectifier.  That drops the DC offset to about half a volt, still enough to keep the whistle relay activated, but also boosts the power for the whistle tender motor, that's at least an amp or two.

Trust me it works, and has been for 70-80 years.

@vroberts posted:

...

These changes did produce a small DC offset in the sinewave, but not much, less than 1 volt, which I do not think is enough to activate the whistle relay in a tender.

...

I believe the relay is designed to pick up at 1.1 VDC maximum and drop out at .5 VDC minimum. I am sure someone here has the official spec.

@gunrunnerjohn, thanks. I missed that point and it solves the resistor value issue!

I have some new information, some good and one point that is confusing.

Based on the comments here I removed the resistor from my simulation of the initial whistle step, and the simulation was a very close match to the measured waveforms, except for one small issue; the polarity in the simulation is reversed relative to the measured waveforms from A and B to U.

The measured waveforms show a positive signal, while the simulation gives a negative signal. See the files:  Lionel ZW Transformer A Whistle Simulated Waveform Rev 3; Lionel ZW Transformer A Whistle Waveform 11 Ohm 220124; and Lionel ZW Transformer A Full Waveform 11 Ohm 220124. I double checked the probe connections and the polarity of the channel input on the scope. They are correct.

I've attached my Rev 3 simulation diagram and the diagram of the ZW I found on the Internet. Both have the cathode of the rectifier connected to terminal 9 of the Compensating Winding and the anode of the rectifier connected to terminal 19. 

So, my working hypothesis is that the rectifier symbol in the diagram is backwards. I measured the conduction through the rectifier used for side A of my ZW using my Fluke DVM when the Whistle/Direction switch was in the "normal run" position, and I get conduction when the positive terminal of the meter is connected to the A terminal and the negative terminal of the meter is connected to what I believe is terminal 19.  This also points to the fact that the rectifier symbol in the diagram is the reverse of the rectifier in my ZW.

In addition, as I prior design engineer myself, I think it would have been natural for the Lional engineer who designed the circuit to have the rectified whistle signal be positive instead of negative, unless there was a good reason to make it negative. 

Some of this may have been discussed in the threads related to replacing the rectifier with a silicon diode, but I have not yet read those. (Why not, the reader asks.)

Finally, I have little interest in this transformer other than keeping it out of the landfill. I'm currently trying to resurrect an old HO setup, and don't have room for both HO and O, in spite of my fond memories of our Lionel trains when I was a kid.  So, now that I know it works, at least sort of, I plan to list this ZW on eBay with all the caveats that it needs a cleaning and serious maintenance. I hope to price it low enough to encourage someone to buy and restore it.

Thanks for all your help. 

Or. You can put it up for sale on the OGR buy and sell forum.

Just curious why expend all this effort on analyzing this if you never intend to use the transformer anyway?

E-UNIT-79, Thank you for that suggestion.

gunrunnerjohn, that's a good question. In the beginning it was to determine if the ZW was working or not, so I knew how to list it for sale. Then I became intrigued with how the whistle system worked, which lead to getting the diagram, and running more tests, which raised more questions about whether not the ZW was working in whistle more.  The simulations were part of understanding how the whistle system worked.  The information you provided helped me resolve the question of whether or not the ZW was working or not.

My ZW had the discs replaced with new disc and the rollers replaced before I got it. Air whistles do fine, however; modern tenders and Lionel Sounds only start and stop the bell when the whistle lever is activated. How can I change the polarity on the whistle lever to work on modern tenders?

Please point me to another thread if needed.

@Santa Bill K posted:

My ZW had the discs replaced with new disc and the rollers replaced before I got it. Air whistles do fine, however; modern tenders and Lionel Sounds only start and stop the bell when the whistle lever is activated. How can I change the polarity on the whistle lever to work on modern tenders?

Simply swap the connections of the two wires going to the track, i.e. move the wire connected to the center rail to the outside rail and the wire connected to the outside rail to the center rail.

Mike

I should have mentioned this in my post.

My layout is somewhat large and complex with the outside rail being common everywhere. I would have to rewire the entire layout to switch the polarity that way. I did disconnect some of the track and swamped the connections as you suggested and it worked, but I want to avoid rewiring the whole layout.

Any other way around this problem?

Swapping the middle rail/ outside rails wires going to the track from the terminals on the back of the ZW would accomplish the same thing.  Of course the track connections that you’ve already swapped would need to go back to the way they were before.

In other words, it should be much easier to swap the wires at the transformer connection instead of at the actual connections at the track.

I think that’s what Mike meant.

..

@vroberts posted:

..one small issue; the polarity in the simulation is reversed relative to the measured waveforms from A and B to U.

Depending on the service info you used to build the simulation, you probably have the diodes in the simulation reversed. The positive offset is correct, the diagrams for several transformers in the service documents show the wrong rectifier orientation.

This didn't matter from 1934 until 1972 when the electronic whistle was introduced, as the old relays were not polarity sensitive.

@Santa Bill K posted:

My ZW had the discs replaced with new disc and the rollers replaced before I got it. Air whistles do fine, however; modern tenders and Lionel Sounds only start and stop the bell when the whistle lever is activated. How can I change the polarity on the whistle lever to work on modern tenders?

The person who did the modification followed the erroneous schematic and installed the diodes backwards or used cathode-to-case diodes instead of the correct anode-to-case diodes.

@Santa Bill K posted:

I should have mentioned this in my post.

My layout is somewhat large and complex with the outside rail being common everywhere. I would have to rewire the entire layout to switch the polarity that way. I did disconnect some of the track and swamped the connections as you suggested and it worked, but I want to avoid rewiring the whole layout.

Any other way around this problem?

My take on this, let's fix what is messed up.

Did you follow the convention that U on a ZW goes to outside rail and ABCD are center rail or accessories?

If that is how your track to transformer is wired- then whoever refurbished the transformer may have somehow reversed the polarity of the disk diodes.

And why this matters- well the U posts are all common on the ZW, and typically, people use common wiring of the outside rails with accessories and that makes switching the wires at the transformer a bigger problem.

So, while simple swapping the wires for a test track to determine- is your ZW actually correct for troubleshooting- perfectly OK as a test.

That said, more permanent swapping wiring between the common U post and ABCD posts can be problematic in a complicated layout where common is used and expected on outside rail to "work around" a backwards whistle diode.

Again, one of 2 scenarios:

#1 your wiring from the transformer to track was wrong from day one

OR

#2 whoever did the refurb- somehow got the whistle diode wrong

@SteveH posted:

In other words, it should be much easier to swap the wires at the transformer connection instead of at the actual connections at the track.

I think that’s what Mike meant.

Yes indeed. That's exactly what I was trying to imply.  You can flip them at either the track end or the transformer end, whichever is more convenient.

Mike

Great feedback. Thanks for the perspectives. Check what I have added below an see if there is something I am missing.

The ZW convention of U on the outside and A D on the middle rail was used everywhere. No diodes just discs. I attached the Block Controls and Layout which is the physical size shown on the attachment. The circles are 15 amp SPST switches to control each of the Blocks. BL 1-4 is on ZW A, BL 5-8 is on ZW D, BL 9-12 is on a 1033. A second 1033 powers all the Turnouts. All transformers have been phased. The outside common is continuous throughout the layout. only the center rail is switched for the Blocks.

The connections used to test reversing the wiring was returned to the initial state.

If the U and A-D are swapped, the outside rail will be hot and the center will be common. Right?

Could the disc be in backward? would that cause this?