ZW Transformer diode replacement

A 6 volt Zener diode more closely matches the voltage drop of the original rectifier disk.

That means there will be less arcing of the contacts on the whistle switch. 

I have been using the NTE 5815 HC diode for about 20 years without one problem.

He is talking about amps not zener voltage. And third person, he is talking about zeners, not just a diode.

16 amps is plenty. What zener voltage are you thinking?

OP Here

Thanks for all the responses

The Zener diodes I found online are 1N3311B  Zener Diode, 50 watt, 12v

They can be seen here:

Zener Diodes

The specs do not state amps just watts and volts

Thanks again!

Dave

About 4 amps. 

Dave, this web page contains "tried and true" information for replacing the ZW diodes.          Tranz4mr’s ZW Page   (lots of photos)

Kent uses  a stud mounted, anode to case rectifier diode. Model 1N1190AR

Is there any reason to replace the original copper disc rectifier if it is still working?

"Is there any reason to replace the original copper disc rectifier if it is still working?"

OP Here the one disk appears damaged so if I'm replacing one I thought I'd do both

Susan thank you for the link!

Dave

PDG,

    You can also drop Jim Lawson, here on the OGR, some private e-mail and ask him exactly what he uses when he rebuilds all the ZW's.  Jim is a great guy and will probably help you with any questions you might have on the full rebuild/upgrade.  I know his choice of Diodes works perfectly, I own several of his rebuilt/upgraded Transformers.  They work better than brand Spanking New.  I recommend his work 100%.

PCRR/Dave 

If appropriate here: A ZW I received in a lot of Lionel stuff, had the wires on two terminals so tight that I broke the posts loose from the bakelite housing trying to loosen them. How can I re-fasten them to the housing. Pretty tight to get to the brass ground bar without removing the ore Thanks.

Wally

 Wally,I believe your best bet is to buy replacement binding posts equipped with a nut that locks it to the body.

Dewey:  Do you have a source for the ZW posts?  Thanks.

   Wally

Henning's Trains Tel: (215) 362-2442

They sell the replacement posts.

Greetings, I actually joined this forum to reply to this thread and share some insight into restoring these.

In the interest of full disclosure I do have a train shop that remains from our antique businesses so there are a couple thousand pieces in my life, and of course I ran into my share of ZWs.

This is not my real job, nor even my real hobby - I restore audio equipment as a hobby and as such generally know which wires not to lick.

that being said, I have 6 zws and found this page when googling info on the copper disc diodes and what people use to replace.  I know its a 3 year old thread, but when I fix a stereo, if there is an old thread I append to it for a simple reason:  I found it thru a google search for info so some day, somewhere, someone else might as well.  new or more info is always good.  The 'net can have good uses too.  ;-)

So, onto the tech data:

The expen$ive replacement$ using the 16A stud diode is fine and good if you want to throw those bones, but not needed.

MOx diodes (metal oxides, like the target copper or the seleniums I see far more of) have very limited voltages and even far more limited currents.  its hard to find any tech data on our copper discs, but using similar seleniums, we can glean:  the single disc is able to withstand the wide open 19.xx OC voltage of the zW.  however, consider the 1 inch 6 plate selenium.  while it can handle the 120+voltages, a 1 inch selenium is good for roughly 150ma of current.  lets say our discs have 2sq inches of usable MOx, I would estimate they pass 300ma of current but suspect more like 200.

the posts by martin above using the NTE 8515 (fwiw I rarely use NTE parts - too generic and questionable sourcing) is very interesting and I essentially copied him.  squeezing the anode in a crimp spade or lug and tying to the metal plate was a REAL GOOD idea.  you can then just ditch the wire from the old disc retention 'nut' and simplify things.

I use on smaller transformers (like the 1033) the 1N5405 (500v, 3a) and I just ordered a passel of OA60 6 amp diodes from parts express at about .46 cents each.

This should appease anyone that buys my refurb ZWs but in reality, I have a ton of 1N4007s that should do just fine.

Why?  the DC offset triggers a specially wired solenoid in the tenders which then powers the whistle or whatever from track current.  ya, the ZW can source a lot, but the reality is, the whistle will run all day on a few hundred mA.  To draw AMPS would require an air raid siren in the tender.  and the newer digital stuff draw almost nothing to detect the whistle signal.

we have breakers in the ZW which are a thermal type, they require a lot of current for a lot of time to activate and we all know, in the meantime wires can and do catch fire.  IIRC the wisdom years ago was to use a 3-5 amp breaker inline with the track wiring and loco, lit cars, caboose, whistle, never tripped it.

The last zw I just did used 2 5405's in parallel I an have no doubt it would live long and prosper.  What about a derailment short while blowing the whistle?  the 5405 has a 200 amp surge rating EACH.  so I should survive at 400A.

Also consider, MOx diodes have a crap ton of resistance.  I use as a rule of thumb when doing for example a tube radio with a 6 plate Se rectifier, replacing the (usual) 22ohm inrush resistor AND the Se with a 120-150ohm 5-10w cement, then this teeny tiny 1N400x diode.  That being said the resistance of the Se is 100ohms plus, or about 16+ ohms per place.  Our copper should be similar.  when the whistle is running the wires from the throttle has for all intents and purposes 0 ohms and the whistle control 16ish.  meaning all the running current will come from the throttle.

I have 4 more Zws on the shelf to do and I will do the next one with 1N4007 and blow the whistle all day just to test it...

thats all for now.

ps:  does anyone know how to start MTH PS1 equipment that will make idle sounds but when you try to leave neutral on the electronic E unit, they make a single 'tink' noise and repeated trys will go tink-tink up to 8 of them, then cycle back to 1.  what sort of state is it in ?  I have 4 never run locos doing that....

Quaddriver posted:

...the whistle will run all day on a few hundred mA.  To draw AMPS would require an air raid siren in the tender...

How about the current draw through the rectifier on a ZW when blowing the horn at full throttle on a 2343 A-B-A with 6 illuminated 2500 series passenger cars? What do you estimate that to be? With remote switch's bulbs and their controller bulbs on the layout, it could be 7-9 amps pretty easily.

that is an interesting point....if I am reading the zw schematic correctly (and it is written HORRIBLY) the 5v compensating winding is put in parallel with the fixed 8 and variable 12 and is dc - it is a popular, old fashioned way of doing signalling (like ringing a phone for example.  the Cu rect has a 1.5 ohm resistor in series with it which means the absolute current on a shorted rectifier is 3.33 amps.  BUT, the working Cu rect has its own resistance.  I cannot find a spec on it, but its at least 1 ohm.  (simply because the old general electric test was to forward bias it with .5v and read .5a and reverse bias it with 2m and read at most .0025a (2.5 mA)

a silicon rectifier removes this internal resistance to one that is negligible.  I need to think a minute what this can mean when the AC voltage is low but Im thinking since the fixed secondary is always 8, its not a problem.

keep in mind the piv of the Cu rect is 8v.  apply more than 8v on the reverse and it burns up...

The compensating winding is put in series with the 20v(total) windings.

I had to poke and probe another 'virgin' zw with the lionel service data to understand the schematic.  you are right, the compensating oil is in series, its the diode itself that is in parallel with the resistor.  I am going to try and import an image of 3 schematics, 1st is ZW in normal mode, 2nd is ZW with whistle button pushed, 3rd is what we should convert them to...

here goes...

hopefully that shows up, I can send the souce 'photo' or the tinycad file.

But anyways, I am measuring between 200 ohms and 800 ohms fwd or rev across 2 Cu discs I have, both measure a Vf drop of 1.7v dropping quickly to .8v then slowly dropping... and reversed biased about 70mv Vf (meaning it leaks)

So we can now conclude that a stock factory Cu disc has small current flowing across it.

The problem is the Si diodes we now use have no current limiting.  if you follow the second schematic and impose an AC signal across it we find that when A+B+C transformers (8v, variable, compensating) are pushing the positive sine wave towards terminal "A", the diode is reversed biased and supposedly no meaningful current flows across it and all the current flows thru the shunt resistor.

However when terminal "U" has the positive part of the sine wave, with the internal resistance of the Cu Disc, a small DC current flows and the majority of the positive sine wave flows thru the shunt.  classic line signalling.

Replace this with silicon and part 1 is the same, when the sine wave is positive going to the A terminal the new silicon is reversed biased and zip flows and all current flows thru the shunt up to 16 amps as set by ohms law. 

But when the opposite is true and the sine wave is positive at the U terminal, then the current divides nicely down the shunt and the biased diode.

The value of the current flowing is going to be determined by the effective resistance seen, the motor in the loco, all the lights in cars, track etc.  On the 90 watt? 1033s with a 3 amp diode - likely safe. 

from my 3rd schematic, to avoid all of this we need a resistor anywhere from say 10-200 ohms to start, to find the largest resistor you can use to activate the weakest whistle solenoid you can find.  (If you run modern equipment like MTH PS, they require so little to fire off...

I need to find a whistle tender to glean what the solenoids require to fire - that would be the easiest way

One last thing, pay attention to the ORIENTATION of the diode.  I lifted this from the lionel service data.  the A+B+C output AND the 1.5ohm wire attach at the cathode, which, is the metal frame it is mounted to.

Anode cased 16amp diodes and the NTE 8515 and my 1N5405/OA60 diodes...we have hooked them ALL up backwards!

(but no one is complaining stuff has blown up!)

lastly, from the condition of contacts I see, there is no arcing.  Im betting with silicon in there, no inrush resistor, you get some great sparks when engaging the whistle.  No one is making replacement contacts anymore - just sayin....stay tuned.

I think with several hundred ohms in series with the diode and the low value resistor in parallel, you'd lose virtually all your DC offset.  The buffering resistor would have to be considerably lower to allow enough DC to keep the whistle going.

Quaddriver posted:

Anode cased 16amp diodes and the NTE 8515 and my 1N5405/OA60 diodes...we have hooked them ALL up backwards!

While the polarity does not matter until 1972(and then going forward), it does then when the new whistle will only respond to a positive-to-center-rail signal. Later, the bell was designed to be triggered by the negative signal.

A closer inspection of Lionel schematics will reveal, as you have discovered, that the schematics are inexplicably wrong, showing the incorrect rectifier alignment/polarity.

The pictorials are correct, and the anode-to-case rectifier is the correct ZW replacement.

Also, your explanation omits the intermediate step of the whistle activation where the entire load of the center rail track output is circuited through the rectifier momentarily to lift the shaded pole relay into the energize position before the majority of the current is then shunted around it, providing a small DC "holding" current to the relay.

gunrunnerjohn posted:

I think with several hundred ohms in series with the diode and the low value resistor in parallel, you'd lose virtually all your DC offset.  The buffering resistor would have to be considerably lower to allow enough DC to keep the whistle going.

well kirchoffs law tells us how each endpoint of each component behaves.

The DC value will vary by the current thru it and the current will always be proportional to the current thru the shunt.  If we had an ammeter in series for some locomotive load and found out what the engine drew, we would know that every wire has that amount, and in the parallel section the diode (if equipped with a dropping resistor) would see its share of current in the reverse ration of its resistance to the shunt.  this would be a pain to measure and the resistances I took from the Cu discs should not be considered the final answer...from my reading it appears the mono-crystalline structure of CuO, while it acts as a semiconductor, does not behave like a typical semiconductor so a lot of my tests have questionable value.

Quaddriver posted:

 

ps:  does anyone know how to start MTH PS1 equipment that will make idle sounds but when you try to leave neutral on the electronic E unit, they make a single 'tink' noise and repeated trys will go tink-tink up to 8 of them, then cycle back to 1.  what sort of state is it in ?  I have 4 never run locos doing that....

When you power up PS1 you are in reset not neutral.  The clinks and clanks are used to set different settings.  If it never moves it is the battery.  The batteries are old enough that they need replaced or a BCR installed.

If you continue to have problems you should start a new thread.

Gene Anstine

This thread reminds me of the Ham Radio rush to use semiconductor diodes in vintage communication transceivers. The power supplies were using vacuum tube rectifiers and the inrush to the storage capacitors was gradual as the tube filaments warmed up. An ideal situation. But with the semiconductor diodes the inrush was nearly instantaneous by comparison. There were resistors placed in series with the diodes, which helped some with the inrush and the resultant higher than design voltage reached, but the ideal fix was to leave it alone. And allow the vacuum tube to do it's job.

Not saying that the ZW case requires the same conclusion, but it is reminiscent. 

Chuck, if inrush is really a problem, an inrush limiter component could be considered in the circuit.  I haven't seen this mentioned by any of the folks that have installed diodes, so maybe it's overkill.  You could also add some caps across the whistle contacts to help suppress any arc.

Quaddriver posted:

gunrunnerjohn posted:

I think with several hundred ohms in series with the diode and the low value resistor in parallel, you'd lose virtually all your DC offset.  The buffering resistor would have to be considerably lower to allow enough DC to keep the whistle going.

well kirchoffs law tells us how each endpoint of each component behaves.

The DC value will vary by the current thru it and the current will always be proportional to the current thru the shunt.  If we had an ammeter in series for some locomotive load and found out what the engine drew, we would know that every wire has that amount, and in the parallel section the diode (if equipped with a dropping resistor) would see its share of current in the reverse ration of its resistance to the shunt.  this would be a pain to measure and the resistances I took from the Cu discs should not be considered the final answer...from my reading it appears the mono-crystalline structure of CuO, while it acts as a semiconductor, does not behave like a typical semiconductor so a lot of my tests have questionable value.

I must have missed your analysis of the circuit.

ADCX Rob posted:

Quaddriver posted:

Anode cased 16amp diodes and the NTE 8515 and my 1N5405/OA60 diodes...we have hooked them ALL up backwards!

While the polarity does not matter until 1972(and then going forward), it does then when the new whistle will only respond to a positive-to-center-rail signal. Later, the bell was designed to be triggered by the negative signal.

A closer inspection of Lionel schematics will reveal, as you have discovered, that the schematics are inexplicably wrong, showing the incorrect rectifier alignment/polarity.

The pictorials are correct, and the anode-to-case rectifier is the correct ZW replacement.

Also, your explanation omits the intermediate step of the whistle activation where the entire load of the center rail track output is circuited through the rectifier momentarily to lift the shaded pole relay into the energize position before the majority of the current is then shunted around it, providing a small DC "holding" current to the relay.

Believe me I agonized over this...every ones wiring picture CLEARLY show the anode of silicon touching the metal frame and the cathode going to the switch.  current flows from Pjunction->Njunction so matter solved, but!

you test Vf of a diode by putting the positive lead on the anode, the black or negative lead of the DMM to the cathode.  I got the .8v Vf when the oxided side of the disc was touched by the black lead and this is the one that touched the metal frame.  meaning the cathode is clearly on the metal frame....

Try as I might, I cannot execute any test that invalidates that.  The service data I got from lionel matches the guts of the ZW...takes a while to figure out how they notated the switch opening and closing

Also, I noted that the diode is connected by the 2 fingered side of the switch material to the A post but I didnt see clear evidence 'to overturn the ruling on the field' so to speak, of when the "top" of the switch makes the connection to the other end of the shunt.  It does have to break the compensating coil first or else you risk a short...

gunrunnerjohn posted:

Chuck, if inrush is really a problem, an inrush limiter component could be considered in the circuit.  I haven't seen this mentioned by any of the folks that have installed diodes, so maybe it's overkill.  You could also add some caps across the whistle contacts to help suppress any arc.

Quaddriver posted:

gunrunnerjohn posted:

I think with several hundred ohms in series with the diode and the low value resistor in parallel, you'd lose virtually all your DC offset.  The buffering resistor would have to be considerably lower to allow enough DC to keep the whistle going.

well kirchoffs law tells us how each endpoint of each component behaves.

The DC value will vary by the current thru it and the current will always be proportional to the current thru the shunt.  If we had an ammeter in series for some locomotive load and found out what the engine drew, we would know that every wire has that amount, and in the parallel section the diode (if equipped with a dropping resistor) would see its share of current in the reverse ration of its resistance to the shunt.  this would be a pain to measure and the resistances I took from the Cu discs should not be considered the final answer...from my reading it appears the mono-crystalline structure of CuO, while it acts as a semiconductor, does not behave like a typical semiconductor so a lot of my tests have questionable value.

I must have missed your analysis of the circuit.

Oh sorry. I didn't mean that inrush was a problem with the ZW, just that there could be a rush to fix something that isn't broken. Is the diode a failure point?

cjack posted:

Oh sorry. I didn't mean that inrush was a problem with the ZW, just that there could be a rush to fix something that isn't broken. Is the diode a failure point?

I don't know that the diode is an issue, one issue mentioned was arcing at the contacts of the whistle switch with the hard edge of the diode turn-on vs the rectifier disk.  I was just suggesting a couple of possible remedies.

cjack posted:

gunrunnerjohn posted:

Chuck, if inrush is really a problem, an inrush limiter component could be considered in the circuit.  I haven't seen this mentioned by any of the folks that have installed diodes, so maybe it's overkill.  You could also add some caps across the whistle contacts to help suppress any arc.

Quaddriver posted:

gunrunnerjohn posted:

I think with several hundred ohms in series with the diode and the low value resistor in parallel, you'd lose virtually all your DC offset.  The buffering resistor would have to be considerably lower to allow enough DC to keep the whistle going.

well kirchoffs law tells us how each endpoint of each component behaves.

The DC value will vary by the current thru it and the current will always be proportional to the current thru the shunt.  If we had an ammeter in series for some locomotive load and found out what the engine drew, we would know that every wire has that amount, and in the parallel section the diode (if equipped with a dropping resistor) would see its share of current in the reverse ration of its resistance to the shunt.  this would be a pain to measure and the resistances I took from the Cu discs should not be considered the final answer...from my reading it appears the mono-crystalline structure of CuO, while it acts as a semiconductor, does not behave like a typical semiconductor so a lot of my tests have questionable value.

I must have missed your analysis of the circuit.

Oh sorry. I didn't mean that inrush was a problem with the ZW, just that there could be a rush to fix something that isn't broken. Is the diode a failure point?

is there a way to not quote EVERYTHING?  I just wanna reference cjacks material...

anyways....from all the ZWs I have moved, aside from the burnt GE type 51 bulbs, the whistle not working have need the only complaints...there are a LOT of websites on the problem.  I fear a fix was found at some point and it went from being 'a fix' to 'the fix'.  Like I said, I joined as my own work showed too much was being spent.  I at first only focused on repairing the 'diode-ness' of the circuit and the unintended consequence was now you have a circuit path where previously, not much flowed thru it and now, a lot can.  Is that inherently bad somewhere?

As an aside, I have been part of the 'rush' to swap out tube equipment rectifiers for silicon....but not the tubes like 35W4 - cuz when you do that, they ran at 125 or 250ma filament current and that is a LOT of watts to dissipate in heat over a cement to keep the string equivalent.

We would swap out the Se rects - which had their own inrush already as they had very low current ratings.  Generally on the sets using a hard rectifier, you had at least 3 voltages for different sections and each was done 'mechanically' - thru a large cement.

I dont have any old whistle tenders to test on...if anyone does, what is the resistance of the coil and what is its minimum operating voltage...we can derive a close answer with a pencil.  once we know how much current fires the relay, we can throttle the Si rect with a nice fine safety factor...

Quaddriver posted:

is there a way to not quote EVERYTHING?  I just wanna reference cjacks material...

Like this?  Just edit the quoted text.

aha!  I used a GE type 381 diode (appx 150piv, 1a) and a 100ohm 10w cement and clearly can see a 6.5 pulse on the U terminal (using the orientation in the lionel service data) based on track voltage of appx 8.5v

and doing it real slow, I saw the track voltage in AC got from 8.5, to 3.6ish, up to 14.something so there is a point in there where it drops voltage as the top part of the switch breaks/makes

I think a 10w cement is over kill...I need to test more but my DMMs only allow 400ma of testing on AC.  So rules of thumb... what are the years in which certain equipments want the pulse on which terminal?  72 and up it does not matter correct?

gunrunnerjohn posted

Like this?  Just edit the quoted text.

OFSS!  lol thanks....I thot cuz it was grayed out that it was not editable....didnt even try...

On the postwar whistle relay, the instructions for the 5D or 5F testers give minimum pick up voltage, 5.5 VDC and holding voltage, 0.8 VDC.  When I work in a tender I check both of these parameters with a DC power supply. Rarely is there a problem.  When a tender is tested on the track there must be additional load to get the copper oxide rectifiers to work.  I usually use several two lamp passenger cars.  If a whistle relay drops out during this track testing or while be pulled behind the appropriate locomotive, the problem is usually high resistance somewhere between the transformer and the whistle relay.  Cleaning the track and wheels frequently helps.  Gunrunnerjohn had suggested using Deoxit D5 on the roller pin and I have found this works very well to help reducing the resistance in the circuit. 

I do not replace the copper oxide rectifiers with modern technology.  But when reading posts by some who do, they sometimes recommend removing the parallel resistor. They indicate this resistor was included in the circuit to reduce the current being pulled through the copper oxide rectifiers after the relay has picked up.

In many resistor applications Lionel used a nichrome wire in an asbestos sleeve as a resistor. To facilitate installing the resistors they attached a short copper lead on each end with crimped sleeves. I have found that the resistance of these two crimped sleeves may be very high.  I assume that this is the result of corrosion after many years.  When checking these resistors, be sure to include the sleeves in resistance measurements.   

David Johnston posted:

On the postwar whistle relay, the instructions for the 5D or 5F testers give minimum pick up voltage, 5.5 VDC and holding voltage, 0.8 VDC.  When I work in a tender I check both of these parameters with a DC power supply. Rarely is there a problem.  When a tender is tested on the track there must be additional load to get the copper oxide rectifiers to work.  I usually use several two lamp passenger cars.  If a whistle relay drops out during this track testing or while be pulled behind the appropriate locomotive, the problem is usually high resistance somewhere between the transformer and the whistle relay.  Cleaning the track and wheels frequently helps.  Gunrunnerjohn had suggested using Deoxit D5 on the roller pin and I have found this works very well to help reducing the resistance in the circuit. 

I do not replace the copper oxide rectifiers with modern technology.  But when reading posts by some who do, they sometimes recommend removing the parallel resistor. They indicate this resistor was included in the circuit to reduce the current being pulled through the copper oxide rectifiers after the relay has picked up.

In many resistor applications Lionel used a nichrome wire in an asbestos sleeve as a resistor. To facilitate installing the resistors they attached a short copper lead on each end with crimped sleeves. I have found that the resistance of these two crimped sleeves may be very high.  I assume that this is the result of corrosion after many years.  When checking these resistors, be sure to include the sleeves in resistance measurements.   

Interesting.  at 5.5vdc let me ask you a question...can you blow the whistle at rest?  meaning, if you advance the throttle just a titch to get the roller off the frame so the A+B+C circuit completes,  but B (0-12v variable) is at it minimum, will that blow?  since the AC would be A (8v) plus C (5V) thats 13V, since the rectifier is half wave, that gives a DC pulse of .45Vrms...is that enough?  The reason I ask, is I am trying to size the limiting resistor.  Can you get me a representative measurement on the resistance of the coil in the tenders?

About D5, its a heckuva cleaner...I go thru it by the case, but may I suggest using F5 instead or in addition?  Its a good cleaner as well and it contains a bit of lubricant for that metal wheel on that metal pin.  You can even apply D5, dry it then F5.

Now the most interesting part of your reply, when you say they remove the 1.5ohm wire, do they remove that path completely or replace it with straight wire?  If they remove it completely, this is the worst possible solution . the purpose of the parallel circuit is to allow AC to still flow with DC laid alongside.  If you remove the path they when the whistle has engaged, the transformer is now an unfiltered half wave rectified DC power supply with an angry pulse.  Im betting lionel equipment would stop and perhaps modern AC track/DC internal stuff would still run a bit but much much slower as the available voltage for the internal rectifier would be less than half of what it was before you pushed the button.

If however the wire was just replaced I guess it would work but the 16 amp current limiting is not a bad idea...

But the CuO rectifier limits a lot of its own current via its own high internal resistance - when current flows thru the diode it cannot also flow thru the resistor.

 I opened a ZW last night to do my last test and the 62ish ohm nichrome in this one is a carbon comp that says 62 ohm (blue red black) and is 3-5w in size and measures 64.5 so thats close.  If someone told me to modernize that circuit Id think LED and a 1/4w 220-470 ohm, gotta do the math on that...

lastly I am now convinced that going to an Si rect is going to leave a large spark on the ZW whistle switch.  those contacts are NOT hardened, they are soft copper.  It wont take many applications to negate the switch....

There was also a procedure for replacing the rectifier and resistor with a Zener diode, this gave you a constant DC offset for modern equipment.

gunrunnerjohn posted:

There was also a procedure for replacing the rectifier and resistor with a Zener diode, this gave you a constant DC offset for modern equipment.

Id like to see this...we typically use zeners to hold a voltage on the base of a transistor for a fixed voltage power supply (I know for a fact all of your stereos have them...  ;-)   )  but since a zener can conduct in reverse Im curious about the scope of that....also there is no provision in the ZW to have a resistor to regulate the 'zener current'so im doubly curious

Some time ago another forumite and I were working on characterizing the Lionel 167 whistle controller, which uses a somewhat different circuit from the ZW whistle circuit but it uses the same copper oxide rectifiers.

Attached are two different V/I curves for the Lionel copper disc rectifiers. Vertical axis is voltage, horizontal axis is current.

The "Hi" test shows an effective resistance of about 1.25 ohms, the "Lo" test shows about .6 ohms. The test setup did not apply reverse voltage, please ignore the trace to the left of the center (zero) graticule marking.

Based on what I see, it looks like the current would be pretty evenly divided between a 1.5 ohm resistor and the copper oxide rectifier when the rectifier is forward biased. Unfortunately, I gave away the 167 and I don't have any more copper rectifier discs so I can't do any further tests.

Copper oxide rectifiers were widely used as rectifiers in the AC circuit of VOMs years ago precisely because they had a very low forward voltage drop.

Well, the Zener's they mentioned were high power stud mounted models, so I don't think they needed a resistor.  The claimed improvement is no change in speed when you blow the whistle.

Quaddriver posted:

ADCX Rob posted:

Quaddriver posted:

Anode cased 16amp diodes and the NTE 8515 and my 1N5405/OA60 diodes...we have hooked them ALL up backwards!

While the polarity does not matter until 1972(and then going forward), it does then when the new whistle will only respond to a positive-to-center-rail signal. Later, the bell was designed to be triggered by the negative signal.

A closer inspection of Lionel schematics will reveal, as you have discovered, that the schematics are inexplicably wrong, showing the incorrect rectifier alignment/polarity.

The pictorials are correct, and the anode-to-case rectifier is the correct ZW replacement.

Also, your explanation omits the intermediate step of the whistle activation where the entire load of the center rail track output is circuited through the rectifier momentarily to lift the shaded pole relay into the energize position before the majority of the current is then shunted around it, providing a small DC "holding" current to the relay.

Believe me I agonized over this...every ones wiring picture CLEARLY show the anode of silicon touching the metal frame and the cathode going to the switch.  current flows from Pjunction->Njunction so matter solved, but!

you test Vf of a diode by putting the positive lead on the anode, the black or negative lead of the DMM to the cathode.  I got the .8v Vf when the oxided side of the disc was touched by the black lead and this is the one that touched the metal frame.  meaning the cathode is clearly on the metal frame....

Try as I might, I cannot execute any test that invalidates that.  The service data I got from lionel matches the guts of the ZW...takes a while to figure out how they notated the switch opening and closing

Also, I noted that the diode is connected by the 2 fingered side of the switch material to the A post but I didnt see clear evidence 'to overturn the ruling on the field' so to speak, of when the "top" of the switch makes the connection to the other end of the shunt.  It does have to break the compensating coil first or else you risk a short...

a correction, you were right I was wrong on the polarity.  The Vf is was measuring in both directions was because all the ones I have are deteriorated.   so that part is solved, the anode case, the 8515 and my 5405s are all installed correctly.  I found a paper by theodore conti from 1958 by Rider publishing on 'vacuumtubeera.net'

also found that the starting resistance is about 100 ohms at 20*C so my 100ohm resistor was a good guess.

I gotta travel for a few days but my next trick will be to scope it and get both voltage traces and zero in on an inrush limiter and wattage size.  my goal is to keep the ac signal for the motors are pure as possible and the lower dc voltage as stable as possible.

So, there is a lot of technical jargon here that is a little over my head but let me ask this. Is the 1N1190AR Diode that Tranz4mer's web page shows us burning out the contacts of the whistle activation switch. Has anyone seen this. I am going to go over one down at the club layout after the holiday open house shows. It has the 1N1190AR diodes in it. Is this really a problem or are we trying to fix something that isn't broken? 1N1190AR Data.  The OP must be pulling his hair out by now!!!

Forest.