ZW Transformer diode replacement

there is always some sort of arcing on contacts, AC or DC.  harmful AC arcing occurs during high inrush current of an AC device (125vish, 200W or more) and its that bane of many an ALPS switch, which uses hardened contact points such that we generally use a triac.  for DC, arcing occurs at much lower voltages and currents (this is why old points ignition used a ballast resistor, to bring the primary voltage thru the points to about 9v)

The switch in question is the 3 point switch in the whistle control.  Its contacts are copper with no platinum coating.

By going silicon with its insane power handling, you can pass a lot of current thru the contact that is the 'bottom' of the switch where there are 2 legs that join the diode output to the A (or D) lug.  Im only guessing it does this just based on my testing, can someone tell me, does it?  take off the top, run some trains and blow...this switch will never be repairable without a donor so that is why I am looking for an inrush limiter.

once the switch contacts carbon up, you can clean them with your wifes emory board...but the number of cleanings will be finite...

I've had no problems with mine. I've had diodes in there 20 years.

Forest posted:

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.

I think I'd trust what a guy that has been doing transformer repair and rebuild for many years has to offer over most casual opinions of what works and what doesn't work.

I didnt return until later evening so I had limited time to experiment and test but I scoped the zw as promised and if this works, following are 3 freeze displays:

First, the ZW at minimal track voltage.  open load where my RMS DMM told me I had 8.67v, we expect, and do see this:

Note that at 5ms display we see a wave period of 16.7ms.  This was before I put the scope in RMS mode, it said my rms voltage was 8.9v - close enuf

Next I slowly pushed the whistle lever forward until it broke the tie on the upper part of the switch, but had not yet made the diode plus shunt.  we expect to see the effect of a half wave rectifier on a 5.something v ac winding:

The voltage max appears to adhere to the theoretical .45VRMS and the shallow floor I would expect from a coupled probe draining.  I am told the tender relay has 2 copper discs (mini rectifiers?) to keep it from building and emf via its impedance.  recall, on AC a coil has impedance (R+Z) and the transformer has effectively a damping factor of 0 so you want to minimize this as much as possible.

This wave would serve to cause the coil to energize in 1 direction 60 times a second, due to DC inspired emf collapse, the holding current never goes to 0

Lastly, I completed the lever push, now AC and DC are supplied via the diode and the shunt

I had used my dmm  (btw the dc supplied by the diode was fluctuating between 5.9-6.3v - unfiltered as it was - and got 14.28vrms.  the scope is showing appx 20v when you remove the noise offset, 14.14v - close enuf as the interfering DC is going to add a clamping effect (keeps the relay on) and the higher voltage (this is still open circuit) is supposed to run the motors in the loco AND the tender now.  everyone seems to say it works.  some even say things speed up a bit.  oh and the scope could not figure out vrms...with the DC signal helping one side and hurting the other, I guess the comparator was flumoxed.

why do I constantly have a wee  bit of offset?  the scope I am using - velleman pclab 2000se - runs off pure USB DC into this laptop.  the transformer has no ground connection and as I said before, is noisy as ****...it floats a wee bit and while the PO put on a polarized plug, I bet if I reversed the wires in the plug, it would float on the upper end now.  there is no need to put a polarized plug on a zw, primary shares no connection with any secondary and there is no center tap nor equipment ground.

summary:  this is what we expect to see in  properly functioning zw as built by lionel.  I am using the 100ohm 10w cement to limit DC current.  But I have not actually tested a whistle on it.  The mth stuff works fine, but I dont think I have any old school lionel in stock.   I got a couple erie builts with a b unit, will go check if that has something onboard, and I have a korean war berk I gotta grab from the store and check.

my goal is to find the largest resistor that allows the whistle to blow, then I can select the smallest diode to effect this and finally come up with a comprehensive BOM to rebuild a ZW, both sides, including lighting so stay tuned.

PS:  picked up 4 in the box 1044s from an estate sale last night, they will added to the fold.

I followed GRJ advice and vendor source for ZW diode recently with excellent results.  I think it was $4.86 for 25amp stud mount common anode.  Used one on each side of ZW.  Works fine for me, and it turns out the hfe on both sides were pretty closely matched.

Success!  It turns out in an auction I won back in sept was a somewhat more than postwar tender (6 axles, with hooter installed) so I got to test.

First off, the darn coil has a resistance of <15 ohms, so the coil itself can draw some serious current.  I used my DMM on the 10A dc setting and put it in series with ONLY the DC portion of the whistle circuit, meaning between the diode and the whistle switch.  at idle (throttle zero) its not really zero and the whistle will fire and draw about 460ish mA of current.

at full track voltage it rose to just under 900 mA

why?  the dc pulse is in sync with the AC pulse and messes up the RMS voltage of the ac signal (remember how my scope could not figure out the rms voltage when hooting? so when any motor poles are aligned to have the center rail swinging positive, the windings present a load to the diode.

I also tested this with a lit caboose and a dual motor mexican made modern lionel GP9 at the same time - they made no appreciable difference in the DC current draw.

In fact, the tender will whistle on just the pulsing half wave dc, it runs crappily as it only half the info it needs to run the motor is there, but it will run at half speed.

I also tested this by pushing the whistle switch real slow to test the effect of each contact.

when you are running off the diode and shunt via the compensating coil ONLY, current goes to 1-1.6A for a second.

So, as promised, the bottom line.

You can get away with *any* 1N400x diode (1amp, varying PIV) as the surge ability exceeds the momentary effects of a slow button push.  But I dont recommend that.

Instead, I would start with at a minimum a 3 amp diode.

Any of the UF5405/IN5405 (mouser part 625-UF5405-E3 - 3A 300V) or equivalent diodes will do, at appx 50 cents each if you order 10 or more.  But you can also get the 625-UF5408-E3 - 3A 1000V, at 42 cents each.

I ran the aforementioned load with the hooter on for 2 minutes and while I could feel some warmth in the diode, the load was appx 20-25% of rated max.  IF you run the hooter for 2 minutes full on your wife will smack you.

I have used these diodes in the 1033 and 1044 transformers without a pause.  I have been putting 2 in parallel for the ZWs, but we now know this is overkill.  Even with the tender coil resistance 1/6th of what I thought it would be.

If you want a modern, reliable overkill 6A, check out mouser 621-6A10 for 42 cents single, 34.3 cents 10 or more - these are 6amp 1000v diodes.

As for the resistor, because the relay draws so much, use none, it does not reliably hold the field.   if you run exclusively modern MTH-like stuff you can do it, but in the end it makes no difference.

IF anyone is interested I worked out the LED lighting.  I heard a rumor that someone sells bayonet base led bulbs that fit (7-8mm?) but I can work it out for single white or colored bulbs.