High resistance leakage on hand laid turnouts

[GaryHinshaw]

Sorry to be slow here, but I'm not comprehending something.   If the BDL is supposed to trigger on <22 K-ohms, and you finding them to be triggering on ~4 M-ohms, why are you confident that the BDL's are working with no issues?  I don't doubt that you want to eliminate any measurable conductivity in your turnouts, but I'm not following the logic of your present conclusion.

Capacitance is worth considering, but I still don't think the numbers work out.  I'll first assume that the rails themselves have negligible capacitance, because if they did the BDL's would just trigger on any track.  The PC ties have capacitance, so let's estimate their impedance:

Z = 1/(omega*C)

where omega is the (angular) frequency of the AC voltage, ~2pi*10 kHz for DCC, and C is the capacitance of the ties:

C = e_r*e_0*A/d

where e_r is the dielectric constant of fiberglass, e_r=4.8, e_0 = 8.8 x 10-12 F/m is the permittivity of free space, A is the area of the conductor, and d is the separation.  A typical PC tie is maybe 2 mm thick, and there might be a total area (between the rails) of 9 mm x 25 mm.  (You can change the dimensions if you like, but not by a lot.)  With these numbers, I get an impedance:

Z = 3.7 M-ohms.

So this should still be far from triggering a BDL...   

[ednadolski]

I would check it with my ohmmeter first after I'd cut the gaps in my PCB ties and BEFORE solder any feeders or you glue on your wooden ties.  Your ohmmeter should read ZERO at this point and if it doesn't, I'd make sure that the edges of your PCB ties have been scraped to get rid of the minute burr that almost always is present there.

Not to be pedantic, but if I read this right then I think you mean that the ohmmeter should read an open circuit across the PCB gaps, not zero 

(And just to be even more pedantic, a true zero would require a perfect (physically impossible) superconductor  )

I removed all the wooden ties and was left with the copper / rail skelton and the issue still exists.. So I think it is a Flux / copper ie issue..

(Apologies if you've already addressed this, I might have missed it)

How did you cut the PCB tie gaps?   While unlikely, it may be possible for the cutting tool to embed small amounts of the copper dust/filings into the PCB material, creating a slightly conductive (i.e. high resistance, such as you are seeing) path across the gap.  This is more likely if you are using a tool like a file that hasn't been cleaned and may be depositing metallic particles from previous work.   If gaps in several ties are affected, then it creates a very vexing situation because there are multiple paths and the situation will not be fixed until you have addressed them all.

Best practice (IMHO) is to gap and test each tie before building it into the trackwork.  Be sure to use a high resistance scale on the ohmmeter.

(Unfortunately, if you cannot isolate the leakage path(s) (which could be *every* PCB tie) then you may be at a point of diminishing return, where it would be less work to start over.)

HTH,
Ed

[ednadolski]

where omega is the (angular) frequency of the AC voltage, ~2pi*10 kHz for DCC

IIRC that would be for a sinusoidal (pure AC) waveform, but I thought DCC used a PCM waveform, so the approximation is only useful for computing the real (not reactive) component of power?   The impedance model for any transmission line would have to include its (DC) resistance, and both the capacitance and the inductance at a specific frequency.

(Something in my head is chiming in about computing an infinite series as a summation of the harmonics, but I've forgotten a *lot* in the last 30 years...)

Ed

[GaryHinshaw]

You're correct Ed, the DCC waveform is a pulse-code-modulated square wave, which effectively encodes a string of 0's and 1's for the decoders:



This signal has frequency content above the fundamental frequency of ~10 KHz, but the voltage of those harmonics drops off as 1/omega, until they are rolled off by a filter at some point.  Since the impedance also drops as 1/omega, the higher harmonics do contribute to the AC current at a comparable level to the fundamental.  But since the impedance of the fundamental mode is still 100 times higher than what should be triggering the BDL (at least by my estimate), I have a hard time imagining that the harmonics contribute enough to close the gap.

[Chris333]

Some of those PC ties aren't even connected 

[bdennis]

Guys,
Thanks for the updates.. I have made some progress!..

First to answer a few questions.

Yes I mean that I want the multi memter to read full open circuit not Zero..

In terms of cutting the gap, I use a round file and run it back and forward over the tie until a noticable gap appears. Then I clean with a small wire brush. If I have missed any of the copper then it shows as a full short and not resistance and am able to find and clean and fix those issues.. So no issue there.

So now to my testing and the problem.

I cut new gaps in some spare ties and found them to be full open circuit with NO resistance at all. - So all good at this point.
I then placed the tie in the FastTracks jig and applied flux to the tie as I usually would before soldering the rail to the tie.
At this point I did not solder the rail to the tie. I just left it..
I left the tie sit for 10 to 15 min to let the flux dry.
I then removed the tie from the jig and tested the tie.. (Note I have not cleaned the tie at this point)
BINGO.. The tie now has a resistance value.

Here is the Flux I am using.


Here is an Example of 1 of the resistance measurements in 1 of the gaps I cut in the ties.. It is important to note that each gap has a different resistance value compared to the other.. Also Note ALL gaps were tested prior to applying flux and NONE had any resistance on them and were all open circuit.



Now the tie has a resistance value, I switch the Fluke Meter to what I call the conductivity tester (to the icon that looks like a diode) and all gaps including those that now have a resistance value all appear as a open (non conducting) circuit..

So now I know why it is happening.. Now to find a way to reliably clean the turnouts and remove all the flux..
As mentioned I have tried a bath of CRC contact cleaner (Electronic Cleaning Solvent) with no real change..

Any suggestions..??

[Chris333]

Use an old toothbrush. I have used dish soap. Have also used brake parts cleaners.

[bdennis]

Chris.
Thanks for that I will give it a try..

[GaryHinshaw]

Well, I still don't understand how 2.3 M-ohms is triggering a BDL, but if cleaning the flux solves your problem, then it doesn't really matter.  (I suppose if you had ~100 such ties like this in one block, the parallel resistance could get down to 22 K-ohms.)

Please keep us posted on the outcome.

[C855B]

It's the flux residue, which you just proved. My apologies for not exploring this earlier - when I think "flux", my mind immediately goes to traditional rosin fluxes used in electronics. They don't conduct. Zinc chloride - a/k/a "tinner's flux" - is a metallic salt derived from dissolving zinc in hydrochloric acid. Therefore it is going to be conductive, although minimally, as you verified. I think alcohol may be your solvent of choice for a thorough cleaning.

Cleaning after using a zinc chloride flux is more than just a good idea. Because of the acid content, the residue remains corrosive, which is why it is not used in electronics or anything else that can't be thoroughly washed after soldering.

[bdennis]

OK, so, can someone educate me on what % alcahol I shoudl get from local supplier.. I see varying % of alcahol...

Then I will set about cleaning these turnouts..

Looking at then turnouts, even though I have scrubbed them, I see left over residue on the sides of the rails.. So some cleaning is needed.

Oh, and find different flux to use or not use it at all.. As that is clearly the root cause of the problem..

[reinhardtjh]

Oh, and find different flux to use or not use it at all.. As that is clearly the root cause of the problem..

Really?  You have to ask about flux?  You know there is one, and only one to use....  Superior No.30
 Supersafe® Soldering Flux (A product of Cleveland, Ohio, btw).  International shipping is available from the supplier (no clue of the cost, though)   

Also available here http://www.ccis.com/home/hn/ with international shipping.

[havingfuntoo]

It is illegal to send these type of products through the post in Australia and I am sure if they are declared correctly they will get confiscated. A better suggestion is to go to where you bought the flux and ask for some advice or contact CIG directly and explain to them your dilemma.  They can be very helpful, one tip, be concise with your request.   

[havingfuntoo]

It is illegal to send these type of products through the post in Australia and I am sure if they are declared correctly they will get confiscated. A better suggestion is to go to where you bought the flux and ask for some advice or contact CIG directly and explain to them your dilemma.  They can be very helpful, one tip, be concise with your request. I am going to make a guess that some of the copper is being embedded in the fibber of the PC board ties as you file the gap, try cutting it with a saw and check your result, even try brushing some Hydrochloric Acid through the cut to see if that will remove any fillings left behind or embedded in the fibber. There are stronger materials available that can be obtained to remove the copper, again best to ask local suppliers for advise.       Looks like I have clicked on the wrong icon but I am sure you get the idea.

[robert3985]

UH OH!!!  You're asking for trouble with that flux.  Get some non-acid flux or self-neutralizing flux.  This stuff you're using is gonna eat your ties, your solder joints, your ballast in about a year.  Ask me why I know that.

Just to assume you can't get Supersafe #30 Soldering Flux in Australia may be premature.  If you can get it, GET IT.  It is THE BEST soldering flux for miniature metal work I've ever used and will turn you into an expert solderer in no time.  I am not kidding.  Get it....if you can.

Here's a photo of some of my turnout work...three of seven turnouts for my Emory Center Siding LDE...all done with Supersafe No. 30...as all of my turnouts have been in the last 20 years or so:



Here's a close up:



That lovely, clean finish is mostly due to Supersafe No. 30 Flux...which self-neutralizes at soldering temps.  I also wash mine off in warm, soapy water, but I've never, ever had a corroding solder joint from acid using the Supersafe, even on joints I couldn't wash.