Welcome to the forum   
Welcome Guest! To enable all features please Login or Register.

Notification

Icon
Error

Share
Options
View
Go to last post in this topic Go to first unread post in this topic
Offline perz  
#1 Posted : 22 February 2009 19:21:00(UTC)
perz

Sweden   
Joined: 12/01/2002(UTC)
Posts: 2,578
Location: Sweden
One of my previous topics about K-track turnout motors turned into a discussion of contact tracks.

Contact tracks used "as is" can cause problems with the return current for smaller loks, since only the rail on one side provides low-Ohm contact. This problem can be cured by bridging the contact track with a pair of diodes. The problem now is that the signal can not be fed directly into an S88 or other 5 V digital logic. You need an amplifier:

UserPostedImage

I have now designed that amplifier, see schematic below.

UserPostedImage

I made some tests and they confirmed that this circuit is operational.

However, the tests were rather discouraging for the use of contact tracks in general. I haven't tested contact tracks at all before, but I was surprised to see that a small freight car only caused contact very occationally. I tested with 3 different cars, and it was the same with all 3. Looking with an oscilloscope, there are short spikes when the car is rolling. This is sufficient to trigger the S88, since it has a latch function.

But if you want to also be able to detect that a track is empty, it won't be reliable. Chances that a car left behind will stop in a way where the wheels don't give any contact are quite high, according to what I could see.

When testing with loks I have seen no problems, probably because the loks are heavier than the cars.
Offline Tdl  
#2 Posted : 24 February 2009 11:01:24(UTC)
Tdl

Netherlands   
Joined: 30/03/2006(UTC)
Posts: 71
Location: Amsterdam
To improve return on marklin track its is sufficient to connect digital + (brown) to the contact track via a diode.
I use 1N400x.

The contact track then can be fed directly in S88 decoders.

Offline perz  
#3 Posted : 24 February 2009 20:59:43(UTC)
perz

Sweden   
Joined: 12/01/2002(UTC)
Posts: 2,578
Location: Sweden
With only one diode, the contact track will act like a "brake section" if contact to the really grounded rail is lost. This will certainly be sufficient to overcome any contact problems while the train is moving. However, I want to be sure it works also in the case where the lok is stopped on the contact track with contact to the really grounded rail lost.

Offline Tdl  
#4 Posted : 25 February 2009 16:46:01(UTC)
Tdl

Netherlands   
Joined: 30/03/2006(UTC)
Posts: 71
Location: Amsterdam
I have not observed that a feed via a diode on the contact track acivates loc decoder brake functionality.
I will verify this by disconnecting the opposite rail, so that return is via contact track feed diode only.
I will inform you on the results.
Offline perz  
#5 Posted : 25 February 2009 23:13:47(UTC)
perz

Sweden   
Joined: 12/01/2002(UTC)
Posts: 2,578
Location: Sweden
Quote:
[size=1" face="Verdana" id="quote]quote:Originally posted by Tdl
<br />I have not observed that a feed via a diode on the contact track acivates loc decoder brake functionality.
I will verify this by disconnecting the opposite rail, so that return is via contact track feed diode only.
I will inform you on the results.



There are diffent possibilities here. The lok action when using a single diode will depend on the type of diode as well as the type of decoder in the lok. In some cases the brake function is activated (usually not, though). In some cases the lok is controllable as usual, and in some case the lok ignores any commands to it. Several tests by myself and by other people have verified all three possibilities.

In the first and last case, a stopped lok will remain stopped if it has stopped so that the contact to the really grounded rail is lost. In the middle case, it will work just fine.
Offline Tdl  
#6 Posted : 26 February 2009 09:43:37(UTC)
Tdl

Netherlands   
Joined: 30/03/2006(UTC)
Posts: 71
Location: Amsterdam
I have done a test on an isolated track, connected to an Intellibox.
The red wire is connected to the point contacts, the brown wire to both rails via a diode 1N4001 feeding towards the rails.
I have left out the connection to a S88 contact.
The IB is setup to issues Motorola commands only.

On this setup an engine with a marklin 60902 decoder runs with reduced speed. It stops when one turns the IB knob to 0 and it starts when one turns the IB knob away from 0.
However it does not change direction reliably.

An engine with marklin mfx/SDS drive stops, starts and changes direction reliable.

Also an engine with ESU Lokpilot 3 stops, starts and changes direction reliably.

I have not checked the brake section setting of the mfx and Lokpilot decoder.
Basically I do not care about it, because I don not use this functionality on my layout.

I use feed diodes on about 10 out of over 300 contact track sections on my layout, because this remedies that engines with poor rail contact stop on these specific sections.
Most of these diode feeded sections are in turnouts. I use contact track sections in turnouts to monitor the presence of axles on the turnout.

Perz, when there is a better solution - more reliable, suitable for a wider range of decoder types - then I am interested to know.

Offline efel  
#7 Posted : 01 March 2009 13:06:41(UTC)
efel

France   
Joined: 23/02/2005(UTC)
Posts: 800
Tdl,

If you re-do the test with a 5V zener diode, (anode connected to brown wire), your problem will probably disappear.
In fact, the voltage on the rails will then be -0.7V and +5V and the loco decoder voltage about -19.3V and +15V.
When using a S88 connected to the insulated rail, and the zener between the insulated rail and grounded rail, the S88 input voltage will be 5V if the track is not occupied, and there will not be a false detection.
I've been using that on my layout for 3 years without any problem, while a simple diode leads to problems, as Perz indicated. I have not noticed any slow down of the loco either, even at full speed . (test not done with mfx for I don't have such decoder)
Note: the zener diode must be an 1A type, and with a standard S88 detector, it's better to use a &gt;6.2V diode, for the bias current from the S88 is very low (in my case I use a home made detector, and a 5V zener is enough.).
Of course you can also use the very good Perz schematic above.

Fred
Offline perz  
#8 Posted : 01 March 2009 13:49:53(UTC)
perz

Sweden   
Joined: 12/01/2002(UTC)
Posts: 2,578
Location: Sweden
A zener diode is a quite nice solution, didn't think about that. You need a zener diode with rather high Ampere and Watt rating, but besides that, I see no problems with it.

The voltage drop in one direction will probably not have any significant effect, since the problem with a single diode is the quality of the digital signal, not the power supply to the decoder. As single diode is capable of supplying enough power for the decoder.

The exception is non-upgraded non-SDS C-sine mfx decoders, where the sensitivity to track voltage is severe. But it is better to get those loks upgraded I think. Märklin has, at least so far, done these upgrades for free. (Maybe I would hesitate a little bit to send items for upgrade now when Märklin is bancrupt, but I guess the situation there will stabilize).
Offline xxup  
#9 Posted : 09 May 2010 05:16:27(UTC)
xxup

Australia   
Joined: 15/03/2003(UTC)
Posts: 9,457
Location: Australia
So just to clarify.. Confused

We replace the three 1N4004 diodes with zener diodes? Is there a specific part number for the zener?

You need to make the circuit for each contact? I have 40 or so on my small layout..

Why do you need the diodes between the rails? (I am not sure how I can solder these on m-track.)

Adrian
UserPostedImage
Australia flag by abFlags.com
Offline perz  
#10 Posted : 09 May 2010 17:17:24(UTC)
perz

Sweden   
Joined: 12/01/2002(UTC)
Posts: 2,578
Location: Sweden
xxup wrote:
So just to clarify.. Confused

We replace the three 1N4004 diodes with zener diodes? Is there a specific part number for the zener?

You need to make the circuit for each contact? I have 40 or so on my small layout..

Why do you need the diodes between the rails? (I am not sure how I can solder these on m-track.)



No, you just have to replace the middle diode, and remove the rest of the components. Then connect the "isolated" section directly to the S88. This much simpler solution has some theoretical drawbacks (as already mentioned in my previous post) but should be good enough in practice. If you really want to "gold plate" it, go for my original solution. I do not know the part number of the zener diode but it should be a 5 V zener diod (standard values are 4.7 V and 5.1 V, both will do) and the power rating should be rather high, I would suggest 5 W.
Offline efel  
#11 Posted : 10 May 2010 00:58:12(UTC)
efel

France   
Joined: 23/02/2005(UTC)
Posts: 800
xxup wrote:
So just to clarify.. Confused

Why do you need the diodes between the rails? (I am not sure how I can solder these on m-track.)



Hi Adrian,

You may have a look here:
https://www.marklin-users.net/fo...aspx?g=posts&t=10513

You just need 1 zener (6.2V, 1A) for each S88 input. The anode to ground and the cathode to the S88 input

Fred
Offline GSRR  
#12 Posted : 10 May 2010 20:04:21(UTC)
GSRR

United States   
Joined: 01/03/2009(UTC)
Posts: 1,339
Location: USA
Adrian,

for the zener diode see if this might be correct: 1N4735 6.2v 1A

http://knol.google.com/k...is/zener-datasheet-1.jpg




Perz,

If it is not to much trouble could you possibly make a graphic with just the zener diode connection?


Regards,


Thomas


ETE UserPostedImage ECoS iTrain TouchCab C-Gleis German Era Id & IIIb USA Era IIIb SBB Era III SJ Era IV GC Era V
Offline xxup  
#13 Posted : 11 May 2010 00:39:02(UTC)
xxup

Australia   
Joined: 15/03/2003(UTC)
Posts: 9,457
Location: Australia
Thanks guys.. Yes another diagram will make it certain for me... I will see if I can pick up some 1N4735 diodes on the weekend from Jaycar...
Adrian
UserPostedImage
Australia flag by abFlags.com
Offline efel  
#14 Posted : 11 May 2010 01:32:07(UTC)
efel

France   
Joined: 23/02/2005(UTC)
Posts: 800
Hi,

1N4735 are 1W zener, not 1A.

1N5341B are more convenient for a permanent layout. I've used diodes of that type for more than 3 years without any trouble.

Fred
Offline perz  
#15 Posted : 11 May 2010 23:33:33(UTC)
perz

Sweden   
Joined: 12/01/2002(UTC)
Posts: 2,578
Location: Sweden
1 W is too little in my opinion. The 1N5341B is 5 W so that will be good. I think 6.2 V is more than necessary voltage. Why did you choose that instead of 5.1 V that I assume would be the most optimal value? I think the corresponding 5.1 V diode has type number 1N5338B.
Offline efel  
#16 Posted : 12 May 2010 01:40:08(UTC)
efel

France   
Joined: 23/02/2005(UTC)
Posts: 800
perz wrote:
Why did you choose that instead of 5.1 V that I assume would be the most optimal value?

Hi,

The nominal zener voltage is given for a test current of 240mA.
When the diode is used with a S88, its bias current will be only some tens of microamps, for the bias resistance inside the S88 is 100 kOhm. With such a low current, the zener voltage of the 5.1V diode is only about 3.5V.
Considering furthermore a +-5% spreading in the nominal value, that leads to a noise margin which is, in my opinion, too low for a reliable operation of the cmos logic.

It's the reason why I suggest to use a 6.2V zener (which is also better in dynamic resistance.)

Fred
Offline perz  
#17 Posted : 13 May 2010 14:02:59(UTC)
perz

Sweden   
Joined: 12/01/2002(UTC)
Posts: 2,578
Location: Sweden
efel wrote:
perz wrote:
Why did you choose that instead of 5.1 V that I assume would be the most optimal value?

Hi,

The nominal zener voltage is given for a test current of 240mA.
When the diode is used with a S88, its bias current will be only some tens of microamps, for the bias resistance inside the S88 is 100 kOhm. With such a low current, the zener voltage of the 5.1V diode is only about 3.5V.
Considering furthermore a +-5% spreading in the nominal value, that leads to a noise margin which is, in my opinion, too low for a reliable operation of the cmos logic.

It's the reason why I suggest to use a 6.2V zener (which is also better in dynamic resistance.)

Fred



OK, I think you are right. I haven't looked at the data sheets for the diodes. I might come to the same conclusion as you if I do.
Offline xxup  
#18 Posted : 13 May 2010 14:19:44(UTC)
xxup

Australia   
Joined: 15/03/2003(UTC)
Posts: 9,457
Location: Australia
perz wrote:
efel wrote:
perz wrote:
Why did you choose that instead of 5.1 V that I assume would be the most optimal value?

Hi,

The nominal zener voltage is given for a test current of 240mA.
When the diode is used with a S88, its bias current will be only some tens of microamps, for the bias resistance inside the S88 is 100 kOhm. With such a low current, the zener voltage of the 5.1V diode is only about 3.5V.
Considering furthermore a +-5% spreading in the nominal value, that leads to a noise margin which is, in my opinion, too low for a reliable operation of the cmos logic.

It's the reason why I suggest to use a 6.2V zener (which is also better in dynamic resistance.)

Fred



OK, I think you are right. I haven't looked at the data sheets for the diodes. I might come to the same conclusion as you if I do.


Blink Okay professors.. So do I buy 1N5341B zener diodes????
Adrian
UserPostedImage
Australia flag by abFlags.com
Offline perz  
#19 Posted : 13 May 2010 15:07:03(UTC)
perz

Sweden   
Joined: 12/01/2002(UTC)
Posts: 2,578
Location: Sweden
Yes, you can go for 1N5341B.
Offline xxup  
#20 Posted : 13 May 2010 15:37:43(UTC)
xxup

Australia   
Joined: 15/03/2003(UTC)
Posts: 9,457
Location: Australia
Thanks Per..

But.. Jaycar doesn't stock them.. aaaarrrrgggggghhhh...

Can anyone else in Australia please recommend another component retailer.. It seems that neither Tandy or Dick Smith stock components any more...Crying

They are available on eBay.. about $2 each and another $12 to ship 30 of them to OZ from the USA.. The trouble is that the seller does not have the best reputation level.... OhMyGod

Adrian
UserPostedImage
Australia flag by abFlags.com
Offline efel  
#21 Posted : 13 May 2010 16:12:38(UTC)
efel

France   
Joined: 23/02/2005(UTC)
Posts: 800
Hi Adrian,

They are available at Elfa for 0.58 €:
https://www1.elfa.se/elfa3~eu_en...FA_EU-EN#item=70-066-04;

I don't know, though, if they can send them to OZ ?

Fred
Offline efel  
#22 Posted : 13 May 2010 16:25:52(UTC)
efel

France   
Joined: 23/02/2005(UTC)
Posts: 800
Offline GSRR  
#23 Posted : 14 May 2010 01:00:19(UTC)
GSRR

United States   
Joined: 01/03/2009(UTC)
Posts: 1,339
Location: USA
Adrian,

I checked both http://www.jameco.com/ and also http://www.mouser.com/ here in the US. Although they list for about 25¢ to 50¢ apiece these are out of stock until about July. Hence it seems those who have them in stock are raising prices to the $3.00 range.


Regards,


Thomas



ETE UserPostedImage ECoS iTrain TouchCab C-Gleis German Era Id & IIIb USA Era IIIb SBB Era III SJ Era IV GC Era V
Offline xxup  
#24 Posted : 14 May 2010 01:05:28(UTC)
xxup

Australia   
Joined: 15/03/2003(UTC)
Posts: 9,457
Location: Australia
efel wrote:


Except they don't have any Australian stock either... See below:

UserPostedImage

At least they list them, so I will contact them to see when the next shipment will arrive.. If I order 60 of them they might make the effort..

Thanks for your help Fred.


Adrian
UserPostedImage
Australia flag by abFlags.com
Offline xxup  
#25 Posted : 14 May 2010 01:08:14(UTC)
xxup

Australia   
Joined: 15/03/2003(UTC)
Posts: 9,457
Location: Australia
GSRR wrote:
Adrian,

I checked both http://www.jameco.com/ and also http://www.mouser.com/ here in the US. Although they list for about 25¢ to 50¢ apiece these are out of stock until about July. Hence it seems those who have them in stock are raising prices to the $3.00 range.


Okay - who got the inside information and has now cornered the world market for 1N5341B zener diodes. This is the turnip economy again! OhMyGod
Adrian
UserPostedImage
Australia flag by abFlags.com
Offline GSRR  
#26 Posted : 14 May 2010 01:36:36(UTC)
GSRR

United States   
Joined: 01/03/2009(UTC)
Posts: 1,339
Location: USA
Adrian,

I went to the Newark website which is a US affiliate of Farnell. I found this.



http://www.newark.com/nt...1n5341b/diode/dp/33C7798


I notice while the 1N5341B is $1.28, the 1N5341BG is only .22 cents, and out of stock.



Regards,

Thomas


ETE UserPostedImage ECoS iTrain TouchCab C-Gleis German Era Id & IIIb USA Era IIIb SBB Era III SJ Era IV GC Era V
Offline Pimanel  
#27 Posted : 01 September 2016 20:20:37(UTC)
Pimanel


Joined: 12/02/2011(UTC)
Posts: 37
Location: Lisbon, Portugal
Hi to all

I know that this topic is dead for a long time but I saw this and got curious.

I was using the 24994 s88 contact tracks but after a while some of them started to fail. The plastic bits got a bit loose and with that the sensor keep on after the loco passes.

I saw this S88 Diode Trcik and began a couple of tests.
But then I saw this thread and apparently the diode trick is not enough for a reliable detection.

You suggest to use 1N5341B zener diodes. Is the wiring the same (just replacing the 1N4002 diode)?
This also works for the Viessmann V5217?
Can you please post a schematic draw of the connections?
Or, maybe is there another alternative?

Thanks in advance

Pimanel
Online H0  
#28 Posted : 01 September 2016 23:22:24(UTC)
H0


Joined: 16/02/2004(UTC)
Posts: 15,254
Location: DE-NW
Hi, Pimanel!
Originally Posted by: Pimanel Go to Quoted Post
I was using the 24994 s88 contact tracks ...
The 24994 is a circuit track. The diode trick is for contact tracks where the loco wheel sets close the contact.

Regards
Tom
---
"In all of the gauges, we particularly emphasize a high level of quality, the best possible fidelity to the prototype, and absolute precision. You will see that in all of our products." (from Märklin New Items Brochure 2015, page 1) ROFLBTCUTS
UserPostedImage
Offline Pimanel  
#29 Posted : 02 September 2016 17:29:00(UTC)
Pimanel


Joined: 12/02/2011(UTC)
Posts: 37
Location: Lisbon, Portugal
Hi again

I know that the way the 24994 works is different from the diode trick. The 24994 uses the pickup shoe to close a circuit that sends current to the S88, and as you said, the diode trick uses the loco and the wagon wheels to close the contact and send current to the S88.

Now, forget the 24994. :)

I already tested the diode trick (http://members.ozemail.com.au/~rossstew/rms/pdf/the_diode_trick.pdf) and track pieces with one side of the rails isolated and wired to the S88 and it works fine, apparently.
But in some posts of this thread people say that the diode trick is not reliable. It is better to use the zener diode.

My question is how is the wiring in the case of the zener?
It is just replacing the 1N4002 by a 1N5341B, maintining the rest of the wiring scheme published by this fellow (and obviously taking in account the correct orientation of the diode)?
Can you post a schematic draw of what you have done?

Although I have a PhD, I am a biologist, and electronics is far beyond my scope. :)
I need everything carefully explained.

Thanks again
Pimanel
Offline river6109  
#30 Posted : 19 September 2016 14:09:49(UTC)
river6109

Australia   
Joined: 22/01/2009(UTC)
Posts: 14,636
Location: On 1965 Märklin Boulevard just around from Roco Square
you always can buy a optocoupler circuit board which fits between sleepers and is activated by the slider or pickup shoe.


John
https://www.youtube.com/river6109
https://www.youtube.com/6109river
5 years in Destruction mode
50 years in Repairing mode
Offline Pimanel  
#31 Posted : 21 September 2016 15:17:46(UTC)
Pimanel


Joined: 12/02/2011(UTC)
Posts: 37
Location: Lisbon, Portugal
Hi John

I already read your post about optocouplers.
I have been folowing your youtube channel for quite some time.
I think the idea is very good but I also want to detect wagons, so for me it is not suitable enough.

That is why I was interesting in the diode trick but reading this post I got confused and asked for help about the zener diode.

But this post died and I think no one will help.

Thanks anyway
Filipe
Users browsing this topic
Forum Jump  
You cannot post new topics in this forum.
You cannot reply to topics in this forum.
You cannot delete your posts in this forum.
You cannot edit your posts in this forum.
You cannot create polls in this forum.
You cannot vote in polls in this forum.

| Powered by YAF.NET | YAF.NET © 2003-2024, Yet Another Forum.NET
This page was generated in 1.050 seconds.