Improvement for ORD1 Signal Output

Post Reply
bbernhard
Posts: 1229
Joined: 23.02.2008, 19:07
Location: Villach/Kärnten/Austria
Contact:

Improvement for ORD1 Signal Output

Post by bbernhard » 31.01.2009, 21:57

Hi

in one of my last post I wrote about the ORD1 which was working out of the box.
Unfortuneately next day I had to recognize that some of my diy accessory decoders dont work any more.

Driven from this thread http://forum.rocrail.net/viewtopic.php?t=634 I did some investigations.

I logical "1" gives this siganls

Image

Trace 1 shows Input to ORD1 measured directly on SUBD Connector
Trace 2 shows Output on Booster connector
Trace 3 shows Output to rails (without load)

We can see input signal Trace1 (RS232) is clipped and as result of this DutyCycle of booster output Trace3 is not symetric
There is 44uS high and 60uS low.
But acc. NMRA the difference between high periode and low periode should be less than 3uS, but we have 16uS which is far away from spec.
http://www.nmra.org/standards/DCC/stand ... 004-07.pdf

This is no problem of the power stage, because output on booster connector Trace2 shows same problem.

Altough as shown in Trace1 the input signals DutyCycle is exact 50% which is OK.
Overal periode is 103us which is very short. This is even discoussed in the other thread.
More or less Booster input signal can be considered as acceptable.

----------------
Then I played around a little bit and changed resistor R42 from 330R to 1k5.

Now signals looks like this
Image

We see input signal is now no more clipped, and also the Duty Cycle of booster output is now within spec (difference High/Low periode <3uS)

Question: Do you see the similar problem, or is this a result of weak driving capability from only my computers serial port ?
If yes, maybe documentation for ORD1 should be adapted.
Probably this could be a fix for others who suffer from non working decoders with ORD1.

Best of all: Now all of my decoders accept the ddc commands as they did with the Intellibox.

best rgds
bb

btw.
If we check the datasheet for PC817 optocoupler it might be advicable to rise also R40 and R39 (eg. 2k2) to avoid problems in case the optocoupler itself is on lowest edge of its spec.

ron&bram
Posts: 2460
Joined: 11.06.2008, 19:34
Location: Heemskerk, Netherlands

Post by ron&bram » 31.01.2009, 23:08

Hi Bernard,

My compliments, nice research work (I can imagine that a "scope" is not in everybody's toolkit for modelrailroading).

Best regards, Ronald

Gerald
Posts: 163
Joined: 22.02.2008, 15:04
Location: Weener / Germany

Post by Gerald » 17.02.2009, 20:07

Hi,
I had some problems with this decoder:
http://www.wagenlehner.net/Decoder/WDEC03i.html
ORD-1 and WDEC03i didn't like to work together.
I increased R42 to 1,63 K (no other value in stock). Now, all is working fine. Bernard, thx a lot for your hint. It's really useful.
Regards,
Gerald

Ol_Nick

Post by Ol_Nick » 29.03.2009, 10:47

Tobbe wrote:Hi Bernard,

thanks for your work!

Yes, It sounds very possible that serial ports have different drive strength. I have never seen this behaviour, but I have only tested on a couple of computers, lets say five (out of millions of PCs... :-) )

I will try change the suggested resistors and then make changes to the design.

Best regards,
Tobbe
I think people should take the "if it ain't broke, don't fix it" approach with this. I have just built an ORD-1 from scratch, and I changed R42 to 1K5, and the ORD-1 didn't work properly (D14 permanantly on, LED's not flickering). Changing it back to 330R restored operation (but early days yet !).

Maybe I should have read this entire thread before jumping the gun ? :oops:

Arpke
Posts: 89
Joined: 17.09.2009, 09:25
Location: Lehrte - Germany

Re: Improvement for ORD1 Signal Output

Post by Arpke » 10.03.2016, 15:12

I built three ORD-1 from scrach five years ago. But I don't finished the job.
During the last 2 weeks I finished all three. But all of them don't work.
Diode D14 is always burning.
The PC with Rocrail was an old one with a single core cpu and RS-232 and parallel port on the main board.

I have no oscillloscope - so I can only measure with an multimeter, read blogs in the forum and think whats going wrong.
I recognized that the problem must be the input of optocoupler O9.

This article gave the impulse to changed R42 from 330 to 1R5 - but I had the same result - all ORD don't work and one of die Diodes D12/D13 does not work too.

Then I changes the PC (old iron with Lubuntu) to an old PC with XP Professional.
Result: D14 off. Rocrail Power on. It works with R42 = 330 Ohm - but it don't work with 1R5.
So I changed R42 back to 330 - all seems to be OK and I tested it with a loco. ORD-1 runs.

I lost days because I testet it with a computer which will not serve.
Lesson learned: Errors can be anywhere. Think !

Norbert

Liviu M
Posts: 949
Joined: 03.12.2011, 20:44

Re: Improvement for ORD1 Signal Output

Post by Liviu M » 10.03.2016, 15:46

Hello Norbert,
Arpke wrote: from 330 to 1R5...
It works with R42 = 330 Ohm - but it don't work with 1R5
I suppose it's just a typo error, but the resistance should be 1k5 (1.5 kilo Ohm) and not 1R5 (1.5 Ohm).

Best regards,
Liviu

Arpke
Posts: 89
Joined: 17.09.2009, 09:25
Location: Lehrte - Germany

Re: Improvement for ORD1 Signal Output

Post by Arpke » 10.03.2016, 19:13

Thanks Liviu,
you are right !

I tried to change R42 fron 330 Ohm to 1500 Ohm.

Norbert

Post Reply

Return to “DIY Hardware”