开云体育

Is there a good reason that the "Output Matrix Mast" can only control 6 lamps? A german combined exit and distant signal can have up to 9 lamps (signal system DB HV 1969). Would be nice to control all 8 aspects via one signal mast.

Eckart

The Output Matrix Mast is designed to control a signal mast with 32 aspects. This is more than enough for a German mast looking at the DB HV 1969 system in JMRI.

If the mast is broken down in to two heads - Main and Distant, it simplifies things a little.
The main head can show five different aspects:-

1. HP00/SH0 (two red lamps)
2. HP0 (one red lamp)
3. HP0/SH1 (one red lamp and two white lamps)
4. HP1 (one green lamp)
5. HP2 (one green lamp and one yellow lamp)

The distant head can show four aspects:-

1. VR0 (two yellow lamps)
2. VR1 (two green lamps)
3. VR2 (one green and one yellow lamp)
4. Dark

I guess that if I was to use the matrix driver the drive the lamps directly I would just drive the six key lamps with the matrix driver, and then use LOGIX to drive the remaining lamps which are really redundant (the ones where you know what the aspect is even when they are not showing!).

The top head green lamp is always on when the left red lamp is off (and vice versa), this is a simple LOGIX rule that can be used to drive the green lamp. You just have to drive the two red lamps, the pair of white lamps together, and the yellow lamp, on the top head using four of the matrix outputs.

On the distant head you can use the remaining two outputs to drive the lower yellow and top green, and use LOGIX to generate the remaining green (on when green on and yellow off), yellow (on when yellow on and green off) and white (on when either green or yellow are on).

So to recap, the matrix driver will drive the following lamps directly with the six outputs:-

1. Left red lamp
2. Right red lamp
3. Double white lamps
4. Yellow lamp (top head)
5. Lower yellow lamp (distant head)
6. Upper green lamp (distant head)

Logix will drive the remaining lamps:-

1. Green lamp (top head) and white lamp (distant head) (on when left red lamp off)
2. Upper green lamp (distant head) (on when lower green lamp on and upper yellow lamp off)
3. Lower yellow lamp (distant head) (on when upper yellow lamp on and lower green lamp off)

Suzie x

A bit of a faf, but will work if you do not have a signal mast decoder that will drive all ten lamp groups.
Clearly if you don't have the HP00 and SH1aspects and signals are not close spaced outputs will be freed up so you should not need all the LOGIX and things will be simpler to implement.

I think it’s just a bit of history.

Let me take a look at what it would take to go to nine. Are there any signals that need more?

Bob

On Jan 10, 2020, at 9:47 AM, mey-jmri@... wrote:

Is there a good reason that the "Output Matrix Mast" can only control 6 lamps? A german combined exit and distant signal can have up to 9 lamps (signal system DB HV 1969). Would be nice to control all 8 aspects via one signal mast.

Eckart

--
Bob Jacobsen
rgj1927@...

A position light signal has the max possible of 9 different marker light
options. I think some would be mutually exclusive, I've seen pictures with 3
or 4 different markers. Now combine that with it being a dual position
light, that does give a max of something like 26 in count between light
aspect (vertical, horizontal, slant right, slant left) plus the two sets of
nine. Way overkill, beyond prototype, but just trying to guess a worst case
ever.

A number of outputs like that should cover the worst case of feathers for
the BR signals, you'd have a whole chicken up there! But Suzie would know
best on that case. Thinking in a crazy way gets us the large number, then
question then becomes what internal strife would such a large number present
to the code or users? Think you could spell the alphabet with that one.

-Ken Cameron, Member JMRI Dev Team
www.jmri.org
www.fingerlakeslivesteamers.org
www.cnymod.org
www.syracusemodelrr.org

B&O Position Light signals will need ten lamp groups for a fully populated mast, and I think some Pennsylvanian double head position lights might need ten as well. I can't think of any scenario off the top of my head with more than ten lamp groups on a single mast.

The original history of this driver was to drive a hardware decoder that converts aspects to lamps, rather than drive the lamps direct, hence 64 aspects was considered overkill at the time. The Canadian system with its 32 aspects was considered to be the most complex system to implement this way. Upping the capacity to ten outputs does extend the the use of the matrix driver and reflects how it is (unexpectedly) being used in the field so will probably be a boon.

The BR system with feather could require eleven without getting in to the complexities of route indicator boxes, but 'The whole chicken' just does not happen in practice, and a model is very unlikely to approach Nuneaton levels of complexity:-



This mast (one feather short of the full chicken, the most complex I know of) would require 14 lamps, it has six high(ish) speed routes and three restricting speed routes.

Thanks for the info.

The code is a bit of a mess, so I just extended it to 10 outputs and cleaned up a bit. Perhaps somebody will clean it up a bit more when somebody has a need to go to more.

The change is in PR 7926 and hopefully will be in a development release later today. I’d really appreciate testing of it. (Did I mention the code is a bit of a mess?)

Bob


--
Bob Jacobsen
rgj1927@...

Thanks! Will test next week.

Eckart

JMRI 4.19.2 should be able to handle up to 10 outputs.

I’d very much appreciate people checking that to make sure it works for them.

Bob

On Jan 10, 2020, at 9:47 AM, mey-jmri@... wrote:

Is there a good reason that the "Output Matrix Mast" can only control 6 lamps? A german combined exit and distant signal can have up to 9 lamps (signal system DB HV 1969). Would be nice to control all 8 aspects via one signal mast.

--
Bob Jacobsen
rgj1927@...

Checkbox "Reset previous Aspect" seems not to be saved into the XML file. Have?nt tested outputs so far...

Eckart

Bob

I don't see that there is any need to increase the number of outputs to 10. I use only four outputs into my own brew matrix decoder and that gives me 16 different combinations of the eight lights in my signal mast. Six inputs to a decoder could? potentially drive 64 separate lights but that would need a huge chip or one that properly drives a matrix. With a 4x4 matrix, 16 lights could be driven. Three rows and four columns of a matrix with five inputs could drive 12 lights in 32 combinations in a 14 pin chip. In a 14 pin PIC chip without using a matrix with say 4 inputs allows eight lights to be driven in 16 combinations. With a 20 pin Pic, no matrix, six inputs would allow 12 lights in potentially 64 combinations.

Eckart, if you are wanting a decoder circuit board to meet your requirements then I have some PCBs for another signalling project that could be used. This is how the Matrix signal mast is designed to be used. My board is set up for nine outputs and nine inputs but not all of those inputs have to be used. You seem to need nine outputs: Main head red1, red2, two white, green, yellow, distant head yellow1, yellow2, green1, green2. I could also create the firmware for the pic to drive what you want as it doesn't look very complicated. From what Suzi has written above you would need only four inputs. The pairs of white lights could be paralled to use a single output as long as each has its own dropper resistor and that the current is not huge. If you, or anyone else, is interested in this please get back to me.

Fraser

--
Petr ?ídlo
Czech Republic

Maybe (!) there is a misunderstanding - perhaps most of you know the history of the Matrix driver and why it was invented. I don?t. I?m new to JMRI. I just wanted to control 9 Lamps directly connected to GPIOs of a very simple controller, no extra hardware, just 9 JMRI "Turnouts" for 9 lamps. When trying to control this by a signal mast, i found that the matrix driver was exactly what I want (besides the fact that it has only 6 outputs for 6 Turnouts). No tricks, no scripts, no hardware, just one pane to configure. So I asked a simple question and was 1. positivly surprised of Bob?s very fast update and 2. astonished about that much counter-reaction (if this is the right word) - no one is forced to use the 4 new outputs :-).

Again, thanks Bob.

Eckart

One of Really Nice feature of model railroading is that there are _lots_ of approaches to building and operating a layout. JMRI is (partly) oriented toward that: If somebody wants to use something in a different way, that’s fine. That lets it be useful, even valuable, to lots of people. Extending it to capture a different way of doing things grows the party.

At the same time, too many options can be confusing. Sometimes telling people how to do things, or maybe two ways of doing things, gets them to a workable solution faster.

There’s some tension between these, and different people have different preferences. That’s OK: Even about those things, JMRI tries to allow lots of approaches to discussions. But then some people prefer just one approach, and the cycle repeats…

With luck and ongoing brilliant work by the moderators, JMRIuses will keep doing this and providing solutions for a long time…

Bob

On Jan 14, 2020, at 4:34 AM, Eckart Meyer <mey-jmri@...> wrote:

Maybe (!) there is a misunderstanding - perhaps most of you know the history of the Matrix driver and why it was invented. I don?t. I?m new to JMRI. I just wanted to control 9 Lamps directly connected to GPIOs of a very simple controller, no extra hardware, just 9 JMRI "Turnouts" for 9 lamps. When trying to control this by a signal mast, i found that the matrix driver was exactly what I want (besides the fact that it has only 6 outputs for 6 Turnouts). No tricks, no scripts, no hardware, just one pane to configure. So I asked a simple question and was 1. positivly surprised of Bob?s very fast update and 2. astonished about that much counter-reaction (if this is the right word) - no one is forced to use the 4 new outputs :-).

--
Bob Jacobsen
rgj1927@...

Was that ever stored? I don’t see any vestige of it in the code.

If it wasn’t there, I’ll add it, but it might take a couple days.

Bob

On Jan 13, 2020, at 11:54 AM, Eckart Meyer <mey-jmri@...> wrote:

Checkbox "Reset previous Aspect" seems not to be saved into the XML file. Have?nt tested outputs so far...

--
Bob Jacobsen
rgj1927@...

Hi Eckart
Apologies if you thought I was complaining about you. What I was attempting to do was save Bob doing any more work on making the matrix signal mast use 10 outputs. As it has been in existance for three years now and so far it has not been necessary to have more inputs, it would seem to be OK as it is. You did not state in your original post that you didn't want any extra hardware so that may be where the misunderstanding was.

You say you don't want more hardware but in your scheme you will use 9 outputs to drive 9 lamps. If you can use an inexpensive and simple piece of hardware to allow you to use less than half the number of outputs required to 4 then that to me sounds like a good trade. You may feel otherwise of course.
Fraser

I think the “reset prior” option is a hold-over from the Turnout Controlled Mast. That has the same option.

There were two use cases for that:

1) For some uses, just the new turnout command has to be set. For other uses, the new one has to be set _and_ the old one has to be reset to prevent the two appearances being mixed.

2) Some people think that a moment of dark in between the new and old appearances is more prototypical.

I’m not sure that either of those is relevant for Matrix Masts, but I don’t have a lot of direct experience.

Bob

--
Bob Jacobsen
rgj1927@...