BLI Paragon 3 Mars light

[Cajonpassfan]

I’ve been playing with my BLI PA’ set tonight, and got my settings tweaked mostly the way I like it, but for the life of me I can’t figure out how to reduce the intensity of the Mars light (L1). From prototypes I’ve seen and would like to emulate, the light cycle should be mostly dim, then ramping up, then a quick FLASH, ramping down, and dim again. The “dim” phase should be at least a half of the cycle that lasts 1-1.5 sec.? The pertinent page from the manual is below. Any ideas on how to make the Mars behave as described?
Thanks, Otto

[tehachapifan]

Well, that's certainly not easy to follow! Like many DCC instructions, I will often need to stare at the piece of info for upwards to about an hour (maybe more) before something finally clicks. This looks like it could be one of those.

That said, if using LED's, the lights may not go as dim as perhaps you would like or as they may if using incandescent bulbs. Some decoders have a setting if using LED's so that they behave more like incandescents, but some people just don't like lighting effects when using LED's. However, I see a lot of models (where incandescent bulbs were used) where the Gyralite or Mars effect goes completely dark between flashes. This is not prototypical as the lights are always on at the same brightness on the prototypes, it's just a matter of where the reflector is focusing intensity at any given moment as it oscillates. You probably know this already, but....

[Cajonpassfan]

Well Russ, I’ve had more than an hour...still nothing
The irony is that I have the headlight just below the Mars light dim quite nicely, per Rule 17, so I know these LEDs are capable of it. And at its dimmest, the Mars is okay, it just doesn’t stay that way anywhere near long enough. Shortening the ramp up/on/ramp down cycle just makes the cycle pulsate way too fast. I’m thinking I’m missing something basic here...
Otto

[jdcolombo]

I’m thinking I’m missing something basic here...
Otto

Or you may not be.  A decoder that has no way of programming the headlight to be on in both directions may not be capable of an accurate Mars light effect.

John C.

[jagged ben]

Do they honestly mean milliseconds and microseconds?  I definitely don't get it.

[Cajonpassfan]

Yea, that’s my problem, I don’t get it either, after rereading it multiple times... And I thought maybe it was just me.
The terminology is baffling to me. There are references to “period” , “counts” in ms and usec. A period is 127 counts, “actual period time” (what?) is 11 ms. 11 milliseconds?
Otto

[tehachapifan]

I've read it a few times over since my original reply. Still completely baffled! May be the most confusing DCC-related instructions I've ever run across.

[peteski]

This is super-confusing the way it is presented. It  is not just that it is written for a computer expert - it is confusing to everybody!

As I read this, single count unit is 11.6ms (milliseconds or thousands of a second).  So a value of 10 (in one of the specific CVs) would equal to 116 ms (or 0.116s) and a value of 100 would be 1100ms (or 1.16s).  But I'm still confused about which CV does exactly what.

[jagged ben]

...

As I read this, single count unit is 11.6ms (milliseconds or thousands of a second).  ...

Nope, that's the period.  A period is 128 counts.  That much is clear.  I think counts are fixed at 90.7 microseconds, since that is a count 'increment/decrement', and the math works out, rounded off.   But I'm less clear what 'increment/decrement/ refers to.

I think we need to see the previous page of the manual though, since the final examples have a reference to CV 160, whose purpose is not explained in what we've been shown.  Maybe CV 160 determines how many periods long the total effect is?  If that were correct then the ditchlight example would have the total effect length at 1.87 seconds, which seems about right for left and right ditchlight to cycle on and off.

I also feel like this business of periods and counts may have something to do with the nitty-gritty of using pulse-width modulation to determine brightness.  That's the only way I can make any sense of needing to consider micro-seconds. 

[jagged ben]

Looking at the examples might be a better way to deduce the meaning of the language than trying to understand the language directly.   So might calling BLI and saying 'can you translate this?'   

[tehachapifan]

The person who wrote this probably thought it was as clear as it could possibly be. Anyone who has written instructions has probably experienced this at one timer or another. You already understand the material so all the steps make perfect sense. Shows the value of someone else proofing it before releasing it (preferably by someone who doesn't already know the steps inside and out too).

[Cajonpassfan]

Nope, that's the period.  A period is 128 counts.  That much is clear.  I think counts are fixed at 90.7 microseconds, since that is a count 'increment/decrement', and the math works out, rounded off.   But I'm less clear what 'increment/decrement/ refers to.

I think we need to see the previous page of the manual though, since the final examples have a reference to CV 160, whose purpose is not explained in what we've been shown.  Maybe CV 160 determines how many periods long the total effect is?  If that were correct then the ditchlight example would have the total effect length at 1.87 seconds, which seems about right for left and right ditchlight to cycle on and off.

I also feel like this business of periods and counts may have something to do with the nitty-gritty of using pulse-width modulation to determine brightness.  That's the only way I can make any sense of needing to consider micro-seconds.

Okay, here’s the preceding page dealing with CV 160. It allows one of three types, as I read it, single modulation, double modulated, and ditch. It still doesn’t tell me how how to reduce time on without shortening the cycle. I’d be happy with a 1.8 second cycle, if the “time on” could be very brief.  As it is, the light is mostly on full brightness, see pic below.
And yea, calling BLI may be the answer, but I’d like to understand this better so I know what questions to ask.
Otto

[peteski]

LOL, that doesn't help. 
Maybe a call to BLI is needed for clarification (working this out through email might be frustrating).