Author Topic: Stay Alive Circuitry  (Read 57258 times)

Alan

  • Conductor
  • ****
  • Posts: 1073
    • LK&O Railroad
Re: Stay Alive Circuitry
« Reply #15 on: July 26, 2016, 06:48:54 PM »
Is the desire to attach the keep alive after the module's rectifier, thus necessitating the LM3 ground pin, just to avoid the 0.7V drop of an additional diode? Or is there another reason?
Alan

LK&O Railroad website

When I was a kid... no wait, I still do that. HO, 28x32, double deck, 1969, RailPro

SD90

  • Fireman
  • **
  • Posts: 38
Re: Stay Alive Circuitry
« Reply #16 on: July 26, 2016, 08:13:20 PM »
RailPro's keep alive will be very cool. Maybe they will just build it into the LM4 and LM4S!

melarson

  • Guest
Re: Stay Alive Circuitry
« Reply #17 on: July 26, 2016, 10:29:12 PM »
Is the desire to attach the keep alive after the module's rectifier, thus necessitating the LM3 ground pin, just to avoid the 0.7V drop of an additional diode? Or is there another reason?

I only mentioned the diode drop in relation to connecting a battery to the module.  Tim's suggestion would bypass one of the diodes in the path to the module's internal power supply lines (henceforth referred to as the IPS).  I like that because that puts more of the battery's voltage into the module.  Since the battery output is only 12vdc, I would like to avoid as much unnecessary loss as possible.

But the keep-alive is to be wired to the 6-pin's new negative lead (the IPS negative line) and the 9-pin's existing blue lead (which is actually the IPS positive line).  This actually bypasses the entire rectifier (and according to Tim, any filtering, regulating, and protection circuitry) and as such would avoid a two-diode drop of around 1.3v, but that is not the reason the KA is to be wired that way.  The KA must be wired the way Tim instructs because it is polarity-sensitive, so connecting it to the IPS (a point of known, unchanging polarity) is essential.  There are other benefits, but that is really the big one.

The following is pure conjecture.  Because connecting a KA directly to the module's IPS bypasses the module's internal protection circuitry, I believe the Ring KA will likely have some sort of output protection built in, something I don't believe the DCC KA's have.  Which would explain the technical reason for Tim's disclaimer about using DCC KA's at our own risk.  Of course, the disclaimer could just be marketing.   ;)

Alan

  • Conductor
  • ****
  • Posts: 1073
    • LK&O Railroad
Re: Stay Alive Circuitry
« Reply #18 on: July 27, 2016, 07:27:05 AM »
So avoiding the voltage drop is the only reason. Then the usual KA cap/bridge/resistor setup can be used on the supply lines. Could you add another cell to make the battery 15.7V (+3.70V) and then the diode drop would be just right.
Alan

LK&O Railroad website

When I was a kid... no wait, I still do that. HO, 28x32, double deck, 1969, RailPro

William Brillinger

  • Dispatcher (Admin)
  • Conductor
  • *****
  • Posts: 1342
    • Precision Design Co.
Re: Stay Alive Circuitry
« Reply #19 on: July 27, 2016, 08:23:46 AM »
Quote
Then the usual KA cap/bridge/resistor setup

Can you elaborate on "the usual" please?
- Bill Brillinger, RPUG Admin

Modeling the BNML in HO Scale, owner of Precision Design Co., and RailPro Dealer.


Alan

  • Conductor
  • ****
  • Posts: 1073
    • LK&O Railroad
Re: Stay Alive Circuitry
« Reply #20 on: July 27, 2016, 09:15:34 AM »
ka.png

Bridge resolves track polarity issue and prevents capacitor from discharging to the wheels.
Resistor limits capacitor charge/discharge rate.
Capacitor supplies current only when none coming from wheels.

Could eliminate the resistor if not too many locos on the layout. Large amounts of capacitance may make it difficult or impossible to startup the power supply/breakers.

Effectively a filter too so not usable with DCC. Would wipe out the digital packets. This is RailPro. We play with pure steady state DC on the rails. Filter doesn't have anything to filter out.
« Last Edit: July 27, 2016, 09:25:09 AM by Alan »
Alan

LK&O Railroad website

When I was a kid... no wait, I still do that. HO, 28x32, double deck, 1969, RailPro

William Brillinger

  • Dispatcher (Admin)
  • Conductor
  • *****
  • Posts: 1342
    • Precision Design Co.
Re: Stay Alive Circuitry
« Reply #21 on: July 27, 2016, 10:07:00 AM »
Wouldn't the diode bridge prevent feedback of power from the capacitor to the rails anyway?

Other than using an existing Keep Alive, can you suggest a simple capacitor to use for this?
Or do you really need a bunch of caps to store enough energy for 2 or 3 seconds of travel?

I proposed this some time ago, but never acted on it.

http://model-railroad-hobbyist.com/node/22101

- Bill Brillinger, RPUG Admin

Modeling the BNML in HO Scale, owner of Precision Design Co., and RailPro Dealer.


Alan

  • Conductor
  • ****
  • Posts: 1073
    • LK&O Railroad
Re: Stay Alive Circuitry
« Reply #22 on: July 27, 2016, 10:13:52 AM »
Quote
Wouldn't the diode bridge prevent feedback of power from the capacitor to the rails anyway?

Answer

Quote
Could eliminate the resistor if not too many locos on the layout. Large amounts of capacitance may make it difficult or impossible to startup the power supply/breakers.
Alan

LK&O Railroad website

When I was a kid... no wait, I still do that. HO, 28x32, double deck, 1969, RailPro

Alan

  • Conductor
  • ****
  • Posts: 1073
    • LK&O Railroad
Re: Stay Alive Circuitry
« Reply #23 on: July 27, 2016, 10:30:03 AM »
Quote
Or do you really need a bunch of caps to store enough energy for 2 or 3 seconds of travel?

Yes, you need a lot of capacitance. The advent of low cost super caps make it practical in size. https://en.wikipedia.org/wiki/Supercapacitor

If you promise not to reverse the track polarity during the test, put a capacitor (correct polarity) across the rails of an isolated section of track. Run the loco. Kill the power. See how far the loco moves. Electrolytic capacitors experience RUD, a SpaceX term  ;D, when polarity is reversed. Rapid Unscheduled Disassembly can be very dangerous to you and your train.
Alan

LK&O Railroad website

When I was a kid... no wait, I still do that. HO, 28x32, double deck, 1969, RailPro

melarson

  • Guest
Re: Stay Alive Circuitry
« Reply #24 on: July 27, 2016, 11:32:09 AM »
So avoiding the voltage drop is the only reason. Then the usual KA cap/bridge/resistor setup can be used on the supply lines. Could you add another cell to make the battery 15.7V (+3.70V) and then the diode drop would be just right.

First, I stated, in the case of connecting a KA, that avoiding the drop was NOT the reason to bypass the module's internal rectifier.  It is a polarity issue.

Second, yes, you could add an external rectifier and put the KA after it, and then connect that to the red/black rail input wires of the module (as Bill has demonstrated in his post).  But then you are introducing yet another 1.3v drop before power ever gets to the module (where if you will recall it has its own internal 1.3v drop).  The more voltage you lose, the more you are going to start flirting with under-voltage alarms from the module.  For the LM-3 it is best to use the 9-pin blue (+) and 6-pin yellow (-) to connect the KA.  I agree that for LM-1's and 2's there is no choice; an external rectifier would be required.

And lastly, I think there is some confusion as to how I am implementing battery power.  It is my intention to run the module on battery power exclusively, with no power on the track, AKA dead rail.  The battery is not intended to act as a keep-alive.  Implementing KA's or Batteries are two very separate issues.

That said, the battery system I'm using is from Dr. Neil Stanton of S-Cab.com.  The battery itself is a single cell LiPo 3.7v pack with protection circuitry.  It is connected to an S-Cab BPS circuit board, which steps the 3.7v up to 12v.  Adding another battery in series to either the existing battery or the BPS output is not an option.  12v is what I have to work with, so if I can save a diode drop here or there I will do so, because I don't want to get under-voltage alarms all the time.  The BPS also acts as the charger when a supply is applied through separate charging inputs, and it has all the requisite output current and voltage protection built in.

Alan

  • Conductor
  • ****
  • Posts: 1073
    • LK&O Railroad
Re: Stay Alive Circuitry
« Reply #25 on: July 27, 2016, 12:32:32 PM »
Yeah, I think we have two interspersed conversations going on in this thread - battery power and keep alive.

I get why you don't want additional voltage drop. 12v is marginal to begin with (assuming HO). On the keep alive side of this conversation, there is no real problem implementing without waiting for a new pin on a module providing you start with sufficiently high power supply voltage i.e. power supply other than Ring. Of course Ring is not going to recommend that. No businessman would.

It appears your constraint is the S-Cab board restricting you to a single cell. If you could replicate or find a similar battery management board that works with multiple cells then you wouldn't need post-rectifier ground connection. But that opens up the whole dark side of LiPo so it is totally understandable why S-Cab is single cell.

Even the wall power group suffers from an occasional low voltage warning if your layout is of any size. Our power supplies have no external sense line so wiring inductance causes voltage regulation to be slow and sloppy with lots of over/undershoot. Fortunately, RP seems very forgiving of the condition. I have yet to notice any change in train operation when the warning appears. I think the warning trip is so fast it is catching the ringing in the supply line due to load change. The physical mass of the loco driveline totally masks the voltage swings so the voltage ringing isn't observable in the loco itself.

I installed a ballast load to calm down the ringing. Wasted power as heat but does the job.



Supply rail ringing before:



Supply rail after:

Alan

LK&O Railroad website

When I was a kid... no wait, I still do that. HO, 28x32, double deck, 1969, RailPro

Alan

  • Conductor
  • ****
  • Posts: 1073
    • LK&O Railroad
Re: Stay Alive Circuitry
« Reply #26 on: July 27, 2016, 01:45:48 PM »
Anecdotal support:

Years and years ago on a layout far far away I had a helicopter that I wanted to spin up/down realistically. On hand I had four enormous capacitors that had been removed from a broken beefy stereo amplifier. One of those old All Electronics 99 cent gearmotors drove a shaft that went up through the layout to the blades. I put the four huge caps in parallel across the gearmotor terminals and fed 6V through a resistor diode series. Essentially a keep alive all day circuit!  :P

The effect was awesome. The caps took so long to charge and discharge with the corresponding slow voltage curve across the gearmotor that the helicopter blades moved very lifelike. It took them several minutes to wind up to speed (or down). A poor man's substitute for pulse width modulation. I wish I had pictures to share but alas it was a time before the luxury of free digital picture taking.
Alan

LK&O Railroad website

When I was a kid... no wait, I still do that. HO, 28x32, double deck, 1969, RailPro

melarson

  • Guest
Re: Stay Alive Circuitry
« Reply #27 on: July 27, 2016, 02:23:11 PM »
1: I get why you don't want additional voltage drop. 12v is marginal to begin with (assuming HO).

2: On the keep alive side of this conversation, there is no real problem implementing without waiting for a new pin on a module providing you start with sufficiently high power supply voltage i.e. power supply other than Ring. Of course Ring is not going to recommend that. No businessman would.

3: It appears your constraint is the S-Cab board restricting you to a single cell. If you could replicate or find a similar battery management board that works with multiple cells then you wouldn't need post-rectifier ground connection. But that opens up the whole dark side of LiPo so it is totally understandable why S-Cab is single cell.

4: Even the wall power group suffers from an occasional low voltage warning if your layout is of any size. Our power supplies have no external sense line so wiring inductance causes voltage regulation to be slow and sloppy with lots of over/undershoot. Fortunately, RP seems very forgiving of the condition. I have yet to notice any change in train operation when the warning appears. I think the warning trip is so fast it is catching the ringing in the supply line due to load change. The physical mass of the loco driveline totally masks the voltage swings so the voltage ringing isn't observable in the loco itself.

5: I installed a ballast load to calm down the ringing. Wasted power as heat but does the job.

1) Yup, HO.  I'm still considering, against Tim's advice, connecting the BPS output to the 9-pin blue and 6-pin yellow, but tonight I will be taking some measurements to see what is there under normal conditions.  I'm hoping it's reasonably close to 12v.  Fingers crossed.

2) I agree, and it may not be a problem even with Ring's PS.  I don't know what the output voltage level is so I will be measuring that tonight too.  I may be worried for nothing.  Trust me, it's happened before.

3) I think I have two options with the BPS.  The best option is to contact Dr. Stanton to see if he's willing, on a special order basis, to adjust the converter output to, say, 14 volts.  That is unlikely but it never hurts to ask.  The other option is to place and external DC-DC converter to step the 12v output to 14v.  Problem solved but only if I can fit yet another component inside the engine.  Argh!

4) I get low voltage alarms frequently when running trains on a loop of Bachmann EZ-Track.  Granted the loop is typically around 60 or 70 feet of track, and they do occur more often at the opposite side of the PS connection, but I do get them.  When I have the info screen up I can see the fairly wild fluctuation of the track voltage, almost regardless of where the train is located on the loop.  And I agree, the train keeps going, but an alarm I can not clear and just have to wait for it to clear itself is kinda frustrating.  Guess it's just my personality.

5) I like it.  And your results!  Just when I thought terminating resistors were a thing of the past.  But actually they are still used to this day on lines like ModBus and DeviceNet and probably many other communication lines.  Those resistors can dissipate a considerable amount of power; what resistance are you using?

Alan

  • Conductor
  • ****
  • Posts: 1073
    • LK&O Railroad
Re: Stay Alive Circuitry
« Reply #28 on: July 27, 2016, 03:13:55 PM »
Quote
Those resistors can dissipate a considerable amount of power; what resistance are you using?

Ha! Funny you should ask.

They are 20 ohm 100 watt units. At 15V that's about 11 watts of heat. Well within the resistor power rating (25W no heat sink, 100W large heat sink) however after particularly long sessions they get quite hot. In excess of 200 F. I don't like that. So, on the bench right now are replacements - three parallel 100 ohm 100 watt resistors for a net of 33 ohms. At 15V that's about 7 watts but spread over much more aluminum.

The ringing is so well dampened at 0.75A I am banking 0.5A will be good enough. China eBay didn't have 60 ohm available and stateside suppliers want $10 each. I bought a lot of 10 for $28.

DSC_0963.jpg

Alan

LK&O Railroad website

When I was a kid... no wait, I still do that. HO, 28x32, double deck, 1969, RailPro

KPack

  • Conductor
  • ****
  • Posts: 784
Re: Stay Alive Circuitry
« Reply #29 on: July 27, 2016, 05:59:59 PM »
Lol, I'm so confused right now.  I'm catching bits and pieces but by and large this terminology is way over my head.

-Kevin