The chuff input on analog/DCC capable sound cards like the Phoenix, Tsunami or Digitrax all need to see the track signal (on DCC) - or - the track voltage (on Analog DC).
When they see a track voltage on the chuff input that exceeds 3 to 5 volts, they chuff ONCE. (voltage dependent on the brand of card)
They will not chuff again until there is NO VOLTAGE on the chuff input for AT LEAST 10 milliseconds.
Got that?
Read it again.
Another way of saying this is:
If there is NO VOLTAGE for at least 10 milliseconds on the input then the sound card is cocked and ready to produce a chuff.
As soon as the voltage on the input EXCEEDS the trigger voltage (3 to 5 volts) the sound cards makes a single chuff sound and will not make another until the card is once again "cocked".
Notice that there are two states here:
1) Zero volts for longer than 10 milliseconds creates a cocked state,
2) Trigger state, which starts a chuff sound as soon as the voltage exceeds the trigger voltage
The trigger voltage can be either positive or negative with respect to one of the sound cards track power inputs. It can be plus DC or minus DC. It can even be AC voltage or a digital signal like DCC.
For all the cards listed above, the card can be cocked, and trigger another chuff BEFORE the previous chuff completes. This allows long chuff sound for slow operation yet lets another chuff start as soon as needed when the loco is moving fast. The loud attach overriding the long low decay of the previous chuff.
Just for kicks, I installed a 2K2, an Tsunami, and three different Digitrax cards in my Kay while leaving the Kay set up for Analog DC
Then I did the same five cards again, but this time using a couple different DCC decoders.
I tried all of the procedures found on the Bachmann, Phoenix and Digitrax sites as to the transistors and such.
In some ways, all the installs worked,,,,
well, sort of worked,,,
well, at least some worked until they burned out something.
I started with something simple.
FIRST ATTEMPTS
I connected the five different cards track inputs to the track pickups, connected a speaker, and connected the chuff input to the Kay's optical signal.
No good! The Kay's chuff signal never goes to zero volts with respect to either track unless the power is off feeding the track. The Kay's optical signal is always 1.2 volts above zero due to the two .6 volt drops across the diodes in the dummy card's bridge and the huge filter capacitors on the boards in the tender and loco. The sound cards only chuff once when they first see power and the optics sends a signal. after that, the sound card is never "COCKED" because it never sees zero volts except on dirty track. You will get a cuff at power up and every once in a while when running on dirty track, thats it.
SECOND ATTEMPTS
No good! The Kay now chuff's perfectly in time with the drivers. The chuff works at all speeds from about 5 volts up to a full 24 volts dc without missing any chuff's or erratic chuff's, even on dirty track. Since one of the sound card track inputs is connected to the ground of the dummy cards DC output, and so is the optical circuit, the sound card now sees zero volts when the Kay's chuff signal is at zero volts. The 2K2 when equipped with the battery pack worked all the way down to zero volts, but the chufing was not in time with the drivers until track voltage reached 3 or 4 volts. The meant that the Kay produces standing sounds until well after the loco was moving. In this arrangement, the battery does not power the Kay, a battery could be added to power the Kay logic and produce reliable chuff's all the way from 0 to 24 volts.
The real objection to this second method has to do with the polarity of the track signals applied to the sound cards. Since the cards are now connected to the output of the bridge on the dummy card, the polarity is always the same. All these sound card depend on the polarity to sound either two toots for forward or three toots for reverse. Connect the track leads to pins 6 and 7 and the loco will always give you two toots no matter which direction the loco starts moving. Reverse the wiring and you always get three toots. These cards also depend on polarity changes to set up triggers for other sounds like the coupler crash, cab chatter and reverser lever squeaks or the drains on the cylinders.
Worse still, the sound cards depend on the voltage for some of the effects or controls on DC. For example, you can blow the crossing signal on the 2K2 by running at a constant, fairly high speed for a while, then lowering the voltage just slightly. the crossing signal will sound. However, the large capacitors and voltage regulators interfere with this. sometimes you get a crossing signal, sometimes you don't. Worse, you may get a crossing signal every time you top a hill or exit a curve. In the brief testing I did, all of the voltage depended controls were so erratic and undependable as to be useless. One other problem has to do with the fact that the motor gets 1.2 volts before the sound card sees any voltage. Three bridge rectifiers in series is required in the motor circuit to get the loco even close to having any effects like the single toot on stop or steam release on stop to work correctly.
THIRD ATTEMPT.
I connected the five different cards track inputs to pins 1 ans 12 of the socket, connected a speaker, and connected the chuff input to the Kay's optical signal and installed a transistor to "invert" the signal. This arrangement sort of works. chuff at low speeds in nonexistent or erratic. from about 8 volts to half throttle, the chuff is reliable, but cuts out at half throttle and once again becomes erratic. below 8 volts the chuff sometimes works, but may cut out if you place your hand near the tender, or touch one rail or place a lighted car on the track. I used a scope to analyze what was really accomplished by the transistor. Once again, the chuff input was not going to zero volts to "COCK" the cuff sound. Instead of dropping to 1.2 volts, it now dropped to .6 volts or the forward bias of the transistor. Basically, track pickup noise has a better chance of dropping the voltage to zero when there is only one forward junction to contend with. I then tried adding bias and dropping resistors to tweak the circuit. That was successful, so much so that I was able to remove the transistor and simply use a resistor between the one rail and the chuff input. This was still a little tricky, If the resistor is too large, the chuff doesn't work reliably. If it is too small, it burns out the bridge in the sound card or the bridge on the dummy card. I then tried a power transistor instead of the little plastic ones. That burned out another sound card. so running low on cards, and wanting to save some for my DCC attempts, I stopped trying to improve the Analog DC experiments.
My conclusions for Analog DC, The Phoenix 2K2 comes the closest to working, If you have the programing cable and disk, You can disable most of the sounds and special effects that are polarity or voltage dependent. Wire the track inputs to pins 6 and 7 of the Kay socket. Then use reed switches and track magnets to trigger the bell and whistle as is done on LGB. This will give you a very accurate chuff exactly timed to the driver positions and will sound good if you have shimmed or replaced the CW's. It will even give you that wheel slip rapid chuff's when starting up with a heavy train or climbing a steep grade. You can NOT get drifting or labored chuff's unless you sacrifice the bell and whistle and use track magnets to operate the labored/drifting (or even Doppler) effects.
FOURTH and more ATTEMPTS
I also performed these same steps using DCC in conjunction with three different decoders. Since the Analog DC experiments only took a couple days and the DCC ones took several weeks, there isn't room here to cover the DCC aspect of the Kay installation. After all, The question above was only about the 2K2 and sound on DC.
For both DC and DCC installations I tried to use th optical chuff. For DCC I wasn't much more successful than with DC. Some methods and tricks yielded a reliable chuff but had other drawbacks. And, remember that a DCC equipped loco should also run on analog DC. That I believe would be an impossible feat for the optical chuff and all the regulators and circuitry in the Kay.
For DCC I ended up abandoning the optical cuff. I epoxied a reed switch in each cylinder where the optics had been hounded. I epoxied a magnet to the light interruption flag The two reed switches were wired in parallel. One side of the switches connected to the left rail pickup in the loco. The other side of the reed switch connected to the draw-bar between the tender and loco with the chuff input connected in the tender to the draw-bar post. All other wiring, circuits, diodes, lights, boards, and plugs were removed. The only boards remaining in the Kay are the two pickup boards above the first second and fourth drivers. Those were re-wired with braided copper ribbon. There are no connections between the loco and tender other than the draw-bar.
The sound card is installed in the tender. Two wires to the tender truck pickups, two to the speakers, one to the draw-bar for chuff input.
I added a backup light to the tender as well as two lanterns and hanger brackets at the rear. Each lantern is fitted with three 12 volt bulbs, (white, red, green) and driven by a function decoder mounted to the rear wall of the tender shell. The two track wires from the function decoder have a two pin plug so that I can remove the tender shell, unplug the two conductor cable and set the shell to one side. The sound decoder stays in the floor of the tender with the speakers.
For the boiler, I used the same method. A two conductor cable with a plug in the middle so I can easily separate the boiler from the motor rand frame. I use a function decoder on the smoke box for the three colored 12 volt bulbs in the marker lights plus one in the headlight. A second decoder function near the firebox operates bulbs for the flicker. The motor decoder is mounted in the frame 1/2 inch to the rear of the flywheel. screw terminals on the decoder connect the track pickups and motor. Wires to the motor are kept short in this location to insure the best possible speed regulation under varying load. A fan and the smoke generator are connected to a plug on the decoder as well as a slow motion turnout motor used to operate the reversing lever.
All in All, I am happy with this arraignment. It is very clean. any wire over an inch long is one of the two conductor track cables and has a two pin plug for easy disassembly. There are no wires between the tender and loco and I have all sorts of options for lighting. I the stupid rock concert in the firebox is replaced with constantly varying realistic fire.
It works well, but it falls far short of what I had hoped to do. I had hoped that I could make up plug boards with a variety of different sound cards, Decoders, and even RC setups. I had looked forward to being able to swap from DC to DCC to Air-wire, to DCS, to RCS and back with the simple swap of a board and perhaps throwing a switch or two. I really wanted a loco that could easily swap and compare various sound boards or decoder. Seems such a system is still a long way off.
For your DC system with a 2K2, I'd say, there is no need to strip the entire system. Just install reed switches and magnets. You could retain all the other lighting, regulators and such. On the other hand, I'm sure there is much that you could improve by re-wiring the entire loco. Most likely, from what I've learned, you will eventually end up riping it all out as you experience various failures. Or, you could get lucky and never have a single problem with the pickups, short circuits, or blown transistors and voltage regulators.
B0B
" border=0> 
