TEAM DIGITAL, LLC

Improving the world of DCC

5/29/09
The purpose of this test is to see if running long lengths of wire side-by-side to the BlocD8 can cause crosstalk (BlocD8 shows a block occupied that is not occupied). The wires were wrapped together with masking tape in several places. The BlcoD8 sensitivity was set to maximum (3 ma). The resistive load was increased in one block until the Zephyr circuit breaker tripped (a little over 2.5 amps). The bus wire that supplies power (rail A or plus) to these two blocks is also 30' and has been moved around to see if it has any effect. The rail B or negative wire from the booster to the blocks is short perhaps 3'.

Next a NCE 5 amp booster was used so the load could be to increased. The load is resistive and was increased to almost 5 amps. Still no crosstalk.

Reference: Test track with DBD22 block detectors with about 2' to 3' of wire. No pictures were taken of the DCC track voltage but as might be expected the scope showed less ringing at light load and less smoothing of the DCC wave form at high loads.

5/29/09
Testing with locomotives as load. Jim Betz makes a good point. His says - " I realize that the BlocD8 senses load and changes in load - but the fact that the load needs to be "real trains" is important. "

"I do not know why this is true - but I do know that the electronic 'signature' of the load seems to be important. You may find that using a simple decoder such as an NCE DASR will produce failures where a static load will not. Our experience showed that older non-sound decoders were 'worse' (more likely to produce a false occupancy state) than a QSI decoder and that a TCS decoder was the best of all (least likely to produce a false occupancy). But that actually running the train was key to the false occupancy (speed step 1 - probably the worst case) and that even the TCS would produce multiple blocks reporting occupied when the train was clicked from step zero to step 1."

The three sound locos mention under the wave form picture are:
1. Athearn Challanger with Tsunami decoder
2. Rivarossi FFE with Tsunami decoder
3. Rivarossi Big Boy (1995) with SoundTraxx decoder
Locos with TCS and Digitrax decoders were tested also.
The 'worst' looking DCC wave form does not nessarily mean most likely to cause crosstalk. The voltage transients are coused be the loco motor/decoder. It appears the farther the detector is away from the loco the smaller the transients.

6/1/09
Addition testing was done with several 'older' Athearn diesels. So far no crosstalk has been obsevered.

BlocD8 Testing

Click to enlarge

DCC voltage signal at block with three sound locos. Current about 1.5 amps. Of the locos tested these gave the 'worst' looking wave form. Osc 10v/div

DCC voltage signal at block. Almost 5 amp resistive load in block. (NCE 5amp booster). Osc 10v/div

DCC voltage signal at block. No load in block but slight load on booster (NCE 5amp). Osc 10v/div

Diesel engines used in testing. Older Athearn.

Example of incorrect (black) wire routing. This caused crosstalk after the current (resistive load) was increased to almost 4 amps (only one LED should be lit).

One way to get 30' of wire in a small space.

Stream engines used in testing.

The wire is 14 ga stranded. The black wire provides power to both blocks.

This is a proto SRC16 and BlocD8 connected. The clamp is used as a strain relief for the wires.

This is the other end of the 30' wires going to the blocks on the test layout. A clamp is used for a strain relief here also.