9-19-16 rush to get and understand data + more on ESU sound decoder comparison mod 9-23-16

Down to 12 days until we head to the Condo. This year I will be away from the trains for just shy of four months. Because do this tests have been listed and prioritized. The goal is to finish four separate engine series. This represents at least 12 separate tests. Multiple motors, and DCC comparison to DC on one. There may be more, but time is short.

Today is a good example of how each test tends to lead to more work which is time. One of the four engines is a P2K E6A Southern DCC ready model. Originally this was to be a three-step electronics variation. With DC & DCC comparisons included. Before it was finished, the interest in ESU lock sound decoders surfaced on MRH among other places. So the series was increased to include a Lok sound decoder. It all sound simple. Well, last week the Lok sound version was tested on DC. The results brought up as many questions as it answered. As shown in the following figure:

The high voltage side indicated that the top speed may not have been achieved in the run length to the speed measuring plane. The other question has to do with the acceleration rate, the slope of the data on the figure. It is much shallower than the other runs on this engine. This implies a number of possibilities. The module must be controlling the motor differently than the other decoders that have been examined by this author. This is particularly true of this engine series. The other area of interest was at the start. It is clear that in this region the decoder is sequencing through a sound startup program, then it starts the motor with a small spike of power. The unit settles to a crawl to the measuring plane. This is typical of decoders. What is not typical is that over the next 0.5 to 0.6 volts, the same thing happened. With the velocity at the measuring plane the same 1.6 smph for each of these settings. This was repeated three times at each setting. For a total of fifteen readings. One velocity measured 1.5 smph the rest were all 1.6 smph. a very interesting result indeed. I think it likely will yield a more stable run condition for the sound decoder, or any decoder for that matter.

So what to do about the high voltage speed uncertainty? This weekend I pondered over this issue. The simple thing was to figure out how to add some distance to the run. Not very easy, no more than about a foot, which I did. Still indications that the max speed has not been achieved. In addition, it is apparent that the decoder is controlling the power to the motor so the speed at a given distance is the same for all external voltages above the voltage that will peak at that speed.

Looking at the data that I had taken, it struck me that additional data was needed, of course. For the speed impact, the data should be taken at several measuring plane locations from start, to get an indication of the acceleration and how close to the peak velocity the unit is running. In addition, the original testing left some doubt in the maximum draw bar force measurement. For expediency, a gage has been found to give as reliable a measurement and the weight and pull technique. The gage is much faster, so it has been used for the last several hundred tests. Because of the way the gage makes its measurement and the acceleration of this decoder, the gage reading appeared to be low. Here the solution is to go back to the weight and pulley measurement system. In addition, the velocity will also be measured at a given measuring plane. This distance is different from the nominal tests, because the test track is only 44 inches off the floor. This limits the run to that length. Runs are made at increasing weight from 40 grams until a maximum is achieved, defined by the unit not able to complete the distance. In several cases, it started, but then stopped before the measuring plane. This pins down the max draw bar force data.

Because of the questions about the maximum velocity being achieved in the length, it struck me that another option was to increase the grade. It has been on the agenda to look at the impact of increasing the grade for some of these runs. So here is that opportunity with a an additional purpose in mind. So today’s testing pertains to running additional data a the no grade position, then running the measuring plane variation and the draw bar force variation at 2.5 percent and 4 percent grade position. Below are the results of these tests.


These charts are first cuts and a bit rough. However they show some surprises. First, the unit accelerates almost identically over the entire run at the three grades. Yes, the maximum distance speed does slow by 3 SMPH at the steepest grade, but the slope with distance indicates the decoder is still accelerating the motor at all three grades by about the same amount. That surprised me.

The draw bar force results also require some thought. First, the limited run distance leads to a similar result as the voltage-velocity curve. Here the maximum velocity achieved at all three grades is defined by a load that intersects that speed level. Below this level, all three grades have virtually the same speed result. Again, that surprises me.

The maximum load does vary with grade. However, there is more change between no grade to 2.5 percent grade then between 2.5 percent and 4 percent grade. This may be explainable. At no grade, there are two slopes at velocities below the maximum measured. The first is the tractive effort curve. The second is the wheel adhesion curve. The wheels are spinning on this line. At the larger grades, the where’s start to spin before the tractive effort curve is achieved. If the run distance was longer, the tractive effort curve should drop below the adhesion line. That is not the highest draw bar force, so for now it is of only academic interest.

The impact of this is the need for more data. Clearly data needs to be taken between no grade and 2.5 percent grade. It may be illustrative to find the grade where the adhesion and tractive effort curves intersect. An additional grade between that grade and no grade would help describe the playing field.

Additionally, because of these results, additional data of this type needs to be defined for the TCS decoder module and the original light board only options. The acceleration data for these other options is expected to show more difference with grade change, but will it? Will the tractive effort and adhesion curves have the same position? The original data implied differences. Will it be?

As implied, this will take some time to get these data. For now the goal is to get the data before the end of the month.

Edit: added 9-23-16
To further understand these results, I have run several additional tests.

These tests include:
1. Acceleration velocity and current measurements at 12 volts.
a. These were done at several increments along the existing test track.
b. The starting point is always the same. The measurement plane was moved.
2. Draw bar force and velocity measurements.
a. Using a weight and pulley set up.
b. Measuring plane at a fixed location, limited by the height of the test track.
c. Weights were increased until the engine would no longer lift the weight over the distance.
3. Steps 1 & 2 were repeated at 2.5% and 4% grade.
a. The expectation is that the acceleration would be impacted by the grade level.
b. I always wanted to make this examination and this gave me an excuse.

I have spent some time with these results. In this case, I am not picking on anyone. If fact, the TCS decoders are amongst the best that are currently available. My interest is in what they do to the performance. Are they a source of the “power” issues that seem to exist. For now, I’ll let you decide for yourself.



I will be posting more thoughts and info as they become available.

Please be patient with the ads in these posts. To do this work and to report it on this site requires expense. As indicated in an earlier post, I’m attacking it on three fronts. Two different types of ads and direct down loads. In the case of the ads, clocking on the ad helps me. What else you do is up to you. In the case of the downloads, that payment goes to me. The recent engine test reports have been made as downloadable PDF files. The S-4 down load is initially free. In part as a test to see if the downloads will actually work. So if your interested, please take advantage of the free download. Also, what price point do you think the reviews are worth. Hopefully, you response will be objective.

Leave a Reply