Post for Use and Interpretation of TIU Tester Results

Ok, Interesting results.  First the LED can be on Solid, Flash rapidly and brightly, or Flicker (faster flashing I guess, but dimmer brightness).

So here are the results and there seems to be a difference between variable channel output and fixed channel.  These test were done after the repairs I note on each channel, and the MTH new TIU TVS Mod completed.

VAR1 which had no repair has both LED go on solid then Red Flicker and Green Flash before going out.

FIX1 which I replaced the Transmit chip (see above; I should have done a signal strength test before replacing).  Solid no flashing.  Real slight green flicker.  Effectively the same before the chip change.

FIX2 No repair, since I got the same result with as FIX1 above.  Solid with a more pronounced flicker on the green.

VAR2 the original damaged chip, replaced.  Solid then red flashes and no green.  If I dial down the pot from about 1 O'clock to 12-1230, the green will flicker slowly.

The signal strength test was done on all channels with PS-2 test fixture and PS-2 5V board.  All channels returned all 10's.  G

 In the long post, someone, Adrian?, pointed out that the on-board signal test only shows packets not the actual strength.

I posted a video there to show what my TIU does on power up, and different button pushes.

I considered adding a momentary switch to jump the 5 volt to calibrate the unit with as Stan pointed out. I'm not sure what's the best? I should have compared the 5 volt setting to the good TIU channel setting on the pot.  I only did the "known good channel" calibration because I had already mounted the tester inside a case.

https://ogrforum.com/...99#80973530281550499

One thing to consider is make sure the variable channels are configured in fixed mode for the test.

Did you try Stan's calibration method using the 5V supply?

Channels were in fixed mod and I calibrated with my know good TIU.  Regardless of my calibration, these channels are giving different variations on the LED.  Why do some not flash or flicker at all, even the red LED.  Why some flash and some flicker.  Your looking at 2 with chip changes and 2 without and all give solid 10s, and are lighting LEDs.

I am going back later to my test TIU to check the var channels, but seems interesting how the variables respond differently on the tester than the fixed.  I have my pot as far CW to just get the green on with a know good TIU.  I can not go any further CW on any TIU channel I tested without loosing the green LED.

So why a solid Flash on some channels and other a barely visible flicker, is that a lower output?  Must it be a solid bright flash of the green and red?  Why do FIX light both LED solid with no flashing, but seem to work.  G

Use a pure sine wave transformer to feed the TIU, the tester doesn't work properly with any chopped waveform transformer.  I just tested mine using the Z1000 and the Lionel 1033.  Works perfectly on the 1033, all sorts of oddities using the Z1000 (with controller).

In my video, that particular channel feeding #1 fixed on TIU #1 (with a wall wart powering the TIU) is from my Z4000.

I'll test my other TIUs using their power sources. My G scale is DC from the Bridgewerks packs. Will this work on DC?

My calibration was with DC Source, my testing was with Z-4000 at 18VAC.  Calibration seems right.  So with a perfect TIU DCS output what should the output of the leds.

Seems the flash once at power up.  Then both go solid for a second or two, than both flash brightly and rapidly?  Is that normal?  What would red staying solid and green flickering mean?  G

Ok, I went back and tested my Bench TIU.  This was bought new as REV L and DOES NOT have the TVS mod as it was an early TIU.  Channels F2, VAR1 and 2 have never been used, and F1 only used to load Sound Files with a DC power source.

I have a 12VAC power supply when I test TIU that I use in the Aux power port so I can just move the Z-4000 lead to each channel and test.  I did the test with this combo.

IF THE TIU is already powered via the Aux port and you apply power to any of the channels, the LED will IMMEDIATELY start flashing together with the watch dog signal being broadcast.  Bright flashes.  All channels did this including VAR channels set to Fixed Mode.  I then moved the pot CW a small amount and all channels still flashed both, but the Green was dimmer and more like a flicker.  The turn CW was just a nudge.

When I remove the AUX power and just tested F1 with the Z-4000 input, the LED Lite solid for a second or two, than started to Flash together as the WATCH dog signal was activated.  Using a jumper from FIX1 input to power the other channels, they all responded the same way.  You get the LED On solid first, than they flash together.  Seems this makes sense in that as power is first applied you need to get the TIU processor started.  SO you must get a constant DC signal until the processor sends the Watchdog, and once that is completed the LED go out.

Also, it seems that if the Green LED just flicker and is more dim in brightness the Signal level is less than the calibrated standard.

Now, I guess I may add the new TVS mod to Fix2 to see if it has any effect on the output.  G

gunrunnerjohn posted:

Use a pure sine wave transformer to feed the TIU, the tester doesn't work properly with any chopped waveform transformer.  I just tested mine using the Z1000 and the Lionel 1033.  Works perfectly on the 1033, all sorts of oddities using the Z1000 (with controller).

Does the Z4000 produce a pure sine wave?

Engineer-Joe posted:

 In the long post, someone, Adrian?, pointed out that the on-board signal test only shows packets not the actual strength.

 

Yes, I remember reading this as well. Also, pretty sure that Adrian posted earlier in this thread, or one of his other DCS troubleshooting threads. The TIU tester is doing a different test (actual DCS signal strength) than the DCS self test (number of packets sent/received reliably). I believe he also posted about the maximum DCS signal strength with a brand new TIU and the minimum DCS signal strength required for his test layout to perform reliably. This is really about all I know for certain (I think?) and I'm gonna stop here, before I dig this hole too deep...

I'm sure a number of us have pointed out many times that the DCS signal test is simply the ratio of packets received vs. packets sent, not an actual numerical analysis of the signal strength.

This tester is looking at the amplitude of the signal, if calibrated properly it should give a decent indication of a good channels vs a weak or bad channel.

Again, only with a pure sine wave transformer.  I didn't test with the Z4000, but I believe that's close enough to a pure sine wave to do the trick.  With no load at full throttle, it's not exactly a pure sine wave.

Although it didn't sound like it, I was referring to Adrian's DCS problem solving posts (I took Joe's post that way as well?). Never meant to infer the DCS signal test (MTH's) had not been posted around here before, as I do remember reading about it here, in other threads. (I did a little editing above.) It's also in the DCS Bible (aka Barry's Book, DCS Companion).  

The added info and your thoughts on the Z4000 is good info to have. I knew the Z4000 was 'electronic sine wave' as also discussed here before (not actually true sine wave like the older xfmrs), but I had no idea what would be 'close enough' for use with the TIU tester? So it's good to know that too, I was also curious on that one. Thanks for clearing that up too.

Z-4000 is a very good approximation.  Remember it was build to operate PS-1 engines which were designed using Pure AC from older Lionel transformers.  Even some Z-1000 control boxes can have issues with being good approximations to Pure Sine wave. 

Anyway, that is not the issue with using this tester.  I figured out why the tester displays a different pattern on a good TIU.  The real issue is anomaly results when trying to repair a damaged TIU.  There are other components beyond the typical Transmit chip, and TVS diode that can have issues. 

Also, remember this is only a transmit tester.  There is a receiver chip that matters too for real control of a DCS engine.  The good news is that it doesn't fail much, especially with REV L TIUs.  G

Is there any way to clean up the sign wave if a guy doesn't have a full wave transformer?

No "sure fire" way, but one way that helps with PS/1 engines is dropping a 47uf NP cap across the outputs.  Unfortunately, that also affect the DCS signal, so that's probably not going to work in this case.

Adrian had floated the idea of changing the filter variables, but since I don't have a blank board to populate and tinker with that option, I can't say how well that would work.

Thanks GRJ. I dug through my boxes of stuff and found 3 old Lionel transformers. Hopefully one of them still works. Would all of these put out a full sine wave?

Pretty sure all of them put out a pure sine wave, so use any one of them.  I used the one on the left for bench testing, the 1033.

Thanks again, John.

Hopefully I will get some time this weekend to build mine and see how it fares.  Thank you for all the comments.

What is the CORRECT testing protocol?  I made up this one.

Before test, use Remote to set TIU Var Outputs to Fixed.

Signal Test Conditions (no remote involved):

Power from rebuilt postwar ZW -
  TIU Aux powered with ZW A post at 17.5 V AC
  TIU Inputs individually powered with ZW D post as tests proceed through TIU Outputs
ZW is powered on and off with plug strip switch.

Test sequence (test for watchdog signal)
  Check Aux Input = 17.5 V AC (leave plugged in for tests)
  Check TIU Inputs = 17 V AC  (move Input for each Output test)
For each output -
  With Power off:
 Attach Input power, Power On, Check Output V at TIU  (My meter is a Fluke 113 Multimeter that starts looking for DC and switches to AC when AC detected.)  I detected 1 DC voltage output put of 16 outputs.
  Sw off Power, attach test leads to Output, Power on while watching tester LEDs.

Test Results have me bumfuzzled (Do results make sense?).

Rev L TIUs with USB:
  #1  Purchased May 2016 and used as Main TIU, repaired Sept 2018 and used to calibrate tester.
      At Power Up,  Var Outputs Blu and Red light at same time then Red flutters and goes off.
      At Power UP, Fixed Outputs Blu on then Red on with 2 slow Red flashes
.
  #2  Purchased August 2018 and used as Main TIU while above TIU was repaired.  Quit shortly after put in service.
      At test power Up, all Outputs Blu on, then Red and Grn give 1 Fast Blink and go off.
    Just for grins I opened TIU case and reseated the loose transceiver board.  No Joy.  All output voltages are normal but no signal.  Anything else I should look at?

  #3  Purchased Feb 2018 and used much less than May 2016 TIU
      At Power Up,  Var Outputs Blu and Red light at same time then Red flutters and goes off.
      At Power UP, Fixed Output 1  Blu and Red on then 1 Slow Red flash.
      At Power UP, Fixed Output 2  Blu and Red on, Red slow flashes then flutters.

Rev.13A with no USB saved for spare
  #4  Purchased Jan 2009
      At Power Up,  Var Outputs Blu then Red light on then Red flutters and goes off.
      At Power UP, Fixed Out 1 Blu on then Red on, 1 slow Red.
      At Power UP, Fixed Out 2 Blu on then Red on, 1 slow Red then Red blip.

If the tester is calibrated correctly you should get Green and Red Lite and fluttering.  Red only means your getting a marginal signal out.  Or the tester is calibrated to a too high level.  G

Thanks for the feedback G.   Since I have no scope I calibrated my tester using stan2004's Calibration Video.  I couldn't tell what to connect my tester to for calibration so I connected it to a TIU Output that had a Input of 17-18 volts AC.

stan2004, was that the correct power source for the calibration?

As I noted in the topic, "Design of a $10-20 DCS-TIU Port Tester Tool?", I cheated on my calibration and backed off ccw enough so the tester would cause the red and green LEDs to slow flash twice on my best Output.

Susan Deats posted:

As I noted in the topic, "Design of a $10-20 DCS-TIU Port Tester Tool?", I cheated on my calibration and backed off ccw enough so the tester would cause the red and green LEDs to slow flash twice on my best Output.

Here are the Test Results of my six TIUs after re-calibrating the tester.  I'm also sharing a blank Excel Spread sheet for your use.  My testing method only indicates the presence of the startup watchdog signal.

Here is my "bench test" setup.

Comments and/or questions would be appreciated.  All my TIUs under warranty appear to have the Tiny TVSs.