The club has a bunch of old 28 speed step decoders that I found in a few locos and I got some "new" ones from the club. The one page data sheet for these decoders is pretty sparse. It says "System One" at the bottom of the page which means that it was supplied by Wangrow (out of business) and probably manufactured by NCE as Wangrow didn't make anything themselves and NCE made most of the stuff that they sold. The Wangrow "Basic" DH-155 decoder definition in DecoderPro seems to be able to handle these decoders.
I couldn't find anything currently on the net with a Google search so I scoured through the WayBack Machine to find wangrow.com. It was in the archive. I picked some dates around the year 2000 and dove through the archive of the SystemOne site. I found a SystemOne manual and on pages 2-43 and 2-44 I found a better data sheet for this decoder. It really does appear to be a DH155.
These aren't the greatest decoders, Stanley Ames (a DCC old-timer and on the committee that wrote the DCC specs for the NMRA) asked me how much I paid for these decoders. I told him that they were free to me. He said I paid too much.
I've gone through 8 of them so far:
I had a lot of trouble figuring out how to make these work, primarily because I was working with ones that didn't appear to work due to programming track issues. When I stumbled across a "good" one, it at least responded on the DCS100 programming track as expected and even though it's not a great decoder, it did work.
This is a 28 step decoder so to make it work with momentum and a speed table with a Digitrax command station, the address needs to be "status edited" to tell the command station how to talk to it. It will also respond in 128 step mode but only in a "direct drive" mode. When commanded in 128 step mode any programmed acceleration, deceleration, Vstart, and the speed table will be ignored.
It should respond on a programming track in page mode without any fiddling with the command station. The 4 good ones do that, the other four do not. The "bad" ones all report 255, NoPg or some arbitrary value on the DCS100 programming track.
The decoder supports 5 functions, but they appear to be just on or off. Even the headlight is non-directional. They do support 4 digit addresses, analog conversion, and at least Vmin, when running at 28 steps. Vmid and Vmax are not supported. Speed matching requires the use of the 28 step speed table. The input current rating is 1.8 amps, 1.5 available for the motor, the balance to functions. I have not stressed them yet beyond about an amp. The motor driver uses a low frequency PWM scheme so that the motors hum some, especially at low speeds. They don't respect the NDOT bit in CV29 nor do they have any attributes to any lighting function. They do have a 28 step speed table.
Neither the manufacturer's ID or a version are encoded into these decoders, they will return 0 in CV7 and CV8 so that you have to explicitly pick the type in DecoderPro to enter the Service or OPS mode programming functions. They do appear to respond to limited OPS mode programming.
I was using a locomotive as a "test fixture" but sometimes I had them all buttoned up and it was a pain to open one just to test a decoder. So I built this test fixture from a board removed from ATSF 2790. The "motor" is a 39 ohm resistor, I can observe the voltage on this resistor with an oscilloscope if necessary. The three LEDs represent the 3 functions wired to the 8 pin plug. This makes it easy to test an 8 pin decoder by connecting the red and black wires to a programming or running track with clip leads.
The four remaining decoders that I had (the club has more) characteristically respond to the request for the value of ANY CV with 255 (an illegal value in some CVs such as the 2 digit address) or with a NoPg indication on the throttle. This is on a Digitrax DSC100 programming track. One of them is sensitive to the programming track voltage. At higher programming voltages, it will respond with values, but never the same one twice in a row. A Digitrax DH123 plugged into the same fixture responds appropriately and it lights the lights appropriately when connected to a running track. The four remaining Wangrow decoders did not respond to the running track so I assumed that they were toast.
However, these four DO respond to the SPROG on it's programming track. The SPROG has more programming power so apparently these decoders are touchy but they do function when properly programmed in the proper way. The SPROG has enough soup to actually cause a loco to bump along as it draws current spikes to signal the command station. It also has enough soup to read a Tsunami. I did try one hard to program loco on the club's programming track and it did not respond to Service Mode programming.
At least one of the touchy ones (the one in the PB1) also responded to the programming track driven by a PR3 and JMRI on a Macintosh. It also bumped the loco along some, but not as forcefully as the SPROG. The PR3 would not read back a Tsunami so it appears that if a programming track can read a Tsunami, it can read these Wangrow decoders as well.
The decoders are wired a little funny too. There is no reversing headlight so that there is no F0(R) function, the yellow wire. Wired onto that pin is F1, the green wire. F2, the violet wire is connected to pin 3 which is optional. Many decoders come wired with only F0(F) and F0(R) connected to the plug with pin 3 unused. Some come wired with F1 (green) connected to pin 3.
In the NMRA design, if the plug is inserted backwards, the red/black pair are reversed, this ok for the track contacts as DCC doesn't care about track polarity unless the loco is analog converted. The orange/gray pair are also reversed, this makes the motor run backwards. However, when analog converted with both the motor and track reversed, the motor will run in the correct direction. With the standard NMRA wiring and the plug backwards, the blue wire would be connected to a socket that doesn't go anywhere and the the headlights would be reversed, but since there is no power for them they won't work anyway. No harm, no foul, reverse the plug and all gets well.
The problem is that, in one instance, I found an "NMRA" socket wired in a non-standard configuration that can be dangerous. Life-Like had jumpered pins 3 and 7. This was probably a feature in that it connects the blue wire to TWO pins. When there is no function connected to pin 3, the socket jumpers the blue wire (+ common) to pin 3 and the lights will work, but they will be reversed too but they will function. The problem is that if a decoder wires a function pin to pin 3 AND the plug is plugged in backwards, the connection will place + Common on that function. As soon as the function is turned on, it tries to short the + Common to ground and it will probably fail instantly.
I have started to check the continuity of all Life-Like sockets by measuring the resistance between diagonally opposite pairs in the center four pins. If I find continuity, something has to get fixed. Either the cross connection needs to be cut on the PWB or the wire to pin 3 should be clipped off the plug.
I also got some Digitrax decoders from the club, a DH140 went into the B&O RDC and a DN142 went into Amtrak 346. I put a DZ123 into GL&W 56. The club also has some small MRC decoders. I have not been overly happy with MRC decoders in large scale. We'll see how these work in HO.
I picked up some more decoders from the club stock, all old ones. There are:
There are a total of seven locos left to convert. The GP35 will get the Wangrow SystemOne decoder that is in UP 1614b now. Then the two UP FA/FB units will probably get the MRC350 decoders as at least there are some attributes to the lighting and the FA could use something It'll be easier to speed match the pair if they use the same decoder. The MRC332 will go into the SW12 provided that the owner of the loco gives the club permission to modify the loco. The three remaining Wangrow decoders will go into the GP38s. Amtrak 227 needs a 1 amp or better N scale decoder and the club doesn't have one now.
I used one of the Wangrow decoders in SP 4823 and another died in that loco during testing. The one that was in BNSF 8630 was toast as delivered. It needs another kind of 8 pin decoder as even a "good" Wangrow decoder won't run in that loco.
I've consumed the three MRC decoders, one Wangrow System One decoder was used to replace one that failed. The other two have been tossed at the request of the club. Lloyd has ordered 10 TCS TX4 decoders. When these arrive the board will determine which locos that already have Wangrow decoders in them will get "upgraded" so I or somebody else will get those back for more work. It should be easy work. A few of the locos have the 8 pin socket so it will be a simple matter of unplugging an old decoder and plugging in a new one. The majority of the older locos were wired so then it will be a slightly more complex cut and splice operation matching wire colors EXCEPT that the Wangrow green wire will go to the TCS yellow wire.
The club has decided NOT to certify any locomotive with a Wangrow decoder in it. Instead, these locos will be tested uncertified to see if the loco is really useful. If so, then the Wangrow will be pulled and replaced with a better decoder. The Wangrow decoders are "odd" in that they are difficult to program on the club's programming track and the users must remember to status edit the loco's address to 28 speed steps or any speed matching that was done will not be enabled. These "features" will cause problems in the future. The decoders were donated to the club so that they have no value other than as test decoders. If a loco eats one, no great loss. That makes it clear that the club should not spend any more effort or money on that loco. If the loco doesn't perform well enough during testing to deserve a decoder upgrade, then the Wangrow decoder can be pulled and used for testing in other locos.
After testing every loco that had a Wangrow decoder in it at the club last night, a pattern emerged. Every one had intermittent motor operation. It appears that even slight power dropouts that will not bother other decoders cause the Wangrow to stop and they take some time for the motor driver to wake up. This can cause loco to sit on the track with the headlight burning (it comes back on immediately) but it takes a few hundred milliseconds or so for the motor to start running again. This makes the locos balky and some can appear to stall with the headlight burning or flickering. This behavior is the death knell for these decoders, it makes locos essentially unusable.
As a test, I did substitute a DH123 for the Wangrow decoder in SP 3343, the Trainmaster. All of that weird behavior went away. The LED headlights did flicker a little in spots, but the loco kept moving. The Wangrow decoders truly are junk and should be simply discarded.
The club supplied some more Digitrax decoders, these came from Lloyd's personal stock. There are 6 DH121 (all wired), a DZ143 (wired) and a couple of DZ120 wired decoders. The DZ143 is current production, the others are quite old. The old ones are 28 step, have low motor drive frequency and do not have BEMF. The DZ143 is a silent drive, 128 step BEMF decoder. As soon as I got them I proceeded to install them in locos that had Wangrow decoders so that we could actually properly test the locos at the club. The two older Z scale decoders will be held back for install in space constrained locos. The DZ143 went into Amtrak 227 which did not have a decoder installed already.
MRC Decoders. The three MRC decoders that I have used (one is actually in a certified loco, UP 1614b) have proven to be more trouble than they are worth. There is no DecoderPro decoder definition for these decoders even though they are actually current production units. This makes them not fit well into the club's decoder documentation scheme. Further, they have very limited speed matching capability. UP 1614b was speed matched, but with a value of 20 or so in CV5. This is crazy because reducing it any further causes the loco to simply not run. I was not able to get UP x1614 to speed match at all. By the time it almost got slow enough, it stopped entirely. I have pulled two of them. The one in BN 248 got replaced with a current production DZ143. The one in UP x1614 got replaced with a DH121 test decoder but it needs a 4 function decoder instead, such as a DH163. I propose also to replace the remaining MRC decoder in UP 1614b with a more modern decoder. A DH123 would probably work fine as we'd turn off the BEMF in the B unit anyway if the decoder had it. However, the three MRC decoders CAN be of service as temporary test decoders as they can allow testing on the LAMRS layout to verify the other performance characteristics of a newly converted loco. If the loco eats one, no big deal.
A club member wants to use one as a function only decoder in a dummy unit. An AD350 might work here as it does have headlights, a MARS light, ditch lights and a strobe. The link is to a manual for the AD350.
Last night I picked up 10 TCS T4X decoders and 3 TCS-A4X decoders. These are electrically pretty much the same decoder. The T4X is a 9 pin JST version. The A4X is an Atlas "drop in" decoder which fits some Kato and other locos as well. I don't much care for the drop in designs, they are quite a bit of work to install.
Overall, the TCS decoder is a good one. The BEMF works very well. Most locos crawl along at speed step 1 without having to mess with any programming at all. I had no difficulty at all with any of them. There are good DecoderPro decoder definitions for them and for what little programming I did, they responded as claimed.
SFX064D. The SFX064D is a sound/function decoder, it has no motor driver. This is a very basic sound decoder that comes with a selectable GP38 or generic steam sound program. I've used these before in large scale and they don't make enough volume for use outside, they are marginal inside. The new SFX0416D with more memory for a larger sound program makes even less volume, it is hardly suitable for HO. The older SFX064D decoders came with 32Ω 1" speakers. The newer ones come with 8Ω 1" speakers. To my ear, the one with the 32 ohm speaker is a little louder. These decoders DO NOT respond to a programming track without a load on a function output so that you cannot read their contents without a load. They will ACCEPT programming though on the programming track. This is done by design so that the decoder will not interfere with a motor decoder wired in parallel. Overall, I found it easier to use the dual address trick (see below) and program them in OPS mode.
The sound file is quite basic in 8 bit sound. The chuff cannot properly keep up with high chuff rates so it is best to run them below 4 chuffs/turn to keep the rate down. The chuff can be set with a cam or by DCC speed control which actually works fairly well. The whistle is a little hard to "play" but if you get the timing just right on the F2 button, it will play a short whistle sound. You actually have to hold the button down a little longer than you might expect to get it to play a short toot. Note that there is a "secret" CV to set the chuff rate that is not in the instruction booklet. See the SFX064D page on the Digitrax web site for the details.
[ Top ]
At the business meeting a few days ago, the club authorized the purchase of 11 decoders for the locos that were currently approved for a decoder upgrade. These are:
I selected DH163 decoders in appropriate configurations for each loco as this particular type of decoder works extremely well with full function control and BEMF. The average cost was $25 each. The RS1 gets a DH165AO decoder as the DH163AO has been superseded.
For the last batch, we used TCS T4X decoders. These have proven to be entirely adequate where they were used. I selected the Digitrax version this time mainly based on cost issues and that the TCS decoder didn't come with a short harness at all. The Digitrax versions were overall slightly less expensive. On many of these locos, the short harness will fit better. One needs the 3" harness.
The 8 pin harness itself is a relatively expensive part of the decoder. It resulted in a $5 increase in cost over the 7" wired harness. This is because the 7" wired harness can be automatically assembled from a wire loom. The 1" and 3" versions add the $5 because they require hand soldering at the 8 pin connector.
Installation will be easy as each loco has a decoder in it already. For 10 of them, it will be an unplug/plugin job, either at an 8 pin medium NMRA connector or at the 9 pin JST connector at the decoder itself. The RS1 uses a drop in decoder so that will involve moving several wires from the board that is in the loco already to the decoder board.
It's been a year since the last entry and the DCC conversion of all the acceptable LAMRS locos is complete. We've used mostly Digitrax decoders with a mix of TCS. The only real differentiator is that we could get a slightly better price on Digitrax. The other difference of note is the color of the plastic wrapper on the decoder.
This page has been accessed times since Dec 2, 2010.
© 2009-2011 George Schreyer
Created 24 Dec 09
Last Updated November 10, 2011