What's the ideal track voltage?

[gavino200]

I just realized that my Digikeijs Controller has a variable power setting on the power supply. I was trying to get an answer on the Digikeijs forum as to what it 1should be set at for n scale. Up till now it's been on the lowest setting 15 V. The best answer I could get as to what it "should" be set at is "about 2V above what you want your track voltage to be".

 

I'm a bit confused. I don't really have a "wanted" voltage. I just want my trains to run without either being sluggish or going up in smoke. Can anyone tell me how I can learn what to "want" my track voltage to be?

[Kiha66]

12 to 15V should be good for N scale, I believe that's the NMRA and Digitrax recommended setting.  Per NMRA, compliant decoders should be able to handle up to 22 volts, but I'd be warry about running my trains with pricy decoders at a higher voltage.

[gavino200]

On 2/13/2022 at 12:24 AM, Kiha66 said:

12 to 15V should be good for N scale, I believe that's the NMRA and Digitrax recommended setting.  Per NMRA, compliant decoders should be able to handle up to 22 volts, but I'd be warry about running my trains with pricy decoders at a higher voltage.

 

What would  you think about trying 17V at the Power Supply/ 15V at the track? Is that likely to be problematic?

 

Currently, mine is set at 15V at the PS/ 13 at the track. I can't get an answer from Digikeijs about how it should be set or if there is an ideal Voltage for different gauges.

 

How much should the gauge/scale figure into this?

 

These Decoders can take a wide variety of Voltages then? 12 to maybe 22V. So just increasing it a bit until you get a level where your trains run fine shouldn't be a problem? Just a trial and error thing? Obviously I'd want the lowest level where everything runs optimally. But there isn't any danger working this way, right?

[Madsing]

I’d agree with this statement. Use the lowest voltage that works. Going higher would just generate more heat anyway. Strangely enough I cannot find any voltage specification (min or max) for most of the decoders I own. Some of them are limited to 22 volts, some (from NGDCC) to 16 volts. I would never go that high anyway. The output voltage of the Z21 is programmable (by software). I have set it to 13 volts as I loose about 1 volt in the feedback modules (current detectors) and cables, so I should end up with about 12 volts on the tracks.

That said it’s not easy to measure the actual track voltage anyway, unless you have a dedicated device. Standard multimeters won’t work (even in AC mode).

[gavino200]

I got some interesting answers to my question today on the Digikeijs forum. See below. There's not exactly agreement but some interesting ideas. MemberB is a guy who's also on the iTrain forum where he's a mod. He makes some interesting points. MemberC just points out a minor point where MemberB mis-spoke. MemberD makes a good point about heat. Other than that he doesn't say too much that's interesting. However, I'm interested in his point about USB galvanic isolation. I bought one of those isolators but it was a dud and the USB wouldn't work. How much real danger is there connecting a DCC control station to a computer without galvanic isolation. Nothing has blown up so far.

.

Gavin: I just noticed my Digikeijs PS has a variable voltage setting. I've had it set at 15V. What's the ideal voltage for n-sale? Is there a voltage I shouldn't exceed.

 

MemberB:  Is that 15v on the track or 15v from the SMPS? I run at 17v out of the SMPS which is around 15.5-16v at the track

 

Gavin: Is that for HO or N that you use 17V at the PS. I'm interested in trying 17V at the Power supply. Is this safe to do with N scale? Or if anyone can refer me to a source where I could get an answer that would be great.

 

MemberB: I run my H0e - which are effectively n-gauge chassis with H0 bodies at the 17v input box which gives me ~16v at the track. Absolutely no issues, in fact I am convinced my very stable and reliable running is because I use a higher track voltage.

 

Remember that 12dc peaks at 17v on an unregulated DC system - which many still are and that hasn’t caused anyone any issue - though I am sure that someone will come along and say different 😉

 

MemberC: : Yeah, I will come along and say something different. As an electrical engineer for more than 15 years, This statement makes absolutely no sense. Did you by chance mean that 12V AC peaks at 17V DC? If so, that is true if the AC signal is measured in Root Mean Squared (RMS) voltage.
The power supplies recommended by Digikeijs are not putting out unregulated AC voltage. They are high efficiency regulated AC/DC switching power supplies, and have a constant voltage output from no load to full load, so I fail to understand what you are trying to say.

 

MemberB: Agreed, I got the figures the wrong way around, my bad – it was early and brain hadn’t yet arrived ☹

 

Add to that the reality is that many powers supplies which were supplied with DC ‘train sets’ that somehow gave arise to the ‘12v’ belief were in actual fact powered by 16v AC which meant a peak of ~23v DC in the unregulated designs of the time. This may also be why decoders work up to 24v? As an example my old DC Hammant and Morgan actually supplies an unregulated 16v DC, yet it says it is 12v, whereas my ‘new’ Gaugemaster also states 12v DC on the output label with a 16/18v AC supply – it measures 17v at max throttle and I use it to power the locos when cleaning the wheels. The DC only controllers are used by many, many modellers in the UK running N-Gauge without issue.

 

Given that the motors used in model railways are DC motors and the majority of modellers are DC modellers where motors don’t ‘burn out’ at 16v why should the same motor burn out when driven by DCC when a track voltage of 16v will be less to the motor given the losses within the decoder?

 

MemberD: There is no difference. Personally I would be more concern about excessive heat and mechanical design of the model rather than few volts above the voltage motor rating.
Actually it is the current what may eventually damage the motor not the voltage.

Now in general for the power supplies used in MRR. I would follow these rules:
1) Safety requirements for MRR/electric toys. It does include requirements for isolation, max. voltage, current, output power of the PS, set of built-in protections and mechanical design.
The best bet is to look for the e-lok sign on the PS label. 2 concentric circles are not enough...There are not many of them on the market but it is your and your friends safety.
2) The requirements for the DCC voltage and decoders are described in the NMRA standards / MOROP (NEMxxx)
3) The requirements for DC layouts are described in the MOROP standard (NEMxxx).

Common pitfalls:
- Use of PS for the laptops which do not comply with the requirement ( EN 62115 and power supplies specifically with the standard EN 61558-2-7) for electric toys.
- Use of AC PS (on the secondary side e.g. 16V) without proper protection against the dangerous circuit created by 2 AC power supplies (typically classical transformer) connected together on the secondary side creating the 230V on one of the plugs.
- Connecting the layout to PC (e.g. running MRR SW) via USB without galvanic isolation.

 

 

 

 

[Madsing]

Ok. I only have one thing to add: Good luck 😀

I'll keep my Z21 set to 13 volts for now. No real reason to change.

...

I was wrong, I have something to add (not really related). A few years ago, I noticed that my trains were running faster in some blocks (sections), or (who would have guessed?) slower in other sections. I figured out that this was due to a track voltage difference between sections, just because the supply wires of some sections go through the Digikeijs current sensor and some don't. I contacted Digikeijs support about that and they promptly replied, agreed, and sent me the schematics of a circuit to insert in the  supply path of blocks that are not connected to current sensors. This circuit is made of diodes that create the same voltage drop as the Digikeijs current sensor. It worked and I forgot about it. I can find Digikeijs reply again if someone is interested.

 

Marc

 

[gavino200]

Actually, I'm interested. But zero hurry. I have a lot of new information to take on board right now.

 

Do you use one of those galvanic isolators between your computer and the DCC controller?

Edited February 15, 2022 by gavino200

[inobu]

DCC uses Pulse Width Modulation to drive the motors. This is equivalent to you tapping the front wheel of a bike when it was upside down.

The number of taps was the duty cycle and how hard you tapped it was the voltage.

 

There are different motor sizes with different voltages associated with them. Z and N has its ranges as well as HO, O and G scale. Your track voltage

should be in the range of your Scale which should be driven by the motor specs.

 

Tapping the motor with 12 volts verses tapping it with 15 volts has its effect which is heat related as mentioned. Tapping the motor lighter more times

maybe better than tapping it harder fewer times. 

 

Which comes to the root of your issue. Running at a lower voltage may give you more control of your speed steps as the torque has less of an effect on your forward motion. The gains and losses in the track voltage will have a more adverse effect. The higher base voltage changes the speed table curve in relation to

speed and distance covered over time.

 

iTrain is using a time variable to control the start and stop. If there are voltage drops iTrain has no way of detecting the variance. This effects the stop

location based on time. Dirty track will induce voltage drop which slows the train down. This changes the distance covered over time.

 

3 or 4 motor rotations is a few MM.

 

Inobu

 

 

 

 

 

[Madsing]

Thank you @inobu, very interesting remarks!

[Madsing]

13 minutes ago, gavino200 said:

Do you use one of those galvanic isolators between your computer and the DCC controller?

No. But my "computer" is a small Raspberry Pi sitting next to the Z21 and connected to the same power strip.

[chadbag]

I read the posts on the Digikeijs list as well. One if the members mentioned that most modern decoders allow you to set the maximum output voltage to the motor so no matter the input to the decoder, the output to the motor will be the same (given the input to the decoder is "high" enough to support it).   This makes me think it shouldn't matter as much what you set the incoming to the CS to or the track voltage to if all your decoders support that feature and you have it set to something like 12V.

 

 

[inobu]

4 hours ago, chadbag said:

I read the posts on the Digikeijs list as well. One if the members mentioned that most modern decoders allow you to set the maximum output voltage to the motor so no matter the input to the decoder, the output to the motor will be the same (given the input to the decoder is "high" enough to support it).   This makes me think it shouldn't matter as much what you set the incoming to the CS to or the track voltage to if all your decoders support that feature and you have it set to something like 12V.

 

 

Those responses seems to be off. The CV's that controls the DC motor input from the Orange and Gray wire of the decoder is CV 3 Start, CV6 Mid and CV5 Max.

These values are a percentage of the Max Voltage. This is a duty cycle value of the Pulse Width Modulation I mention before.

 

The decoder will rectify the DCC track voltage to derive its DC voltage. This DC voltage is a derivative of the track voltage and will be slightly lower.

 

The example being a Track voltage of 15 volts will yield a pulse of 13 or 14 volts. CV5 Max Voltage will establish the maximum duty cycle of the

PWM. The voltage does not increase only the frequency of the pulses.

 

This goes back to the bike tire example. As you tap the wheel it spins. The faster you tap the faster the wheel spins. You are not tapping the wheel harder

but tapping it faster. This is the duty cycle.

 

The decoders circuitry will dictate its operating capacity. This is how a N Scale decoder can be used in some HO application. It is the circuitry that

the NMRA standards specifies. People are mixing this up with the operation of the decoder.

 

The higher the applied voltage the more work the motor will generate. The more work the hotter the motor will get. The hotter it gets the faster it will fail.

 

The bottom line is the track voltage should be within the range of the motors used on your layout. The best reference is on Kato's site

https://www.katousa.com/N/Unitrack/accessories.html

 

12v Power supply for N Scale

16v Power Supply for HO Scale.

 

The difference here is the DC Throttle is varying the Track Voltage from 0 to 12v or 0 to 16v to control speed depending on the scale.

With DCC the Track Voltage is the same but the number of pulses or duty cycle is changed to control speed. It is reduced or increased. 

 

Inobu

 

 

 

[gavino200]

The DCC wiki recommends a track voltage of 12-14 for N scale, with disadvantages for going either over or under.

 

Below quoted from DCCwiki

 

If you use lower input voltages, you will notice performance issues with your trains. If track voltage falls below ~10 volts, the DCC signal on the rails becomes unreliable. Also, the booster's short-circuit protection will not work without sufficient current. However, if you were to use higher voltages, the booster will run hotter than necessary, possibly shortening its life in the process. The excess voltage will be lost as heat, heat kills electronics.

 

Track voltages higher than those shown will drastically shorten the life of any lighting effects (functions) installed. As you have probably already figured out, the higher the voltage, the more heat is generated - the higher heat can lead to premature booster or decoder failure, and/or thermal shutdown.

 

Lowering the track voltage provides no benefit. Many believe it is better, yet it only results in more current draw. Which places additional demands on the booster and the decoder. The booster has the advantage of free air for cooling where a decoder does not. If a slower running locomotive is the goal, use the speed table.

Edited February 16, 2022 by gavino200

[inobu]

That doesn't seem to give a definitive answer as it looks like the NMRA answer minus 1 volt. I think I mentioned

years ago the importance of the track laying and wiring. If not don't correctly it can have you chasing ghosts.

 

Inobu 

[gavino200]

Basically track voltage 12-14

Anecdotal evidence of trustworthy people using both 15V and 17V at the PS with no problems.

 

Risks going to far in either direction.

 

Low voltage can cause poor train/locomotive performance.

Low voltage can damage electronics by drawing too much current, overheating components in the process

 

High voltage can lead to boosters overheating

Can lead to system shutdowns during operation

Can lead to premature failure of the booster

 

I feel confident enough about trying both 15V and 17V PS settings on my small practice layout to see empirically if I notice any difference.

[gavino200]

A related question I have is about the "quality" of the DCC signal.

 

I was recently troubleshooting some automation problems on the iTrain forum. The quality of the Digikeijs DR5000 was brought up by a few people as a possible variable. Some suggest that it's a cheap, jack-of-all-trades. A few people suggested that Digikeijs use cheap materials and the equipment can be problematic. Others attest to using it with no problems. I actually didn't buy it for the price (which is good) but because of good reviews and because I like the interface and support.

 

So I asked people for specifics. Some pointed to a cheap USB, and other connectors. The most interesting reply relates to the quality of the power supply. This opinion comes from MemberB in the conversation above, who's a very smart and helpful guy. This is his opinion.

 

"The DCC signal produced by the DR5000 as supplied is poor, looking more like a noisy sine wave than a clean square wave, and this can, and often does, result in poor control. A significant improvement in the DCC output can be achieved by purchasing a decent quality SMPS to replace the DK supplied SMPS"

 

I'm curious about this. I wonder if it relates to a problem I was having. In iTrain some of my train routes were working for my E5 and E6, but others weren't. All were working with two different locos. All trains could be driven manually all round the loop. And the train routes all worked for all locos in demo mode. So I was sure it was a decoder problem.

 

I was wrong. It was a track problem. I hadn't focused much on the track connections as this is a temporary practice layout. I noticed that if I gave the trains a little push with the manual controller they would then execute the train route. So I check the section of track and replaced the unijoiners with new ones, and the problem when away.

 

The trains ran in manual without stalling but they weren't getting commands. The problem was fixed by improving track connections. How could that be? I wondered if my track voltage might be on the low side. And I thought about replacing the PC for the above reason. That's when I noticed it was a variable voltage supply. Anyone have any ideas?

 

When I build my new layout I want to focus on track connections and minimize voltage drop. This page talks about how to measure DCC voltage. I'd start out using a multi meter. It seems that an oscilloscope is the only way to look at the quality of the waveforms themselves. They mention that it's possible to make a very cheap oscilliscope using a cell phone as the display. I think I'll look into this.

[cteno4]

Gavin,

 

inexpensive oscilloscopes are at hand these days. Even can get some kits for under $50. I think it a wise investment for you to look at voltage and waveform, especially with just loose test track, on your test/learning system so you can better diagnose issues and not spend a lot of time running down false leads.

 

jeff

[inobu]

The truth to the matter is some track are good for a few snap cycles. After a while jointers open up and create poor connections.

We are under the allusion that the track can be connected an infinite number of times. That's not true. This is why

you have Kato 24-815. After a while you should replace them.

 

On your final build use new jointers that way you will have new and solid connections.

 

I think I have posted this before but its worth posting again. It starts with track geometry. If the track plan is geometrically sound then

the  connections will be sound. My N Scale layouts that I build has to be exact down to Zero tolerance. 

 

This is a O Scale layout with solid geometry. This layout ended up being 350+ linear feet. It is only powered by 2 power supplies.

Listed to the wheels and Rail there's no clacking. The only clacking is from the switches.

 

The coal car is drawn by gravity only. When the track is sound you will not have to chase ghosts. 

 

By the was this is Atlas 3 Rail Track

 

 

 

15/17 volts can be ran but it has its cause and effect. The motor will run hotter causing the plastics to expand.

This expansion will cause the motor to run rough over time.

 

The internet are full of experts (Me included) that will offer up opinions its up to you to discern the

different between noise and facts. There are always underlying factors that one has to identify and scrutinize.

 

Inobu

 

 

[MangakaRailfan]

I have a hoj layout with a16v controller. Does anyone know of i could run japanese models on that?

[gavino200]

Yes, I understand why the problem would be caused by janky joiners and poor geometry. You convinced me of this before and I haven't forgotten.

 

What I'm puzzled by is why there was sufficient voltage for the train to drive fine, but that it would affect information transmission through the track.

 

The DCCwiki information would suggest that higher voltage would cause overheating in the booster, but that lower voltage would cause overheating in the decoder due to higher current draw. What do you think about that?

Edited February 17, 2022 by gavino200

[gavino200]

4 minutes ago, MangakaRailfan said:

I have a hoj layout with a16v controller. Does anyone know of i could run japanese models on that?

 

For HO the recommended range is 14-16V. So that should be fine.

[Madsing]

2 hours ago, gavino200 said:

A related question I have is about the "quality" of the DCC signal.

Last year, I have built a small DCC control station based on an open source project from France (it’s sitting on my desk and I use it to program decoders). I have initially faced a lot of issues with the quality of its DCC signals, luckily, I have a good oscilloscope/logic analyzer (check www.saleae.com) and was able to debug it. DCC signals can be tricky. Unfortunately I don’t have access to a Digikeijs control station so I cannot comment, but I really doubt that it would be selling so well if its DCC signal looked like a noisy sine wave. 
As @inobu said, I would look at the quality of the connections, not just the tracks but also the locomotives.

[inobu]

1 minute ago, gavino200 said:

Yes, I understand why the problem would be caused by janky joiners and poor geometry.

 

What I'm puzzled by is why there was sufficient voltage for the train to drive fine, but that it would affect information transmission through the track.

 

The DCCwiki information would suggest that higher voltage would cause overheating in the booster, but that lower voltage would cause overheating in the decoder due to higher current draw. What do you think about that?

The track serves two purposes for DCC. Data communication and a power source for the decoder.

 

Data communication comes by the 1's and Zero's created by the Command Station. These 1's and Zero's are voltage duration that formulates the

data packets for the Decoder to read. If there is an interruption/drop out in the transmission stream an incomplete packet will be received by the decoder and it will discard it. 

 

Power is derived from the track as well. The decoder will rectify or extract the voltage from the DCC signal is use it to power itself, the lights and motor. The same drop out

or signal interruption occurs so fast we cannot detect it with our eyes but the data stream will lose a lot of data.

 

Lower voltage will cause the motor to draw more current to make up for the low voltage. This is why you want to dial in the track voltage. The closer you

get to the correct voltage you better off you will be.  The motor will convert the energy to work (Physic) and minimal energy is lost to heat.

 

Inobu 

 

[gavino200]

19 minutes ago, Madsing said:

 I have a good oscilloscope/logic analyzer (check www.saleae.com) and was able to debug it. DCC signals can be tricky. Unfortunately I don’t have access to a Digikeijs control station so I cannot comment, but I really doubt that it would be selling so well if its DCC signal looked like a noisy sine wave.

 

That's really nice! A bit pricy though.

 

Quote

As @inobu said, I would look at the quality of the connections, not just the tracks but also the locomotives.

 

Yes, absolutely. I learned about that from inobu. I've been a believer for a long time. I wasn't paying close attention to connections on this layout as it's just a throuw-together practice layout and there were no dead spots. It's a real eye-opener for me though about signal quality loss.

 

Edited February 17, 2022 by gavino200

[gavino200]

I'll need to read about scopes a lot more before I really know what I should look for. But it looks like it might be an interesting Arduino/Pi project. At very least I'd learn some about scopes and increase my Pi fu at the same time.