Pacific Stay alive fitment without major surgery

[John-357061]

Has anyone thought about 3d printing a tool to do this?

take something like https://cults3d.com/en/3d-model/tool/ho-train-wheel-gauge and scale down. Probably needs to be a resin printer to get the accuracy.

Will try later.

[Rallymatt]

if you have the tech to do it it’s an interesting possibility. I would want to put a chamfer in to ease the dimension to the one we are looking for. I’m not sure why DCC Concepts went for the minimum standard and not the Target? 3D printing has really transformed the hobby in so many ways 😁

[John-357061]

I do. Cost 52p. Trying now. The Z size was 14.4mm for HO. Changed to 10.45mm.

As the author states for FDM printers you need to experiment to get the actual size.

[John-357061]

Printed on both FDM and resin printer. Does seem to work.

[Rallymatt]

Good work John, I think to fine tune this, file a small chamfer into the two parts either side of the opening, done to say 10.1mm then pushing this tool in should open up the wheel set quite easily. Have you tried it out yourself?

[John-357061]

I did try it, you can push it in, first at a slight angle. Although could be designed to work better. Hopefully will fix the issue I was having with one of the locomotives, every so often would stall or once the front boggy derailed.

[Les-345151]

Worth adding that a 10.5mm steel slip gauge can be used for re-gauging, preferably with slightly chamfered edges. Particularly useful where access is limited.

[Rallymatt]

Good suggestion Les 👍

[ateshci]

Hi all,

here's the stay-alive I fitted into my "Night Hawk".

First of all, the recess housing the sugar-cube came quite handy to take two 0.3F/2.7V supercaps after the speaker's removal. I use the speaker that comes with the HM7000. See first picture.

Buffering is done via an ultra-small ( 9x9mm² ) pcb a friend of mine makes. It comprises both a step-down ( for charging ) and a step-up ( for buffering ) converter. See picture 2. The delivered backup voltage is 10.5V out of an effective capacity of 0.15F/5V. This is enough for ~ 2secs buffering time with bells and whistles. As a side effect, the decoder is no longer prone to jerky behaviour when in BT mode. The only butchering necessary was to remove some material inside the tender shell in order to accommodate the speaker which is held in place by the HM7000's rear antenna extension. I have not yet noted any detrimental effects to RF performance.

Pic1:

Pic2:

The step-up's output are the green and brown wires that are soldered to the decoder's V+ and GND tabs, respectively. Cellotape is used to insulate the components from the chassis..

Regards, H.

[96RAF]

On 24/09/2023 at 01:39, ateshci said:

Hi all,

here's the stay-alive I fitted into my "Night Hawk".

First of all, the recess housing the sugar-cube came quite handy to take two 0.3F/2.7V supercaps after the speaker's removal. I use the speaker that comes with the HM7000. See first picture.

Buffering is done via an ultra-small ( 9x9mm² ) pcb a friend of mine makes. It comprises both a step-down ( for charging ) and a step-up ( for buffering ) converter. See picture 2. The delivered backup voltage is 10.5V out of an effective capacity of 0.15F/5V. This is enough for ~ 2secs buffering time with bells and whistles. As a side effect, the decoder is no longer prone to jerky behaviour when in BT mode. The only butchering necessary was to remove some material inside the tender shell in order to accommodate the speaker which is held in place by the HM7000's rear antenna extension. I have not yet noted any detrimental effects to RF performance.

Pic1:

Pic2:

 

The step-up's output are the green and brown wires that are soldered to the decoder's V+ and GND tabs, respectively. Cellotape is used to insulate the components from the chassis..

Regards, H.

The HM7K decoder has a built in time lag of 1.5 - 2 seconds before charging starts. The decoder MCU triggers this event, which is specially introduced to preclude the programming inrush current problems. The external components in the PB simply react to that trigger and the associated component values control the charging rte and of course the discharge rate upon seeing a dead track. No more difficult or easy than that. The PB has 3 x 1F caps in series each rated at 2.7v.

[ateshci]

So it's a rather bulky affair that charges to 8V. The rather simple regulating components ( one transistor, one or two resistors and a zener plus a reverse bypass diode ) are in the capacitor pack. The decoder just has a delayed switch to prevent the charging current from giving wrong acknowledge signals when programming.
The usable voltage range is maximally just 3V , the decoder will not operate at less than 5V inner system voltage. So one can use <1.49 Wsec out of 10,56 Wsec that are stored in the capacitors. Take one of the step-up units, it could be hooked up the same way as Hornby's ( provided you find that Hirose or JST connector with wires already attached ) and you have 0.47F @5V, the usable voltage being somewhere in the 3V range again, and you get a constant(!) output of 9V..11V. The available energy is 2.1 Wsec out of 5.9 Sec at a space requirement of roughly 2/3. Even the unit that I use is more than sufficient to compensate for contact loss and it provides 10,5V for  ~ 2secs. There are 0.7 Wsec out of 1.9 Wsec to be used. The one I use is small enough to find room even in the Arnold/Hornby HN9062 Koef, I have successfully installed more than one.

 

[ateshci]

I think I found what type of plug is used for the stay-alive socket : MOLEX 501330 0300. If I could only find one with wires already attached...

Edited March 28 by ateshci

[96RAF]

1 hour ago, ateshci said:

I think I found what type of plug is used for the stay-alive socket : MOLEX 501330 0300. If I could only find one with wires already attached...

They are JST. I found a supplier in China that would make them and speaker cables up but even for an order of 100s of each they worked out at £2 each wired.

Hence SUHR - 2V or 3V.

[ateshci]

The decoder's pcb has two pads( under the receptacle above ) labeled C+ and C- where one can attach a stay-alive with AWG 26 (0.14mm²) wire. This is an easy-to-find size.

Edited March 28 by ateshci

[Too Tall]

I got some TCS KA-N1 stay alives a while back, I have yet to fit one, but the small size and  listed run time made me think they could be ideal for this, and possibly any small tank locos we may get in future, and here is a picture next to my index finger, actual dimension also on the attached image above the circled text :

 

 

[ateshci]

The photo shows one(!) SMD electrolytic capacitor rated 100µF/10V on a small pcb that contains the elements I have already shown in a previous post. Don't expect much of it - 100µF can store a charge of C=0.01Cb. Half of that can be used, because you start from 10V and your decoder will stop at ~5V. So that's 0.005Cb, which gives you a buffering time of t = C/I. Running light, your loco may draw 50mA, so the time t will be 0.01 sec and will be reduced to 0.003 sec under 150mA. Admittedly, in most cases it is just enough to overcome that loss of contact, but will be insufficient when crossing insulated point frogs at shunting speed. The price asked for it is - well compare with other offers. There are other makes around that start from 16V ( yes, there are SMD caps rated 330µF@16V ( e.g. Vishay 594D337X9016R2T, 7x7x3 mm³) and provide space for up to four in parallel -these are the ones that really just make sense in terms of ease of installing and space requirement. Anything in capacity <1000µF only provides marginal improvement and is wasted money IMHO.

Edited March 31 by ateshci

[ateshci]

Sorry for this one, but the editing time is just too short.
When you take the quoted time of 1..2secs and the capacitor shown on the photo, you end up at an allowable maximal current consumption of motor and decoder of 5mA for a moving loco! So this claim is ...take your choice.

 

[ateshci]

Yet the final post:

This is the charging circuit used: 

 

[Too Tall]

Like I say, as yet untested but bought with the intention of using on tender locos where its more a case of that milisecond interuption causing the BT decoder to reset.

Given my initial fitment of the 2x220 tantalums has proven to be reliable, even though their is no noticeable run time in them. I saw this claiming to have 1-2 secs run time on an N gauge with lights, and in a tiny package, so ordered a couple to try. If it proves no good, no sweat, I will chalk it up to experience 🙂