Product for use with a diode matrix?

[haasmarc]

Thanks, Pete.  My thought is this.  As I understand the situation, the OP has an installed base of a bunch of those boards working well already, and was originally looking for something that performed an "equivalent function".  So if you *have* to make something, why not just make more of those?  I bet those boards are simple enough that all we need is a close-up photo of each side (or one in hand that actually works would be awesome) to draw up the PCB plan (I'm a big fan of kicad), and then have them made.  My experiences with Aisler is that for a small board like this, it should only cost about $4 each or less, and of course, if you go the route of a China maker, it will be literally less than dollar a board.

I will volunteer to draw up the PCB if somebody can get me a board or the photos.

Like I said near the beginning of this thread, I have the actual Tractronics boards he is using if he wants more of the same.....   Some of the boards for Tractronics products were printed in Mainline Modeler Magazine if you have those issues.

Should I use my invisibility for good or evil....

[mmagliaro]

Like I said near the beginning of this thread, I have the actual Tractronics boards he is using if he wants more of the same.....   Some of the boards for Tractronics products were printed in Mainline Modeler Magazine if you have those issues.

Should I use my invisibility for good or evil....

I did see that haasmarc, sorry.  I guess I assumed the OP wanted a solution that couldn't still one day "run out".
All it would really take is somebody to copy one of those boards into a new PCB layout, so they could be manufactured at will by anybody.  I volunteered to do that, and I'd even make up one of them, to be sure it works.   I know... electronics scares people.  Honest folks, once you have the PCB layout file, you can upload it on a web page, and like magic, beautifully finished PCB boards show up at your door.  You just have to be willing to solder some parts on them.

[jbonkowski]

Blank circuit boards and full kits for the Stall Motor drivers John pointed to earlier are available at http://www.circuitous.ca/556Stall08.html

You are mistaken. See the circuitous.ca main page. He shut down the business in July.

Jim

[jbonkowski]

Thanks, Pete.  My thought is this.  As I understand the situation, the OP has an installed base of a bunch of those boards working well already, and was originally looking for something that performed an "equivalent function".  So if you *have* to make something, why not just make more of those?

Actually, the Tractronics SwitchLock boards don't work THAT well. The ICs on them tend to die (which is why they are in sockets, to make them easy to replace). I don't know if this is due to the IC choice or if our layout is over-stressing them somehow. So I was hoping to try something different.

I wanted to see if there was any commercial product out there that does the job and isn't too expensive, and it appears there is not. I've now done a PCB drawing with my own solution to see if I can make something better. I still need to select the connectors/cabling.

While I'm at it, I'm going to do a diode matrix PCB, basically a "perfboard" but with some wiring already built in and places for connectors.

Jim

[peteski]

You are mistaken. See the circuitous.ca main page. He shut down the business in July.

Jim

Here I go again - not  reading things carefully enough. 

Perfboard is ideal  for home brew circuits like a  diode matrix.

I would be curious to see a schematic of the SwitchLock board to see why those components blow up.  I was under the impression that you were hoping to expend your current setup using the same circuitry already being used by you.  Sorry that I misunderstood.

[mmagliaro]

...
...
I would be curious to see a schematic of the SwitchLock board to see why those components blow up.
...
...

ME TOO!  My thought exactly.  Maybe that inductive kickback from the Tortoise motors is causing this. 

[jbonkowski]

ME TOO!  My thought exactly.  Maybe that inductive kickback from the Tortoise motors is causing this.

[ Guests cannot view attachments ]
The Circuitron SwitchLock uses a flip-flop made from two NAND gates with the output fed to an op amp (x4 on the chip).

The op amp has a maximum rating of 570 mW and has 4 amps on one chip, so can drive four Tortoises at once. If each Tortoise is 16 mA at stall current and 12V, that's 768 mW. The instructions for the SwitchLock has a lot of detail, but they don't state their stall current assumption, so maybe they underestimated? One the other hand, they sold these for a long time, so the problem could easily be at my end.

I'm going to try a circuit made from LM556 timers. They can handle supply voltages up to 18V. I'm hoping they are harder to kill.

In the end, I'm really trying something else for the experience. It only costs a couple of dollars to replace all of the ICs on a SwitchLock, so the cheapest thing is just replace them when they fail (which is only one every couple of years).

Jim

[peteski]

While it has been a while since I used LM324 Quad Op-amp in a circuit, or looked at its specs, going by my experience I would not have considered it for directly driving a DC motor (in a Tortoise).  It is a low power analog amplifier, not designed as a motor driver.

I just did look up LM 324 specs, and while it can source 40mA and sink 20mA, it is not really best choice to drive an inductive load (a DC motor).  I guess that might explain why it gets damaged.  You might minimize the failures by adding a 0.1uF capacitor across the motor leads. That was mentioned earlier in this thread discussing the high voltage spikes. Also lowering the power supply voltage (thus reducing current), would likely help in reducing the failure rate of that chip, but will slow down the motor's speed.

[jbonkowski]

While it has been a while since I used LM324 Quad Op-amp in a circuit, or looked at its specs, going by my experience I would not have considered it for directly driving a DC motor (in a Tortoise).  It is a low power analog amplifier, not designed as a motor driver.

The owner of the circuitous.ca website agrees, if you dig enough. Here is the last note he has about his stall motor control boards:

The 556 timer was chosen for these circuits because OPAMPS such as the LM324 or LM358 were found to be too unstable if the inputs are allowed to float and have a much lower current capacity.

Jim

[peteski]

The owner of the circuitous.ca website agrees, if you dig enough. Here is the last note he has about his stall motor control boards:

The 556 timer was chosen for these circuits because OPAMPS such as the LM324 or LM358 were found to be too unstable if the inputs are allowed to float and have a much lower current capacity.

Jim

Good to know. Thanks Jim!  I'm also a bit leery of using outputs of the 55X timers to directly drive the motor, but it seems that it has been done and it works.  At least the timer's outputs are designed for full-on, full-off states.

I am also curious about those high voltage spikes Rasputen showed earlier.  Next time I get my hands on a Tortoise (when my friend returns from an extended vacation) I'll have to connect my scope to it and see what surprises I discover.

[mmagliaro]

Golly... they are driving the turnout motor directly from the opamp ?   For lack of a better word, ... EWWWWW.
I don't know how game you are, but it should be possible to just tack small driver transistors onto the outputs of the opamps, and then drive the motors off of those.   Heck, little transistors are beyond cheap and could handle this with ease.  They could be patched onto the board itself so there would be no mess.

Something like this:


When Vin is +,  the stall motor (Rsw) will have +11 ... GND across it. When Vin is -, it will have -11...GND across it.
Each transistor only has to dissipate about 20 mW, and the load on the opamp is something like only 200uA.
So if you drove 4 Tortoises, we're talking 80mW on a transistor that can easily dissipate double that, and still less than 1 mA load on the op amp.

Yes, it means 2 transistors and a resistor on each opamp output, but the cost is minimal.