MERG Servo point control

[96701]

To continue a discussion on MERG Servos from SI modelling thread, does it make a huge difference? I only ask because this is the third one that I have built, and they seem to work okay as they are.

[Pannier Tank]

Hi Phil,

 

In the other thread I said I recently built 3 of CAMIO SVOs but I did the mod of replacing the 16 MHz Resonator with a 16 MHz Crystal and 2 Caps. The mod kit is available from http://rmeuk.com/buycbuspcbs.html at bottom of page.

 

The reason for the modification was to overcome some issues with the 16 MHz Resonator that affected some CANMIO's but not all. As a precaution I built the 3 CANMIO's using the 16 MHz Crystal.

 

 

 

[Ian Morgan]

From the early days of the CANMIO-SVO, some users started reporting that the firmware suddenly got corrupted after a while. It took a lot of investigation and experimentation, but it was eventually discovered that, on powering down, as the supply voltage dropped (which is not instantaneous), the resonator could stop working (keeping the clock frequency stable) before the processor stopped carrying out its instructions. If this happened at the precise moment the processor was writing data to its EPROM, it could corrupt the data. On next powering up, the program found gobbledigook in its data and would not work. This problem has, so far, only been identified with the combination of PIC processor and resonator used in the CANMIO-SVO.

 

As can be seen, this precise combination of events, and timing, may only happen very infrequently, and many users may never see the problem occur, but it could be guaranteed to happen at the most inconvenient moment.

 

Fortunately, two alternative cheap fixes have been identified:

1. Replace the resonator with a crystal and pair of capacitors. Not so long ago, crystals were much more expensive than resonators, which is why they are used in many Merg kits. Now, the price of crystals has dropped, and there is not much difference. The crystal will keep the clock frequency stable until the supply voltage completely drops to zero.
2. A second alternative is a voltage monitor chip that halts the processor as the supply voltage drops, before the resonator stops working. This is now a simple kit from Merg that just plugs onto the ICSP pins on the CANMIO-SVO PCB. The kit cost just over £1.

So, you may be lucky and your CANMIO-SVO might run faultlessly for as long as you want it to. However, for a few pennies more, you could add one of the above fixes to insure the problem never happens.

 

[96701]

On 04/04/2020 at 18:18, Ian Morgan said:

Fortunately, two alternative cheap fixes have been identified:

1. Replace the resonator with a crystal and pair of capacitors. Not so long ago, crystals were much more expensive than resonators, which is why they are used in many Merg kits. Now, the price of crystals has dropped, and there is not much difference. The crystal will keep the clock frequency stable until the supply voltage completely drops to zero.
2. A second alternative is a voltage monitor chip that halts the processor as the supply voltage drops, before the resonator stops working. This is now a simple kit from Merg that just plugs onto the ICSP pins on the CANMIO-SVO PCB. The kit cost just over £1.

So, you may be lucky and your CANMIO-SVO might run faultlessly for as long as you want it to. However, for a few pennies more, you could add one of the above fixes to insure the problem never happens.

 

I've installed the second alternative on my existing CANMIOs thanks Ian, also successfully built a daughter board that is mounted adjacent to (not on top of) the MIO to do the frog switching.

[96701]

Well, one board has a MIO driving 2 servos and a daughter board operating relays off the midpoint event of the servos to do the frog switching. I have successfully fitted a MIO to another board, and this one drives 4 servos, so the daughter board has 4 relays fitted, so next step is to set Events up for the 4 new servos and to get the relays to operate from the mid point servo events. Should be quite interesting because the event numbers on the MIO for 2 points are replicated on the MIO for 4 points. I'm banking on the Event numbers being MIO specific. I may well find out soon.

[96701]

Well, that didn't go well at all. I got the servos set up and moving, but one of the relays of the daughter boards was permanently energised, and I could not set an event up to de-energise it. Started testing the MIO outputs and managed to blow it up!. Never mind, I have another MIO, so clone it, write the NVs from the old board to the new one, and now only one servo works, the other 3 cannot be controlled. I am now officially sick and tired of servos, MIOs and daughter boards.