Speedset will not adjust

[IDBoating]

Note: It's been privately pointed out to me that this thread may appear to suggest that Medallion is at fault somehow. That is not true. Medallion isn't the bad guy here. Indeed, I think poor Medallion did a fine job accommodating someone else's substandard hardware design as mentioned above.

The problem people are having here is with the touchpad. That touchpad is not made by Medallion, but by a little subcontractor outside of Seattle WA. I don't know who designed it, but both Chris and I have questioned the overcomplexity of the circuitry and (especially) the packaging choice to use a flex circuit. To review: The functionality of this board could have been achieved with a single logic IC and with a standard nonflexible PCB (snap domes work great on standard PCB's). Instead, the overly complex circuitry adds cost while the flex circuit reduces reliability.

Finally, we've established that these panels existed before Medallion OR Murphy started supplying helm electronics to Tige. When Medallion came along, they had to accommodate these panels and their existing signal behavior. Nobody in the electronics biz likes to deal with inherited problems caused by someone else's design decisions, but Medallion did a fine job here. It's not their fault that they had to essentially synthesize momentary behavior from an artificial toggle switch!

Just wanted to clear up any misunderstandings.

[IDBoating]

use 3 momentary buttons plus a simple circuit that does the same function as the MCU, basically a DIY clone of the speedset.

A simple D flip flop would duplicate the toggling functionality. A 4013 would run directly from +12VDC, eliminating the voltage regulator, and no programming would be required. $0.51 in quantity one from multiple sources. Would need to be wired up of course, not sure it's available in DIP package these days, but even as SMD it's solvable.

[ChrisSnow]

As far as I can tell it just comes straight off the ignition supply on mine.
If you think about it the battery should prevent any particularly nasty transients, I do agree that if it were my design that for the cost of a TVS and a poly-fuse you would include something!

glad to hear you were able to confirm it was as simple as I thought :-) I do have scopes etc but was busy all yesterday, yep the pin you needed comes out on the harness :-)

I have successfully fixed mine now using silver ink and using a FPC socket, I had to cut the old connector off, carefully scratch the solder resist off, then add a layer of silver ink to bulk it up just in case.

I then put a layer of conformal coating on the PCB and resembled, notice that I have put the connector on the top and not the bottom, there is now a much more sensible bend radius, but at the expense of weather protection.

here's some photos
IMAG0010.jpgIMAG0012.jpg


With regards to people options I see 4 main options.

1. Repair as per my solution, fiddly but leaves you a working original unit.

2. use 3 momentary switches and reuse the main PCB (which is probably fine)
I would use something like these
http://www.rapidonline.com/electroni...switches-81479
you would need to do some delicate ish disassembly and soldering but doable for anyone with electronics experience.
you could even drive the LED's directly from the PCB, I was thinking of using blue switches lit from the backlight supply for up and down and a red lit switch for the enable driven from the status signal.
Or you could use a separate panel indicator and 3 of the same switches fed from the backlight supply, personal choice really.

this solution is rather neat, it gives you a reliable system with exactly the same functionality and method of use. I would probably get an oval bit of aluminium or stainless laser cut to the same dimensions as the original speedset and go from there.

3. reuse nothing
You would need 2 momentary buttons and a latching one, if you bought the matching Deutsch connector it could be and entire plug in replacement.
another variation of this is to reuses the dash mounted cruise control switch and just have the up and down switches, this would only work for the direct ECU style control on older non medallion boats.
you may have issues as WABoating mentioned with regards to accidentally leaving it in cruise control, but in reality I think the ECU will stop this actually being an issue.

4. as with 3, but you would use 3 momentary buttons plus a simple circuit that does the same function as the MCU, basically a DIY clone of the speedset.

Depending on how many people actually need a fix it might be viable to get a batch of metalwork and or PCB's made up for relatively little cost.

That will be a question of current. I don't know if the line coming from the ECU or Medallion is a current source or sink, and how much current it can handle. More experimentation - but this will need to occur on an actual boat. An LED almost certainly could be driven, and today's higher brightness LED's would probably be an improvement. An incandescent bulb... there might not be sufficient current capacity.

On the basis that the PCB has 220 ohm resistors you must have at least say 50mA to play with,(enough to light a pair of modest LED's in parallel as per original speedset) the ECU \ medallion must be a current source as the cathode of the status LED is connected to ground on the FPC.

All in all excellent news for anyone with a broken speedset :-)

[IDBoating]

Anyone know if the +12VDC powering this panel, and other helm circuitry, has transient protection? I was surprised to see absolutely no protection on this panel for load dump, for example. That little voltage regulator is not going to like +/- 80V spikes when the alternator's magnetic field collapses. There isn't even a back biased diode. That's a pretty big risk to take, so I'm wondering if perhaps there's a "protected" +12VDC that's powering the underdash electronics.

[IDBoating]

There will be some very happy Speedset owners. One benefit would be that the switch could be lit to clearly indicate the cruise is engaged. Wasn't that a complaint on early Speedset? A small led that was really hard to see in the sun?

That will be a question of current. I don't know if the line coming from the ECU or Medallion is a current source or sink, and how much current it can handle. More experimentation - but this will need to occur on an actual boat. An LED almost certainly could be driven, and today's higher brightness LED's would probably be an improvement. An incandescent bulb... there might not be sufficient current capacity.

[IDBoating]

Sweet! So what exactly has gone wrong with his control module then?

The weak spot in the design seems to be the ill-advised use of flex circuitry, and the even more ill-advised decision to bend it so aggressively. That material has a minimum bend radius spec, and while I haven't looked it up, I bet *folding* it violates the spec! The flex traces then tend to break, you lose continuity, and stuff stops working.

Sounds like Chris is planning an attempted repair to his actual traces with silver ink, etc. I'm going to skip that because I don't want to tear into JM's panel any more than I have to. Instead, I'm going to attempt using very fine (30 gauge) wire to substitute for the flex trace that has failed. This will require removing some of the potting material, and doing some really delicate work on the flex circuit, but I've done that sort of thing before so I think it will work.

The good news is that if it works, this should be a permanent fix for this broken trace. The bad news is that there are seven other traces just waiting to fail like this one did. Bad designs tend to hang around and haunt you indefinitely.

At some point I suspect Chris or I will write up a DIY plan for broken panels. That's not necessary in this case - I hope!

[Stingreye]

Duh... I just realized I don't need to plug JM's panel into a boat to scope it out. So I hooked it up to a bench power supply, dialed in 12.25VDC, and started testing.

Sure enough, it appears that the MCU, and its voltage regulator, and its discrete FET, are there solely to turn a momentary contact switch into a toggle switch. There appears to be zero intelligence on the signal; it's flatline DC on my scope. The MCU does insure that the line always powers up high (+12VDC), after which successive actuations of the center snap dome toggle the output between ground and +12VDC. I was going to include photos but trust me, it's really boring.

Several dollars of electronics to fake a toggle switch. And extra code in Medallion's modules to DE-toggle (edge trigger) the switch for menu operations! Wow. My guess is that Tige had this panel designed for Chris's era of boats, where the panel talked directly to the ECU. When Tige added Medallion, they probably wanted to keep using this panel (the NRE was already paid for!) so Medallion had to accommodate the fake toggle behavior in their code.

Given this, it's almost certain we can DIY replace failed panels with three switches: Two momentary contacts and a toggle switch (the latter can be any switch style that alternates positions). You run the slight risk of starting the engine with the replacement toggle "cruise" switch in the on position, but since our engines won't start unless the throttle is in neutral anyway that really doesn't matter. That would make cruise work on older boats without Medallion.

For Medallion-equipped boats, a somewhat more elegant button might be desirable since Medallion's code now expects the edge triggered toggle behavior. Perhaps a locking pushbutton, which locks in a slightly depressed position on alternate actuations. We could replicate this toggling behavior with a bit of electronics, too, but I still say it feels like overkill.

Probably also need a simple RC filter to debounce the center button too. Side buttons shouldn't need it since they're already raw outputs, any debouncing is happening in hardware or software on whatever this panel plugs into (ECU or Medallion). Most likely the center switch is similarly debounced, but a resistor and cap would be cheap insurance.

There will be some very happy Speedset owners. One benefit would be that the switch could be lit to clearly indicate the cruise is engaged. Wasn't that a complaint on early Speedset? A small led that was really hard to see in the sun?

[Timmy!]

Sweet! So what exactly has gone wrong with his control module then?

[IDBoating]

Duh... I just realized I don't need to plug JM's panel into a boat to scope it out. So I hooked it up to a bench power supply, dialed in 12.25VDC, and started testing.

Sure enough, it appears that the MCU, and its voltage regulator, and its discrete FET, are there solely to turn a momentary contact switch into a toggle switch. There appears to be zero intelligence on the signal; it's flatline DC on my scope. The MCU does insure that the line always powers up high (+12VDC), after which successive actuations of the center snap dome toggle the output between ground and +12VDC. I was going to include photos but trust me, it's really boring.

Several dollars of electronics to fake a toggle switch. And extra code in Medallion's modules to DE-toggle (edge trigger) the switch for menu operations! Wow. My guess is that Tige had this panel designed for Chris's era of boats, where the panel talked directly to the ECU. When Tige added Medallion, they probably wanted to keep using this panel (the NRE was already paid for!) so Medallion had to accommodate the fake toggle behavior in their code.

Given this, it's almost certain we can DIY replace failed panels with three switches: Two momentary contacts and a toggle switch (the latter can be any switch style that alternates positions). You run the slight risk of starting the engine with the replacement toggle "cruise" switch in the on position, but since our engines won't start unless the throttle is in neutral anyway that really doesn't matter. That would make cruise work on older boats without Medallion.

For Medallion-equipped boats, a somewhat more elegant button might be desirable since Medallion's code now expects the edge triggered toggle behavior. Perhaps a locking pushbutton, which locks in a slightly depressed position on alternate actuations. We could replicate this toggling behavior with a bit of electronics, too, but I still say it feels like overkill.

Probably also need a simple RC filter to debounce the center button too. Side buttons shouldn't need it since they're already raw outputs, any debouncing is happening in hardware or software on whatever this panel plugs into (ECU or Medallion). Most likely the center switch is similarly debounced, but a resistor and cap would be cheap insurance.

[jmhatchett]

I feel like a little kid at church when i was young... I had no idea what the grown ups were talking about!! But power to you guys and this foreign language you're speaking! Hope we can find a simple fix or DIY fix.

[IDBoating]

I am going to try and use a combination of conductive silver paint and silver wire to hopefully repair it, though to be honest I'm leaning towards a complete replacement with switches as it's probably going to fail again in future!

I presume you have the usual tools (scope, meters, etc.). It would be extraordinarily valuable to look at the signals - especially those associated with the center button and that MCU - with the unit connected and powered up. Given that you've de-potted yours, you're in a unique position to do this. Please consider it!

[Timmy!]

Good luck!!!

[IDBoating]

I agree from your photos that the side switches appear to have their commons going to +12V. That was my impression from JM's panel too, but I couldn't be certain without disassembly. As you know, that's an odd choice... normally folks don't use passive pulldowns. Not unheard of, just unusual which is why I wanted a second pair of eyes on it.

FYI, here is the page of the PCM schematics that show how their ECU-based cruise is controlled:
PCM_Engine_Connections.jpg
You can see the four cruise-related signals running straight from the ECU to the boat wiring harness. Note that the panels we're working on would not be compatible with this scheme!

I'm hoping to have JM's panel reworked today. More soon.

[ChrisSnow]

couple more photos,
IMAG0007.jpgIMAG0008.jpg

[ChrisSnow]

Seriously? This is the first I've heard of this about your boat. As I mentioned, this three-wire ECU based cruise control does exist but are you saying that this panel plugs directly into your ECU? Does not pass through any sort of helm box? I theorized they could do this given the pinout on this module but didn't know they had actually used it that way early on.

that is correct, I have no additional electronics in the helm, check out the schematic's I sent to you a while back, it involves the electronic throttle control module obviously but certainly nothing else.

You know, I just can't quite get past the "it's too complicated" feeling. They have 12VDC available, and there are logic family choices that run at that voltage. If the only thing they are doing is turning a momentary switch into a toggle, they didn't need two IC's and a discrete FET and the PCB real estate and the cost and testing and decreased reliability from more parts just to do that. Seems like someone, somewhere along the line, would have realized they could save money and hassle. Makes me think we're missing something. Hmmmmm.

The part isnt expensive and i guess minimises BOM count, also may have given them flexibility in future.

Also: On boats with Medallion the center button does not enable/disable cruise. There's a dash toggle for that. Meaning that on Medallion boats the entire two-IC's+FET setup is utterly useless, just bringing out the snap dome would have sufficed. All they need is momentary contact to talk to Medallion's menu system. In fact, it makes it worse because (if our theory is right) they now have to do edge triggering in software to OVERCOME the toggling "feature" and turn the button back into a momentary contact!

I'm going to dig out my ECU cruise documentation and see if I can confirm that the on/off signal is level triggered. If the ECU itself handles the toggling function then we still don't have an explanation for the onboard MCU. It's not like the ECU doesn't have the horsepower to deal with it.

I also have the dash toggle, see schematic. i agree, though if they didn't do it it wouldn't work with my ECU.
Do they even customise the ECU code or just tune and enable features?

I don't see this. The LED signal comes in JP2 and goes straight to the keypad. Where's this decoding of which you speak?

sorry this was with reference to the Faria gauge.

Did you ring this out? I believe fixing JM's panel is simply a matter of restoring the lost common side of those switches, but I need to make certain I'm going to the correct side (+V or ground). I could tear his apart to find out, but since yours is already open you can confirm without me having to damage his. How did you make certain of this?

here's a few more photos that should help you (hopefully) get to the same diagram as I drew. mines in a bit of a sorry state at the moment so optical assessment is easier than electrically!
you can see how the traces make their way from the connector to the switches, it's single layer so it's dead easy to do :-)

I am going to try and use a combination of conductive silver paint and silver wire to hopefully repair it, though to be honest I'm leaning towards a complete replacement with switches as it's probably going to fail again in future!
many of the tracks have cracked right by the connector, due to the FPC being bent 180 degrees in a very tight radius.

Overall a pretty dire assembly! I think it was actually designed with the FPC being put in from the top of the board, it's plenty long enough and even has a guide of sorts for it, this however would have made it a pain to pot and waterproof, but would have made the FPC much less likely to break
IMAG0002.jpgIMAG0004.jpgIMAG0005.jpgIMAG0998.jpg