5 Channel Thermocouple Gauge/Amp w Analog Out & More

[PRY4SNO]

For me, my intent is to mount in the ashtray of my coupé. Not sure how universal the packaging would be for that with the large chassis (and others).

Ideally I'd like a p'n'p solution but recognize that's not always possible.

Sent from my SM-G900W8 using Tapatalk

[speeding-g60]

i am lost......

for someone like me, i could/would monitor each cylinder separately and then monitor like i normally do now for the 5th channel and feed that signal into the ECU.

so what is the total cost, including thermocouples and all related?

[loxxrider]

For me, my intent is to mount in the ashtray of my coupé. Not sure how universal the packaging would be for that with the large chassis (and others).

Ideally I'd like a p'n'p solution but recognize that's not always possible.

Sent from my SM-G900W8 using Tapatalk

For mounting the LCD, I won't have a plug and play solution. That just isn't feasible due to the variety of applications this will be used in. It is possible to do that later on though, if you help me understand your specific needs (like the dimensions and method of mounting you would like to do).

I will probably end up putting the buttons on the LCD board. I will also include solder points in case someone doesn't want them on the board. That way, the part of the board with the buttons could be cut off and the buttons placed somewhere else if desired. I think that would please everyone.

i am lost......

for someone like me, i could/would monitor each cylinder separately and then monitor like i normally do now for the 5th channel and feed that signal into the ECU.

so what is the total cost, including thermocouples and all related?

Not everyone will want/need a solution like this. Are you saying you don't care to monitor all of the channels at the same time? That's fine, but I like to be able to see what they are all doing, especially with the alarm because it could potentially save you from a meltdown in the event of what would otherwise be a minor failure. This is basically an extra layer of protection for your engine. If something is wrong with any of the cylinders, the alarm will be set off, you'll get a visual indication, and the ECU can halt things as soon as it sees a problem with ANY of the channels. That's the real value here.

In fact, you don't even need extra channels on VEMS to react to an "over temp" situation. That can be triggered by the alarm feature which reacts to any channel if there is an over-temp condition. That is something you might not see with just one thermocouple.


On the subject of cost... I haven't quite pegged that yet just due to the fact that everything is still in development. I'm hoping to keep it all under $200 though. Of course, the final price depends on how many channels you want. I can accommodate 0-6 channels fairly easily at this point. More is possible, and I do intend to make up to 10 channels available. Maybe even 12 if anyone wants one. I figure the lambo and R8 guys would like the 10 channel version. Anyway... cost estimate...

The major electronic components cost around $100. That includes microcontroller, TC amp, and LCD/backlight. If you just want the TC amp, I can provide that alone for around $60. Then it's yours to do what you want with. I can also provide the TC amp and microcontroller in a nice case without the LCD if you don't want that. Various levels of backlight and LCD combinations will have different costs, but without an LCD and backlight there is roughly a $30 difference in hardware (not including the cost of the circuit board). Without the LCD, the advantage you get is simply that you have an alarm, decent packaging (hopefully), and an option to upgrade to an LCD display for fairly cheap.

I'll have to get two fairly simple circuit boards made, get various connectors, terminations, etc. made, a case, and more... not to mention labor. Those are small costs, but they could easily eat up $100 if not done carefully. It's hard for me to give a good estimate at this point, but the GOAL is to come in under $200.

[loxxrider]

In my spare time this week, I've been researching the hell out of what I want to use for ingress and egress of inputs and outputs for this thing. I'm trying to make it as clean and quality as possible, while keeping size and wire routing in mind as well as overall cost.

I have been ordering a lot of parts to test out to make sure the quality is where it needs to be. So far, I have the thermocouple inputs covered with a plan and parts are ordered. I bought a punch for making the square holes for those, so they should come out very nice.

I'm still testing the power jacks, but those are pretty easy since they're so common and mostly come round in shape. I'm just testing for quality. The plan is to use a 2.1 mm x 5.5 mm barrel terminal because the plugs are super easy to find around the house in case you need a spare one or want to save a dollar or two on the price of this thing.

It took me a long time to come up with a solution for the output ports, but I think I finally found what I need. Quality panel mount terminal blocks are not easy to come by. The only problem now will be the cutouts for them.

And most interesting of all, I am going to attempt to use an HDMI cable to power and control the LCD and it's buttons. Why? Because it's now becoming one of the most common cables in the home, they're cheap, and the ends are quite small, so running wires is much easier than with a larger connector like any D-sub. The only drawback of using HDMI is going to be figuring out how to get it out of the panel and make it look nice. That's easy with a D-sub connector because punches are readily available for that. Not so for HDMI.

Now that I've got all the connectors sorted out, I need to pick out a box. If it's going to be cheap ($10 or less), it'll have to be plastic. I may provide more than one option for this if there is ever a demand for it (maybe the V10 lambo and R8 guys will want that ). Everything is being done with scalability in mind, although for more than 6 cylinders, I think a different enclosure will be required to house the bigger boards required for all of that.

[audifreakjim]

What about RJ45 or RJ10 like zetronix uses? HDMI is kind of bulky in my opinion. How many wires do you need?

[loxxrider]

I forgot to mention to Aaron, that thermocouples will not be included with this. The reason is because there are large differences in cost and quality of thermocouples out there. You can choose whatever you want, from $10 on eBay to $50 or more from some of the best manufacturers. I don't really have time to ship and receive things like that either since I'm away from home about 27 days a month usually.

What about RJ45 or RJ10 like zetronix uses? HDMI is kind of bulky in my opinion. How many wires do you need?

That's the problem. I need 10 wires as a bare minimum to control the LCD and two buttons. I could drop that down to six wires for a single color LCD with no buttons, but I think that's compromising a bit too much.

I don't think HDMI is very bulky, especially compared to any other connector with as many pins that I know of.

[audifreakjim]

Ok, that makes sense. Places like Monoprice sell just about every kind of adapter imaginable for tight installs.

[loxxrider]

Yeah, you could use a 90 degree angle adapter if you needed some space in whichever direction the connector is facing.

In which dimension do you think you will need the most space behind the display? Straight out the back? To one side or the other? I tend to think straight out the back is the best way to avoid stress problems on the connection and that most people would have room in that direction if it is mounted near a radio or something like that.

[loxxrider]

Jim,

You'll be happy to know that we have peak and hold now! The gauge also has the ability to display in Celsius now if you feel so inclined. I also found out that if a thermocouple is unplugged or broken, the TC amp will detect it and send the maximum value to the output. That's a pretty nice failsafe.

I'm not sure how other peak and hold functions work, but right now this one will retain the highest value encountered since power to the gauge was cycled last. I could make it retain that value even after the ignition has been turned off, but that would be a pretty bad use of the EEPROM memory (memory that stays even when the gauge is off). It has a lifecycle of about 100,000 writes (really, it's more like 1,000,000, but they state it conservatively). That means you can only write to the permanent memory about 100,000 times before it starts to get corrupt and unusable in that particular location. Reading from EEPROM memory is unlimited.

I can add the ability to let the user save the values at the press of a button if you want it. That wouldn't be a waste of memory at all unless you save new peak values like 50 times a day for the rest of your life lol. To me, it isn't really very useful to retain values between runs. I can imagine instances where it would be though, like if the car shut down unexpectedly or if you forget to look at it before turning the ignition off. I don't know... maybe I could have it write a new value into the permanent memory every 5 minutes automatically or something if you think it should be implemented. What do you think?


Up next, I will program auto-detection of plugged in TCs and a few other small tweaks.

I spent a lot of time last night trying to find a nice box to put this all in. I found a few decent options, but it's hard to tell what will be best to use until I get a few of them in my hands. One option is only around $5, so that's pretty good!

Speaking of cost, I just did the most thorough cost analysis I've done yet and I'm still under $200 in parts, but I haven't designed the circuit boards yet. They could push it up around that mark if they aren't made with some volume. With scale, some of the prices for things can come down. Also, there are some cost savings to be had if you don't want certain things because you already have some at home (like an HDMI cable or 2.1 mm cable for power). A single color LCD will also bring cost down by about $15 for anyone who is truly trying to be cost conscious. Anyway, the point is that I still think the $200 price point or below is possible.

[audifreakjim]

Cool, I have always liked out of the side best, having it out the rear forces you to drill a hole somewhere and you will be limited to where you can actually drill holes.

I am having a few second thoughts about keeping the probes on the car all the time. It seems exposed tip is the way to go, but man it just seems like a disaster waiting to happen. Anyways my point is maybe mounting doesn't matter as much because it would be a tuning tool more than a permanent item for me if I get one.

[PRY4SNO]

I am having a few second thoughts about keeping the probes on the car all the time. It seems exposed tip is the way to go, but man it just seems like a disaster waiting to happen. Anyways my point is maybe mounting doesn't matter as much because it would be a tuning tool more than a permanent item for me if I get one.

Agreed. My plan (perhaps flawed, would love some feedback if so) was to dial in the tune using the EGT and AFR. After a solid tune is achieved I plan to monitor AFR only. Udner special circumstances the EGT sensors could be reinstalled pretty easily.

[loxxrider]

Why are you worried about leaving them in? You think one is going to break off and enter the turbine or something? There's about zero chance of that with a high quality probe, especially one with a sheath made from inco. The one I bought a few months ago is meant for use in top fuel... And even if the probe did go bad, they will crack at the junction and you'll know about it on the gauge. In the unlikely even that the tip of the probe did come off, I doubt your turbo would even notice.

Or are you concerned about something else?

I don't see why you wouldn't want to run them all the time if you have the hardware for it honestly. Especially for someone like you Jim, using the car for the track often. Sure, pyrometers are good tuning tools, but they're really good at indicating potential failures too, sometimes before you'd see it with AFR.

[audifreakjim]

I don't want to pollute this thread too much, but here is why. viewtopic.php?f=11&t=73&start=2875

Just a little paranoid at the moment, but I agree, there is value in having as much data as possible all the time. I have two of the exposed tip sensors from the sensor connection here, they are pretty robust, and if the controller can sense the failure of a probe before it snaps off, that's really cool.

[loxxrider]

Sorry you had that QSV failure man. That's rough, especially on an EFR. I can understand why you'd be cautious, but this is a pretty different situation.

I think it's easy to see why subjecting a steel plate like that to broadside impulse forces of a backfire at high temperatures would cause a failure. The momentary forces under such circumstances would be tremendous. Not so with a TC probe. The forces are so much lower due to the smaller cross sectional area and coefficient of drag of a circular leading edge. Also, being made of inconel means the material retains its strength AND ductility at extreme temperatures.

If it would make us all feel better here, I can call up The Sensor Connection and ask them for some failure data. Nothing wrong with having more data right?

[alxdgr8]

Have you heard of Phoenix connectors? They're nice removable terminal block style connectors.

Sent from my Nexus 6P using Tapatalk

[loxxrider]

That's exactly what I'm using Alex

They're not cheap though.

[loxxrider]

Does the dimmer switch output 5V on most cars? I think it does, but need to make sure. Wouldn't want to let the magic smoke out of any nice backlights. I'd check if I had a car at the moment, but I'm in Texas.

[loxxrider]

Sorry if this is boring, but this serves as a nice little development blog for me to refer to and to "talk" things out with myself.

I've really been hammering this thing with code for the past few days. I really can't think of anything else that anyone would want for it at the moment. If you guys want something, let me know. I'll give a brief rundown of the functionality in a bit.

I was thinking the other day, and I realized that I never implemented a way to silence the alarm when it starts going off. The user should be able to acknowledge it and go on with their lives until the next alarm even comes about. It isn't really as simple as just silencing it though... because it should be able to activate again when the temperature goes over the limit again. The problem is that usually when the temperature goes over the alarm setpoint, it will take a little while to come back down. Even though these thermocouples react quite quickly, there is some considerable thermal mass surrounding them, and if you get the exhaust REALLY hot, the alarm might just come right back on again as soon as you silence it. That would be infuriating!

I could have just set a timer so that the alarm stays deactivated for a period of time (like a minute) so that the sensor would have a chance to cool off before being set off again. BUT that would kinda suck because if you silence the alarm and then decide to keep driving hard, the alarm would come right back on again after one minute.

So I decided that it should only come back on after a minute IF all of the temperatures drop below the alarm set point and then go back over it. I think that is a pretty solid scheme.


So, the fully-implemented feature set to date is:

-Menu with several features easily accessed and adjusted with only two buttons

-Automatic thermocouple (TC) detection (will list the number of TCs connected and their locations. Also only displays temperature readings on the LCD from TCs if they are connected)

-TC failure detection (will not be detected at startup if failed at or before startup, and will go to the max value of 1832F (1000C) while under operation (thus setting off the alarm)

-Peak and hold function - just press one button while the display is in normal operation to see the peak temperatures from the run (peaks are not stored after a power cycle... if there is a need or want for this, I can make that work easily). Just press the same button again to exit the peak and hold feature.

-User alarm temperature can be set anywhere from 1000 - 1800 F (537 C to 982C) in 50 F (~27 C) increments. I'll probably make the C temperatures more user friendly (more normal numbers) next. If you want more increments or a wider range let me know, but it is already 16 presses to cycle through those temperature ranges. This alarm sets off a 5V output to an ECU, buzzer, or anything else you want to control with it. As mentioned above, it is now able to be acknowledged by the user also to avoid having to just set the temperature higher to get the alarm to turn off in the event of an over-temperature condition.

-Celsius or Fahrenheit selectable for entire gauge functionality (display in F or C and alarms automatically carry over to the correct temperature in C or F)

-Backlight color selectable using the RGB scale (if you get order an RGB backlight with it). The user can select values for R, G, and B anywhere from 0-255 (the full scale for each color) in 25 point increments. This should be enough to make any color you want. If more adjust-ability is needed, this can easily be changed later.

-Overall maximum brightness of the display is user settable (1-10) independent of the RGB settings. So you can have it any color you want, at a clamped maximum brightness. That means if the display is too bright or dark compared to the other things on your dimmer switch, you can dim it in the software as needed, all while retaining the color you've selected for the backlight. That's easier said than done. If 0-10 isn't enough, I can add more adjustability, but I tried to keep things easy and quick to cycle through.

-Contrast is selectable (0-20). This makes the LCD easy to view from ANY angle.


I suppose that's it. It's amazing how much work goes into making a "simple" gauge like this! I knew I was in for a challenge, and I knew I could do it, but I didn't really figure on it requiring so much code haha.

All of the stuff I bought to package this thing hardware-wise has shipped, so hopefully I'll be able to bring it to the rig with me and work on it there (I've been in Houston for the past 5 weeks for training).

[Aktapod]

Good work Chris! I like the solution for the continuous alarm system. You could also maybe complement the alarm with something visual (say, flash the LCD brightness on half-second intervals) for the duration that the condition exists. That way, even though you won't be pestered by the buzzer continuously, you still have an indication that the problem has yet to be resolved. Similar to the way the autocheck system works when it pops up the warning symbol in the cluster display.

[loxxrider]

Good work Chris! I like the solution for the continuous alarm system. You could also maybe complement the alarm with something visual (say, flash the LCD brightness on half-second intervals) for the duration that the condition exists. That way, even though you won't be pestered by the buzzer continuously, you still have an indication that the problem has yet to be resolved. Similar to the way the autocheck system works when it pops up the warning symbol in the cluster display.

I definitely considered that. Right now, it flashes the values which are over the alarm temperature on the screen on and off. For example, if only T4 is over temperature, it's value on the lcd will blink so that you know that there is an issue and you know immediately on which cylinder the issue is happening.

I think that's a good idea though. Maybe acknowledge the alarm, and turn off the alarm output (buzzer, etc.), but keep the visual indication on the screen?

[audifreakjim]

I like that idea. Just might be annoying if you have a sensor go bad and are a lazy prick. Or I guess a bad sensor would log a different error, not overtemp right?

[loxxrider]

Currently, if the sensor is bad at startup, the unit won't even recognize it as plugged in and won't use it for any alarms. If it goes bad during your run, it will set off the alarm (which I think is a good thing). When you silence the alarm at that point, it actually won't come back on again until you cycle power to the amplifier/gauge. That's becuase in order for the alarm to be set off again, all of the active thermocouple temperatures have to drop below the alarm setpoint and it has to be after the allotted time (currently 60 s).

I can make it handle this differently if there is a failure. For example, if the TC fails it would only base the alarm reset off of the remaining active sensors. What would you prefer? I could also have the display rear error on that channel or something if a sensor reads the full scale (1832 F) which is what happens when the TC fails or becomes unplugged while running.

[ringbearer]

This is awesome Chris!
Good work

[audifreakjim]

As-is sounds good. If a power cycle shows it as failed and keeps it from alarming, that's good to go.

[loxxrider]

Last week I got the chance to try making some hardware fit in a nice little box. Unfortunately, I didn't account for a few things and it doesn't quite fit everything in the configuration I wanted it to. I will have to find a slightly bigger box. It was a good learning experience though.

I had to use a dremel to cut most of the holes out unfortunately. I will come up with a way to avoid that in production models (templates, jigs, waterjet, etc). I have a die for the thermocouple connectors, but it doesn't quite work for the outer connectors, and it definitely doesn't work for the HDMI, USB, etc.


Here are some pictures. Please ignore how rough everything is (especially that floppy TC port lol). This was just me experimenting with different layouts, trying to make the smallest enclosure possible work.

TC ports

2016-06-04 22.13.10.jpg (175 KiB) Viewed 11242 times

Analog outputs, power, and HDMI out to LCD display (HDMI connector will probably not stick out like that)

2016-06-04 22.13.42.jpg (243.05 KiB) Viewed 11242 times

Chap-stick for scale. The box in the background is another one that I ordered for consideration, but it is too big and expensive.

2016-06-04 22.11.36.jpg (273.32 KiB) Viewed 11242 times

Over the past few days, in-between 18 hour shifts and despite the most contradictory and conflicting documentation I've ever received with an electronic component, I somehow managed to wire up the RGB backlit version of this display. I only had to fix a line or two in the code that I'd already written but had been unable to test and it worked. It had a strange "beat-like" flicker to it though, and it turns out that one of the PWM pins I was using runs on a different timer than the other two. Not the most fun kind of issue to chase, but I managed to track it down in a reasonable amount of time. That would have been super annoying to have to deal with! It's perfect now though

2016-06-04 22.17.39.jpg (366.08 KiB) Viewed 11242 times

2016-06-04 22.15.27.jpg (492.13 KiB) Viewed 11242 times

I haven't taken it in the car to find out what settings produce the stock instrument color yet, but I do plan to do that and send them out with that as the default setting.



Up next, find and order another box, test fit hardware in new box, test running the LCD via HDMI, design and procure PCBs for both the LCD and box components, and figure out how to optimize the manufacturing process.