More idea's... Like you need them..


 
M

Marc Strickland

TVWBB Member
Just thinking out loud, What is the voltage drop/resistance of something like a quad bilateral switch? Something along the lines of a 4066. Was thinking of a way to put 4 thermocouple sockets and the 4 mono jacks for thermistor probes and let the selection you make when you define the probe, connect the correct one.... The down side I can see is needing 4 ports to run the 4066 but someone may have a serial driven idea that would use less ports...
 
You're not gonna make the change from thermocouple to thermistor with software alone, you will need to reconfigure the hardware as well... The thermistor probe needs a pullup resistor and the thermocouple does not.
 
Hence the reason for a 4066 quad bilateral switch or something like it to do the hardware switching or maybe a latch and hold so you would only need to hit it during boot and it would be set.
 
then you would need some spare logic from the ATMega to tell the electronic switch what to do, and code in the HM software to do it... I'm pretty sure the ATMega is all used up in the HM at this point, unless you go with the SMD ATMega that has a couple extra analog pins, though I'm not sure if they could be used to control a switch like that or not...
 
I've considered this as well, even considered using something similar to form a switch network so that only one thermocouple amp would be needed and it would switch between the pairs of wires. Of course the cold junction compensation wouldn't be 100% correct, but all the connectors would be pretty close together in a box and therefore should be very close to the same temperature so it might not make a difference.

If you mux it just right you can use a single ADC channel too and just switch the inputs to get the probe you're measuring into the ADC. Then you could use the other ADC pins as digital IO to run the switches.

The problem with the idea is the required board space to fit another 16+ pin DIP chip, and of course the IO needed to select. There's only 1 unused pin on the ATmega currently.
 
My question is... Why do you need to do this in an automated fashion? Your probes aren't gonna change on their own, you know what you've got plugged in where, so why not just use a regular switch that you can flip manually to select the probe type?
 
lol, You say why, I say why not... I mean, we're building a computer to run a grill.. I can grill without one too. Not trying to be a smarta$$, but if there is enough computing power to do so, why add physical switches.. The other problem I see with physical switches are the contacts themselves, over time they will corrode and then change the resistance we are trying to read.

So if there is an unused pin, is there a way to do serial io with it (might have to be software based if it is not a serial port.) to run another shift reg.? That would give a min. of another 8 physical outputs to drive the switching.

Or... Use a 16 bit shift reg. instead of an 8 bit... Then the part count and io pins used would be the same as what we have now. But we would have 8 more outputs....

I need to stop thinking while I'm typing... This sounds way more promising than trying to run 2 shift registers...
 
I guess we can agree to disagree. I think the why not is the software and hardware modifications that this would require just to flip a switch which once it is set might never flip again, or very rarely at best.
 
Disagreement is good... That's how things get worked out... I will point out that if you were to change probes even rarely, you still have to change the probe profile in software, so you are already switching it in one place (software) just seems to reason that it you already need to change it in one place, why also add a switch. Besides you would also add the price of 4 switches when you could probably get a 16 bit shift register for the same price as the 8 bit one already being used.
 
4 switches assumes you have 3 more thermocouple amps somewhere on the HM, something which I've never heard raised as a possible path forward with the HM. It would increase the cost by quite a lot since the TC amps aren't cheap, and would require a bit of real estate on the board that doesn't seem to be available, and multiply the amount of SMD soldering that would be required to build the board.

It would seem your idea would be better integrated into the SMD version of the HM where you would have the room for all the extra parts to include the option of 4 TC amps and the switching for them. Some advanced users might really dig it but I don't expect the average joe would need 4 thermocouples and the ability to switch between probes types more easily.
 
Marc Strickland said:
lol, You say why, I say why not... I mean, we're building a computer to run a grill.. I can grill without one too. Not trying to be a smarta$$, but if there is enough computing power to do so, why add physical switches.. The other problem I see with physical switches are the contacts themselves, over time they will corrode and then change the resistance we are trying to read.

So if there is an unused pin, is there a way to do serial io with it (might have to be software based if it is not a serial port.) to run another shift reg.? That would give a min. of another 8 physical outputs to drive the switching.

Or... Use a 16 bit shift reg. instead of an 8 bit... Then the part count and io pins used would be the same as what we have now. But we would have 8 more outputs....

I need to stop thinking while I'm typing... This sounds way more promising than trying to run 2 shift registers...
Click to expand...

The SMD version of the Heatermeter I have built, uses the 32pin version of the Atmega328. It gives you another 2 analog inputs.
 
Using your SMD version, and switching the shift register from 8 to 16 bit would do two things, one you could move the third led from a port pin to the shift reg. like the other two led's, freeing up another port. Giving you an additional 7 outputs to drive switching for whatever.... Could the software handle driving another servo? For grills like a Big Green Egg or Primo or any of the other Komodo type grills. It could come in handy to not only control the bottom vent but to be able to run the top vent as well.

As for the cost of the TC amps, yes it would add expense, but they could always be added in the future. It's not real difficult to solder the couple of parts needed. At $5 each, one could add them if they wanted. Just because they can does not mean they would have to. As for board space I would envision that end of the board having a small daughter board under it. Each having two TC jacks and two Thermistor jacks on each. Over all size would not change too much.

I like TC probes because they are so much more durable than thermistor probes. I've been through a pile of thermistor probes over the years with different remote type sensors but my TC's are still going strong. I like versatility..
 
That sounds like a more viable plan, being integrated into the SMD HM version solves most of the isssues you would bump up against with the through hole HM.

The one thing I don't fully agree with is the idea of adding TC amps after the fact. Soldering SMD chips isnt the easiest thing, and is better done when the board can sit flat and stable while you work, before you solder all the probe jacks and bigger stuff on the board. That said, it wouldn't be "impossible", and if this was taken into account when designing the board you could place components in a way that would make installing them after the fact easier (keep other stuff that might interfere with your soldering in that area away).

The daughter board idea would be helpful in this regard, though I would suggest that perhaps the TC amps be made in individual modules that would solder across the standard probe jacks one at a time, probably on the underside of the board. This way you could choose how many TC modules you want to go with, and add more by soldering in another module under a standard probe jack.
 
I was thinking of using a daughter board for the obsolete button, on a version of the smd. But, with any changes to the heatermeter that requires different pins to do something different, that would need to be programmed, which I dont have the knowledge of. I am willing to do a revised board in eagle if i knpw the parts used
 
You must log in or register to reply here.

 

Back
Top