JCDammeyer's 42 projects

[Susquatch]

And as the photo shows after changing the two resistors and the scale/offset values to calculate a new resistor value based on voltage I'm not so sure the temperature readings at this high a value is actually accurate.

That's kinda what I was getting at earlier when I said thermistors are not really that accurate anyway.

A thermocouple can be cheap, but they also sell small 8 pin chips (same as a 555) that have onboard regulators that will measure temperature very accurately. Not sure how to use that in your application though.

Edit - I see there are new LM34 LM335 AND LM35 three lead linearized chips now. That would be easier for your application.

 

[jcdammeyer]

That's kinda what I was getting at earlier when I said thermistors are not really that accurate anyway.

A thermocouple can be cheap, but they also sell small 8 pin chips (same as a 555) that have onboard regulators that will measure temperature very accurately. Not sure how to use that in your application though.

Thermistors are fine when used in the middle of their temperature range. The DS1822 one wire devices (TO92 package) I use all the time. Again, depending on the suffix good to within a degree or so.

The best are the RTD parts but they are more expensive and also require pretty accurate parts to measure the devices.

Understanding RTDs

An RTD is a sensor whose resistance changes as its temperature changes. The resistance increases as the temperature of the sensor increases.

www.te.com

I am perfectly happy with +/-3C and don't need down to 0.1C resolution. What I didn't expect is that the temperature on the outlet would go as high as it does. The IR reflective never showed more than about 45C on the tubing and only the actual pump had temperatures approaching 100C. So at this point, until I stick in something that can measure up to 400C or better I won't know if the thermister is just plain wrong at that high a temp or if it's really 163C.

Inquiring minds want to know. Won't be able to test this until tomorrow afternoon. Meanwhile I can look into how to mount and run wires for everything else.

 

[Susquatch]

@jcdammeyer - you are too fast. I think you might have missed my edit which added the LM34/35 devices.

 

[jcdammeyer]

Edit - I see there are new LM34 LM335 AND LM35 three lead linearized chips now. That would be easier for your application.

The problem with either the LM34.. etc (which I have here) is that the body won't fit through the 3/16" tube fitting.
I could remove the existing copper T and replace it with something much larger that can take a 1/2" tube copper fitting. Then I can put pretty well anything in there. There's a second one wire channel available on the Left side connector so using the DS1822 sensor could work but unlike the LM35 it goes only to 125C not 150C.

 

[jcdammeyer]

@jcdammeyer - you are too fast. I think you might have missed my edit which added the LM34/35 devices.

@Susquatch I'm always 42 seconds ahead or behind.

 

[Susquatch]

@Susquatch I'm always 42 seconds ahead or behind.

Well at least that's better than me! I'm usually 42 days ahead or behind!

Sometimes I feel like it's years......

I knew the 8 pin wouldn't fit, but I had hoped the LM34/5 would.

If you have lots of them, you could try grinding it down. You never know. Its already close.....

 

[jcdammeyer]

I knew the 8 pin wouldn't fit, but I had hoped the LM34/5 would.

If you have lots of them, you could try grinding it down. You never know. Its already close.....

The problem is the brass fitting. To stop the tube from going down to far into it when clenching the collar it's only got a hole the same size as the tube ID. About 3mm. Can't drill it out because then the threaded 1/8" pipe ends up too thin for mechanical support.

So if I were to start over...
I would use larger fittings on the T side. Something that could pass the TO92 package.

 

[Dabbler]

I think I will have to change to thermocouples rather than thermistors.

Yes!!

 

[jcdammeyer]

I did find this LM235. With some work it might be possible to fit it into the air flow. But at the moment I think an LM135 would be better. Max output would be 4.23V at 150C (423K) so I'd have to configure the unit's resistors again for a different divide ratio and the mX+b scaling factors to change the A/D value from 0 to 4095 into degrees. But that part is easy.

 

[jcdammeyer]

An LM135Z is $6.13 from Mouser plus the parts to fit it etc.. The max temp is above the max temp for the LM235 I have on hand.
I clicked the shot just as the compressor stopped and the display flipped from 155C to 152C. That's running from about 50 PSI tank pressure up to 125 PSI. Outlet temperature climbed from 23C to 25C.

 

[jcdammeyer]

The metal brackets to hold the gold box could have been built a lot faster with just the drill press but it's been so long since I used the edge detection probe with LinuxCNC on the mill that I first drew up the assembly, printed a drawing and then issued G-Codes to move the part in the vise for each of the drilled holes for the different sizes and even to line up the 3mm tap.



The end result is a couple of free powder coated aluminum scraps, one cut into two pieces to make a set of brackets to hold the gold box.


Only had to drill and tap two 5mm holes into the main plate. The intention is to isolate the electronics with some foam rubber pads.

 

[slow-poke]

Well here is another option.....

Quite a while back I made an AC load, this thing would generate a fair bit of heat across some forced air heat sinks. Depending on the voltage and power applied different sections of the heat sinks would heat up, so I used quite a number of DS18B20 one wire sensors and then reported the highest temperature on the display. IIRC I used four boards daisy chained.

The PCB image shows two flavors of the same chip TO-92 and uSOP (that is only 1mm thick). The image shows the TO-92 sitting on top of the uSOP, except in one location I removed the image of the TO-92 so you can see what I'm talking about. This was to allow either package to be used if one was not available. These sensors work very well.

 

[jcdammeyer]

Well here is another option.....

Quite a while back I made an AC load, this thing would generate a fair bit of heat across some forced air heat sinks. Depending on the voltage and power applied different sections of the heat sinks would heat up, so I used quite a number of DS18B20 one wire sensors and then reported the highest temperature on the display.

The PCB image shows two flavors of the same chip TO-92 and uSOP (that is only 1mm thick). The image shows the TO-92 sitting on top of the uSOP, except in one location I removed the image of the TO-92 so you can see what I'm talking about. This was to allow either package to be used if one was not available. These sensors work very well.
View attachment 49048

I like the way you laid out the holes so the TO-92 package could lay over the SOP. Nicely done.
I just put mine beside the other one.

 

[jorogi]

The metal brackets to hold the gold box could have been built a lot faster with just the drill press but it's been so long since I used the edge detection probe with LinuxCNC on the mill that I first drew up the assembly, printed a drawing and then issued G-Codes to move the part in the vise for each of the drilled holes for the different sizes and even to line up the 3mm tap.

View attachment 49042

The end result is a couple of free powder coated aluminum scraps, one cut into two pieces to make a set of brackets to hold the gold box.

View attachment 49043
Only had to drill and tap two 5mm holes into the main plate. The intention is to isolate the electronics with some foam rubber pads.

The only thing missing is a cup holder.

 

[PeterT]

@jcdammeyer completely random YouTube drop. Just seems like it would be your kind of thing. Lets call it...project #43 for now LOL

ðð ð£ðð ðð

I started a company: https://www.rackrobo.io/ If you are a Chinese tool distributor, and want to send me cool stuff, go right ahead.

www.youtube.com

 

[jcdammeyer]

Good news. After all the mucking around with the air compressor and new tubing I found it leaked down very quickly. A couple of hours and air pressure in the tank dropped to 40 PSI.
Out with the foamy soap squirt bottle and I tested every plastic quick connect. Every connection. No bubbles.
Then I looked at the outlet water separator. The Sears one from the compressor I bought in 1974 that is now gone but I kept the parts.



The plastic bowl won't come off but boy does it leak from around edge by the metal top.

I pulled it out. Connected the outlet hose to where it went before. Now I dropped 5 PSI tank pressure over 5 hours. That's with three blow guns plugged into quick connect couplers. One of them very old. Although I didn't see bubbles from it... likely there are still teensy leaks.

 

[jcdammeyer]

Yahoo. I now have a box of 2mm screws and 2mm taper and plug taps. My smallest tap wrench is still a hair too large.


And now I've been able to mount the RH sensor inside my 3D printed container. I could have used 2-56 screws but didn't have any short ones.


I've redone my electrical drawings after some thought on how to mount everything. Time to take it all back into the lab, rewire stuff and test that the software does what I asked.

 

[jcdammeyer]

Printed a small box (not happy with print quality) to hold the K Type Thermocouple ==> SPI interface board. Used the 2mm taps again.



Sometimes I think the hardest part of this type of project is deciding what colour wires to use and how to incorporate connectors etc so that it can be serviced.

I'll put a hole in the side of the display module and mount this module to the side of it. The RH+TEMP sensor module goes on the bottom of the display module. The only extra wires running to the display module is the One_wire interface for the RH+TEMP and the serial input to the computer (MISO). The rest are all part of the interface that is already there.

 

[jcdammeyer]

Progress... sort of. I think it's a bad wiring connection (crimp) but not sure. Frequently the top temperature now measured with the K Type Thermocouple goes to 0. Look at the data stream with the scope and sure enough the Data Out pin is zero. And then suddenly when it works it's got data. And the trouble is move something or look at it cross eyed and it changes. Sigh...
Anyway, the 26.3 is not really ambient but I put my fingers on the sensor to warm it a bit to see it change.

In either case... progress...

 

[slow-poke]

Not my favourite thing to do (chasing intermittent issues), like you said, just look at it and it will start working, connecting a scope probe will fix it (temporarily), every time