To avoid these ads, REGISTER NOW!

Between 265 & 485 SQ/FT Woody's Works Garage

Workspaces sized between 265 and 485 squarefeet.
To avoid these ads, REGISTER NOW!

nicholam77

Well-known member
Joined
Dec 18, 2016
Messages
2,655
Location
Minneapolis, MN
Frigate has really impressed me.. so any kind of detection on a camera seems kind of irrelevant now, ha. All of my current cams are POE. You basically feed the lower res “sub” feed to frigate for motion detection/recognition, and then it records the main feed when it needs to. You can configure 24/7 recording but the ability to tune motion, add motion masks etc makes motion recording very reliable. I have been recording on SD cards on the cameras using the Amcrest app to view...but going forward that will just be a backup. After just a day playing with frigate I’m sold. It presents all the “events” from the entire camera set and provides a thumbnail of the actual detected object, zoomed in. Basically a live index that you can quickly drill into for full res 4K video. If you’re trying to find an event it’s super quick/easy to review it..and lightning quick. I’m impressed.

That's awesome! I'd look into it myself, but I only have the video doorbell. Seems overkill for one camera since the only thing I'm really missing is package detection. Maybe if I add more cameras later.

I've noticed that there is a WIDE disparity in effectiveness despite companies marketing "induction" capable for their cookware. Then I got to thinking, "How much does one save by going with a cheaper pot, and how do you even compare them" ????

See, this is the kind of interesting "Dennis science" I knew you would apply to Home Assistant! I love these niche use cases for smart home stuff.

I'll slowly bring over automations. I wanted to give things a test drive though and see how good the HA platform is for graphing, and well, cool automations.

Spoiler alert: HA is awesome at graphs AND automations. And the dashboards are extremely customizable.

I see you're well off to the races with automations and dashboards. The induction-power-boiling dashboard you set up looks great. The sky's the limit, really.

And if you develop a more permanent dashboard eventually, you can heavily style it with themes and custom cards and custom CSS. The built-in look does the job, but just know you can really get expressive with it if you want to!

A month or so back, I was approached by Reolink to do a review on their video doorbell. This is a new thing for me, requiring some thought about whether or not I wanted to go down this road with my channel. I figured I'd give it go as I was aware of this company's products, particuarly in the world of standalone solar powered security cameras. I was also looking for a video doorbell, so this kind of worked out. In any case, the company sent out their video doorbell (battery) to review. If you consider the time it takes to set up, shoot and edit a video, I'd be working at about $3/hour on this project, but I figured it was worth checking out this side of things.

I think that's pretty cool!

I thought the doorbell video was great, and the sub-10min runtime is a nice spot to be. Information dense, but doesn't drag on too long.

Great points about RTSP with the battery model. I think for anyone wanting to go outside the Reolink ecosystem with it, the WiFI or even better the PoE models are better for home automation like you said.

Having multiple models with different features is awesome though, Reolink basically sell 4x different doorbell options. I love that, and I love how they build in multiple ways to power and use each model. They are always a favorite with HA users because they are relatively affordable, well-thought out, open to 3rd party interactions, and don't force a subscription.

Doing a product review might not be a huge cash grab, but I think the way you think about and give information on products, it could easily fit within your channel and could possibly be a place to grow bigger sponsored videos. I think as long as you feel the product is good and is legitimately interesting to you, no harm in doing reviews.
 
OP
D

Denwood

Well-known member
Joined
Sep 22, 2014
Messages
4,184
Location
Thunder Bay, Ontario, Canada
That's awesome! I'd look into it myself, but I only have the video doorbell. Seems overkill for one camera since the only thing I'm really missing is package detection. Maybe if I add more cameras later.



See, this is the kind of interesting "Dennis science" I knew you would apply to Home Assistant! I love these niche use cases for smart home stuff.



Spoiler alert: HA is awesome at graphs AND automations. And the dashboards are extremely customizable.

I see you're well off to the races with automations and dashboards. The induction-power-boiling dashboard you set up looks great. The sky's the limit, really.

And if you develop a more permanent dashboard eventually, you can heavily style it with themes and custom cards and custom CSS. The built-in look does the job, but just know you can really get expressive with it if you want to!



I think that's pretty cool!

I thought the doorbell video was great, and the sub-10min runtime is a nice spot to be. Information dense, but doesn't drag on too long.

Great points about RTSP with the battery model. I think for anyone wanting to go outside the Reolink ecosystem with it, the WiFI or even better the PoE models are better for home automation like you said.

Having multiple models with different features is awesome though, Reolink basically sell 4x different doorbell options. I love that, and I love how they build in multiple ways to power and use each model. They are always a favorite with HA users because they are relatively affordable, well-thought out, open to 3rd party interactions, and don't force a subscription.

Doing a product review might not be a huge cash grab, but I think the way you think about and give information on products, it could easily fit within your channel and could possibly be a place to grow bigger sponsored videos. I think as long as you feel the product is good and is legitimately interesting to you, no harm in doing reviews.
Thanks Nick for taking the time to provide all that feedback! Ai has made it a LOT easier to jump into the pool, knowing that you have a highly technical and tireless assistant at your side. I figure we’ll all become flash bags, or just a lot more productive as the tech progresses. Ha.

I had a guy come by today to quote on a tree removal ($$$$) but his eyes lit up when I said had a drone, and gear to provide cool marketing material for him. He has a very unique machine with a 100ft boom that can grab and then cut tree sections from the driveway, back lane etc and lift them over buildings etc. So perhaps there is hidden value to be found in getting competent at least with media!

IMG_3754.jpeg
 
OP
D

Denwood

Well-known member
Joined
Sep 22, 2014
Messages
4,184
Location
Thunder Bay, Ontario, Canada
More interesting "science" on the induction cookware testing. This Tramontina "Signature" 2.8L pot set a few records in my latest test. It boiled a liter (quart) of water the fastest so far, and used the least power doing it, about $.016 worth. What I'm finding interesting is that the various pans/pots I've tested vary a lot with respect to how much power the induction coil can actually deliver to them, even on the same power setting. It varies with the cookware composition, and the size of the cookware base, again depending on which "element" you are using. For all my tests, I'm using the center (large) induction element. This one soaked up 3000 watts, whereas the worst was only taking about half of that power. This Tramontina 2.8L pot also required over 4lbs on my magnet pull test, on par with my good old cast iron pan, which also "took" over 3000 watts.

1779984384361.png

The cookware set can be found on Amazon, and a few of the big box retailers as Tramontina "Signature" series. It looks like they sell some cheaper lines too. The signature stuff looks like it's rated for 500F in the oven, (stainless lids) but I'm not exactly sure how it differs from the cheaper stuff otherwise.

IMG_3771.jpegIMG_3772.jpeg
 
Last edited:
OP
D

Denwood

Well-known member
Joined
Sep 22, 2014
Messages
4,184
Location
Thunder Bay, Ontario, Canada
Whipped up a magnetic pull test rig to compare the magnetic holding properties of different induction pans. Fun little project using a few 2x3s, carriage bolts, snowblower control spring and cheap fish scale. Rattled around in the cranial cavity for a few days and came up with this…works awesome :)

Simple pull bar setup. Top bar is fixed and adjustable, bottom bar slides freely as I took it down about 1/32” on the planer.

IMG_3785.jpeg

Test pan on the rig:

IMG_3783.jpeg

Fish scale :

IMG_3784.jpeg

Very strong little rod shaped magnet in this plastic holder.

IMG_3786.jpeg

Love this tool…

IMG_3787.jpeg
 
Last edited:

drivesitfar

Well-known member
Joined
Oct 23, 2013
Messages
36,006
Location
Pacific Northwest
I certainly don’t understand all your fun testing and results but I do appreciate the effort and thanks for posting them. I use only a 10 and 12 cast iron pan these days. Not sure if efficient or if I clean and season correctly but I’m still above dirt and eating well.

I’ve never heard of that brand name but I agree having a planer and good woodworking tools is great.

Have a grand day!!

Nice pic of you biking with your bro and that dog is in dog heaven running with you two in the woods.
 
OP
D

Denwood

Well-known member
Joined
Sep 22, 2014
Messages
4,184
Location
Thunder Bay, Ontario, Canada
I certainly don’t understand all your fun testing and results but I do appreciate the effort and thanks for posting them. I use only a 10 and 12 cast iron pan these days. Not sure if efficient or if I clean and season correctly but I’m still above dirt and eating well.

I’ve never heard of that brand name but I agree having a planer and good woodworking tools is great.

Have a grand day!!

Nice pic of you biking with your bro and that dog is in dog heaven running with you two in the woods.
Cocoa is living her best life for sure. She LOVEs going on the trails and is perfectly behaved with other riders and dogs. When my brother throws his bike in the truck she naturally knows what's going on, and gets all fired up. We tend to scale back rides a lot for her (she's getting older) keeping speeds down, and making sure we're looping back to rivers/streams so she can go in and cool off for a bit.

On the 10" and 12" cast iron, good news there if you plan on going induction as they perform amongst the very best. I think I got up to about 15lbs of pull on the magnet test with my cast iron pan. On the boil water test it was up there as the fastest, and pulled nearly 3600 watts from the induction cook top.
 
Last edited:
OP
D

Denwood

Well-known member
Joined
Sep 22, 2014
Messages
4,184
Location
Thunder Bay, Ontario, Canada
Just about 5 years ago when doing the kitchen reno, I also replaced the appliances. The "new" fridge is an IKEA Valgrundad (704.621.58) . Like all of IKEA's appliances, these are usually made by companies like Whirlpool...but this fridge is definitely made by Midea. I deliberately chose a simple fridge for reliability...hmm.

vaelgrundad-bottom-freezer-refrigerator-stainless-steel__0956609_pe804786_s5.jpg

The boss noted that the bottom chest freezer seemed to be defrosting randomly, and this seemed to be getting worse. In any case, I enlisted good old GPT and ran through a bunch of tests. I fed it some thermal images of the freezer back wall, pulled out an access panel to check the evaporator, and checked out the compressor etc. at the back.

I used some of my automation bits to get temp probes in the freezer and noted that it would cool just fine for six hours after a reset...then just stop cooling, and temps would rise. The evaporator fan, door functions etc. all worked fine. I finally drilled things down to the control board's compressor relay. The control board is this unit is not expensive, under $100 - Midea Main Control Board — part #17131000010802 The relay is clicking on to send power to the compressor, but the switch contacts inside are not sending power through the relay. I jumped that relay on the board, and then ordered a new board. For a few days until the new board arrives, good old Home Assistant will control the frigde. Power off (via a smart plug at the wall) when temps drop below -15 F for 20 minutes (temp probes in freezer) and power on when temps rise above -10 F for 20 minutes.

You can see below that temps are dropping and the auto on/off automations have not started yet as temps are not outside the set points. Fridge is pulling about 90 watts, so actually pretty efficient. I figured @nicholam77 would be happy I'm making good use of HA :)

1780424016207.png

Must be a pretty common part as the local appliance repair shop will have it here for me in 3-4 days. Maybe I should order two... The compressor relay is the black box top left, under the red wire. I checked voltage across the output contacts (RED and WHITE wires top left), and found that after I heard the relay click on, there was still 120V across these contacts..and there should have been zero volts.

IMG_4009.jpeg

I picked up the TopDon TC001 MAX thermal camera (for iOS) to replace my old FLIR ONE and it came in handy for quick diagnostics. This is the back of the freezer compartment which pretty much tells you the compressor had not been running for a while as the evaporator is behind this panel and should have been much colder.

1780424604475.png

1780345776431_100.png
 
Last edited:
To avoid these ads, REGISTER NOW!
OP
D

Denwood

Well-known member
Joined
Sep 22, 2014
Messages
4,184
Location
Thunder Bay, Ontario, Canada
You are a genius to figure that out!
I wish I could take all the credit, but I'm not an expert on fridge repair by any stretch. That said, ChatGPT more or less led me in the right'ish direction to come up to speed on the how things likely should work. At the end of the day, that "click" from the control board followed by crickets when the compressor should have fired up sealed the deal after some testing with the voltmeter. You do need to be careful though as for example that larger black component ChatGPT ID'd as a transformer which made no sense to me looking at the wiring. After some prodding we agreed it was a relay and then the voltmeter told the rest of the story.
 

drivesitfar

Well-known member
Joined
Oct 23, 2013
Messages
36,006
Location
Pacific Northwest
I love/hate how cheaply made some of the components on appliances are cause if the regular person without your skills or the ability to ask for help on forums or YouTube ends up buying a new replacement usually within 10 years of purchase. They are very cheap to fix on your own cause parts are usually very inexpensive but there is more than a little skill involved in figuring out the issue and repairing it.

Very nice find and I hope your part keeps yours running too too for another decade.

Good luck!!
 
OP
D

Denwood

Well-known member
Joined
Sep 22, 2014
Messages
4,184
Location
Thunder Bay, Ontario, Canada
I love/hate how cheaply made some of the components on appliances are cause if the regular person without your skills or the ability to ask for help on forums or YouTube ends up buying a new replacement usually within 10 years of purchase. They are very cheap to fix on your own cause parts are usually very inexpensive but there is more than a little skill involved in figuring out the issue and repairing it.

Very nice find and I hope your part keeps yours running too too for another decade.

Good luck!!
Thanks for the confidence vote @drivesitfar :)

Talk about a crash course in fridge operation/logic..see below!

I noticed, based on reported power use from the fridge smartplug, that the evaporator defrost logic seems to use an elapsed time counter that is not reset when you fully power off the fridge. You don't want the fridge compressor running and cooling the evaporator, while the fridge is also trying to heat/defrost the evaporator coils! So rather than cycling the fridge power on/off, I added in this Aeotec G2 relay that I had kicking around. It usually is used to control a light (makes a dumb light switch, "smart"), but apparently works fine for a very low draw compressor like this one. Essentially I've replicated the broken relay's control of the compressor by monitoring power use (it goes up to 300 watts if both compressor and defrost is on), and also the temp in the freezer. So we have a working fridge until the new board shows up.

Home Assistant Logic is basically this:

If power use is above 100 watts, this signals that defrost is running, so the compressor needs to be turned off.
If temps are below -15C for 20 minutes in the freezer then turn off the compressor.
If temps are above -10C for 20 minutes in the freezier, then turn on the compressor.

I guess you could run your fridge on HA if you added a few more relays in there...ha. The ultimate DIY serviceable simple fridge??


IMG_4017.JPEG
 

drivesitfar

Well-known member
Joined
Oct 23, 2013
Messages
36,006
Location
Pacific Northwest
Wasn’t sure if I needed to send love or a lol but happy you figured it out. Maybe you should maybe design a new fridge cause I haven’t heard of great one since Kelvinator.

I’d prefer a 50’s fridge design with all the chrome and cool compartments if you do design one I’ll come pick it up in person.

Have a grand day. It was 85 here today which is almost too hot for this old guy but it’s a wonderful evening now. Last nights sunset here from my yard. IMG_1129.jpegIMG_1128.jpeg
 
Last edited:
OP
D

Denwood

Well-known member
Joined
Sep 22, 2014
Messages
4,184
Location
Thunder Bay, Ontario, Canada
Some interesting insights maybe for the next person troubleshooting one of these IKEAVÄLGRUNDAD refrigerators.

It looks like the compressor runs at about 100 watts, and takes a significant amount of time (like 4 hours) to take the freezer from -7 to -20C. You can also see the power spikes, about 200 watts, every 6 hours which looks to be a standard timed defrost cycle. I suspect that powering off the fridge does not reset this 6 hour timer..it's based on actual run time of the fridge. The evaporator heater/defrost runs for about 20 minutes. I could also see in the data that from time to time the control board is actually working as the zwave switch I added has turned off (so the compressor should be off) but the fridge control board relay is running the compressor to meet it's programmed cooling temp targets. I just need to this to work for another 3-4 days...

At the end of the day, I had wondered how "modern" fridges manage to keep free of frost, vs the upright freezer in our basement which requires old school thaw/melts every 4-5 months. I also had no idea there was a fan inside the fridge (blows cool air from the evaporator) and a motorized damper that throttles air flow up into the fridge to keep it warmer. I feel a bit more educated now...ha.



1780543884814.png

This is the HA YAML code that controls the compressor off/on logic based on the temperature sensor I have in the freezer, turns off the compressor if power over 150 watts is detected at the wall plug (which means the fridge is defrosting) and also captures timestamps for the dashboard. This is where AI is super impressive. It generates code like this in seconds, although you have to massage it a bit (via prompts) to get the outcome you want. I have coded automations like this in Hubitat but it's taken me a LOT longer to debug, and test to get stuff working. ChatGPT and Claude can generate efficient code in literally seconds, and walk you through where to set up helpers, scripts, automations and the dashboards themselves. It's not perfect, but jeeps creeps it's kind of crazy how much faster/easier it makes setting up systems like this.

Code:
alias: Freezer - Compressor Control Simple
description: Controls compressor relay only. Fridge outlet remains powered.
triggers:
  - entity_id: >-
      sensor.erv_exterior_intake_air_temp_fgk_10x_sensor_1_2_5_firmware_temperature
    above: -18
    for:
      minutes: 10
    id: too_warm
    trigger: numeric_state
  - entity_id: >-
      sensor.erv_exterior_intake_air_temp_fgk_10x_sensor_1_2_5_firmware_temperature
    below: -21
    for:
      minutes: 10
    id: cold_enough
    trigger: numeric_state
  - entity_id: sensor.christmas_light_4_aeotec_smart_energy_switch_power
    above: 150
    for:
      seconds: 90
    id: high_power
    trigger: numeric_state
  - entity_id: sensor.christmas_light_4_aeotec_smart_energy_switch_power
    below: 130
    for:
      minutes: 3
    id: high_power_cleared
    trigger: numeric_state
  - minutes: /2
    id: safety_check
    trigger: time_pattern
actions:
  - choose:
      - conditions:
          - condition: or
            conditions:
              - condition: trigger
                id: too_warm
              - condition: trigger
                id: high_power_cleared
              - condition: trigger
                id: safety_check
          - condition: state
            entity_id: switch.fridge_relay_temp_aeon_micro_switch
            state: "off"
          - condition: numeric_state
            entity_id: >-
              sensor.erv_exterior_intake_air_temp_fgk_10x_sensor_1_2_5_firmware_temperature
            above: -18
          - condition: numeric_state
            entity_id: sensor.christmas_light_4_aeotec_smart_energy_switch_power
            below: 130
          - condition: template
            value_template: >-
              {{ (now() - (states('input_datetime.freezer_compressor_last_off')
              | as_datetime | as_local)).total_seconds() > 300 }}
        sequence:
          - target:
              entity_id: switch.fridge_relay_temp_aeon_micro_switch
            action: switch.turn_on
          - target:
              entity_id: input_datetime.freezer_compressor_last_on
            data:
              datetime: "{{ now().strftime('%Y-%m-%d %H:%M:%S') }}"
            action: input_datetime.set_datetime
      - conditions:
          - condition: or
            conditions:
              - condition: trigger
                id: cold_enough
              - condition: trigger
                id: safety_check
          - condition: state
            entity_id: switch.fridge_relay_temp_aeon_micro_switch
            state: "on"
          - condition: numeric_state
            entity_id: >-
              sensor.erv_exterior_intake_air_temp_fgk_10x_sensor_1_2_5_firmware_temperature
            below: -20
          - condition: numeric_state
            entity_id: sensor.christmas_light_4_aeotec_smart_energy_switch_power
            below: 130
        sequence:
          - target:
              entity_id: switch.fridge_relay_temp_aeon_micro_switch
            action: switch.turn_off
          - target:
              entity_id: input_datetime.freezer_compressor_last_off
            data:
              datetime: "{{ now().strftime('%Y-%m-%d %H:%M:%S') }}"
            action: input_datetime.set_datetime
      - conditions:
          - condition: trigger
            id: high_power
          - condition: state
            entity_id: switch.fridge_relay_temp_aeon_micro_switch
            state: "on"
        sequence:
          - target:
              entity_id: switch.fridge_relay_temp_aeon_micro_switch
            action: switch.turn_off
          - target:
              entity_id: input_datetime.freezer_compressor_last_off
            data:
              datetime: "{{ now().strftime('%Y-%m-%d %H:%M:%S') }}"
            action: input_datetime.set_datetime
mode: single
 
Last edited:
OP
D

Denwood

Well-known member
Joined
Sep 22, 2014
Messages
4,184
Location
Thunder Bay, Ontario, Canada
I guess my stint as a fridge repair guy went well, ha. The new control board arrived today and is doing it's thing :) The "hacked" former control board has been replaced with the new OEM version. I was controlling based on a -20C to 18C window, and the OEM board is running the fridge to maintain -24C. It's doing a good job with -24 to -21C being maintained so far.

ikea fridge fix.jpg

For reference, the fridge is the very popular IKEA VÄLGRUNDAD (still looks to be the same 5 years later) and the control board is under $100, reference part is made by MIDEA, part number 1713100001407

On the "failed" orginal control board, the compressor relay (seen below with the "520" sticker on top) was the failed component so let's hope the corresponding component on the new board below is an upgraded component!

IMG_4281.jpeg
 
Last edited:
OP
D

Denwood

Well-known member
Joined
Sep 22, 2014
Messages
4,184
Location
Thunder Bay, Ontario, Canada
This is my last fridge/freezer post, I promise...ha!

Below is the data plot with the new controller board running things for 24 hours. My home assistant code ran the fridge pretty much identical to the control board! For someone just interested to see how a "modern" efficient fridge works, you can pretty much see that in the graph. The fridge and freezer are set for "2 snowflakes" for temperature on the fridge/freezer controller, which for the freezer maps to -2˚F (-19˚C). The graph shows the temps inside the freezer varying from -18C to -24C while the compressor runs at around 90 watts for 45 minutes, then turns off for around 45 minutes. The expected defrost cycle does indeed happen every 6 hours where the compressor is off, and the fridge pulls around 200 watts for 20 minutes to heat and defrost the evaporator core. It looks like some of that power is also used by a PTC heating trace in the drain tray at the back of the fridge to evaporate water that has collected there. PTC wire conducts power based on temp (more when cold, less when hot), so is a simple way to control defrost heaters.

On to the next project...

ikea fridge fix2 .jpg
 

Attachments

  • vaelgrundad-bottom-freezer-refrigerator-stainless-steel__AA-2215509-4-1.pdf
    9.6 MB · Views: 0

drivesitfar

Well-known member
Joined
Oct 23, 2013
Messages
36,006
Location
Pacific Northwest
Very nice work Woody!!

To keep an ailing fridge up and running while waiting for parts should earn you some brownie points with your bride at least. Thanks for including us in your adventures!!
 
To avoid these ads, REGISTER NOW!
Top Bottom