Fridge temp control fix, for AC and DC operation. (ANY model, not just N300.3)

[rickst29]

I started a long(!!!) Thread on this, in which I changed my mind several times. At the end, though, we have a solution which creates provides daily temperature variation in the range of single-digits. (The former Thread was here.) This solution does NOT work for Propane operation, only AC and DC. But the recommended DC circuit design bypasses the front-panel selector switch. So, when boondocking with either a generator or a high-capascity Solar system, you could run the fridge on propane for most of the load, with DC providing "fine tuning" as needed. In the afternoon and early evening, when temperatures are still warm, the DC heater would be working quite hard. As you approach dawn, and the lowest temperatures of the night, the DC heater would run less and less (and maybe not at all). Then, in the daytime, you recover the used battery capacity via solar panels or generator.

This solution has also been confirmed to work by ThePair, although he uses it only for 120VAC. The total parts cost somewhat more than $100, depending on the number and type of "bypass" switches you install. (I installed, but then removed my "bypass" circuitry; if the electronics ever fails, it's very easy to just disconnect power leads from one of the relays, and connect them together with a wire nut or push-in connector Union. Here are links to the necessary temperature control parts to handle both the AC and DC heater circuits:

The electronic "brains", which uses the input temperature measurements to operate the output "SSR control" circuit:
http://www.lightobject.com/JLD7100-D...oller-P66.aspx

The thermometer: http://www.lightobject.com/Premium-S...aded-P681.aspx
Do not try downgrading the Thermometer to type K, because ThePair had bad results using one of the cheap ones.

An electronic switch ("SSR"):
http://www.lightobject.com/25A-Solid...C-Out-P61.aspx

A 4-pin automotive relay, sized for 40A. (These are typically used for turning on headlights, stereo, car alarm, etc.) Available at any auto parts store for a bit more than $5, here's an example:
http://www.ebay.com/itm/Car-12V-40A-...db8562&vxp=mtr
You can buy them with either the pins, or with a wired socket already attached. I recommend a "pinned" version, even though you must use a push-on crimp connector, because the pair of 'power side' wires built into most of those sockets are very small. (Perhaps AWG 14?) IMO, your new DC "power" wires should match TM wiring - 10 AWG stranded copper.
Vibration-Proof, weather-proof wire connectors.
All wire used behind the refrigerator should be THWN, because of the heat which the refrigerator generates.

If using DC independent of the Norcold selector switch, as recommended: ON/OFF/ switch for the PID-DC power wire path. You might consider installing this on the exterior of the TM, using a marine switch or a weather-proof switch cover.

Note: You will be using the existing fuses (the new circuits are wired into the "hot" sides of both existing circuits, 12VDC and 120VAC, after the fuses.)
-----

I mounted the PID device in a new wall plate (a blank plate, with 1/16th DIN cutout made by hand tool). If you care to reach that far up, and don't want Cable TV at RV parks, then you could pull out and replace the existing wall plate with CATV and 2-wire telephone connectors. (We nearly all use cell phones now.). Otherwise, cut your 1/16th DIN hole into the paneling at least 3-4 inches away from the vertical side of the TM interior, and lower down. (You NEED those 3 inches, because there is a big bundle of wires in the TM along the vertical edge. You don't want your new "wall plate" mounting screw going into that bundle of wires.)

This position allows the Thermocouple to easily reach the fridge interior, from the back. I drilled into a place a few inches down from the fluid tank, slightly to the right of the air tube (looking into the back of the fridge). In order to make the thermocouple reach all the way into the interior, you will need to trim off the hex-shaped mounting edge (the one which matches the nut). Use a grinder to make it round, matching the size and shape of the solid stainless column directly behind it. The Thermocouple attaches to the PID as follows: terminals #6 and #7 get wires with the blue wire wraps, and terminal #8 gets the red one. (Pay attention to the plastic wire wraps, NOT the colors of the actual wires.)

Don't dream of modifying the PT-100 thermocouple lead wires or jacket in any way; PT-100 measurement is based on resistance, and properties of the wire leads are critical.
-----

The Thermocouple attached to PID terminals 6, 7, and 8. Terminals 9 and 10 are not used. Next, we wire the PID terminals #1 through #5:

#1 and #4 are tied together, and both get +12V, from any TM "hot" wire in the vicinity.
#2 and #3 are tied together, and both connect to any TM 12V "grounded" conductor.
These two pairs of leads do not have to be #10 wire, they only carry small "controller" current loads. (I used #10 anyway, because I had plenty of it.)

#5 is the +12VDC control output from the PID internal Relay. It provides "sense" voltage to both of the refrigerator power relays in parallel. Create a "Y" from this wire behind the refrigerator, and connect the resulting wires to the "+" terminals of the SSR (for AC power) and the automotive relay (for DC power). Note that the correct terminal is marked with "+" polarity.

And finally, we connect the "-" terminals of the two refrigerator relays to any TM 12V "Grounded" conductor.

We're done with the control side, handling 12VDC and AC. Now Jump to post #6, in which we wire the power leads.

[ThePair]

Quick note: I have not worked with the DC side for PID control, as I melted mine and decided it wasn't worth the effort/replacement!

AC side SSR, if used on its own, does not get hot enough to require a heat sink (tested via finger). I cannot speak to the setup if the DC is included.

[ThePair]

Originally Posted by rickst29

The MOSFETs in the switches consume 1-2 Volts of output potential on the 'Load' circuits. Our AC heatert, without an SSR in line, consumes 1.8A at 120V. So, 2V lost at the SSR @ 1.8A is only about 5 watts.
The DC heaterside, however, runs 10.7 Amps at @ 14VDC. The 2V voltage drop creates a significant loss of power on this path: At the 12 Volt net potential after 2V loss in the SSR, we will run at only 9.2A (yielding just 110 Watts, instead of 150). This is below Norcold's minimum DC voltage limit of 13.5V, so I could need to add a Buck Boost DC regulator as well.
Anyway - back at the SSR, the 2V voltage drop times the 9.2 Amps is creating almost 20 watts of damaging heat. (roughly 4x the amount of heat generated on the AC circuit). At least the small heatsink would seem to necessary, in order to prevent overheating and burnout.

As I said, I successfully melted my DC side without a heat sink! But the AC side, on its own, has been fine sans heat sink. So, the math corroborates reality, at least in this instance.

Although there were many false starts, I cannot tell you all how much happier we are now that the fridge self-regulates.

[Bill]

Originally Posted by rickst29

The MOSFETs in the switches consume 1-2 Volts of output potential on the 'Load' circuits. The DC heater side, however, runs 10.7 Amps at @ 14VDC. The 2V voltage drop creates a significant loss of power on this path: At the 12 Volt net potential after 2V loss in the SSR, we will run at only 9.2A (yielding just 110 Watts, instead of 150). This is below Norcold's minimum DC voltage limit of 13.5V, so I could need to add a Buck Boost DC regulator as well.

Rick -

That may be overly pessimistic.

First, I seem to recall that in some thread, it was pointed out that the "spec" of 13.5 volts is a typo by Norcold, and that the real number was 11.5 volts. I'll see if I can find that comment.
EDIT: Found it. See posts 3 and 4 here

Second, my guess would be that the minimum voltage spec applies only to the voltage that operates the electronics, not the voltage applied to the heater. If the heater voltage is low, it would not get as hot, which would slow down the cooling process, and 110 watts vs 150 might be a significant hit. But I'm not sure it would affect anything else.

Please remind me why the MOSFET switches, with their 2-volt hit, can't be replaced by a relay? Or maybe a pair of relays, a small one driven by the PID, to operate the power-handling relay? I think you discussed it at one point, but I'm not finding it.

Bill

[rickst29]

Originally Posted by Bill

Rick -
Please remind me why the MOSFET switches, with their 2-volt hit, can't be replaced by a relay?
Bill

Great idea! I did that, and the revised instructions use an automotive relay on the DC side.

[rickst29]

Next let's wire the AC switch and power leads.
In the back of the fridge, cut into the "hot" leg of the 120VAC wire, following the 3A fuse. (Leave most of the length on the end leading into the heater unit, because the Norcold wire is more heat-tolerant than the wire you're probably using.)

If you want to create an "emergency bypass switch" for AC:
Connect the Norcold "hot" lead to the common (middle leg) of an ON/ON switch. This is the kind of switch which supplies power to ONE of the other two legs at a time, never both.

Make a connection from one of the free switch legs, stripping the other end for a red wire nut connection with two other wires. Also strip the "free" end of the wire you previously cut (the Norcold "hot" wire into the heater core). The third wire of the wire nut will be from the SSR.

Connect the other secondary side of the ON/ON switch to a one of the switched "AC power" terminals of the SSR. Connect from the other switched "AC power" terminal back to the wire nut, and tighten it with 3 wires. (One wire from the SSR; One wire coming directly from the ON/ON switch; and the Norcold wire into the "hot" side of the heatger core.)

How they bypass switch works:
Set the Norcold selector switch to "AC". When you want to run TM-controlled, because "Rick's newfangled thing broke", then you flip the switch to the non-SSR wire: Current reaches the heater wire directly through the "Y". (When bypassing the SSR, you will need to turn down the Norcold temperature dial to an appropriate level.)

But the "new normal" mode is to use the SSR: Flip your switch to the activate the leg which groes through the SSR on the way into the "Y", and turn the TM temperature controller to maximum ("6"). The PID will de-activate the circuit when the temperature reaches it's low limit, and re-activate the circuit when it sees that you are about to reach the high limit.

If you wire directly, with no bypass switch, you simply turn the interior Norcold selector switch to "AC", and turn the Norcold Temperature dial to maximum ("6"). You should double-check to be sure that your DC On/Off switch is set to "Off", because it is possible to run PID-Controlled DC at the same time as propane or AC. Since electrical power from the Norcold dial is either fully "on" or fully "off", DC and AC together creates double the heater power - and might damage your ammonia solution, or even create a crack in the sealed Refrigerator. (DC together with the Propane burner at less than full power is an intentional use of this design. DC with AC is a very bad idea.)
- - - - -
My fridge stays between 34.6 and 34.9 degrees (note that I have the thermocouple installed in a hole near the coldest part of the fridge - near the bottom, and closer to the side with the propane heater behind it).

[rickst29]

Next, we wire the DC side (if you want to implement it). Things are different over here, because one Power Path is made independent of the Norcold selector position. An "ON/OFF/ON" switch, just before the heater input +12V wire, is used to select from Relay and Norcold output. (It's after the relay, whereas the optional AC switch was placed before the SSR. It's also mandatory to have some kind of switch either ON/OFF or ON/OFF/ON, in order to turn off DC power.)

We create an entirely new path to "power up" the DC Heater, and it starts with with a new "Y" connector (Red twist connector), inserted immediately after the big fuse which Norcold installs on "their" +12V input power line from the 8-pin "straight through" set of connectors (4 by 2) between Norcold and TM wiring. Use this wire as the 12V source, because it is protected by double fuses (one at the TM power distribution panel, and another immediately before your Cut).

3 wires in the red connector are: (1) "Hot wire" from Norcold interconnect bus, with the fuse; (2) The other end of the wire you just cut, leading up to the Norcold selector switch (and other Norcold wiring); (3) A new wire, connecting to a "load" terminal on the automotive Coil-based Relay. (This is the kind which makes a little "click!" noise when it turns on or off.) Go ahead and make these connections now.

Next, Now cut the Norcold DC heater +12V lead, quite far away from the heater unit, and rewire into the ON/OFF/ON selector switch as follows:

• Pin 1 connects a new wire from the other power terminal of the Relay;
• Pin 2 is Common, the cut wire which goes to the DC heater coil;
• Pin 3 is the Norcold +12V "Power to Heater" wire (The other end of the wire which you just cut).

Whenever the Selector switch has the Relay circuit active, you WILL have DC power going to the fridge heater (even with the Norcold selector switch set to "Off", or "Propane". Even in storage. And even in an RV park where you have turned the interior selector switch to "AC" - you will be running both heaters, and that is a really bad thing(tm).

But it has big benefits. First, and maybe most important, it gives some Boondockers, with lots of Solar power available, a chance to use propane via the Norcold switch, and augment it with Solar-sourced DC power to achieve perfect temperature control. If you mount the switch in place of your CATV input port, then you can also take care of the "I'm towing all day and forgot to turn the Fridge to DC!" scenario by popping the front shell up and opening the switch cover - that's a lot less work than opening both shells, assembling the door, and then walking in to change the fridge panel setting. But most important of all: No more frozen beer, even on DC. Because it's controlled by the same PID, 12VDC power achieves the exact same temperature management as 120VAC: in my case, 34.6 to 34.9 degrees in the coldest part of the fridge.

[ThePair]

Could you post the settings you used for the PID? I know we discussed some initial thoughts, in the past before you made the mod yourself. Now that you've got it running, what specifically are you using?

[rickst29]

I wasn't thinking about it, and threw away my sheet without providing those numbers here. I'll poke at the PID and get a listing. Although I did create "always off" and "always on" settings, they are never reached. The key values are the mode (cooling, of course), "target temperature" and something with a name like "maximum deviation desired".

But your older one might not have the same firmware as my newer one - these values did not exist in the manual at the time you bought yours. Interestingly, the PID never allows the Fridge to reach "target temperature". Instead, it keeps temps very tight, just slightly above "target" minus "maximum deviation".

One thing I do recall: Since it uses "kitchen-toaster-like" resistance heaters, it's desirable to have a VERY slow rate of On/Off switching (less stress on the wires). It also turns off the power and "thinks" for about 3 seconds per control period startup; you want to minimize that down-time in extremely hot weather. And, depending on where you installed your Relays and PID, you could hear noises during those the switch-off, think, switch-on cycle. So I set the control interval just below the maximum value, at either 198 or 199 seconds.

Photos of my install are coming, too. I might, in the future, replace the 12V line from the TM distribution panel with a new connection, tapped directly from the line which connects the Battery to the WFCO Distribution panel. That would avoid a voltage drop of about 2% along the Battery-to-Panel and Panel-to-Fridge +12VDC wiring. Along the lines of ScrubJayNests's mod, but with an "extra" 20A fuse in front of the supply line which now drives the Fridge: here.

(The TM wiring has one fuse at the WFCO Panel, before going into the Fridge wiring junction. My DC implementation currently depends on that fuse, when bypassing the Fridge Fuse for the power connection in either "DC DIrect" or "Propane+ DC Direct" Operation. If going directly from the battery, I must replace that fuse with another.)

[ThePair]

I'm amazed your temperature values are so tight. If my wife opens the fridge and pokes around in there a bit getting food for a meal, I'll see a 2-4 degree rise on the temp, at least, followed by the appropriate drop when the door is closed and the heaters on the back kick on full time. With the door closed I'm seeing an approximate 2 degree swing, probably around my setpoint, so that's not an issue.

I'll be very interested in your settings, even if they're somewhat different. You seem to have changed some of your initial ideas viz: the settings, since you originally felt the SSR could handle fast switching, and now you're thinking the time should be longer? Or is that just a difference in the programming of the two PIDs?