- Nov 11, 2020
- 4
- 3
We are new to the world of food smokers, so my wife strongly suggested
that a budget of only $200 would be appropriate. I chose the Masterbuilt
MES 130B electric, which was $184 from Amazon and was delivered for free
in two days.
My bottom line: it works ok. We've made some great meals with it. But as
an engineer I have a few quibbles, some of which I can have fun fixing.
Forgive me if this is all obvious to you experienced smokers.
(1) Opening the door every 30-60 minutes to check the internal food
temperature really disrupts the cooking.
(2) The chips are heated in a small tray over the heating element, so
when the thermostat turns off the heater to let the temperature
gradually come down, smoke generation diminishes and then stops until
the heater turns on again.
(3) The smoker has a large thermal mass, so their simple thermostat-like
control ("heater off when the temperature goes above the set point, on
when it goes below") results in huge overshoot and undershoot, by 10-15
degrees in each direction, with a cycle time of about 15 minutes.
(4) The display is all but unreadable in sunlight.
Being a tinkerer, I can't resist thinking about causes and fixes for
each of these:
(1) Problem solved by using an Enzoo 4-probe temperature sensor with
thin wires that can go past the gasket on the door. No need to build
anything here.
(2) Problem solved by putting a couple of small 100W heating elements
(wired in series, so 50W power dissipation) in the bottom of the chip
tray, and turning them on periodically whenever the main heating element
is off. To do that I tapped into their 4-wire controller cable (ground,
5VDC power, heater on, and temperature sense) and connected a little
circuit with a small solid-state relay that controls power to the
additional heating elements. For design details and photos see
https://github.com/LenShustek/smoker_mods. It seems to work well.
Lots of smoke almost all the time.
(3) One common solution to feedback control for an inertial system is a
PID controller, which I see many of you recommend as an add-on, wired
into the 120 VAC heater power. Did Masterbuilt's engineer not understand
PID? Maybe not; I looked at what is in their controller: a Zilog S3
8-bit processor that has only 4K of program memory, 208 bytes of
registers, and no other read/write memory. It would be a challenge to
implement a PID algorithm with that chip, in addition to what it's doing
now. So why did they choose that chip? Because it's cheap: in quantities
of 1000 it costs only 31 cents!
(4) The story is similar for the display: their 4-digit 7-segment LED
costs less than 40 cents. Something that works well in sunlight would
cost a lot more.
I'm thinking about four possible solutions to the temperature control
problem (3). All would be compatible with the chip tray heater.
(a) Discard the Masterbuilt controller and design my own. I'd use an
Arduino-like processor good enough to implement a PID algorithm, and
include an e-paper display visible in bright light, which also solves
problem (4). I'm guessing the the parts bought new would cost about $50,
but I have most of them already. No changes to the smoker's 120V wiring
would be needed, since it just replaces the stock controller and plugs
into the existing 4-wire cable coming out of the hole in the top. It
would use the same temperature sensor, which incidentally isn't a
thermocouple; it's an NTC (negative temperature coefficient) thermistor
with a resistance of about 50K ohms at room temperature. (The advantage
of a thermistor is that there is no reference junction temperature to
deal with; the disadvantage is you have to correct for its
non-linearity.)
I wouldn't be entirely plowing new ground; here are two guys who
experimented with building a Masterbuilt replacement controller:
https://github.com/meyercm/masterbuilt_controller
https://github.com/rshields9093/SmokerMC
(b) Discard the Masterbuilt controller and use the inexpensive Japanese
Berne Rex-C100 PID unit with a new thermocouple, and a big solid state
relay with heatsink connected into the 120 VAC wiring of the unit.
Others have posted about doing this. At least with the MES 130B that
won't require drilling out rivets, as it apparently did with older
models, because there is a screwed-on access plate in the back.
(c) Discard the Masterbuilt controller and use the Rex-C100, but with a
little interface circuit that controls the "heater on" signal that the
original controller provided. That means no big relay, and no mucking
around with the AC wiring inside the unit; just plug in the 4-wire
connector that originally plugged into their controller.
(d) See if I can reprogram the CPU in their controller and shoehorn in a
PID algorithm. (I've been writing software for 50+ years and am pretty
good at space optimization.) The CPU's memory is rewriteable without
removing it from the board by using a auxiliary programmer that connect
with 5 wires, and the Zilog S3 development tools (assembler, etc.) are
available to anyone for free. It would be great if I discover that
Masterbuilt didn't set the "hide program" bit, so I can read out and
disassemble their program as a starting point. The advantage of this
approach: no new hardware is needed. The disadvantage:
reverse-engineering machine-language software is hard.
Comments? I'll let you know which path I go down, if any.
that a budget of only $200 would be appropriate. I chose the Masterbuilt
MES 130B electric, which was $184 from Amazon and was delivered for free
in two days.
My bottom line: it works ok. We've made some great meals with it. But as
an engineer I have a few quibbles, some of which I can have fun fixing.
Forgive me if this is all obvious to you experienced smokers.
(1) Opening the door every 30-60 minutes to check the internal food
temperature really disrupts the cooking.
(2) The chips are heated in a small tray over the heating element, so
when the thermostat turns off the heater to let the temperature
gradually come down, smoke generation diminishes and then stops until
the heater turns on again.
(3) The smoker has a large thermal mass, so their simple thermostat-like
control ("heater off when the temperature goes above the set point, on
when it goes below") results in huge overshoot and undershoot, by 10-15
degrees in each direction, with a cycle time of about 15 minutes.
(4) The display is all but unreadable in sunlight.
Being a tinkerer, I can't resist thinking about causes and fixes for
each of these:
(1) Problem solved by using an Enzoo 4-probe temperature sensor with
thin wires that can go past the gasket on the door. No need to build
anything here.
(2) Problem solved by putting a couple of small 100W heating elements
(wired in series, so 50W power dissipation) in the bottom of the chip
tray, and turning them on periodically whenever the main heating element
is off. To do that I tapped into their 4-wire controller cable (ground,
5VDC power, heater on, and temperature sense) and connected a little
circuit with a small solid-state relay that controls power to the
additional heating elements. For design details and photos see
https://github.com/LenShustek/smoker_mods. It seems to work well.
Lots of smoke almost all the time.
(3) One common solution to feedback control for an inertial system is a
PID controller, which I see many of you recommend as an add-on, wired
into the 120 VAC heater power. Did Masterbuilt's engineer not understand
PID? Maybe not; I looked at what is in their controller: a Zilog S3
8-bit processor that has only 4K of program memory, 208 bytes of
registers, and no other read/write memory. It would be a challenge to
implement a PID algorithm with that chip, in addition to what it's doing
now. So why did they choose that chip? Because it's cheap: in quantities
of 1000 it costs only 31 cents!
(4) The story is similar for the display: their 4-digit 7-segment LED
costs less than 40 cents. Something that works well in sunlight would
cost a lot more.
I'm thinking about four possible solutions to the temperature control
problem (3). All would be compatible with the chip tray heater.
(a) Discard the Masterbuilt controller and design my own. I'd use an
Arduino-like processor good enough to implement a PID algorithm, and
include an e-paper display visible in bright light, which also solves
problem (4). I'm guessing the the parts bought new would cost about $50,
but I have most of them already. No changes to the smoker's 120V wiring
would be needed, since it just replaces the stock controller and plugs
into the existing 4-wire cable coming out of the hole in the top. It
would use the same temperature sensor, which incidentally isn't a
thermocouple; it's an NTC (negative temperature coefficient) thermistor
with a resistance of about 50K ohms at room temperature. (The advantage
of a thermistor is that there is no reference junction temperature to
deal with; the disadvantage is you have to correct for its
non-linearity.)
I wouldn't be entirely plowing new ground; here are two guys who
experimented with building a Masterbuilt replacement controller:
https://github.com/meyercm/masterbuilt_controller
https://github.com/rshields9093/SmokerMC
(b) Discard the Masterbuilt controller and use the inexpensive Japanese
Berne Rex-C100 PID unit with a new thermocouple, and a big solid state
relay with heatsink connected into the 120 VAC wiring of the unit.
Others have posted about doing this. At least with the MES 130B that
won't require drilling out rivets, as it apparently did with older
models, because there is a screwed-on access plate in the back.
(c) Discard the Masterbuilt controller and use the Rex-C100, but with a
little interface circuit that controls the "heater on" signal that the
original controller provided. That means no big relay, and no mucking
around with the AC wiring inside the unit; just plug in the 4-wire
connector that originally plugged into their controller.
(d) See if I can reprogram the CPU in their controller and shoehorn in a
PID algorithm. (I've been writing software for 50+ years and am pretty
good at space optimization.) The CPU's memory is rewriteable without
removing it from the board by using a auxiliary programmer that connect
with 5 wires, and the Zilog S3 development tools (assembler, etc.) are
available to anyone for free. It would be great if I discover that
Masterbuilt didn't set the "hide program" bit, so I can read out and
disassemble their program as a starting point. The advantage of this
approach: no new hardware is needed. The disadvantage:
reverse-engineering machine-language software is hard.
Comments? I'll let you know which path I go down, if any.