MES Controller Reverse Engineering

Discussion in 'Electric Smokers' started by foragefish, Apr 10, 2013.

  1. I've had my 40" since November, use the AMNPS, and have had good results with PB, pork ribs, turkey, and fish. I've often wondered why the MeatProbe (MP) reads about 20degF high and BoxProbe (BP) read about 30degF higher than a $20 Taylor (which agrees within 1F to a Omega thermocouple meter using a type-K probe).

    I took the cover off the controller, and figured out the wires to the harness. I think both the MP and BP use thermistors which are probably spec'd at 230-250Kohms at 25C(71F). The top of the thermistors are biased to the supply voltage of the controller (which is about 4.3VDC, coming from the PCB at the bottom of the MES with the relays to drive the element and the LEDs. Basically it's the output of a 7805 thru a diode, why I have no idea). The bottom of the thermistors go to 20Kohm resistors to ground, and the junction point goes to the controller chip (no idea if it's a custom chip or something like an Atmel micro)

    I recorded Taylor, MP, BP, and Omega readings with all the probes in a bottle of boiling water that was allowed to cool, and other runs with thermocouples in both the MES air and attached to the BP, and also recorded the junction voltage to the MP thermistor on the controller. I think maybe Masterbuilt assumed that they were using 250Kohm thermistors, but got some that were on the low side; I think I read somewhere that thermistors might have as much as 5-10% error from an ideal curve.

    Anyway, I wanted better accuracy in the 140-270 range, so I calculated that 20Kohm resistors to ground should be more like 15Kohms. I tacked 60Kohm resistors in parallel with the 20K to generate 15K total, and now the MP reads  6-8degF high compared to the Omega, and the BP reads 12degF high. The downside is that now at ambient the MP and BP read about 10degF low, but I don't really care about that region.

    Here is a picture of what I did:

    https://docs.google.com/drawings/d/1qeJaeZR7FoE5DYUAuYJ1dGpkpmIrOtDK0qOasatYZQg/edit?usp=sharing

    I wouldn't recommend this mod unless you are comfortable playing around with electronics. If you goof it and short out the resistors, then the controller will never turn off the element, and you'll be depending on the Klixon to eventually cut off the heating. I have yet to read about the Klixon working, but lots of horror stories of the box overheating.

    I plan to replace the controller someday, perhaps with an Ardunio.

    Mike
     
  2. daveomak

    daveomak Smoking Guru OTBS Member SMF Premier Member

    FF, morning.....  Are the  parts to make the MES controller more accurate, that much more expensive than the current parts they use ??

    Do you figure the parts they use are obsolete and free or something like that ????  

    By The Way, you are way over my head on this electric stuff.....  

    Thanks for the description, for those members that are in the know....  

    Dave
     
  3. The PCB and its components look OK. More expensive components on the PCB wouldn't help. The temp inaccuracy may be a design issue, where they didn't take into account the range of values the thermistors might come in, or there is a poor calculation of temperature from the measured voltage.

    I do wonder why they chose 250Kohm thermistors, that seems higher than necessary. The Taylor uses 30Kohm, which seems like a more commonly available value.

    Mike
     
     
  4. daveomak

    daveomak Smoking Guru OTBS Member SMF Premier Member

    FF, You want to start a business of rebuilding MES PC Boards ????   HAHAHAHA
     
  5. geerock

    geerock Master of the Pit SMF Premier Member

    Yeah, you'll have plenty of clients that's for sure.
     
  6. tjohnson

    tjohnson Smoking Guru Staff Member Moderator Insider OTBS Member SMF Premier Member

    I'm in!
     
  7. dward51

    dward51 Master of the Pit OTBS Member

    That mod is less than a nickel, way less for a company buying resistors by the reel.  You think MES would have better quality control on their end.  I would bet the design was for one brand/model of thermistor and the supply changed later and no adjustment was made to the other components to compensate.

    I'm sort of surprised there was no trim pot for re-calibration in their design.  That means they are not calibrated at all, just put together and shipped out.
     
  8. geerock

    geerock Master of the Pit SMF Premier Member

    Maybe the intent of the design was lost in the translation to chinese.
     
  9. eatemup

    eatemup Newbie

    Bumping an old thread but I used it as a guide for my mods -

    My Gen1 MES40 had similar issues as the OP's, with Maverick probes showing ~25 degrees lower temps than the MES probe that controlled the heating.  In hopes of making the same mod as above, I cracked my controller open but my circuit board was a bit different.  Instead of a DIP-package controller chip, mine has a TQFP package controller, and instead of a single 7-pin cable connector, it has two, one 5-pin and one 2-pin.  After probing around and figuring things out I determined that the controller board followed the same principles but with different values and locations.  A few keys:

    1)  My MES40 has 63k resistance thermistors, not 250k as above.

    2)  My MES40 uses 5.1k biasing resistors to ground to sense temperature from the thermistors,  not 20k as above.

    3)  The 5-pin cable includes the 5V supply, GND, the internal temp probe bias and sense lines, and one other net I didn't figure out.

    4)  The 2-pin cable includes the meat probe bias and sense lines.

    I decided I wanted to be able to take the MES up higher than 275 actual temp so I biased the thermistor past what would have made it accurate.  The bias resistors for the internal temp and meat probes are reference designators R18 and R19 in the silkscreen on the board, and they are 5.1kohm.  I placed an 8.2k resistor in parallel with the existing 5.1k to create an equivalent 3.14k biasing resistance.  Subsequent use shows this to be about just as I desired - it reads about 20 degrees lower than actual at ambient and 25 degrees lower than actual at 200-275F.  If desired to make the probes read more true to actual, a resistor in the 15k range in parallel with the 5.1k existing resistors should be about right.  The Gen2/3 may be significantly different, but this works at least for a later Gen1.

    Hope this helps folks out there if they have similar lower-than-actual temp issues.
     
  10. daveomak

    daveomak Smoking Guru OTBS Member SMF Premier Member

    Thanks for the technical update...   I'm sure others will find you work worthy and appreciate what you have done...  

    BTW, [​IMG]  ...    smoke some grub and show us in pictures, if you have time..  

    Please feel free to stop in again...  Dave
     
  11. eatemup

    eatemup Newbie

    Thanks Dave!  I've been reading for quite awhile and learned a bunch from you and everyone else that's graced us with their experience.  I'll get that Central Texas brisket dialed in this spring, with pics certainly to follow...

    Already turned on at least 3 Q'ers to Mr. T's Smoked Salmon in the meantime though...
     
  12. daveomak

    daveomak Smoking Guru OTBS Member SMF Premier Member

    Hey....  There is a whole lot to learn from our members...  These folks really know what's happening in the smoked meat world..    

    I'm lookin' forward to your "dialed in" Brisket...  Brisket can be tough to figure out...   It's a different critter compared to it's neighbors..    I'm gettin' to really like cheeks and tails...   just keep trying to figure out different ways to cook / smoke them....

    Enjoy......
     
  13. smoker21

    smoker21 Meat Mopper

    Sold!

    JD
     

Share This Page