Announcement

Collapse
No announcement yet.

Thermologging the Sumbeam EM6900 thermoblock

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

  • Sparky
    replied
    Re: Thermologging the Sumbeam EM6900 thermoblock

    Yeah, once the group heats up its stable as. I took a bit longer to run the last shot and you can see the group had cooled a bit more. The real deal EM6900 with PID control and some water preheating would be ideal, as the group is part of and so heated by the thermoblock.

    I have some improvements in mind, but am fighting the thermal gradient between the thermoblock and the group. Im going to try a warming flush before the shot to see how that helps things.

    If Sunbeam would lend me a real EM6900 and a spare PF so I can cut out the bottom and make it naked, then I reckon Id be able to achieve Synesso-like stability and repeatability.

    I have actually seen a flat profile when doing a 30 ml pour. So it looks like at half the flow rate the heating and cooling balance each other.

    Still its fun to play around with these things and gain a greater understanding of how the technology works.

    Leave a comment:


  • luca
    replied
    Re: Thermologging the Sumbeam EM6900 thermoblock

    Thats pretty freakin amazing, Mark. Any chance of you getting to thermolog the real (unmodified) deal?

    Leave a comment:


  • Sparky
    replied
    Re: Thermologging the Sumbeam EM6900 thermoblock

    I have now controlled the preheater with a PID controller. It hasnt been optimally tuned for the application but it serves to limit the preheating temperature and doesnt seem to be too critical. This is a good thing for using passive preheating.

    The lower preheating temperature has dramatically reduced the intrashot variation to around 2 C. However you can see from the shot sequence that the group starts cold and needs one or two blank shots to heat it. This is an artifact of my design. The Sunbeam EM6900 has a group that form part of the thermoblock and shouldnt need too much heating.

    You can see from the plots that the temperature still rises during the shot. This seems to be due to the cold group and hot thermoblock. The preheater has a small effect as well as it usually switches on about half way through the shot. I have a few ideas about flattening these profiles even further. However one thing I notice is that once the group is heated (first two shots) the repeatability is remarkable. This is really what a thermoblock is all about. Getting this level of repeatability with a boiler is considerably harder due to the large thermal gradients within a boiler.

    Cheers,

    Mark.

    Leave a comment:


  • Sparky
    replied
    Re: Thermologging the Sumbeam EM6900 thermoblock

    Originally posted by Javaphile link=1161475339/0#2 date=1161476703
    Have you thought about the possibility of hooking up a 2nd thermocouple by the exit from the thermoblock and using its reading in conjunction with the one on the input side to regulate the heating? If that can be done it would eliminate the need for the preheating setup.
    Yep preheating is not optimal if it requires twice the number of gadgets. For the prototype, the PID pre-heater is the best way to go forward. As for the control point, I had planned to move it to the thermo-block/group interface to more accurately control the brew temperature (similar to what I achieved with the Faema). I dont think using two sensors to control one element will give you anything unless the control algorithm is significantly modified beyond what PID controllers are capable of. That would require some control theory considerations. With the PIDed preheater, Ill be able to test a range of configurations to see what works best and what is the simplest cnfiguration that gives adequate control. Thats the role of this prototype.

    Ultimately Id like to use passive pre-heating from a steam boiler to reduce the load on the thermoblock. Then the PID controller should have no trouble effectively adapting and controlling the load. If I can manage that, it will feature as one of the groups in my Bezzera "Hydra" project.

    Cheers,

    Mark.

    Leave a comment:


  • Javaphile
    replied
    Re: Thermologging the Sumbeam EM6900 thermoblock

    Looking good there Sparky! I love the project!!

    Have you thought about the possibility of hooking up a 2nd thermocouple by the exit from the thermoblock and using its reading in conjunction with the one on the input side to regulate the heating? If that can be done it would eliminate the need for the preheating setup.


    Java "What a great project!" phile

    Leave a comment:


  • Sparky
    replied
    Re: Thermologging the Sumbeam EM6900 thermoblock

    Even though the EM6900 thermoblock seems to work pretty well, Im not satisfied. This technology should allow far more control of the brewing temperature. In order to achieve this Ive added a preheating boiler based on an old Silvia boiler that I had lying around (fitted with a replaceable element). This allows me to run the thermoblock control in PID mode. The PID parameters for this thermoblock are incredibly fast compared to a boiler. This allows for very rapid response and tighter control (in theory). Currently the preheating boiler is only thermostat controlled using a 100 C thermostat. This is too hot for the thermoblock feed and causes the thermoblock to heat up. Here is a series of three 70 ml shots with the PID controller and pre-heater. You can see the rising temperature profile again. This time the thermoblock element is NOT running and the heating is purely due to the preheated water. Therefore in order to get better intra-shot performance Ill need to run the boiler at or slightly below brew temperature. Therefore my next step is to use a PID controller on this boiler as well so I can vary to set point and see what works best. So theres a few things to do yet, but the initial results are very promising.

    Leave a comment:


  • Sparky
    started a topic Thermologging the Sumbeam EM6900 thermoblock

    Thermologging the Sumbeam EM6900 thermoblock

    I have started looking at my EM6900 thermoblock based machine. It is based on the Sunbeam EM6900 stainless-steel lined thermoblock. I have mated it to a Bezzera BZ35/99/02 group and installed a 3-way solenoid valve. This is slightly different to how Sunbeam have used this thermoblock. In the EM6900 machine, the thermoblock forms the top of the group giving rapid heating of the group. It also uses a separate exhaust port, rather than a 3-way valve. While these are excellent design features, they were not practical to implement in my machine. Nevertheless, my machine is useful in exploring the core technology and capability of this thermoblock.

    The first thing I noticed is how fast the thermoblock comes to temperature. It only takes a few minutes. In the actual EM6900 machine with the bottom of the thermoblock forming the top of the group, this means the machine will come up to heat far quicker than anything else out there.

    I have started using this thermoblock in a mode very similar to the Sunbeam heating mode. That is a K-type thermocouple mounted in a thermowell next to the input to the thermoblock. This means the temperature control will respond rapidly to the cold feed water. In practice the controller is in on/off thermostat mode and it responds instantly to the cold feed water by turning the element full on during the course of the shot.

    Here are a series of 3 successive 60 ml shots taken as fast as I could log them. I am using a thermofilter-type device which has a K-type thermocouple probe mounted just under the shower screen and an adjustable flow valve in order to simulate single shots, double shots or anything in between. It is important to measure the brew temperature of a thermoblock at proper brewing flow rates as the thermoblocks ability to heat is very dependent on the flow rate through the block. What I noticed is that the thermoblock responds rapidly to the cold feed water and is capable of bringing the water up to brew temperature at a reasonable brewing flow rate. In fact the temperature actually climbs during the shot, indicating that the element has more grunt than necessary. However this extra grunt results in relatively poor intra-shot stability with a 4 C variation in the shot. In contrast the Silvia is capable of less that 2 C variation, especially when fitted with a PID controller. However, this is at least on par or better than a Gaggia, although the Gaggia profile falls rather than rises.

    So the bottom line for the EM6900 brew thermoblock is that it is capable of maintaining proper brew temperatures during a shot. Add to that the ultra-fast heat up time and the separate pressure exhaust port (which keeps the brew circuit clean) and you have a potentially killer technology.


Working...
X