HELP Voltage / kW / ROR all over the place

[Steve]

Having some very strange behaviour.
Line voltage is pretty solid here usually 235 - 237 can be seen on the roast log its 237. During the roast however it shows solid 184V or 182V can not remember which? while the kW is going over 1.4?
Tried a reset, reinstalled latest firmware and it has done this 3 roasts in a row.

Espresso.jpg (43.1 KiB) Viewed 6081 times

Espresso.klog

(73 KiB) Downloaded 299 times

[Angela]

Have you noticed at times 2:35 and 8:50 there are kinks in the profile curve that you are asking the controller to follow? There may be others. Is this confusing the software?

How does it do on more normal profiles?

[Steve]

Have you noticed at times 2:35 and 8:50 there are kinks in the profile curve that you are asking the controller to follow? There may be others. Is this confusing the software?

How does it do on more normal profiles?

Hi Angela, thanks for your thoughts.

Not sure what kinks your are referring to? The profile is about as straight / smooth as it gets. Its also quite a gentle gradient overall and in comparison to something like what Ninja profile does (thats what I would call a big kink) it is a walk in the park.

Not sure what a "normal" profile is, i am working on spending specific amounts of time in certain segments of the roast to highlight different attributes.

The major point is that in the "technical info" during a roast it was reading 184V or close to that while the mains is actually 235 - 237 with the kW goiing over 1.4 and this roast does not need much over 1.2 to complete.

[kaffelogic]

Having some very strange behaviour.
Line voltage is pretty solid here usually 235 - 237 can be seen on the roast log its 237. During the roast however it shows solid 184V or 182V can not remember which? while the kW is going over 1.4?
Tried a reset, reinstalled latest firmware and it has done this 3 roasts in a row.

Hi Steve,

Strange, but not so strange.

The profile software has an 'intelligent' feature where it tries to guess the best scale factor for the Y-axis. You have designed a profile where the RoR never goes above 25 deg/min, so the software has scaled the RoR values by 10. Actually this is extremely unhelpful of the software because it creates the kind of scary looking graph that you are seeing. The data is perfectly normal, but the presentation of the data is strange. I will fix this in the next beta release to lock the scale factors to the standard values. Meanwhile, now that you know what is going on you can relax.

The software also adjusts the power to be equivalent to power at 20 deg C ambient. The system has determined that the ambient temperature is 28.3 deg C so it has taken the actual power level and increased it for display to be equivalent to what it would be if ambient was 20 deg C. This is intended to make one profile log more directly comparable with another. The problem is that the roaster doesn't have a separate ambient temperature sensor and has to deduce the ambient temperature as best it can. It can get it wrong by overestimating up to 8-10 deg too hot, especially with follow-on roasts where the roaster has been turned off in between roasts. You can tell from the shape of the power curve (black line) that it has not topped out otherwise the line would have distinctive flat topped zones. This means it still has some power in reserve, although potentially not much. Your profile actually needs more power at that stage in the roast than other, faster profiles.

One approach could be to try dropping the fan speed through the parts that need more than 1.3kW. Also, zooming in on your first crack zone there are some exothermic/endothermic things going on that suggest you might benefit from some negative boost zones in the profile leading up to first crack which could also limit the peak power drawn.

I hope these comments help.

Chris

[kaffelogic]

The major point is that in the "technical info" during a roast it was reading 184V or close to that while the mains is actually 235 - 237 with the kW goiing over 1.4 and this roast does not need much over 1.2 to complete.

During the roast 'technical info' will read incorrectly (a known issue). So please ignore supply voltage readings during the roast.

[kaffelogic]

Have you noticed at times 2:35 and 8:50 there are kinks in the profile curve that you are asking the controller to follow?

Not sure what kinks your are referring to?

I think Angela is referring to the little cusps in the RoR line at those points. Mathematically these are 'kinks', but in practice the roaster will follow the profile line and such 'kinks' get evened out and lost in the bigger random things that are going on anyway.

[Steve]

Thanks Chris that is very helpful I will try and tweak it and see what happens.

I guess it's slightly counter intuitive in that I have a shorter / faster roast below with the same sort of straight line approach and it works perfectly, both 100g batch size. I would have thought this would require the higher power to push the roast faster but as you say I may have had ambient temp reading issues (it was quite warm in garage last night).

Cheers

Filter4.jpg (36.36 KiB) Viewed 6062 times

[kaffelogic]

Regarding ambient temperatures: the roaster determines ambient temperature when it is turned on.

It initially assumes that it has been turned off for long enough to allow the temperature probe and internal circuitry to come to thermal equilibrium. If it detects a difference between the temperature probe and and the internal circuitry temperature it assumes it is not at thermal equilibrium and discards that measurement, using the last saved ambient temperature reading instead.

So, if you haven't used your roaster for several hours it will get the correct ambient temperature when you turn it on. If you last used your roaster just a few minutes ago it will detect a difference between the temperature probe and circuitry from retained heat, and will revert to the last saved ambient temperature.

The problem emerges if you last used your roaster 15 mins ago or so. At this time it will have settled to an ambient temperature that looks good to the roaster, but is actually 8 to 10 degrees too high. It turns on and measures this 'incorrect' temperature. This causes it to apply too little actual power to the heating element, and you can see the result when you compare the two logs you have posted above: one shows the roast starting out above the line and settling down onto the line nicely (ambient temperature presumably correctly measured) and one shows the roast starting out under the line and coming up to meet the profile curve from below (ambient temperature presumably measured too high). In general this difference doesn't matter because it gets back on to the line very early in the roast. But, when aiming for a very high level of roast replication it is important for the ambient temperature to be as accurate as possible.

The easy way to do this is to leave the roaster on between batches if you plan several batches in the same session. Of course, if you are using your roaster continuously all day, and the ambient temperature changes dramatically during the day then even this technique won't work. However, it will give you the best ambient temperature measurement you can get. (A separate ambient temperature probe will be available as an accessory in the future.)

[Angela]

When I expanded or zoomed in to your profile - the blue line showed showed sudden direction changes at the times specified. It may, of course be an artifact of the graphing software, but I hadn't noticed it before on my graphs.

If you only created the profile in KLPMS then it is probably an artifact caused by over magnification of the screen view.

[Steve]

Thanks again Chris for that detailed explanation and it actually helps me understand much better!

tweaked the Espresso profile so the ROR was a touch higher at the start closer to 30 and 1st roast perfect.
The most I will do in a session is 5 or 6 in a row with resting time between, I have been turning the roaster off in between roasts though which after reading your explanation may have also been contributing factor, so will now just leave it on or until it says turn me off now

Thanks again for your timely support.

Espressofixed.jpg (37.68 KiB) Viewed 6059 times