Ok, thanks, I only asked for clarification because a PID sends on/off signals to an electrical heating element as well and I wondered if it were different for gas supply. I imagine there is a pilot light loop on the supply side of the solenoid as well. Are there drawings for this or any similar gas powered sample roasters available?

As to the calculation of arbitrary bean weight from previous roasts, you could find out the gas flow settings for a large batch and a small batch then do linear interpolation/extrapolation to know the right heat input for arbitrary batch size. That eliminates heat absorption by the roaster from the equation. Of courese it wont be perfectly liner but it should get you pretty close and the more roasts you do between the high and low batches, the better you'll understand the correlation.

kmills wrote:Ok, thanks, I only asked for clarification because a PID sends on/off signals to an electrical heating element as well and I wondered if it were different for gas supply.

Yes, the two applications would almost certainly be quite different. With electric heaters you can just vary voltage and current to vary the heat input (and they tend to be linear for certain ranges). For a gas burner you'd need to run a proportional valve with an actuator (like maybe a stepper motor).

The US Roaster does have a pilot with auto-ignition and the gas is tied into a safety interlock loop that requires the motors to be running (fan and drum); and has a hi-limit safety shutoff with the sensor in the exhaust path.

kmills wrote:As to the calculation of arbitrary bean weight from previous roasts, you could find out the gas flow settings for a large batch and a small batch then do linear interpolation/extrapolation to know the right heat input for arbitrary batch size.

But I'm guessing that the issue isn't just heat absorption by the roaster.

If you saw Henry's post above, he advised turning off the burner when I drop in the load and letting the drum and beanload equalize to the point where the bean temp reading starts to rise, and then re-igniting the burner. This correlates with my earlier post describing a 4-minute roast to second crack based on a high drum starting temperature.

So, during the first part of the roast I would expect to need a very different preheat for a light load as compared to a heavy load. Also, it seems to me like the heat transfer to beans for the full load is primarily conduction by contact with the drum but it seems to me that as you reduce the load there might be less direct contact (conduction) and more exposure to hot airflow (convection) because beans are free to fly around more and there is more space for air contact as the load gets smaller and smaller. Of course this is pure speculation on my part, maybe the whole thing operates linearly independent of beanload.

There may be some linear extrapolation that might reasonably model the system (or perhaps a combination of linear approximations to represent the different phases of the roast). The only way I'll learn is to install a flowmeter and keep a log as I try different bean loads.

Maybe later. For now I am just trying to get consistent with a more fixed set of conditions

So what valve does your roaster have? Is it easy to tell the make and model or input signal? If it really is just an on/off signal form the PID should it not just open and shut the solenoid when under/over temp? Does the PID actually use the PID algorithem or is it just used as a selectable deadband thermostat? Sorry to drag off topic and congrats on the cool roaster.

kmills wrote:So what valve does your roaster have? Is it easy to tell the make and model or input signal?

No, I would have to partially disassemble the unit to get at the valve.

The gas line is on a solenoid for ON/OFF so the valve doesn't get any electronic signal (it is just a manual needle valve). And as I indicated earlier, the controller is set for ON/OFF signaling based on a programmed setpoint (and it has an alarm that is settable as well). The burner is ON until you reach the setpoint and then it turns OFF. I have the setpoint well above the temp I will finish at, so I'm running the roaster manually.

In order to implement a PID algorithm I would need to figure a way to account for the lag in response due to conduction through the drum, (etc.). I have been thinking about getting a copy of LabView so I can use the laptop as a PLC and that would allow me to preprogram profiles and ramps, etc (after getting a valve actuator that runs from 4-20 mA signal). Again - that will be another time....

Well....now that I've had some time to play with this new machine and become more aquianted with it I can now respond to Henry's super-helpful post and maybe we can resume a dialog.

Note to Germantown Rob - if you see this would you like to chime in with comparitive discussion of the Deidrich IR Lab roaster? I'd love to see thoughts from some of the other gas-fired roaster people....

chang00 wrote:Consider roasting in centigrade, because the rate of change is mentally easier to conceptualize. Aim for about 10C/min before 1st crack, and about 5C/min after 1st crack. Load at ET of 130C, turn off gas. The ET will drop until turning point, then turn on gas. This generally takes about 1:10 to 1:30 min.

Well, I have chosen to stay with Fahrenheit temperatures, and my exhaust temp is an analog gauge so it's probably not useful for controlling the roast. My roaster has a BT probe sitting in a thermal well that I am using, and I have adjusted the readout offset so that I generally get first crack at a reading of 385-ish °F and second crack reads between 417°F and 420°F for me. I do not have an environment temp.

Also, there is a fan to draw the air through and this fan has variable speed control, but there is no damper. Perhaps this is one reason why it takes roughly 60 seconds for a ramp change to be fully realized on this unit, also perhaps due to the heating system and/or maybe the thickness of the drum.

This weekend I did a little test by shutting off the gas completely after first crack went silent and the roast took a full minute to stall so that's how I arrived at the time.

I'm dropping in between 280°F and 290°F and pinching the gas down way low (rather than shutting it off). I see the turn at 1:30 TO 1:40 (usually between 220°F to 230°F) and adjust the gas to a higher level so that I'm ramping at around 20°F/minute.

I want to try to incorporate more of your advice though, and so I'm going to test with starting out at lower heat from the turn and increacing it after drying....and then backing WAY off just prior to first crack.

Up until now I have been pinching down about 50% around 90-100 seconds before starting first crack, but I plan to try pinching down 80-90% about 45-60 seconds prior and then re-establishing a new ramp for the last leg of the roast (for lighter roasts - presspot brewing). I want to find the sweet spot to get 10°F/minute from first crack to the end of roast so I can extend my lighter roasts without stalling.

Thanks again for your help along the way....and I hope to see more from the other gas drum roasters out there!

JonR10 wrote:....and I hope to see more from the other gas drum roasters out there!

Glad to see you have made some progress, Jon.

With luck I'll be picking up my new 1kg Diedrich the end of this week. If the weather or production problems intervene, then it will probably be the end of next week. I don't know how long the installation is going to take, but I could be up and running in less than 2 weeks from today. I have laid the groundwork for an easy install, but unforeseen problems often intervene on projects like this . . . . . the people at Diedrich have been very helpful.

Yesterday I did what I hope will be my last roast with my 1lb sample roaster, which has served me well over the years. My freezer is well enough stocked with roasted coffee that I could go up to 4 weeks without roasting, if need be. If some installation (or other) disaster intervenes, I guess I'll have to consider (gulp) ordering some roasted coffee.

There's an enormous difference between using any larger gas drum roaster vs. a small consumer level appliance. The major thing one can take away from experience with smaller consumer appliances is the general theory of roasting and the sorts of roast time-temperature parameters one wants to achieve. Beyond that, not much is transferable from say, a Hottop or an air roaster, to a larger gas fired drum. I am hoping that the transition from what I'm using now to my soon to be Diedrich, will be smoother than that

ken

Jon, I can give you some tips from my drum roasting, I am quite pleased with the results I am getting now since owning the Ambex for 1 1/2 years...I've had the huge benefit of an experienced roasting friend
(mivanitsky) who has a Diedrich HR1. The main difference with my roaster is Ambex upgraded them about 3 years ago with a digital gas controller "photo here" it certainly would help you get repeatable results, even just the addition of a gas pressure gauge if yours doesn't already have one. The gas control is via a knob on the panel and correlates to 0-9 on the pressure gauge, so you could assume 3 is around 30% gas pressure. I'm glad I did not have to, but if mine was a previous roaster with just a gas on/off toggle I would certainly be roasting my batches at max capacity because it would be too difficult trying to slow smaller batches.

I drop @ 300F and usually see the turn around 1:15 and 165-175F, I have the gas control set at the pressure I will use for the roast until 1st crack, typically "ranges" 3 (decaf or soft bean) to 4.5, I am roasting 50% capacity 1kg loads most of the time. Especially for espresso roasts I will shut the gas off for 45 seconds at 340F then back to initial setting to before 1st crack (I had previously been setting the PID to 340F, but as you noted the lag between when the gas valve re-energizes can be annoying, so I now set my max PID temp to finish and just use the knob to go "off" with the gas...then back up manually). I will typically turn the gas down to 1 or 1.5 before 1st crack then just monitor the rate of rise until finish, many times adjusting a little up or down and then the PID shuts down at a target I set, so I can coast 5 degrees to the finish.

Ken Fox wrote:With luck I'll be picking up my new 1kg Diedrich the end of this week.

Hi Ken,
I am really looking forward to hearing about your experiences with the IR-1

danetrainer wrote:...it certainly would help you get repeatable results, even just the addition of a gas pressure gauge if yours doesn't already have one.

Thanks Pat. Mine does have a fairly precise needle valve with a gauge for regulating flow (actually I read pressure of course, but everything is relative).

danetrainer wrote:I drop @ 300F and usually see the turn around 1:15 and 165-175F, I have the gas control set at the pressure I will use for the roast until 1st crack

Zowie! That is a serious drop in temperature!
What time and temperature do you get for first crack?

danetrainer wrote:Especially for espresso roasts I will shut the gas off for 45 seconds at 340F then back to initial setting to before 1st crack

This is very nearly what I plan to try this week....

Thanks again for the input!!

-Jon

Zowie! That is a serious drop in temperature!
What time and temperature do you get for first crack?

The bean mass probe just sitting out there in the air inside the drum...then a full Kilo of room temp beans brings things down quite a lot. First crack can be achieved at just about anywhere you want, it has power to spare for sure, I am typically shooting for 8 to 11 minutes depending what I'm roasting. I can't think of a roast that I've ever experienced charring or tipping, even when I first had it and dropped around 340F (wanting my turn temp higher), then getting a piece of advice to look for 175F on the turn with the Ambex.

Correspondingly the temp displayed when applying extra heat at the end of the roast can fool you too,
as the gauge will be reading lower yet you will hear the onset of 2C You get a feel for the kinetics of the situation you're in after a while.