Not a member yet?
Click here to register.

Forgotten your password?
Request a new one here.
You must login to post a message.

11/25/2021 1:17 AM
JackedUp and CapnWhidbey, Welcome !

11/21/2021 1:56 AM
Welcome, pimaCanyon and Chris Porter!

11/18/2021 3:29 AM
locotoro, ionpau and gougege, welcome cup

11/16/2021 12:53 AM
RidgevilleRoaster OH, _lg3, Yirgaman, Houncer and AgressivStreetLamp
welcome !

11/15/2021 9:57 AM
Welcome, jasper_8137, goldyy9 and Evermancg !

In Memory Of Ginny

Latest Donations
JackH - 25.00
snwcmpr - 10.00
Anonymous - 2.00
Anonymous - 5.00
Anonymous - 5.00
Users Online
Guests Online: 18

Members Online: 0

Total Members: 7,359
Newest Member: JackedUp

View Thread

Who is here? 1 guest(s)
 Print Thread
Another Propane Gas Control Method
Per request, here are the details of the burner setup being used on my drum roaster build. Which can be found here: http://forum.home...ad_id=4083

This gas control scheme was proposed by allenb a couple of years back (http://forum.home...ad_id=2127) which utilizes a solenoid valve along with needle valves in order to create a high fire/low fire burner setup. With this arrangement he low flame is always on, and the high flame kicks in when the solenoid valve is energized/opened. A PID controller or Arduino TC4 can energize a SSR which in turn energizes/opens the valve. In practice, it works the same as powering/controlling an electric heating element. The two needle valves allow for independently fine tuning the high/low flame states (BTU outputs), and will behave as pre-sets without the need for adjusting once they're dialed in to the roasters BTU requirements.

For my setup I'm using propane torch regulator that mounts to a standard 20lb propane tank. Downstream from regulator is a manifold for the emergency shutoff & needle valves, another for the pressure gauge junction, and a clippard solenoid valve. The tubing runs are made using 3/16OD aluminum "pilot" tubing with SAE flare fittings. Refer to the attached photos & diagram for a better understanding of the layout of the various components.

To determine the ideal burner BTU output I fiddled around with the BTU calculator found here: http://ronreil.ab...ign2.shtml. I settled on a system pressure of 10psi, with an orifice diameter of 0.025" which gives a maximum of 26k BTU output. Thats more heat than I'll possibly need, but the control scheme allows it to be turned down to the desired output.

I'm using this solenoid valve made by Clippard: http://www.clippa.../ET-2-12-H.

To make the burner orifice I drilled a hole in the center of a 1/8"fnpt cap fitting with a #72 bit. Which connects to a 1/8" npt to 3/16" flare fitting. Sandwiched between the two are washers drilled with holes for the adjustable air intake/venturi.

The burner is made from 3/4" x 12" black iron nipple with a 1 1/4" to 3/4" reducing fitting on one end, and cap on the other. The venturi is attached to the reducing fitting with screws threaded into the end of the fitting. A set screw allows the venturi to be locked in place once the optimal air fuel mix is found. To my surprise, it worked on the first try with a stable (blue with small yellow tip) flame at both high & low pressure. The flame holes are slots cut 1/4 of the way into the pipe diameter with a hacksaw.

Here's a video of the setup:

The toggle switch is used in place of the Arduino TC4/PID & SSR in the test video, but you get the general idea. I'll make another video once things are installed in the roaster.

Not mentioned are the safety features of this setup. I'll update this thread with the details once they are implemented.

Lylabrown attached the following images:
venturi_1.jpg setup_4.jpg btu_calculator.jpg controldiagram.jpg
Interesting concept. It will be good to hear how this setup works with the roaster.

I like this set up, should work on my pipe burner. It would be nice to automate the roasting process.
All you need in life is ignorance and confidence, and then success is sure. Mark Twain
I, too, like this concept.

Does anyone have a feel for how frequently the solenoid can be "pulsed" before a problem is caused?

The PWM16 library for the TC4 system presently supports a PWM time base only as slow as 4 seconds. Is that going to be too many cycles for the solenoid?

Hey Jim, I was wondering the same thing. The spec sheet says it's rated for continuous duty with a cycle life of over a billion. There was no recommended pulse or cycle intervals, as is the case for most solenoids. I emailed clippard with that question this morning, here's their response:


The EV line is rated continuous use, you can cycle it as fast at 10ms 24 hours a day.

That's encouraging. I guess it makes sense when considering how low the power draw of the solenoid is (0.67 watts).

Here's a neat thought:

If the valve can do 100Hz (50Hz full on/off cycle) at max, running it at 4Hz would be no problem. If that's the case, 4Hz switching in tandem with the inherent system hysteresis (lag from pressurized gas in the tubing between the valve & burner orifice) would create virtual proportional control. Could that be right?

It would be easier to just do digital proportional control, or PI, or PID control, wouldn't it? My guess is that PWM with a 1 Hz time base would be plenty fast for pretty good temperature control of a roaster.

Of course, I know absolutely nothing about the lag or hysteresis associated with propane valves and piping Shock

You're exactly right Jim. Here are a few test videos of different cycle rates.

1Hz -
2Hz -
4Hz -

I don't think the beans will care either way.

Hey Russ, very nice setup. This should give you excellent control of your roaster. I think what you've accomplished will not only give your roaster perfect control but will show others a way to easily control their gas fired fluidbed with little to no noticeable ET swing.

One thing I noticed when taking a look at my burner setup again is because of the small 1/8" OD tubing I used and the length of it, I have a greater hysteresis lag when the valve switches back and forth between hi and lo states. I'm wondering if one were to experiment with various size orifices or a needle valve located at the outlet of the solenoid so you could dial in an optimum amount of hysteresis. With just the right amount of hysteresis there would be a smoothly varying heat output which would produce the proportional effect you're looking for with using on/off PID.

Edited by allenb on 08/31/2014 11:23 AM
1/2 lb and 1 lb drum, Siemens Sirocco fluidbed, presspot, chemex, cajun biggin brewer from the backwoods of Louisiana
Thanks Allen, I was about to start tinkering with the pressure & orifice sizes when another idea came to mind: You can dampen pressure spikes by increasing volume.

To test the idea - I added about 6' of 1/4' ID "type T" tubing to the system between the solenoid and the pressure gauge manifold (see photo). Here are the videos of before & after adding the tubing:

2Hz Before -
2Hz After -

I'm blown away by the improvement the extra tubing has made. The pressure spikes are now just ripples and the flame is much more stable.

Lylabrown attached the following image:
That's another way to go about it and seems to work well. I had given some thought during my build to incorporate some sort of snubber using a small cylinder with tubing connections on each end similar to a refrigeration system suction line dryer (somewhere around 2" long by 2" dia) and from the looks of how your extra length of tubing worked, it would have most likely done the trick. I would like to see if a restriction, post-solenoid, would work as well. After pondering it more, I'm not certain it would result in increasing the hysteresis but would be worth a try as it would be a cleaner solution.

1/2 lb and 1 lb drum, Siemens Sirocco fluidbed, presspot, chemex, cajun biggin brewer from the backwoods of Louisiana
Something to consider on choosing your pwm time base is the faster you cycle it, even though the valve will handle it, it will have a drastic affect on how long your pressure gauge lasts. But, if there's a large hysteresis they can handle much faster cycling than when there's abrupt pulsing.

FWIW, I've found with my drum roaster that cycling every 2 seconds is more than adequate and the drum surface and inlet air to the rear of the drum don't see any temperature swings during the hi-low pulsing.

Faster pulsing would be advantageous for a fluidbed though.

1/2 lb and 1 lb drum, Siemens Sirocco fluidbed, presspot, chemex, cajun biggin brewer from the backwoods of Louisiana
Russ -

Very neat stuff!

After watching your videos, I might be inclined to speed things up. A faster PWM time base, together with the extra hose length, might smooth out the response to the point where it behaves like a proportional valve.

Maybe going all the way to 60Hz is feasible? I don't want to take this thread into the weeds, but the ICC control used in aCatauiQ and aArtisanQ_PID might work for you.

Hey Allen,
Agreed, the added tubing proves the concept, but isn't the most elegant solution. I like the idea of using a pulse dampener of some sort. The purpose built devices are very expensive, I can hack one together using some 1" black iron pipe and fittings. It would behave like a water hammer arrestor used in household plumbing. As long as it's the low point of the system and purged of all air -- it should be safe in the rare, but possible, event of flash-back behind the burner orifice.

Also agreed, cycling the solenoid 1Hz or slower would be more than adequate for good roaster control. It's fun to experiment with the various options though.



JimG wrote: After watching your videos, I might be inclined to speed things up. A faster PWM time base, together with the extra hose length, might smooth out the response to the point where it behaves like a proportional valve.

Maybe going all the way to 60Hz is feasible?

Hey Jim,
That sounds familiar Grin.
Here's a test using the 60Hz "Fast PWM" intended for the fan in roastlogger:

Video link:

60Hz is slightly faster than the solenoids response time for a full cycle (50Hz). As a result, it sticks "off" below 15% and "on" above 85%. Other than that, it worked fine. The valve didn't even get warm after 20 minutes of testing. Impressive.

I wonder how it would respond with 20 and 30 hz? You guys may have come up with a means of tricking an on/off valve into becoming a proportional valve. Very interesting experiment!

1/2 lb and 1 lb drum, Siemens Sirocco fluidbed, presspot, chemex, cajun biggin brewer from the backwoods of Louisiana
I thought more on the issue of gauges not liking rapid pulsing. Something that is commonly done in refrigeration and hvac is to install an orifice fitting just behind the gauge that effectively cancels rapid needle fluctuations but allows it to react fast enough. Not sure what size orifice would be needed but would probably need to be quite small.

1/2 lb and 1 lb drum, Siemens Sirocco fluidbed, presspot, chemex, cajun biggin brewer from the backwoods of Louisiana
Hey Allen,
It looks like I can get an Ashcroft "pulsation dampener" (http://www.dfs-ga...s/50+1106B) on theBay for under 20 bucks shipped. It uses a little pin that acts like a piston to help cancel out the pulses. Also, I'm currently running the gauge "dry", I'll fill it with some propylene glycol soon. That should help smooth the pulses as well.



Lylabrown wrote:
60Hz is slightly faster than the solenoids response time for a full cycle (50Hz). As a result, it sticks "off" below 15% and "on" above 85%.

Oops. I lost track of your mains frequency ;-)

Try adding a few lines of code to user.h in RoastLoggerTC4. This will give you several more options to try.

I would expect the 49.92Hz setting to work well. I would also try the 25.00Hz setting.

The crystal on the TC4C is normally quite accurate, so these will track mains frequency well. I think tracking the mains frequency more closely will reduce the blocking at low and high output settings.

// choose one of the following for the PWM time base for heater output
#define TIME_BASE pwmN4sec  // recommended for Hottop D which has mechanical relay
//#define TIME_BASE pwmN2sec
//#define TIME_BASE pwmN1Hz
//#define TIME_BASE pwmN2Hz
//#define TIME_BASE pwmN4Hz
//#define TIME_BASE pwmN8Hz

// --------- added code follows -------------------------

//#define TIME_BASE pwmN60Hz // 60.096 Hz
//#define TIME_BASE pwmN50Hz // 49.920 Hz
//#define TIME_BASE pwmN32Hz // 32.018 Hz
//#define TIME_BASE pwmN30Hz // 29.990 Hz
//#define TIME_BASE pwmN25Hz // 25.000 Hz
//#define TIME_BASE pwmN20Hz // 20.006 Hz
//#define TIME_BASE pwmN16Hz // 15.993 Hz
//#define TIME_BASE pwmN10Hz // 9.9968 Hz
//#define TIME_BASE pwmN8Hz // 8.0005 Hz
//#define TIME_BASE pwmN5Hz // 5.0000 Hz

If this continues to show promise, it might be worth doing a little coding to shorten the firing pulses down to 1 ms or so. I think a shorter firing pulse for the SSR would also help with the blocking.

EDIT: Upon further review, the above will probably not help while using the present PWM code in RoastLoggerTC4. What is really needed is to change the control method to "integral cycle control" (ICC) or "burst mode" where short pulses are timed to coincide with the AC sine wave. Sorry for confusing the issue.


PS - thanks goes to Bhante for doing the math on several of these values.
Edited by JimG on 09/02/2014 11:32 AM
Hey Jim,
The additional pwm options for the heater are most welcome. I gave the 50 & 25Hz options a try. On 25Hz there was partial blocking at 5% & 95% but good everywhere between. At 5% the plunger doesn't make a full on/off stroke with the normal 'click' sound. You can hear it in the beginning of this video:

At 50Hz it behaves in the same manner as 25Hz, with the partial blocking at 10% & 90%, and normal in between.

Not that any of this matters for my purposes, I'm perfectly happy to have proportional control in 10% increments. But i thought you'd like to know how this system is behaving at the different settings.

Also, I added my hacked together pressure dampener made with black pipe and fittings. By my calculations it doubles the airspace that the extra hose provided. It further smoothed the pressure spikes at the expense of response time. The pulsation dampener for the pressure gauge also arrived, and it works perfectly. The pressure gauge needle is absolutely steady & free of vibration now.

A big thanks to you and Bhante for the additional pwm frequency options & backend math.

Lylabrown attached the following image:
Jump to Forum:

Similar Threads

Thread Forum Replies Last Post
Is automated Artisan PID control of a drum roaster a "fool's errand"? Drum Roasters 8 08/30/2021 5:47 AM
Roaster Control Wiring Fluidbed Roaster 4 07/22/2021 10:20 AM
Artisan - PID control - setting ROASTING SOFTWARE APPS 3 04/13/2021 5:41 PM
My SR540 with TC4+ Artisan Control Other Roasters 7 03/30/2021 1:07 AM
Help! Blower motor control Fluidbed Roaster 5 03/19/2021 12:07 PM
Homeroasters Association Logo, and all Content, Images, and Icons © 2005-2016 Homeroasters Association - Logos are the property of their respective owners.
Powered by PHP-Fusion Copyright © 2021 PHP-Fusion Inc
Released as free software without warranties under GNU Affero GPL v3
Designed with by NetriX