Andy98

For anyone interested, this is the 3/4" version running with the PC fan at full (the last batch of videos were the 1/2" version). I don't think I can run the PC fan less than about 70-80% and still maintain the necessary output velocity. I believe the 3/4" works better.

Wayne, Buzzkill and Frosty  I see I have argued and speculated you all past your tolerance levels (not to mention having hijacked the thread). 

I do totally accept the stated benefits of ribbon burners and although I see that my first post here does read a lot like I was rejecting that information. It does also make intuitive sense to me - at least I did say that somewhere above.

...but I do also want to understand the mechanics of why they are better. I don't see that as a bad thing, and I don't apologize for that. I realize that some people find theoretical discussions annoying, and I do apologize for being annoying. 

Frosty: I think combustion is voodoo. I don't think I have the vaguest idea of how it works. I do think I have my fluids right (although I'm less sure I have expressed them clearly) but would welcome being told if and where I've gone wrong so I can get it right next time. I'm not surprised burning outside the forge is wasteful, but I am surprised mixed gas/air could get all the way to a forge exit before burning completely (I thought it must mean there wasn't enough oxygen in the forge).

I would love to experiment more but I have family commitments that prevent that.

On 5/13/2017 at 9:03 PM, ede said:

RIbbon burners since they have multiple outlets need higher volume but don't rely on high pressure so the flame fronts in effect are spread out(less concentrated)

Hi - In both cases the burners are required to operate with the gas/air velocity at the nozzle exit being at least the flame propagation speed. That means that a single-large-burner running at the minimum will have the gas/air exiting at the same total (stagnation) pressure as the many-small-nozzles also operating at it's minimum.

The fact that the small nozzles are diffusing the flame is probably important.

It's also possible that the many-small-nozzles allow ribbon burners to run closer to their minimum? For all I know my burner is dumping fuel/air into the forge at several times faster than the flame propagation speed.

Playing this thought out: On surface area a ribbon burner with 26 x 5/16" ports is like a 1.6" single port burner (26x5/16=2sqin, which is the same as a 1.8" diameter circle).

Most people don't have 1.8" diameter burners, they have 1" or 3/4". 

The real world will be more complicated than this, but if a ribbon burner was running at the minimum then a 1" burner would have to run at 2.5x faster flow rate to push the same amount of fuel/air. So that seems like it would mean the single larger burner will have a lot of cool, unburned gas in the forge, and it would make sense that would be bad.

19 hours ago, David Kailey said:

But for lack of equipment and having a "lab" lol. I just tracked how much fuel I used every day

Hi  - Thanks! I saw your post in the other thread with the actual fuel consumption estimations too. Interesting stuff.

18 hours ago, Buzzkill said:

In our case the splashing can end up as excessive dragon's breath, which amounts to fuel burning at the openings instead of inside the forge where we want it.

Is it your belief that the dragon's breath is caused by gas that didn't have enough time to burn in the forge?

I always assumed it was completely caused by fuel that didn't have enough oxygen inside the forge, and thus burned on exit into the atmosphere.

On 5/14/2017 at 0:12 AM, Buzzkill said:

believe there is also a difference in combustion between the two.  The bigger the burner, the greater distance needed for complete combustion.  If the flame hits whatever is opposite of the burner port before combustion is complete then it will be cooler

I think this is the real reason. If you think of the "surface area" of the flame itself, lots of little flames will have more surface area than one big jet. So I think the gas at the centre of a small jet will both slow down more quickly, and heat up faster, thus burn sooner. Total guess. If your belief about dragon's breath being related to time is true, then this would be the second part that makes it make sense (to me at least).

Your other points about pressure and exhaust exit velocity I'm not sold on, for reasons that would take lots of paragraphs of tedious explanations that would just annoy people reading (but if you're interested I'm happy to go on and on and on about it).

30 minutes ago, David R. said:

Is this done with a wood block on anvil or a seperate block? What is the set up?

All I know is what's in the video. He uses a thick wood board placed on his anvil, and just for the initial setting blows. Second heat onward are directly on the anvil or powerhammer.

He was using it for welding an oddly shaped billet made up of many pieces of cable, hence my initial assumption it was for stability. But he also does it when setting the welds on more uniform billets as well.

Most of the comments on the video are in, I think, Russian - so I couldn't gleen anything there (admittedly I didn't try that hard). I might comment and see if he replies.

12 hours ago, Will W. said:

The idea that the burning wood would heat the metal is a fallacy however, wood burns somewhere between 400-500 degrees F

I think that is the autoignition temperature, not the flame temperature.

Per this link the flame temperature in the "continuous" part of the flame is around 1500F and Wikipedia states it as ~1800F.

So not welding heats, but surely hot enough to significantly reduce the rate of cooling. Combined with the insulation effect of the wood, and the helpful pressure distribution the wood would give when you land the tacking blows, that seems useful... 

19 hours ago, Buzzkill said:

since the effectively slower flame speed

I'm confused by this. Why would the flame speed be any slower? Ribbon or not, burners generally run with the gas/air mix exit velocity at close to the flame propagation speed. No?

I'm also not clear why people believe ribbon burners will waste less heat. I'm not saying the do or don't (especially since I've never even seen a ribbon burner in person) but if I input X cfm of unburned gas/air it'll produce Y cfm of exhaust gas, and thus the same exhaust flow velocity regardless of wheather that input was via a single nozzle or a ribbon, right? So unless the ribbon burner is actually running at a lower cfm, I don't understand how the belief that the flame has longer time in the forge can be true.

I guess what it boils down to is if slowing the flame down to keep it in the forge longer works, then why would I not get the same effect turning down a regular burner? Every way I puzzle through this leaves me stymied - although the idea has some intuitive sense.

The more even heat distribution makes total sense. I wonder if people with ribbon burners run then at lower cfms because of the distribution alone?

On May 12, 2017 at 9:01 AM, David Kailey said:

I am able to run the one i have at a bout 1.5-2.5 psi. It saves me thousands in fuel costs...just more efficient.

Are you judging efficiency just by psi? That doesn't tell you the full picture. You need to know flowrate (by mass or volume). Your low-input-pressure burner might have a big orifus and thus still be delivering a high volume of gas.

Hi, continuing my practice of vicarious blacksmithing, I came across,this video:

In the video, the blacksmith sets the initial weld with the workpeice resting on a wood block (happens in the first minute).

I assumed this was for stability, but thinking about it more (while watching Mark Aspery's scarf welding videos) it occured to me that the wood would also help insulate the work piece from the heat-stealing anvil. Further, since that wood is burning it might actually heat the metal  and reduce oxidization.

Is that a crazy thought? Has anyone ever seen this approach before who could explain it?

I couldn't find this posted here before, so I thought I'd post it up. From Lewis Razors: 

He has a build video too, the most interesting step of which is hammering the brass (copper?) tube to narrow the ID sufficiently to hold the mig tip.

Cheers.

Exciting updates:

I tried a few new configurations:

1. 3/4" burner, fan attached to the "drop" of the T, natural gas attached to the "run" of the T. Air-rated stop valve added inline with shutoff ball-valve.
2. 1/2" burner, gas via the "run" and air via the "drop"
3. 1/2" burner: 3/4" to 1/2" reducing T, 1/2" mixing tube approx 4" long, air into the 3/4" run, and NG via stop-valve through the "drop" of the T.

Results:

All 3 are successful!

In all 3 cases I could turn the burner up high enough to make it stutter a bit, but it never burned in the mixing tube.

Having the NG feed in through the run (rather than through the drop) definitely increased the airflow. I assumed it would increase the velocity by about 10% but I think it did more than that. Configuration #2 would feed gas/air fast enough that I'd get a cold spot on the impact wall, whereas #3 did not have a cold spot.

This is configuration #3:

Configuration #3 is running pretty well at the top of it's potential, as far as I can tell, in this video, and seems to stutter just a bit but remained cool. This configuration was sensitive to the backpressure created by the firebrick doors, so I had make sure I didn't restrict the airflow too much.

In this configuration I was able to successfully weld this steel. I say successfully, in the sense that 1/2 the weld worked and the other 1/2 failed and I think that's my fault. This is my second ever organized attempt at forge welding, so lots of potential error on my part.

Question: 

It's not conclusive, but it seems to me that the stop valve is making the difference (and for that: Thanks everyone! I frankly never would have bothered without everyone suggesting this).

I don't understand why it is helping (and would happily take an explanation if anyone has one).

I assumed/thought that for any arbitrary position I put the stop valve in, there would have been an operationally equivalent position on the ball valve. Yet, all 3 configurations above worked successfully with the stop valve. Technically, I've not done a fair comparison (none of the 3 configs above were tested without the ball valve - I'll try that on another date).

Without the stop valve, there were occasions where I smelled unburned NG escaping the forge. So the only guess I have is that I had too much NG all along and the extra pressure drop caused by the stop valve is helping..?

Ok, so I got a chance to run the thing again yesterday. Now that I have a better idea what I'm looking for, it was easier to recognize what was going on. Here is how it's behaving:

State 1:

• Hot forge
• Gas set low.
• I set the fan at 100%

Result: I can get a good roaring burn with no visible dragon's breath (to me, in dim-ish daylight).

State 2:

• From State 1, I increase the gas:
• As the gas quantity increases, the flame goes from roaring, to sputtering, to (I believe) burning in the mixing tube.

Result:

• The mixing tube gets hot.
• The burner is very quiet (but what roar there is has become higher pitched in keeping with the idea of burning occurring in the tube).
• Lots of orange flames come out of the forge openings.

When this is occurring, if I blow into the back of the PC fan then the flame goes roaring again but only while I am blowing into the fan.

State 3:

• While in State 2
• Back down the fuel again.

Result:

• The burner continues to act as if it is in State 2.
• If I blow into the back of the fan, it will toggle back into State 1 and stabilize there.

So this is the state that confuses me. I don't know why the flame position is changing the richness of the burn. I can guess that the flamefront position is affecting the backpressure felt by the fan. I guess if combustion is happening in the tube, there would be a marked increase in backpressure and thus less air from the fan? Perhaps when I have a roaring burn in the forge I'm actually getting some air induction?

I figure that States 1 and 2 could be explained purely by richness changes - but that State 3 cannot. So in combination, I feel like I've confirmed that the air/gas velocity is not sufficient. 

Determining this was my immediate goal, since I do have the option of upsizing the fan.

Side Note: I believe I've confirmed that burner position doesn't seem to matter (I've upgraded my burner mount so I could move the burner around better).

On 4/7/2017 at 1:26 PM, Buzzkill said:

Regardless of the CFM capability of the air source, if it does not provide enough pressure to overcome the pressure created by the ignition of the fuel/air mix then the flame will travel back towards the blower - at least a little bit until an equilibrium is reached.

Yes - that's what I was getting at with the momentum thing. I'm picturing that pressure wave going backwards to the PC fan. I figure for lots of reasons (momentum, flex, etc..) that PC fan will be less capable of keeping the air pressure steady than a bigger blower. Plus, I figure that the more pressurized air (e.g., in the supply duct) the more the pressure drop would be diluted.

Kinda like a pulse jet effect.

Anyway, half the time when I think about this it makes sense, and the other half of the time it seems crazy.

On 4/7/2017 at 9:12 AM, Latticino said:

MHO you are still missing a key step that has been suggested by several parties:  install an appropriate valve that can meter your gas supply adequately.

Apologies if it appears I have ignored this advice. I haven't - I have been looking into it quite a bit, trying to see if I can find a good and reasonably priced option.

I assumed, perhaps incorrectly, that the recommendation to go a needle or globe valve was to help me adjust the flow more easily and not because it would actually meter or regulate the gas flow any better (once set) than the ball valve. I've assumed it I actually want any improvement in regulation, I'd need an actual regulator. Do I have this wrong?

All the fuel gas needle and gate valves I can find are in the $30+ range. There are $7 Water & Air rated stop valves at my local homedepot, but I've assumed it would be unwise to use them? There are cheap valves on ebay that claim to be gas rated, but I half think the "air" rated HomeDepot valve might be more trustworthy than something random and unbranded on ebay.

On 4/5/2017 at 1:34 PM, Buzzkill said:

Your PC fan may be able to move enough volume of air, but it may not have the required static pressure for the burner to function correctly.

New information:

• I did try augmenting the blower with a hairdryer. It clearly increased the airflow, but didn't change the sputter. It was a short test, and it was done while the mixing tube was overheating, but my conclusion at the time was that the extra air didn't help at all.

Random chatter - might be super boring to others:

• I know the burner can produce gas flow rates both above, and below, the flame propagation speed outside of the forge (when I ran it flareless, I couldn't get a sustained burn but I could clearly get it to point where it was too fast and would blow out my lighter, and then too slow where it would light and the flame would go straight into the mixing tube).
• Back of the envelop calcs, rooted in my measured actual value of ~6cfm (no NG) has my fuel&gas velocity at ~800cm/sec, and the flame speed is something like <60cm/sec from what I can research. So even if the burner became 90% less efficient in the forge, I'm still above flame propagation speed when running full out.
• I'd speculate that the draft inducer was running at a much higher rate than my PC fan is capable of (150%-200%?) of what my PC fan can do. It was leaving a dark spot on the impact wall even while the rest of the forge was really hot.
• I am wondering if the momentum of the fan is a problem. The rotating part of the PC fan is really light, whereas the draft inducer is a lot heavier (not heavy, but heavier). If there were pressure shock-waves then the PC fan might be speeding up and slowing down rapidly and imperceptibly?
• It also occurs to me that for the draft inducer, I have a 2" by 3' duct run. That entire duct will be filled with (slightly) pressurized air, potentially acting as a buffer. I do not have that on the PC fan build.
• Or the fan speed control could be misbehaving and not actually running the fan at a constant rate.
• Or it really is something that's happening in the forge itself, and that the effect goes away at high enough flow rates (draft inducer).

My current plan for immediate next steps (hopefully this weekend):

1. Get a reducer to use as a flare and see how the burner performs outside of the forge.
2. Get some high speed video of the fan blades and the flame.
3. Replace my burner mount with something larger so that I can move the end of the burner forward into the forge.
4. Do a better job lighting the burner so I don't let the burner tube get hot.

In any case, so long as I can at least replicate my first run then the burner is working, and I do intend to use it whether I can make it better or not.

7 hours ago, Buzzkill said:

Still, if I were you I'd change one thing at a time to try to nail down where the issue lies.  I'd want to eliminate the forge itself as any part of the equation before moving on, so that's why I recommended trying a flare outside the forge before making any other changes.

I agree with everything you've said, and the course of action you have suggested. It'll take me a while to progress it.

It occurs to me that we do have one more piece of info: The burner performance with the draft inducer (from my Feb 28 post's video) vs the PC Fan. The draft inducer appears to be running sputter free (to me). That is a single (but big) variable change: same burner, mounting, and forge - different blower.

20 hours ago, Buzzkill said:

After watching the videos and listening to the sound of the flames again I'm fairly convinced that the staccato sound happens when your flame is burning inside the tube rather than at the end as it should

I took some of the earlier videos and watched them frame-by-frame and it does look like the flame front is spreading backwards, but it's hard to tell because it happens it happens in just two frames. I did, however, just discover that my phone will do 240fps high-speed video so I should be able to get some good slow-mo of the flame front next time.

So I think you're right  - but I can't riddle out why the burner wasn't getting hot in those original videos. In all those videos I posted, the burner stayed nice and cool (with one brief exception). It wasn't until a test run on Sunday, video of which I have not posted, that the hot-burner became a problem. It still looked to me like the burner was putting itself out rather than sustaining a burn inside the mixing tube.

I guess it's possible it burned all the way back to where the gas/air was not mixed well enough to burn? Then it goes out until the next slug reaches the hot forge and reignites?

20 hours ago, Buzzkill said:

Once that gets hot then the ignition of the fuel/air mix will continue inside the tube even if you turn up the gas and air.

I'm really glad you mentioned that. That is a really good point, and is completely non-obvious to me (but makes total sense). It explains a lot of what happened on my most recent test run. I guess I have to be really careful, especially when starting the unit up when the temptation/need for slower flow is the highest.

If I put all the pieces together then I guess:

• When the gas velocity is lower than the flame speed I get sputtering but the burner tube stays cool since the flame puts itself out.
• When the gas velocity is high then I get regular burning but can blow the flame out.
• When the gas velocity is such that somewhere in the mixing tube it's exactly the same as the flame speed I get burning in the mixing tube and an overheating burner.

...I'm starting to think that the 3/4" burner needs to run at a much higher rate than my forge really requires. All of this does point to a high-power fan being a good thing - it would not only let me have higher velocities in the 3/4", but would surely be necessary if I try a 1/2" burner.

BTW, thanks again for taking the time to watch and respond - I appreciate the help immensely!

On 3/30/2017 at 2:26 PM, Buzzkill said:

have you tried lighting it outside the forge with a flare on the end of it?

Good suggestion. I don't have a usable flare at the moment, which is annoying. I did try lighting it without the flare, and I could not get a sustained burn (no surprise) but it did make me realize that at low fan speeds my gas velocity is below the flame-speed and the flame front moves into the mixing tube, while at high fan speeds I can easily blow out the flame (without a flare).

I presume that's normal for blown burners? The burner gets crazy hot if the flamefront is inside the mixing tube. Would a more typical blown burner have a minimum blower speed that is always above the flame speed?

2 hours ago, Buzzkill said:

One simple thing you could do in the meantime is move the burner tube further in and out of the forge body to see if that changes the behavior of the flame

Unfortunately that's easier said than done. I have to change out the burner support holder - the 3/4" pipe is just slightly too large to fit, so it's pushed in as far as I can push it and that makes the end of the mixer tube end about 1/2" into the outer layer of kaowool.

3 hours ago, Buzzkill said:

I'm fresh out of onions at the moment,

Oops! That gave me a good laugh. Of course I'm looking for opinions but will take onions if they are offered.

3 hours ago, Buzzkill said:

I'm fresh out of onions at the moment, but I do have an observation.  I never saw you adjust the gas flow in any of the videos.  When you speed up the fan you are changing the fuel to air ratio. 

Yes, understood. When I did adjust the gas a bit to see what effect that had on things, but that gas ball valve I am using needs two hands to make a small adjustment, so I didn't get that on video. I was trying to get a handle on what adjusting the gas does to the performance, and then separately what adjusting the air does to the performance.

Once the forge was hot, I did fiddle with trying to get the forge into a good operating mode. I think I can forge at that low idle setting. It was quite hot, plus quiet. Then the 90% setting is, I figure, where I'd need to be to weld. I have some other videos in the middle, in each case I'd tune it to try and get some orange flame out the door.

The ball valve is a giant pain and I'm going to replace it with something better (probably needle valve), just trying to source something reasonable at the moment. I didn't do it already because I'm trying to limit my investment in each step with the (until now) expectation that I'd have to abandon this at some point.

3 hours ago, Buzzkill said:

You can tune by ear once you get used to it.  The flame should produce a continuous roar.  Turn the air flow up and down to find out where the flame is the loudest.  Once you find that, cut the air back just slightly and that should give you a mildly reducing flame.

In the 90% video, would you consider that a "continuous roar"? Or just high-frequency-sputtering? I'm thinking in that video it's sputtering and falls into the "ok" but "not great" category but if you have an onion I'd hear it 

I was able to put all the pieces together and give it a run today. Bright sunlight out, so hard to see what's going on. 

My synopsis:

• Success! 
• I don't think I need a stronger blower. 
• I don't understand why it's sputtering. I think/guess I had it too lean with the blower too high. I think the sputtering is due the gas velocity being too high.

Apologies in advance, I took a lot of video.

First the good: Here is the forge running at a really, really low setting. The noise on the video that sounds like high-rate-gas-flow is actually the fan. The NG flow noise is undetectable. For some reason the camera also seems to be exagerating the fan noise. Running at idle like this, the forge was almost silent compared to the usual street noise around me.

Next up, forge running at about 90%. The exposure on this one was a little wonky. The forge lining was getting into lemon yellow in person, and the refractory floor tile had that slick melty look to it. I think this is hot enough to weld in?

This next one is really long, and unless you are really interested in this you might want to skip it. In this video I am adjusting the flow, trying to get a feel for things. I've noted in the video description what I was doing. Most interesting part is probably around 2:00 minutes when I bring it all the way down to a rolling flame.

So that's the thing running, and I think it's doing pretty well.

What I don't understand, is the sputtering. Here is the forge starting up from cold - note how much it sputters. It actually went out a few times. I could make the forge go out at will by turning up the blower. I didn't know if the forge was stuttering because the gas velocity was too low, or if it was too high. I now have the opinion that it was too high, and the mix was too lean, but I would really appreciate feedback and advice from others.

So, there you go! Obviously I need to package up the electronics, but I think this is a success.Next time I get it in my head to try an electronics related project, I have to remind myself that I hate soldering.

I'd appreciate any feedback, onions, advice, observations others might have. I don't really know what I'm looking at, honestly and can use the help.

4 hours ago, stockmaker said:

 

Hi Andy, nice work on your blower.  I had a thought that you need to be aware of.  With the fan that close to the torch body  it will be susceptible to forge heat when it is turned off after a long run.   The heat will chimney up the torch pipe and could possibly melt or warped your fan blades.  Of course I don't know how you are going to mount the burner or if it will be removable but just watch it during testing.

Yes - good point! 

My burner is side mount, so that should be somewhat helpful. I'm also not going to put an off switch on the fan it, so that should help me remember to do something smart to keep it from melting before I shut off the air.

I probably will end up removing either the entire burner, or the non-metal parts of the burner, after use. Should be straightforward, if a little annoying.

Hopefully this week will be test run #2.

I also wanted to say that Digikey is awsome: They will happily either ship me a replacement fan, or refund me the cost of the fan. I have found an even more powerful unit (this one) that I might replace it with. Will depend on how things go this week.

I've also made progress on the electronics front to control the fan speed. I had two viable solutions:

Approach #1: Off the shelf parts

Note that R1 in there is really a 0-5k pot, but LTSpice doesn't model that.

That circuit will produce a 25kHz PWM signal with a duty cycle between approx 5-98%. To get it to 100% you'd need a trim pot and tweak the R2 resistance up. The trick is making sure you don't get something like a 110% duty cycle, because that will actually cause it to produce something like a 12.5kHz PWM at ~50% duty cycle. It should be relatively easy to tune when connected to the fan, and represents about $4 worth of parts. I post it here for posterity.

Approach #2: Microprocessor

Instead of the circuit above, I built a PWM controller using an ATTiny85 that takes voltage input from a 0-10k resister, and outputs that as a 25kHz PWM wave.

The code came from here - it's straightforward once you decode it. I used an Arduino Uno as the ISP/programmer. Both of these things were pretty new to me, so that was fun to try.

It worked great! On the bench at least, I have a working circuit that, according to a USB logic analyzer I was able to borrow from a friend is giving me a perfect 25kHz PWM signal, with a range of around 1-100%. It doesn't do full zero. I could modify the program to make that work, but I don't think it's necessary so I'll leave it be for now.

This was about $3 worth of parts. There will likely be another $5 worth of stuff required: I'll need a barrel connector to plug the power in, a connector for the fan wires, a regulator to produce 12VDC for the fan, and another one to produce 5VDC for the microprocessor circuit. I would have needed those with the other circuit too.

Lack of (portable) power supplies is preventing me from testing the PWM circuit with the actual fan, and from testing the entire burner in the forge. Frustrating. But I should be able to get that all setup this weekend, and then I'll have to wait for cooperative weather to see how it runs.

I'm optimistic that the result will be about the same as the previous test with the draft inducer. I think the draft inducer in my video was running at about 6-7cfm, which is what the fan can produce. Fingers crossed.

On 2/28/2017 at 4:35 PM, Andy98 said:

So, assuming the manufacturer isn't outright lying about their fan performance specs,

The fan is here. I'm not saying they lied about the specs, but they are at least extremely optimistic.There may be one of these fan units that produces the specified fan performance curve but the one they shipped me does not. I purchased an "R" class fan, and it performs like "P" class curve (the fan is labelled as being an "R" unit):

• In my burner, I measure the unit as producing ~6cfm - spec curve suggests more like ~16.
• I also measure the fan to produce a static pressure of ~1inH20 - spec says 1.8.

All that said, the fan is still sufficient - at least according to measurements in my garage. 6 cfm is close to my original performance spec.

I did make an improvement to the fan: I inserted a layer of tape around the inside of fan's housing (basically sealing the housing to the fan blades a bit better) and that upped the performance to 8cfm.

So, I'm disappointed that I won't have as much headroom with the fan as I wanted but the good news is that it should still work and if it doesn't, then upsizing to 1" pipe from the 3/4" I'm at now should give me significantly more airflow in any case. I'm tempted to buy the parts just to test that out.

Temporary mounting and ugly green painter's tape aside, this could be pretty elegant for a blown burner:

Next post: Electronics.