Porous Moderator Design Tests

[denovich]

I'm on a GluBoost kick lately. Long working time before it sets up. Wondering if you've ever heard of it?

Looks like a decent CA glue. I've been using Starbond recently and always had good luck with stuff from Bob Smith Industries. I have a bit of a chemical sensitivity to CA after building lots of balsawood planes and bridges (Science Olympiad), so it's not my favorite type of glue to use. But it is so handy.

I made a guitar and using CA as the finish on the body... apply using a piece of wax paper to smear it around until it sets. Wet sand to a glass finish. Really nice when you don't want to wait for a typical finish to cure hard enough to sand. Good at reinforcing soft woods too.

 

[WobblyHand]

I've had good luck with the Bob Smith Industries glue. Not very expensive and lasts a while in the bottle. The version I have is very thin and flows like water. I used it to glue a PETG moderator print into a CF tube. Just a couple of drops at both ends at the gap between the two parts and it wicks in well. That insert isn't coming out.

 

[RAMJET66]

0

Looks like a decent CA glue. I've been using Starbond recently and always had good luck with stuff from Bob Smith Industries. I have a bit of a chemical sensitivity to CA after building lots of balsawood planes and bridges (Science Olympiad), so it's not my favorite type of glue to use. But it is so handy.

I made a guitar and using CA as the finish on the body... apply using a piece of wax paper to smear it around until it sets. Wet sand to a glass finish. Really nice when you don't want to wait for a typical finish to cure hard enough to sand. Good at reinforcing soft woods too.

do you have a pic of the guitar you made ? TIA

 

[denovich]

Sure... This was my first attempt at a guitar. It's still sporting some components from a $25 Ibanez that I used as a template. I reused the neck and refinished it. I started down the road of building the new neck, but life got busy.

I built it in 2019, the second photo is current and shows how much the cherry darkened with time. (I bought a truckload of cherry rough-cut lumber at firewood prices about 20 years ago. I really like using it. I don't mind experimenting with it because it's cheap.)

 

[daddypaddy]

cool! i can only dream of having the capability to create something like that from the ground up. my moderators are simple cans with imagination powered guts inside them.. but they work well enough for me . my madness would be printing the second baffle as a non porous section with a steeper cone( like an otoscope cone)?

 

[WobblyHand]

Getting back on topic, I ran a test between two versions of a TPU moderator both 38mm OD and 148mm long, one with 30% gyroid infill internal to the moderator and solid walls, and the other entirely solid infill. Printed with Overture TPU, 0.4mm nozzle. 7 hours for the solid version, about 6 hours and 20 minutes for the internal 30% gyroid.

The solid infill moderator was 3.1 dB quieter, at least from an RMS sense. The spectral shape was slightly different as well. Since I was inside, and needed to avoid saturating the mic, (small room) I used a 2240 as the airgun. This is a relative measurement, made with a Dayton IMM-6 Reference microphone, captured using Audacity. Distance to the microphone was about 0.5m.

Sure it is low power, but this is a test between porous and solid. When I process the audio I get the following spectra.

Thought it might be interesting. Edit: Seem to have problems with the wav files, at least playing it on my Mac.

 

[subscriber]

WobblyHand,

There is a clear difference in sound, but the trace shapes seem similar. I would be interested in seeing the test repeated: Solid; TPU; Solid; TPU. Just to see if it repeats and that something else is not going on.

 

[WobblyHand]

In the time domain, they look different, the phase is not the same. Top traces are at approximately the same scale.

Might get a chance to try again tomorrow. Both parts are entirely printed from TPU. They are self supporting, so no CF tube. The green is the time domain data for the 38mm x 148 with internal baffles made with 30% gyroid, but with solid walls. The red is time domain data for the same part but both the baffles and the walls are solid TPU. (No gyroid infill at all.)

 

[subscriber]

Are the baffles in the gyroid version porous all the way through; or just faced with gyroid? If fully porous, perhaps too much air is leaking through the baffles.

What about the external "tube" wall? Is that porous all the way through to the outside on the gyroid version? If so, it is begging to be sealed off; even with just a layer of duct tape.

 

[RAMJET66]

Sure... This was my first attempt at a guitar. It's still sporting some components from a $25 Ibanez that I used as a template. I reused the neck and refinished it. I started down the road of building the new neck, but life got busy.

I built it in 2019, the second photo is current and shows how much the cherry darkened with time. (I bought a truckload of cherry rough-cut lumber at firewood prices about 20 years ago. I really like using it. I don't mind experimenting with it because it's cheap.)

View attachment 426040

View attachment 426041

Are the baffles in the gyroid version porous all the way through; or just faced with gyroid? If fully porous, perhaps too much air is leaking through the baffles.

What about the external "tube" wall? Is that porous all the way through to the outside on the gyroid version? If so, it is begging to be sealed off; even with just a layer of duct tape.

looks good , nice job..

 

[WobblyHand]

Are the baffles in the gyroid version porous all the way through; or just faced with gyroid? If fully porous, perhaps too much air is leaking through the baffles.

What about the external "tube" wall? Is that porous all the way through to the outside on the gyroid version? If so, it is begging to be sealed off; even with just a layer of duct tape.

The baffles are porous all the way through, everything inside a cylindrical volume is porous gyroid. The special cylinder is a modifier and one can specify the infill (30% gyroid) and perimeters (0) in the region. Outside of the modifier cylinder, which includes the exterior walls and ends are solid infill.

Here's part of a failed print. The exterior walls are solid. The interior is gyroid. The print failed 5mm before the thread region. And yes, TPU is a bit stringy. Getting better at it, but no where as nice as PETG.

 

[subscriber]

Thanks for the clarification, WH.

 

[WobblyHand]

Have a confession, the OD of the tested TPU LDC's is 36.7mm, not 38mm. Was targeting 37mm so I could use 40/37 CF tubing. With a 1.5mm wall thickness, the tubing can be threaded on a lathe.

 

[subscriber]

That's almost a rounding error, WH

Not of much consequence, unless parts have to fit.

 

[OldSpook]

Makes no sense to me. @WobblyHand when you say the one was solid TPU and the other was 30% gyroid do you mean it was as solid as the walls on you 30% in fill unit?

My tests don't show anything like that. Can you post a cross section of your stack, please?

 

[denovich]

I found that when I switched to TPU the effect of the porous walls was less obvious. What was obvious was the lack of torsional rigidity. Because I was threading the insert onto the barrel, the insert needs to be rigid enough to transmit the screw force. The porous walls made it too easy to twist, tearing one of my inserts in half.

I did then try solid outer walls, but with no inner walls (porous on the inside only), but I couldn't hear the difference between that and all solid walls. The resulting print also looked nicer.

@WobblyHand (you probably already know this but do what you can to dry your TPU before use. Moisture in the filament will exacerbate its tendency to be stringy. I had mine at 60C for many hours (fresh after coming out of the bag) before I started printing and was surprised by the print quality. Arguably better than PLA.

@OldSpook does your slicer have a setting for "Polyholes"? It changes how holes are sliced. Instead of circles it uses triangles or pentagons for the walls that define the hole. The shape is rotated for each subsequent layer, so the effect is a round hole. But compared to a regular hole, it should be more precise in its dimensions. This makes having to fool around with hole compensation, or making parametric holes to get precisely sized holes unnecessary. I just found out about this feature and plan to experiment with it.

 

[WobblyHand]

Makes no sense to me. @WobblyHand when you say the one was solid TPU and the other was 30% gyroid do you mean it was as solid as the walls on you 30% in fill unit?

My tests don't show anything like that. Can you post a cross section of your stack, please?

The solid TPU LDC had 100% infill everywhere, baffles, threads for adapter, ends and exterior walls. The 30% gyroid infill, 0 perimeter walls was only for the baffles, not anything else. This LDC was intended to be self supporting, it doesn't need an external aluminum or CF tube.

If one uses 30% infill everywhere, with 0 perimeters, it needs to be mounted inside a CF tube, as one can squish the whole LDC easily, as there is no rigidity at all. My intent was to make a monolithic print, not individual baffles and spacers to slide into a tube. Hope that makes sense.

So in the green print shown a few posts back, the 3mm thick exterior walls are 100% infill with perimeters. But all of the interior baffles and internal structure were printed with 30% gyroid infill and 0 perimeters. (Using a cylindrical modifier that was totally inside the LDC walls and ends.) Clear as mud?

 

[WobblyHand]

I found that when I switched to TPU the effect of the porous walls was less obvious. What was obvious was the lack of torsional rigidity. Because I was threading the insert onto the barrel, the insert needs to be rigid enough to transmit the screw force. The porous walls made it too easy to twist, tearing one of my inserts in half.

I did then try solid outer walls, but with no inner walls (porous on the inside only), but I couldn't hear the difference between that and all solid walls. The resulting print also looked nicer.

@WobblyHand (you probably already know this but do what you can to dry your TPU before use. Moisture in the filament will exacerbate its tendency to be stringy. I had mine at 60C for many hours (fresh after coming out of the bag) before I started printing and was surprised by the print quality. Arguably better than PLA.

@OldSpook does your slicer have a setting for "Polyholes"? It changes how holes are sliced. Instead of circles it uses triangles or pentagons for the walls that define the hole. The shape is rotated for each subsequent layer, so the effect is a round hole. But compared to a regular hole, it should be more precise in its dimensions. This makes having to fool around with hole compensation, or making parametric holes to get precisely sized holes unnecessary. I just found out about this feature and plan to experiment with it.

I keep my TPU in a bag with about 250 grams of desiccant (when not printing) and print from a dry box at 55C. The TPU has been dried maybe 24 hours at least. I think my printer is printing a little too hot, so I could reduce the temperature a bit. Which TPU are you using?

I found a similar thing with all porous walls, I could barely even start a tap in the porous threads, and the unit wanted to twist up like a pretzel when screwing it on the adapter. I was able to try it, but it was kind of loud without walls. I didn't bother with taping it up, simply because I didn't think it would be a long term viable option. I also didn't have a tube to slide it into to try.

 

[OldSpook]

Regarding the difference in improvement using TPU over some other plastic. I saw about 3 dB using the gyroid infilL at about 37% AND ANOTHER ~3 dB when I went to TPU. I guess I was looking at it backwards from you other fellows. I stand by it though. Thats what I'm seeing. Do remember I am building these inside CF tubing and printing my end caps on the resin printer. The core is one piece TPU with a nominal 40% infill except where I use a region to specify some change. I am not attaching the core directly to the CF except in as few places as necessary to keep it stationary WRT the shell.

@denovich when your solid wall is made of the same TPU (not separate from) the TPU which comprises your baffles the TPU CONDUCTS the sound better to the outside shell. You won't hear any improvement because the noise is still being conducted to the outside of the moderator. This is my take on what I am seeing anyway (certainly could be wrong.). I don't have a setting for "polyholes". I am running Prusa Slic3r. I am also seeing some stringing but I am trying a different route to clean it up. I am using regions like @WobblyHand and enabling ironing on top surfaces that should clean up the "fuzzies" where they matter and leave them where they can help out.

@denovich && @WobblyHand I observed today something you should know about. I have a design which is pretty small and was intended for lower power guns in the 25 fpe and below class. I printed it at 40% with no shells. When attached to my 25 fpe 20 CAL Talon is would not keep it's shots on a page of paper at 25 yards BUT when I put it on my 22 CAL DragonFly and was shooting it around 12 foot pounds it was shooting pellet on pellet. The design incorporates two baffles at each end of a hollow space for expansion. I think, don't trust me on this but ... I THINK the baffle cones were not stiff enough to take the pressure at 25 fpe and were closing up around the pellet as it passed through the moderator. There was no observable clipping. I repeatedly performed that experiment with the same results each time. I also noticed that the gun was noticably less pellet fussy with moderator than without it.

Food for thought anyway.

Here is a zipped GCODE file you can look at in your GCODE viewer.

View attachment 22-STANDARD-TPU_1h35m_0.36mm_240C_TPU_CR6SE.zip

 

[WobblyHand]

I re-plotted the two signals on the same graph in both the frequency and time domain. For frequency domain I used a Hamming window of the full length of the signal. Since I am writing the code, I can have any length FFT I want. The FFT length is equal to the signal length. (It happens to be 18843 points). I cut off one of the time domain signals (to make it the same length as the other) so both frequency domain outputs would share a common x axis. I'm not saying that I took the data perfectly, but here it is. In this configuration, the solid one is quieter, it's obvious in both domains.

I did this in python (numpy & scipy) to process and plot the waveforms. I exported the Audacity signals as matlab files, (rather than a wav file) and ran my program on these data files sequentially. The frequency domain signal is slightly filtered with a Savitsky-Golay filter to remove some of the fuzz without changing the shape.

I put substantial bracing on the baffles, as I was concerned about the structural integrity of the weak mesh. The six radial braces hold more than 1/2 the height of the cones. I could see that high power would cause some issues. Haven't had any good days to test accuracy.