1 make sure your printer is elevated off the ground.
2 keep your filament in a warm place in a box with silica granules, and use fresh given the choice.
3 if your room has ceramic walls, keep it at a constant temp for a few days before printing. The thermal mass of the room makes the biggest difference.
4 run a short test print for 20mins before you run the real deal. The second print of the day is always the best as the temps stabilise more in the printer.
5 if you have dynamic setting for ramping temp and cooling fans its worth setting points where it runs slightly cooler after a hour and 2 hours etc as your print looks like it's getting hotter as it progresses.
If your printing PLA, leave any doors on the machine open, and don't enter the room to check on the print while it's running. I can see the layers in my prints where people have entered the room.
6 prints over night often run better as sunlight doesn't change over the print time, and rooms with little or no drafts or Windows are a given.
7 different colours even of the same brand can need different settings to print. Choose a colour and stuck with it. Don't ask me why, but I do better with light colours.
First problem is that your printer doesn't have enclosed chamber.
Let's just say that what I have in mind (for my business) is a proper granular based machine weighing 3.5 metric tons.
No spectogram at the moment, going through a computer switch. I'll see what I can do when it's done.
/Anton
If the DI was low it wouldn't beam like hell
Frankly, I think that such narrow as 50° coverage angle is not really advantageous. Make it 70 - 90° and it will be a lot easier to achieve. And I mean "constant directivity" - beaming is easy and what onni shows is still not a particularly well controlled, IMHO. Probably considerably better than the JMLC though.
IMO that's not proven to be a problem in my experience of using both types, unless you are printing in materials like ABS, even then, the solution is stabilising the room ideally.
If the DI was low it wouldn't beam like hell
Frankly, I think that such narrow as 50° coverage angle is not really advantageous. Make it 70 - 90° and it will be a lot easier to achieve. And I mean "constant directivity" - beaming is easy and what onni shows is still not a particularly well controlled, IMHO. Probably considerably better than the JMLC though.
Ah, right. Did I mean the DI is high if it's narrow then?
Regarding the exact pattern width I should go to, I suppose it depends how narrow I can go without loosing a smooth pattern, as the Harman data shows this to be of high importance. I don't care what happens above 16/17 khz, so can trade this off too. I would like to make it oval so as to keep the vertical dimension of the horn as low as possible for c to C spacing.
I would also like to keep the center section of the horn profile straight to reduce HOMs (if the design lets me, so it's of secondary concern)
Either way, at the moment, I'm of the opinion that the 'excitement' factor of a speaker is down to the proportion of direct Vs reflected sound. I can't find another explanation for it anyway, and others have confirmed similar observations.
It causes problems mating the horn to a bass driver when it's a narrow pattern horn due to a DI miss-match either way.
Perhaps I will look into a cardiod bass design to help this. Not sure if cardioid is possible in a flush mounted corner horn though mind!
At the moment I cannot load the software, as after unzipping, it does nothing when I run the execute file.
Last edited:
The DI is neither low neither high with a beaming horn - it's rising with frequency to very high values. That's what makes it "beam", i.e. to loose high frequencies off-axis and in total radiated energy.
Regarding the C-C spacing, I think this is a common misconception but so deep-rooted that I almost gave up to explain it anymore.
- If anyone provides me raw X,Y coordinates of a JMLC (or similar) horn I can easily run the BEM analysis and show a "standardized" report including the DI.
BTW, just a while ago I implemented this (gnuplot) reporting to a regular version 4.7, it will be available to anyone so we can share our results in more directly comparable manner (without a need to use VACS anymore).
Regarding the C-C spacing, I think this is a common misconception but so deep-rooted that I almost gave up to explain it anymore.
- If anyone provides me raw X,Y coordinates of a JMLC (or similar) horn I can easily run the BEM analysis and show a "standardized" report including the DI.
BTW, just a while ago I implemented this (gnuplot) reporting to a regular version 4.7, it will be available to anyone so we can share our results in more directly comparable manner (without a need to use VACS anymore).
Last edited:
That is because you need to execute it from the command line
Have a look at the post here and the discussion afterwards to see if that helps you to get it running.
Acoustic Horn Design – The Easy Way (Ath4)
Almost all of it is explained in the instructions
Sorry, then to ask you, but what is the misconception?
"the C-C" distance isn't a misconception, just a distance.
I don't think we want to increase that distance, typically.
We do have a trade off, that we can't make the horn too narrow without loss of pattern control by the horn itself.
But the reduction of C-C distance is not mutually exclusive.
For instance the whole point of Danley style Multi Entry Horns is to do this.
Whether a MEH is worth some of the other trade-offs is contentious, but it's not the only option.
The low frequency driver can be nested somewhat into the horn to reduce C-C distance, several makers have done this.
Intuitively it seems reasonable, unfortunately it does complicate the simulation.
Whether this is worth the trouble is a matter of opinion but not a "misconception" surely?
Do I misunderstand you?
Best wishes
David
A good proven starting point is a waveguide about the same size as the woofer, preferably a bit larger. The misconception is the idea that making this waveguide "oval" by reducing its vertical dimension will help the vertical pattern control. In the end it won't because of the worsened radiation pattern of the waveguide itself.
What's missing?Almost all of it is explained in the instructions
I'm finishing instructions for the new release - I could improve on this."The most convenient way of starting the command line for Ath is to create a shortcut to cmd.exe and set the "Start in" property to the Ath directory"
This line in the guide explains what to do but not exactly how to do it. Most windows users are not used to the command line. The screenshot of the shortcut command window seems to help. I think it would be worth adding as the thread is long enough now that most new users won't read through it all.
This line in the guide explains what to do but not exactly how to do it. Most windows users are not used to the command line. The screenshot of the shortcut command window seems to help. I think it would be worth adding as the thread is long enough now that most new users won't read through it all.
An reasonable rule of thumb for an axisymmetric horn/WG.
But to make it oval why not widen it horizontally rather than reduce it vertically?
The horizontal pattern control will be maintained below the nominal frequency but that's no problem.
Around crossover frequency the vertical dimension effectively increases because it's the sum of the horn and the woofer size. (woofer under horn assumed)
This can maintain the vertical pattern control.
The centre to centre distance can be adjusted independently to some extent, if we nest the woofer into the horn or move it away.
So we have more free parameters to tune - more complicated but more adjustabilty usually means a better fit.
Lack of axisymmetry does mean it takes more time to sim, my current machine is a bit too slow.
So I have only done back of envelope calculations but reduced centre to centre looks best so far.
Have you considered or tried to sim such a scheme?
Best wishes
If you don't reduce the vertical height of the guide how can you reduce the centre to centre distance
