Hi,
My first and unsorted thoughts on this....
One might get the pulse duration by measuring the current with a shunt in the loop and a pretty fast DSO. Pulse duration (again referring to the time the spark is "up") will only alter absolute level (think of dipole speaker below the dipole freq as an analogy) but not the relative response captured by the mic. One also might need a series of measurements, apply the integration to them, normalize the results and do a best-curve-fit to get more reliable data. And have the S/N issues addressed etc...
A problem I see, Scott mentioned it, is knowledge or estimation of the wavefront hitting the mic... If you look at the piezo spark (cigarette smokers, anyone?) one can see the spark is not well controlled in shape, which would blur the two doublet pulses... but I guess both pulses should have the same shape exept for polarity (unless the spark changes its shape during its lifetime -- or if build-up and extinction are different). If both shape and diameter of the air column that gets ionized is well controlled all would be fine. And we'd need knowledge about the time constants of the ionization, a tiny spark might have higher acoustical rise and fall times than a bigger one (more mass).
Indeed complicated matters, lot's of major and minor points to consider (ESD impact on the mic's eletronics)... time for both a lot of armchair reasoning (the Sherlock Holmes approach) and "just-go-for-it" measurements, like Scott has already exercised for us.
Any empircial physicist reading this and willing to comment?
- Klaus
My first and unsorted thoughts on this....
With a piezo-stack as the source things are rather uncontrolled I'd guess. But thats the first option for DIYers, thogh we could check out what the tesla coilers and coin shrinkers have at hand...1audio said:
One might get the pulse duration by measuring the current with a shunt in the loop and a pretty fast DSO. Pulse duration (again referring to the time the spark is "up") will only alter absolute level (think of dipole speaker below the dipole freq as an analogy) but not the relative response captured by the mic. One also might need a series of measurements, apply the integration to them, normalize the results and do a best-curve-fit to get more reliable data. And have the S/N issues addressed etc...
A problem I see, Scott mentioned it, is knowledge or estimation of the wavefront hitting the mic... If you look at the piezo spark (cigarette smokers, anyone?) one can see the spark is not well controlled in shape, which would blur the two doublet pulses... but I guess both pulses should have the same shape exept for polarity (unless the spark changes its shape during its lifetime -- or if build-up and extinction are different). If both shape and diameter of the air column that gets ionized is well controlled all would be fine. And we'd need knowledge about the time constants of the ionization, a tiny spark might have higher acoustical rise and fall times than a bigger one (more mass).
Indeed complicated matters, lot's of major and minor points to consider (ESD impact on the mic's eletronics)... time for both a lot of armchair reasoning (the Sherlock Holmes approach) and "just-go-for-it" measurements, like Scott has already exercised for us.
Any empircial physicist reading this and willing to comment?
- Klaus
KSTR said:
Neumann has a few papers in German and a short summary in English by Peus. They use the spark response in a very qualitative "seat of the pants" way to show impulsive and time smearing mic behavior. Several mics show very clear doublet behavior proving a good spark can be made. Reading between the lines I think they get a 60kHz limit. Earthworks claims 100kHz but keep the good stuff secret.
After much prodding Earthworks revealed that the best mike was the Gefell MK301 1/4" mike. They use that as a baseline. They did claim 100 KHz is possible and that they had to work some tricks to do it.
I'll ask around to see if I can access a Manger driver to try. Perhaps its optimized to give a really good transient response with a very fast rise time.
I'll ask around to see if I can access a Manger driver to try. Perhaps its optimized to give a really good transient response with a very fast rise time.
No Anatoliy, Earthworks sure *are* serious guys from what all I know.Wavebourn said:
To get at it we'd need to look at said snare hits (assuming it was a simultaneous recording) in close detail, that is analyse those signals differences in depth. Assuming all else equal, from differences in phase (from their different roll-offs) we should expect to see different waveshapes, even below the 22k CD brickwall. If you could provide the same identical snare hit in both 20k and 50k versions I could try to analize it (WAV or FLAC, not MP3).
But again we're polluting someone's thread...
EDIT: you are fully entitled to pollute your "own" thread, of course
On topic, I'm currently modifying one of my piezo spark gaps and will try to do the experiment myself (with WM61A and another capsule salvaged from an very LoFi old cassette recorder).
- Klaus
KSTR - You're in the correct subject and thread here... non-polluting and about calibration of mics. Although, again I think this is an acoustics, not a Solid State matter... but whatever...
DM, I was not impressed with the Manger I auditioned here - it requires an IEC or larger baffle to get the response flat (I did not use such...) as far as I can tell... it was very beamy in the highs... sounded downright wierd, not bad, wierd. A stereo pair would have told me more, but I only had one available. I suspect that the soundstage would have been "unusual"... but I can't be sure of that. I'll be curious to know your results, subjective and measured... and there is a Manger thread or two over in Loudspeakers, fwiw...
_-_-bear
DM, I was not impressed with the Manger I auditioned here - it requires an IEC or larger baffle to get the response flat (I did not use such...) as far as I can tell... it was very beamy in the highs... sounded downright wierd, not bad, wierd. A stereo pair would have told me more, but I only had one available. I suspect that the soundstage would have been "unusual"... but I can't be sure of that. I'll be curious to know your results, subjective and measured... and there is a Manger thread or two over in Loudspeakers, fwiw...
_-_-bear
That will cause a lot of excitement but this WILL kill you. However it makes a really good spark- 6000V at 3000A 8 uS rise time 20 uS duration. Is there a relationship between that and its potential acoustic waveform (should explain my earlier questions)? I can tell you that its loud!.
Attachments
If you want adjustable sparks feed variable duration voltage pulses from a car battery into a car ignition coil that feeds and adjustable spark gap. By playing with the amount of energy stored (pulse duration) and the gap breakdown voltage (seperation) you shoud be able to get the spark you need. I used to design electronic ignitors for oil field flare stacks and with a good coil, could get sparks to jump 3 or 4 inches. (I might still have a working cct board in the garage.) Your pizo spark will change with humidity air pressure (temperature) and air contaminants. Not sure how exact a spark your looking for.
Besides for the geometrical considerations (perfect ionized air column in all regards), it's the optimium systematic behaviour, the most perfect doublet we can obtain.cbdb said:
I think the point is to make the mental connection to the ideal dipole source and then a Linkwitz drawing shows us what's important:
http://www.linkwitzlab.com/models.htm#A1
The similarity to Scott's plots is striking, isn't it? It dawned to me when Scott mention his trick the first time a while ago but now I'm certain that this is exactly the same thing going on here... with all its consequences....
And, your idea of using a step-up transformer for more control is sure something to check out, thanks! We definitly need a control (measurement) or close estimate of the spark's uptime. Too long a duration and we run into the roll-off.
- Klaus
KSTR said:
Yes I forgot that, same math going on here. One of the references talks qualitively about the spark causing a very sharp (supersonic?) rise with the relaxation (negative transition of doublet) being related to both the acoustic and thermal properties of the air. This I figure creates the soft error at the high end rather than the null as shown in Linkwitz's picture which is a mathematical idealization.
In general we do not want 3"- 4" sparks due to the prop delay across the length issues.
I would love to quantify my fudge factor or even be able to measure the induced current in the spark and predict it. Right now I only know my particular generator needs this fix. I have sucessfully used this to equalize the larger Nak mic capsules which have a bad 5k peak that birders seem to dislike.
With an adjustable gap, you can have any lenth of spark you want, 4" or less. And with a small resistor in the current path, measure the current (the trick there is isolating the scope from the high voltage, 17kv per cm gap if my memory serves (25 years ago) its possible (I believe I used an opto isolater) but I dont remember the details.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.