Homemade Exponential Horn Project

[Watanabehi]

Here are more videos of my home-made Gramophone playing Helen Ward and Billie Holiday. Both are original Brunswick records. I hope you enjoy them

Hideki

[Watanabehi]

Hideki, I surrendered years ago to your charming achievements in gramophone things, your yt videos, your singing, and all that. On me you have an unconditional fan. I already knew your emg-esque handmade gramophone, and it's great! Probably by one of your videos playing it...

Hi Inigo,

I hope these videos work to project the sound.

Hideki

I posted the same videos for youtube.

https://youtu.be/URGnAbdQzI0 Dick Haymes Decca record
https://youtu.be/1bradIqBN0M Campoli UK Decca record

[Watanabehi]

Hi Everyone,

I recorded the sound wave of my Home-made Gramophone using Audacity. It is the beginning of Grieg's Peer Gynt No.2 Symphony London Record. I have no idea what it shows. Thanks.

Hideki

[Inigo]

This is amplitude against time. If you look for the spectrum, which audacity will have, I believe, you'll get a picture of the frequency content...

[Ethan]

This is amplitude against time. If you look for the spectrum, which audacity will have, I believe, you'll get a picture of the frequency content...

That's right--to see the plot spectrum, go to Analyze-->Plot spectrum…; then a pop-up with a graph of amplitude versus frequency should open.

Anchorman, have you looked into the costs and logistics of 3D printing? I considered it briefly, but I presumed it would be quite expensive for something the size of an EMG horn—and I wasn’t sure that any of the printable materials are acoustically inert enough. It would be nice not to have to deal with extracting a curved former, though!

I thought there were some fiberglass resins or epoxies that were at least relatively non-toxic (can’t remember what they would have been—I could be misremembering); regarding itchiness, would sanding it wet work to keep the particles out of the air?

[anchorman]

Depends on the material if it is cost effective. Most likely would have to be made in sections and pieces together. But they could make a nice form upon which to build up the papier-mâché, and left in place the way the aluminum part of the original EMG horns was.

[anchorman]

Most polyester resin used for fiberglass has a high styrene content. hello cancer. And epoxy is almost exclusively manufactured from bisphenol A solids - potent endocrine disrupter. I try to avoid using them if there are other viable alternatives that will get me to my end goal. There are supposedly less terrible epoxies, but my general observation is that the stuff they replace the bad materials with is almost always untested and often as bad or worse than what it is replacing. There is no requirement to show that new materials are safe for people or the environment we live in, and it's very few companies that bother to check.

I've not worked with either in some time, so can't comment on wet sanding, but my experience with such things makes me think that it is almost a bigger pain than the itchiness. wet sanding and most power sanders don't mix well, and sanding something that big by hand seems a little crazy to me.

[Watanabehi]

Hi Everyone,

I recorded the sound wave of my Home-made Gramophone using Audacity. It is the beginning of Grieg's Peer Gynt No.2 Symphony London Record. I have no idea what it shows. Thanks.

Hideki

Here is the another one for the same recording.

Hideki

[Inigo]

This graphic can be analysed to extract conclusions, but to test a gramophone, the ideal way seems to me to be this: to record in the same circumstances several records of different types of music, and then add or join all the wave files in a single one. First you'll have to use, on each one of the files, the audacity facilities to remove the bias or DC component, and then level the files to the maximum volume (maximize feature) proceeding one by one. After that process, you can add all them together, one behind the other.
Then the spectrum analyser can be extracted from all the records, so you'll have s more clear picture of the overall sound of your gramophone... played in that particular environment. To judge only from one record didn't give the complete image except for that particular record.
Of course you can do it also without any pre-processing of the files. It could be interesting to compare both spectra, the pre-processed one and the raw one.
Judging by one sole record, I can see narrow 'holes' in the reproduction at 650-900-2000-3000-5000 Hz in the general decay line from 300 to 11000 Hz. Although these peaks in the high frequency range could be parasites from the recording. Or the recording is a later one, Decca ffrr type post-1945, which had a surprising treble response.
Of course, to test the response of the machine, the best is to use post-1945 recordings, or the most modern ones you have, or a sweep frequency test record, if you have one.
I don't have one, but I have a fixed note recording at the back of one record, and I used that to test (only by ear) my 194. I played it very slowly, then raising uniformly the speed, etc, so I managed to get a kind of sweep. I noticed by ear clear points of high intensity, resonances of the horn, and some 'holes' in the frequency sweep. Curious.
I don't know how these programs calculate the spectrum, but I've always observed this slope from the mid bass to the high treble. Clearly the energy captured goes down towards high frequencies. I think that some kind of 'normalising' would be required to see this slope as an horizontal line, so you can compare different spectra. All the programmes I've used show this mountain shape. Then you go to the books and see the spectra diagrams of different machines, and they show an horizontal line in the 300-4000 region. Clearly the energy is measured in another way. I'm thinking of the Percy Wilson book, for instance.

[Watanabehi]

Anchorman, I used a constant-bore conduit because starting the taper right after the tone-arm would have made the bends through the conduit and swan-neck much tighter, which could have caused worse internal reflections and decreased the treble response—a narrower tone-arm with the same cut-off frequency or a lower cut-off (i.e., slower flare) with the same tone-arm outlet diameter would have fixed that, but the HMV large-bore swan-neck style seemed like the best compromise of tracking alignment, cut-off, and price.

It would theoretically be best to start the exponential flare at the tone-arm outlet, so that the conduit and horn have the same flare, but the consensus seems to be that it isn't extremely important--there’s a video on YouTube of an EMG Xb Oversize horn on an HMV 31 base, and I believe whoever posted it said the sound wasn’t much different from a complete Xb Oversize, even though the horn, conduit, and tone-arm had mismatched flares.

Hideki, does the Credenza tone-arm you’re using have the post-1928-recall crook? I believe they’re supposed to improve the tracking alignment, although it still might not be as good as the HMV re-entrant type.

Can you get Audacity on your computer? It’s a free audio editing program, and it has a spectrum analyzer feature, so if you can get videos or recordings on your computer, you should be able to import them into Audacity and measure the frequency content—it wouldn’t be perfect, but it might at least give an idea of the response.

I’m looking for better materials, myself—I suspect that the Elmer’s glue I used is too resonant and flexible to pass low frequencies effectively. What kind of wood glue did you use? And have you tried water-flour glue? I’ve heard that it can be quite hard, but also prone to mold and bugs if not done right.

Ethan,

The tonearm is the earlier Credenza's model. I downloaded Audacity. I used a glue called "Harder-than wood". No, I have not tried water-flour glue.

Hideki

Ethan,

I compared the length of the early Credenza tonearm and HMV 163 tonearm. The early Credenza one is longer than HMV 163 tonearm, but the reproducer part of the Credenza tonearm has sharper bent than HMV 163 one, which might cause the tracking error. I think the HMV re-entrant tonearm improved the tracking error using much less sharper soundbox part of the tonearm.

Hideki