How do I Tune Bass Reeds that share the same hole?

[Sebastian Bravo]

So... I did what i said. I put all the Arietta Bass reeds on my concerto III reed block (It has individual holes for the reeds)

This is the measure of the Bigger reeds inside the accordion (closed all the other blocks with masking tape on the bottom):

(O = Opening the bellows, C = Closing)

A) O: -10c C: -18c
A#) O: +1c C: -12c
B) O: -12c C: -35c
C) O: -36c C: B+42 (very detuned)
C#) O: -35c C: -30c
D) O: C+48c (very detuned) C: -35c
D#) O: D+44c (very detuned) C: D +48c (Very detuned)
E) O: -15c C: -30c
F) O: -11c C: -7c
F#) O: -10c C: -17c
G) O: -12c C: -37c
G#) O: -4c C: -22c

So i will tune these ones, then the other reeds. I will use the method that Paul De Bra mentioned, correcting the difference measured here.

Thanks to everyone for the support, please wish me success!

Last question: Will the tuning change when i move the reeds from the reed block with individual holes, to the original reed block with the big hole?

 

[debra]

These deviations are absolutely not normal. Do not ruin the reeds by attempting to just tune them!
When a reed is off by a lot in many cases the reeds themselves are not the problem but the valves are. The leather becomes stiff over the years and then the valves either do not open wide enough (note goes flat) or do not close properly (opposing note goes sharp). It is also possible for the valve on the inside to touch the end or sides of the reed block and then not open and close smoothly. So it is imperative that you do some good detective work trying to find sources of large deviations other than the reed itself!
The lowest reeds also have very pressure-dependent frequency. When it starts it may be 10 cents off from where it is after playing for 2 to 3 seconds. You then need to take some kind of average.
As to the question whether the tuning may change when you move the reeds back to the original block the answer is a very likely yes. Moving to a different resonance chamber can certainly influence the frequency.
The overall conclusion is: tuning bass reeds is hard and often unpredictable. (I wish I had a more positive message here...)

 

[Sebastian Bravo]

These deviations are absolutely not normal.

Yes, i thought the same.

I inspected the valves and they are in a good condition. The problem was with the reeds, they were covered with rust on the back part (the blue part of the reed) so i removed all the rust from all the reeds and got this strange tuning.

I will tune the normal ones and try to tune the very detuned reeds but very carefully, like 5 cents up, wait 5 minutes, then tune up again, trying to avoid overheatig the reeds and affect their tone.

I measured the last 3 sets of reeds and they seem to be normal (obviously detuned but no more than 15 cents) so i wont have problems with them.

I will post results in 5 hours!

 

[debra]

These deviations are absolutely not normal.

...
I inspected the valves and they are in a good condition. The problem was with the reeds, they were covered with rust on the back part (the blue part of the reed) so i removed all the rust from all the reeds and got this strange tuning.
...

That the valves are in good condition is good news. That the reeds are rusted is not so good. Depending on where on the reed the rust is the frequency may indeed go up by quite a lot or down by quite a lot. In hindsight it would actually be better to have all reeds on the original block again so you can start making corrections and hear the effect on the pairs of reeds sounding together. Perhaps you should stop making corrections when you are no more than 5 cents off and then put the reeds back.

 

[Sebastian Bravo]

Depending on where on the reed the rust is the frequency may indeed go up by quite a lot or down by quite a lot.

The rust covered all the back blue part of almost all the biggerbass reeds (are those 16 or 32?)

First i removed all the rust, then i got that tuning chart with very detuned reeds.

Today i tuned all the smaller reeds. And half of the big bass block (it is 4 reeds on the bass) i got all the notes on good tuning (almost +/- ~0.3c of detuning on all the reeds) so the bass is sounding really perfect.
Im currently working on the bigger reeds but i have to eat something! So, i hope i finish with the basses tomorrow by the morning, its 22:46 and i dont want my neighbours calling the police because of the reeds/dremel sound :lol: :lol: :lol:

 

[Sebastian Bravo]

<YOUTUBE id=qLdUG6y1v7g url=></YOUTUBE>

this is how it turned after first tuning... some notes are still out of tune, but it is getting better...

The girl asked it for monday so i had to give it to her... the accordion will be back in 5 days because she is going to a small town in the south just to practice on the wild side... she doesnt know how to play basses so she didnt care that much.

D big bass reed is still D - 35 after a lot of work (It was C# +44 before work) but i had no time to tune it better.

 

[Geronimo]

D big bass reed is still D - 35 after a lot of work (It was C# +44 before work) but i had no time to tune it better.

Have you cleaned it thoroughly from rust (using a glass fiber pen and/or some oil which you need to completely remove afterwards)?

If you did, then with that amount of material rusted away from the weakest point of the bass reed, this one is likely to break at some point of time. I mean, right now you can probably mostly file from the weight, but youll have a misfit between reed profile and weight afterwards which is bad for pitch stability under pressure.

If you did not, do that first thing before any more tuning on it.

 

[Sebastian Bravo]

D big bass reed is still D - 35 after a lot of work (It was C# +44 before work) but i had no time to tune it better.

Have you cleaned it thoroughly from rust (using a glass fiber pen and/or some oil which you need to completely remove afterwards)?

If you did, then with that amount of material rusted away from the weakest point of the bass reed, this one is likely to break at some point of time. I mean, right now you can probably mostly file from the weight, but youll have a misfit between reed profile and weight afterwards which is bad for pitch stability under pressure.

If you did not, do that first thing before any more tuning on it.

Thanks for the tip Geronimo, i think what you say is probably the problem. Because i removed a lot of mass from the reed and i couldnt get it to perfect tuning. Maybe another reed would do the job.

 

[Edocaster]

Must be a Hohner thing. I have the same shared hole thing with a pre-war Verdi (plus that was glued in :roll: ).

Tape won't help, as they share more than just a hole, more like a quarter the chamber.

Trying to silence one reed will affect the tuning of the other, I believe.

What I found worked was recording the sound in audacity or audition) and using frequency analysis to see the two peaks.

 

[Geronimo]

Must be a Hohner thing. I have the same shared hole thing with a pre-war Verdi (plus that was glued in :roll: ).

Tape wont help, as they share more than just a hole, more like a quarter the chamber.

Trying to silence one reed will affect the tuning of the other, I believe.

What I found worked was recording the sound in audacity or audition) and using frequency analysis to see the two peaks.

Well, FFT does not work well for that: a time window of 1s gives you just 1Hz of resolution. Stroboscopic tuners are one possibility: you use eye sight to lock onto the period of the respective frequencies. Using visuals for discriminating between various rather constant frequencies and their overtones and noise works reasonably well. LPC analysis might also work but is more sensitive to noise and overtones.

 

[Edocaster]

Must be a Hohner thing. I have the same shared hole thing with a pre-war Verdi (plus that was glued in :roll: ).

Tape wont help, as they share more than just a hole, more like a quarter the chamber.

Trying to silence one reed will affect the tuning of the other, I believe.

What I found worked was recording the sound in audacity or audition) and using frequency analysis to see the two peaks.

Well, FFT does not work well for that: a time window of 1s gives you just 1Hz of resolution. Stroboscopic tuners are one possibility: you use eye sight to lock onto the period of the respective frequencies. Using visuals for discriminating between various rather constant frequencies and their overtones and noise works reasonably well. LPC analysis might also work but is more sensitive to noise and overtones.

Im not an expert on FFT, but Ive found the results more than acceptable. Playing steady notes for two seconds duration, the readings are accurate enough to show differences between push and pull that I cant really hear. Set an FFT size of 65536 (the highest on most software), and for a 48khz sample rate that will use up most of the two seconds.

This is what you see on Audition:


The choice of window function can help how the data is represented. Hamming in this case gives a nice visible peak.

Audition will also let you see the underlying data - i.e. the amplitude at each frequency. In the example above each reading is 0.732Hz apart (48000/65536). But you can easily interpolate from the available values to work out the location of the peak just by looking at three or four values on either side. I reckon 0.25-0.5Hz accuracy is quite possible.

Counter-intuitively, dropping the sample rate to 24khz would also increase the frequency resolution, although then the note would have to be played for over three seconds (I wonder if looping the note would achieve the same result?).

Finally, as this tuning method is only used when two octaves are played together, the best approach would be to use the better available accuracy on the higher octave bass note, and then finally correct the lower octave bass note by listening for beats.

 

[Geronimo]

Counter-intuitively, dropping the sample rate to 24khz would also increase the frequency resolution, although then the note would have to be played for over three seconds (I wonder if looping the note would achieve the same result?).

Looping the note would achieve the same result if you looped after an exact length of the notes period. To do that, youd need to know the exact length of the notes period, and if you knew that, youd already be finished. So yes, it would work if you did it right, but to do it right, youd already need to know the end result.

Finally, as this tuning method is only used when two octaves are played together, the best approach would be to use the better available accuracy on the higher octave bass note, and then finally correct the lower octave bass note by listening for beats.

Accordion tuners (and organ tuners, the people, not gadgets) basically do everything by laying down a central reference octave first, tuning it in fifths tempered with the correct beat frequencies. Then they tune everything to be without beats with regard to the reference octave. That way, the tuning accuracy is invested where one can hear it. Any tuning method working with absolute pitches needs to be done doubly as exact for the same audible accuracy since then the tuning errors for both ends of hearing-critical intervals are independent.

Tremolo tuning is an additional complication.

 

[Sebastian Bravo]

What I found worked was recording the sound in audacity or audition) and using frequency analysis to see the two peaks.

I will try to use that method tomorrow, with a a Hohner Contessa II that came recently for a full restoration (rust removal, cleaning, new valves, wax and tuning)


The girl came back from her vacation and got me her accordion for the final tuning. I recorded this video days ago and forgot to upload it, it sounds really better now.

Thanks everyone for the support, suggestions and everything. I wil make the D.I.Y. accordion microphones tutorial in some days!

 

[Morne]

The girl came back from her vacation and got me her accordion for the final tuning. I recorded this video days ago and forgot to upload it, it sounds really better now.

Well done. That is indeed much better.
How much did the tuning change going from the Arietta block to the Concerto block and back again?
And how did you measure the final tuning inside the accordion? Did you block the reeds somehow, or did you tune by ear?

With the shared hole, glued-in block Im working on now, I find myself weaving a wide strip of paper over the valves and under the reeds in order to silence the one side. Now this probably affects the pressure to some degree since the one side is now no longer releasing air properly. But thats as close as Ill get until I figure out a trick I have in mind with the Tuner app I use.

 

[Sebastian Bravo]

Well done. That is indeed much better.
How much did the tuning change going from the Arietta block to the Concerto block and back again?

It was Ok, just 1 or 2 cents changed, to correct them i removed one of the reeds and put a piece of wood on its place, so i could play just one note. Then the same with the other reed. Got a perfect tuning. I used this chart from a russian forum, to get a bassier sound:



The tuning is OK, treble side is 5hz tremolo on the first notes, then 4hz on the middle, 3hz on the higher notes.

The girl is very happy with her accordion now because it had a really big change. I did all the tuning, valves and wax process in 2 weeks, i think i got this one pretty quick.

Try using Dirks Accordion tuner Trial version. It has some notes missing for tuning, but it will work on most notes correctly to tune in octaves perfectly. For the missing notes on the App, use your method.

 

[Morne]

I used this chart from a russian forum, to get a bassier sound:

bas33.png

While I can make sense of the Russian on the chart, I am not 100% sure if I am interpreting it correctly. Are they saying for a 4 reed bass you would (de)tune the reeds as follows:
- low bass +4
- high bass +2
- low chords -2
- high chords +0

Actually, if you still have the link where you got it from, I would like to have a look at it.

 

[Sebastian Bravo]

- low bass +4 cents
- high bass +2 cents
- low chords -2 cents
- high chords +0 cents

Actually, if you still have the link where you got it from, I would like to have a look at it.

It is correct Morne, try that tuning!

http://forum.mirbajana.com/viewtopic.php?f=10&t=36

This is where i got that tuning chart. There are other 3 versions of it.