What to do if the back plate is still in the bouts, and still attached to the neck?
If you want to improve the sound of a poor-sounding violin (and there are a lot of them!) then taking off the belly or front plate is one thing - the fingerboard and the violin belly plate should ‘pop’ off if half strength animal glue has been used to glue them on. They are designed to do that.
However, taking a back of is not easy - you are likely to cause a lot of damage to the violin and on top of that it probably will not glue back on without distorting, or the gap around the edges will be uneven. This is a job for professionals using custom molds and specialist clamps.
However I have devised a method of estimating quite accurately :-
(a) the Mode 5 and the Mode 2 frequencies of the back plate as if it were free standing, and
(b) optionally the weight of the back plate alone. The low weight of a back is a measure of its quality, but it is not essential to know it.
The first part (a) is by loosely using a method quite similar to Patrick Kreit’s, but I have found that the back plate Mode 5 usually splits into 2 frequencies that need to be combined.
The second part (b) uses a method of estimating the weight of all the parts based on knowing where their centre of gravity is and subtracting them from the overall weight.
Estimating the Mode 5 of the back plate
The Mode 5 of a back plate is reduced by about 15% in the bouts, as can be seen in John E. McLennan’s paper (UNSW) on page 5 & 6 and also in work published by Jansson.
However, usually Mode 5 divides into 2 frequencies up to 55 Hz apart, one either side of 300Hz. So I can get a good estimate of a back plate’s Mode 5 while it is in the bouts and still attached to the neck but with no fingerboard.
I do this by ‘combining’ these 2 ‘modified Mode 5’ frequencies measured with the back plate still in the bouts using a ‘weighted average’ method: a look-up table is shown here on the right.
Just click on it and print it out. The instructions are at the top right.
The Mode 2 of a back plate when in the bouts is only slightly increased, but the neck with no fingerboard unfortunately has a resonant frequency at almost exactly the same frequency, so we have to move the neck’s resonance out of the way to measure it. Get in touch with me if you want to know more.
You will need to look at the ‘hierachy’ charts and figures given on ‘Kreit & parts hierachy’ page because you can then see that all the parts of a violin contribute to the whole. When a belly or back plate is put into the bouts to make up the body its frequency (and weight) are then modified, but modified predictably .....
So the sequence for improving a poor violin's tone is like this.
- Measure the B1+ and B1- body frequencies of the violin before you take it apart!
Carefully remove the fingerboard.
You will then have data of what they are then for the plates before you thin them in step 10 below because you’ll know what the Mode 5 of back and belly plates in steps 5 and 8 below. This can help you decide whether to thin the back or front: thinning either affects both Modes B1- ands B1+.
Carefully remove the belly from the bouts.
Simon Stace showed me that this is the only time a hammer is ever used in making or repairing a violin. If you insert a small this chisel under the fingerboard and get the blade into the gap (!) between fingerboard and neck the chisel can be lightly struck to begin the separation of fingerboard and neck. It is then completed with a thin knife ..... but mind your left hand! Use a leather glove.
If the back has not got 4 corner blocks you will need to put some new ones in by cutting some from spruce or willow. The grain should be vertical. Make sure all 6 blocks are good! Weigh the assembly again.
First carefully break the line of the varnish with a scalpel and then use a thin blunt knife inserted under the belly edges and pop it off, do not cut it off.
Weigh the back in its bouts: this is optional.
Estimate the ‘combined’ Mode 5 of the back plate by measuring the 2 frequencies either side of 300 Hz and then combining them using the chart above.
Optionally estimate the weight of the back plate by ‘subtracting’ the weights of the neck, bouts, linings and all blocks. Ask how to do this.
Estimate the Mode 2 of the back plate. Write to me to ask how to do this. Mode 2 (a measure of the cross-plate stiffness) is ideally just below half of Mode 5.
Do any necessary repairs to the belly plate edges and mend any and all cracks first.
Divide1 the ‘combined’ Mode 5 frequency by ~0.85 to get to the true estimate of the back’s Mode 5 free-plate tap tone.
Measure the Mode 5 and Mode 2 frequencies (tap tones) of the belly plate. The Mode 2 frequency of the belly is ideally just below half Mode 5.
Using the chart and tables on the ‘Plate tuning 4 dummies’ page now choose the back and belly Mode 5 frequencies you want to give a B1- to B1+ difference to suit your taste in violin tone.
Now you can start the process of thinning the back and belly plates as necessary to get the ’delta’ (B1- to B1+ interval) you want, as shown on the chart and tables on the ‘Plate tuning 4 dummies’ page. This involves the thumb plane, measuring and checking in a repeated cycle: measure twice, cut once. Of course you may well be limited in your choice by one of the plates being too thin already! Edge thicknesses will need to be put right first!
Optionally Weigh the belly plate: a low weight near 60 grams is a very good thing, but needs excellent spruce. Yours is more likely to be 75 grams or more.
If need be you can raise the belly Mode 5 a little by putting in a new slightly higher bassbar.
If the back plate is too thin in the centre (often the case) you can raise its Mode 5 and Mode 2 frequencies by gluing in a chalk-fitted patch in the centre of the back plate.
I’d like to tell you that this is an exact relationship of belly and back plates to the B1- and B1+ body mode frequencies, but no it’s not: it’s an estimate, but I often have to remove the belly again to get the B1- to B1+ interval where I want it (say 75 Hz) after measuring what it is in the newly reassembled violin at the first attempt.
Do not make the plates too thin. No part of the belly should be less than 2.5 mm. and not less than 2.7 mm in thickness between the f-holes: it is easy to go it too thin.
Glue the belly and then the fingerboard back on and set up the violin again.
Measure B1- and B1+ and record them for your records. And see how much the violin's tone has improved.
It needs to be an oval shape: the longer dimension length-ways to raise Mode 5 and sideways to raise Mode 2. The patch’s grain direction needs to match the grain direction of the back plate. The patch edges need to decrease to zero as it is blended into the back.
One or more layers of maple veneer there can also work.
Examples of how this sequence was used to improve the tone of a particularly poor-sounding violin is shown on the ‘Examples’ pages.
Get in touch for more details!
Footnote 1: With thanks to Sean Hay, Ajax, Ontario, Canada for pointing out this should read ‘divide’ and not ‘multiply!