5. Kamancheh Tarhu - Cone making, setup and adjustments
Peter Biffin September 2019
There have been 2 cone systems in use for kamancheh tarhu over the past 15 years, with a third concept introduced in mid 2019. This post describes kamancheh tarhus made using type 2 and type 3 cones (see Post #3 for details of these cone types). The third design will be described in more detail when it has been tried and tested a little more.
Type 2 cone
Most of the kamancheh tarhus I have made to date have used this system, including the instruments made for Habil Aliyev and Kayhan Kalhor. The body for this design was 210 mm diameter x 165 mm high.
This type of cone has a fixed bridge and a moulded edge made from polyester cloth impregnated with high-grade epoxy resin. The cone size is: 30 mm x 3 mm button on top of the cone, 48 mm high, the width of the cone itself is 144 mm and the total width including the edge is 166 mm.
Making the edge: The cloth is 0.08 mm thick polyester, which in the West is used as a very light-weight lining for ladies’ dresses. The epoxy resin used is highly viscous, about the same as varnish. After the cloth has been wetted out with resin, any excess is scraped off so that the absolute minimum needed for uniform wetting is used. The cloth is then placed between two layers of thin plastic film and pressed in a mould. The female part of the mould is first turned on a lathe, and then a silicon male mould is cast from that. The groove in the mould is 10 mm wide x 3.5 mm deep.
Making the cone: The formula for making a straight cone is shown below (thanks to Justin Montefiore).
The stock used is western red cedar (WRC) that is at the lighter end of this timber’s weight range – 290 > 320 kg/m3. This needs to be carefully quarter-sawn to ensure maximum cross grain stiffness. Strips are cut on a bandsaw and then thicknessed to 0.7 mm. Once the segments are cut out they are glued together and while still flat, are finally thicknessed (by sanding with the flat cone-blank on a piece of glass). Final thickness is 0.7 mm in the middle down to 0.6 mm on the edge. The outside surface of the flat cone-blank is steamed to make it swell up and form into a cone shape, and the final join is made (reinforced by a thin piece of cross-grain balsa). When the glue is dry, the 30 mm x 3 mm cone-button of WRC is glued onto the apex of the cone. At this stage my cones weigh around 6 gms. The cone can then be glued onto the edge, and then the edge glued on to the perforated plywood cone disc (see below).
Once the whole cone assembly is glued onto the cone disc, the tap-tone should be around 120 -130hz (see notes on using a simple Spectrum Analyser app at the end of Post #3). A resonant frequency in this range for such a small cone is really very low – type 2 cones work a very different way to type 1 cones. If you use a floating bridge on a type 2 kamancheh tarhu cone with a tap-tone around 125Hz, the result will not be useable……and likewise if you use a fixed bridge on a type 1 or 3 cone with a tap-tone of over 300hz it won’t work.
Setup and Adjustment
1. Before fixing the bridge to the top of the cone (by glue), make sure the slot in the bridge that accepts the bridge pin is parallel to the bottom of the bridge-well (make adjustments to the bottom of the bridge if necessary).
2. Once everything on the lower part of the bridge is nicely adjusted, fix it to the cone.
3. Proceed to adjust the top of the bridge in relation to bridge height/string action and bridge curvature. While it is not always possible to set the final height at this stage, it is worth trying to get it as close as possible.
4. Adjust string height at the nut. While this doesn’t have any particular significance for tarhu making, this is a reasonable point to do it.
5. Make an initial adjustment to the string-angle block immediately behind the bridge so that each string has just enough downward angle where it passes over the bridge so that it won’t buzz. This angle is much smaller than you would think if you are comparing it to something like a violin – the angle required is barely perceptible. Adjustments can be done by filing grooves of the required depth in the string-angle block. However, without sympathetic strings in the way, it is possible to install screws in the string-angle block so that adjustments to the string angle can be made just by winding the screws in or out.
6. Adjustment of string angles differs from one side of the bridge to the other – the treble side and the bass side have different roles to play. The first string usually requires the greatest amount of string angle as it needs to maintain firm enough pressure between bridge and bridge pin that there are no rattles. The other strings then have to be adjusted so that the bridge stays level. If screws are being used in the string-angle block, it is good to have enough head on the 1st string’s screw so that if necessary the 1st string can be hooked under the screw head and the string pulled down as required.
7. Too much down pressure on the treble side will make the bridge pin difficult to adjust.
8. Too much down pressure on the bass side will tilt the bridge/cone assembly sideways. The treble side of the bridge can’t be pushed lower by string tension because the bridge pin prevents this, but the bass side can be depressed by excessive down pressure. The resulting tilt in the bridge/cone assembly is most easily detected by looking at the profile of the cone’s moulded edge on both the treble side and the bass side.
9. Adjusting of string angles needs to take bridge pin location into account. The usual starting point for a type 2 cone is to have the bridge pin right on the treble edge of the bridge. When the bridge pin is moved towards the centre more than a few millimetres, the balance changes and the downward string pressure applied by the first string will start to tip the bridge. Fortunately bridge pin position for type 2 cones is almost always right out on the treble side of the bridge, so this requirement is not too onerous……..just remember it is a factor if you want to fully explore bridge pin positions in this cone type.
10. Once you are happy with the bridge pin position, take some blue tack and use this to explore the high frequency part of the sound (blue tack is the sticky putty used to stick posters onto a wall). Start off with approx. 0.5 gms of blue tack and stick it to the treble side of the bridge. This should cause a noticeable change in the high frequencies - experiment with varying amounts/locations of blue tack until an ideal is reached. Putting blue tack on the treble side has much more effect on high frequencies compared to putting it on the bass side (the reason is not because that is where the highest pitch strings are – blue tack on the treble side also reduces the high frequency component in the lowest strings as well).
11. In this cone design it is often necessary to increase the gauge of the 1st (highest) string, both as a further method of adjusting high frequencies and also to create a better balance between upper and lower ranges. If a violin string-set is being used, the top string will be around 0.010” (0.25mm). Explore increasing this to at least 0.012”– Habil Aliyev preferred either 0.14” or even 0.015”. Increasing the gauge this high would only be recommended for someone who was interested in the Azeri sound-palette - a 0.015” first string does not produce enough high frequencies for Persian tastes.
Adjusting a Type 2 Cone.
Because the moulded edge is curved and extremely thin, very little adjustment of the edge stiffness is possible for a type 2 cone. One can explore increasing the stiffness of the edge by adding extra strips of either cloth or carbon fibre to the inside of the edge - however, this approach tends to be “pot luck” – if you happened to guess the right amount of additional cloth, then great – if not and you made it too stiff, it is very difficult to change other than by making a series of tangential cuts in the moulded edge with a sharp fine blade.
This type of cone is usually very even and wolf notes are generally not an issue. If you do happen to have some troublesome notes, you would now have to do whatever is possible to minimise these.
Do another round of adjusting the bridge pin/pillars, making sure they are in the optimum location. If there are problem notes, they will usually be worse with the bridge pin pillars wide open – a compromise often has to be made between preferred position and wolf-note mitigation.
Do a round of fine tuning the amount of blue tack on the bridge. If you find that even with no blue tack on the bridge you still want more high frequencies, then the cone should be removed and some weight taken off the bridge. Because of the bridge being glued to the cone, this has to be done with great care.
If there are still any problem notes, take about 2gms of blue tack and start exploring locations on the cone that could do with a little damping. If there is a soundhole in the cone, a good place to start is directly opposite the hole – the extra weight added by the hole reinforcement can cause non-linear movement in the cone. If weight in this area has a positive affect (or any other area you find), fine tune the amount of blue tack required and squash it out flat, making sure it is really stuck on well. The alternative is to weigh the blue tack accurately and then glue on a cube of hard wood onto the cone in the same location, making sure that the cube weighs the same amount as the blue tack.
Type 3 Cone
These were introduced in 2016. One of the requirements of this design is that the active foot of the bridge needs to drive the cone at its geometric centre. This requires the cone to be mounted in the instrument slightly to one side…..see below.
In order for this to be possible, either the cone has to be of smaller diameter, or else the body has to be slightly bigger. I chose to do a little of both: the cone diameter was reduced to 160 mm x 28 mm, and the body size was expanded to 228 mm x 175 mm.
See post # 1 for instructions on making a type 3 cone. Have a look at the cone profile at the bottom of that Post – I suggest using that profile as a starting point. If the cone was made from 0.7 mm western red cedar (WRC) and is 160 mm wide x 28 mm high with a 15 mm x 3 mm button of WRC on top of the cone, it will most likely have a tap-tone around 400 > 420 Hz when first tapped. Ideally that will now be carefully reduced with successive listenings down to around 330Hz.
The weight of the cone is not so easy to determine because some of the edge will be glued to the perforated plywood cone disc and will therefore not be part of the effective surface area. The effective area of the cone should weigh around 7gms
Initial adjustment of a kamancheh tarhu with type 3 cone
Compared to a type 2 instrument, a tarhu with a type 3 cone is able to go through quite a different process once it is strung up. This phase is influenced by two elements – necessity and opportunity: 1. because of how light, delicate and responsive tarhu components are, full potential cannot be achieved until all these components are brought into balance with one another; 2. because type 3 cones have an edge whose thickness/stiffness can be easily changed, and because of the accessibility to all major components, there is an unprecedented opportunity for fine-tuning of the sound. The tarhu can be listened to, pulled apart without slackening the strings, altered and put back together in under 5 minutes……..with the memory of its pre-change sound still very strong to compare to.
Below is a well-tried series of steps for this process. The first section deals with setting up the strings and the area around the bridge (similar in many respects to that described above for type 2 cones). The second section deals with analytical listening and the third section with adjusting the cone.
Around the Bridge
1. Adjust the length of the bridge stick carefully such that when the bridge is sitting on the bridge pin, the bottom of the bridge is parallel to the bottom of the bridge-well. This is necessary so that the bridge pin will be able to slide right along the bottom of the bridge without changing the bridge angle. If the bridge stick is too long, when the pin moves towards the centre of the bridge the pin will lose contact with the bridge and will rattle. If the bridge pin is too short, when the bridge pin is moved towards the centre the pin will raise the bridge and break the contact between bridge stick and cone, again resulting in a rattle.
2. If the bridge stick is inadvertently made too short, a small disc of wood can be glued on. Because this is always a possibility, it is best to leave the end of the bridge stick unrounded until all adjustments are complete.
3. Once everything on the lower part of the bridge is nicely adjusted, proceed to adjust the top of the bridge in relation to bridge height/string action and bridge curvature. While it is not always possible to set the final height at this stage, it is worth trying to get it as close as possible.
4. Adjust string height at the nut. While this doesn’t have any particular significance for tarhu making, this is a reasonable point to do it.
5. Make an initial adjustment to the string-angle block immediately behind the bridge so that each string has just enough downward angle where it passes over the bridge so that it won’t buzz. This angle is much smaller than you would think if you are comparing it to something like a violin – for tarhus the angle required is barely perceptible. Adjustments can be done by filing grooves of the required depth in the string-angle block. However, without sympathetic strings in the way, it is possible to install screws in the string-angle block so that adjustments to the string angle can be made just by winding the screws in or out.
6. Adjustment of string angles differs from one side of the bridge to the other – the treble side and the bass side have different roles to play. The first string usually requires the greatest amount of string angle as it needs to maintain firm enough pressure between bridge and bridge pin that there are no rattles. The other strings then have to be adjusted so that the bridge stays level. If screws are being used in the string-angle block, it is good to have enough head on the 1st string’s screw so that if necessary the 1st string can be hooked under the screw head and the string pulled down as required (see image above in Type 2 section).
7. The other factor that needs to be taken into account in adjusting string angles is that the bridge stick should only put just enough downward pressure on the cone so that the stick and cone don’t rattle against each other. You can check this by plucking the bass string and immediately pushing upwards on the bass edge of the bridge with your thumb, lifting it up away from the cone. If the sound of the bass string is cut off straight away (because the bridge stick has been lifted off the cone) then there is not enough downward pressure from the lower strings. If the thumb pressure continues to lift the bridge and the sound is not cut off, then there is too much downward pressure. Even though it would be tempting to apply plenty of downward pressure so that rattles are never an issue, soundwise it is important to not subject the cone to any more pressure than is necessary.
8. A way of achieving minimal pressure on the cone without rattling is to use powerful neodymium magnets inlaid into the end of the bridge stick and the top of the cone. The ones I use are cylinders 3 mm diameter x 2 mm long. They weigh only 0.1 gms each, so in that location they have a minimal effect on the sound.
1. The first stage is knowing roughly where the cone is at its beginning point, and roughly where you are trying to go with it. Ideally you would have noted the tap tone of the cone before it was mounted in the instrument. Assuming that the cone has been made using the measurements outlined above , the cone would be slightly over-built, and the primary tap-tone will therefore be higher than the expected final result.
2. Make a guess as to what the most likely bridge pin position will be (refer to Post #2 - it will be more towards the centre of the bridge). Set the pin’s pillars to about half open.
3. Play every note on the instrument, in every possible location on the fingerboard. Build a picture of the sound’s strengths and weaknesses and note down the location of any wolf notes.
4. Get to know any wolf notes intimately: does the wolf note appear on the same scale interval in successive octaves (and if so, make sure you know which is its lowest manifestation as that is most likely to be its principal identity); identify where on the instrument the wolf note makes its presence felt the strongest (usually high up on the thicker strings).
5. Turn to the bridge pin and explore various positions along the bridge and various pillar positions, listening for general sound-types.
6. Combine steps 4 and 5, investigating what effect pin/pillar positions have on wolf notes.
7. Once an optimum version of the sound has been achieved, take some blue tack and use this to explore the high frequency part of the sound. Start off with approx. 0.5 gms of blue tack and stick it to the treble side of the bridge. This should cause a noticeable change in the high frequencies - experiment with varying amounts/locations of blue tack until an ideal is reached. Putting blue tack on the treble side has much more affect on high frequencies compared to putting it on the bass side (the reason is not because that is where the highest pitch strings are – blue tack on the treble side also reduces the high frequency component in the lowest strings as well).
8. Step #7 is a preliminary step only in relation to high frequencies – come back to this step again later. Note that sometimes the application of approx 0.5gms of blue tack to the bridge has an effect on the sound that is really negative (making it too dull and muffled). When even a tiny amount of blue tack is shown to be a bad thing, then most likely the bridge is too heavy. A heavy bridge will reduce high frequency response in a similar way to blue tack. You could explore this a little bit now, taking some weight off the bridge, or else leave it till later. For this type of tarhu, the bridge should weigh no more than 2 gms, preferably closer to 1gm. This is difficult to achieve – watch for a separate post about bridges (coming soon).
9. At the end of this phase it would be good to have determined: where you like the bridge pin/pillars to be; whether or not high frequency damping is desirable; what are the locations of any problem notes and how do these interact with pin/pillar positions; whether or not you need to go back to the first phase and make setup changes/adjustments before proceeding to the cone.
10. Don’t be in a hurry – if possible, it is better to do a little bit of listening at a time, spread over several days. Try and come to it from all sorts of different perspectives; in a good mood, in a bad mood; after listening to an instrument you love, after listening to one you hate; trying to ignore the existence of problem notes, focussing strongly on problem notes; in a live room, in a dead room; played by yourself, played by someone else etc etc.
1. Remove the cone from the instrument and make sure that you know where its principal resonant frequency is (tap tone)
2. Hold the cone up to a strong light – if you have used 0.6/0.7 mm western red cedar for the edge, any thin spots will show more light through. Mark these spots so they can then be avoided in the next phase
3. Place the cone flat on the bench, convex side facing up, use a cork block with 120 grit sandpaper and sand the edge firmly. Use maybe 3 firm strokes then rotate the cone to a fresh section, do another 3, rotate, etc etc. Sand radially, in and out towards the centre on each segment in turn using a cork block.
4. Hold the cone by the very edge of the cone disc and tap on the apex of the cone. Note the tap tone. You would be aiming for dropping the pitch of the tap tone between 50 and 100 cents in a single cycle of going around the cone’s edge once. If you didn’t drop the pitch that much in one cycle, go round again – it needs to have dropped by 50 – 100 cents before its worth having another listen. If it has dropped by more than that, have a listen straight away and go a little more gently on the next round of sanding.
5. Reinstall the cone, check tuning and have a “general impression” listen. Next go to any wolf notes. Even though any wolf notes won’t be the same pitch as the free-standing cone’s tap note, they will usually have dropped by the same amount. Next check wolf note severity – sometimes wolf notes decrease in severity as the cone stiffness is reduced, sometimes not. It is important to know what territory you are heading into……..if they are getting worse, go very cautiously in relation to further sanding. Usually if the cone and edge are initially made of 0.7 mm western red cedar, there will be maybe 300 cents (3 semitones) that is available between the starting point and the most likely finishing point.
6. The finishing point……..how to determine this?.......By building up a knowledge base from what happens to the sound when you go too far. You can use my end-point tap tones as a guide, but these will only help if the cone has the same width and height as mine. The other thing is that I never just reduce the thickness till that tap-tone is reached and then leave it…..the tap tone defines a region for me, not a specific location. From experience I recognise certain qualities emerge in the sound as the stiffness is reduced. I check frequently to see if these qualities are still developing as I proceed further. A point comes where, from past experience, I know that if I keep going I will start losing quality rather than gaining it. However….i only know this because I have gone too far often enough in the past that I now know when to stop.
7. What to do when you have gone too far? Put some stiffness back into the cone edge. If you have been recording the tap tones all the way through, it should be possible to say something like “I liked it the most around 330hz, it is now 320hz”. To re-stiffen the edge of a type 3 cone, add strips of carbonfibre radially along the segment joins.
8. Once the glue is dry, check tap tone. Hopefully it will be just above your target pitch. If not, add some more stiffness – usually easiest to put it in the same places, just increasing the thickness of your initial strips. Once the tap tone is just above your preferred location, have a good listen and decide whether you will accept what you have or whether you will keep hunting. When you decide you have finished adjusting the stiffness, proceed to other final adjustments.
9. As you have been doing the final stiffness adjustments, hopefully you would have been keeping track of any problem notes. You now have to do whatever is possible to minimise these. Do another round of adjusting the bridge pin/pillars, making sure they are in the optimum location. If there are problem notes, they will usually be worse with the bridge pin closer to the centre – a compromise often has to be made between preferred position and wolf-note mitigation.
10. If problem notes remain in a type 3 cone, this design responds well to increasing cross-grain stiffness towards the edge of the cone. This can be done by using either strips of carbon-fibre or WRC that can be glued onto the cone in concentric circles. These strips make it quite difficult to continue reducing edge stiffness, so make sure you have finished that process before gluing the carbon-fibre on. The job is easier if you use lots of small strips end to end, joining at the segment joins. Because these strips are concentric and cross-grain, they don’t change the location of the tap tone very much.
11. Do a round of fine tuning the amount of blue tack on the bridge. While type 3 cones usually have a strong first string, you might also wish to try increasing the gauge of the 1st string, especially if the instrument is to be used for Azeri music. If the high frequencies are still not as you would like them, consider either reducing the weight of the bridge, or if you have too many high frequencies, maybe make a new bridge that is a bit heavier (or use a bit more blue tack)
If there are still any problem notes, take about 2gms of blue tack and start exploring locations on the cone that could do with a little damping. If there is a soundhole in the cone, a good place to start is directly opposite the hole – the extra weight added by the hole reinforcement can cause non-linear movement in the cone. If weight in this area has a positive affect (or any other area you find), fine tune the amount of blue tack required and squash it out flat, making sure it is really stuck on well. The alternative is to weigh the blue tack accurately and then glue a cube of hard wood onto the cone in the same location, making sure that the cube weighs the same amount as the blue tack.
Type 2 versus Type 3
In its upgraded form, with a cone disc added and tuned soundholes added to the back of the body, a type 2 kamancheh tarhu can be a most enticing instrument to play. The lack of wolf notes is a real delight, and one can follow a musical idea right along a particular string without having the musical thread broken by irregularities. Responsiveness to changes in bow position and weight of bow stroke is high. They are, however a bit temperamental - one has to watch that the bridge is not tipping the cone sideways, or that over time the strings don’t pull the bridge forwards. Either of these circumstances, if left unattended, will eventually change the profile of the moulded edge, and the response will gradually change.
Type 3 kamancheh tarhus have a bit more power, right across the range. The first string speaks more easily, and it is less likely that one will be experimenting with different first string gauges. Responsiveness to changes in bow position/stroke weight is high, but not quite as much as type 2. Related to this is that one has to work a bit harder as a maker to provide the same amount of high frequency sting that is available with a type 2 cone. While there are fewer components in the type 3 cone construction, configuring the cone within the body is more complicated because of the need to offset it. The possibility of running into wolf notes is probably this cone type’s least favourite characteristic…….however, the various cross-grain stiffening options could well change that picture altogether, once explored further.