is not the type of silicone you extrude from a caulking gun.
is a bi-component one that found application also in the medical industry.
it is from Prochima, the silicon is the PRO-LASTIX 30.
No acetic acid is involved (or any other corrosives)
Hello,
Maybe with clocks it is no problem but i am sure ( Supersurfer for sure) heard about negative influences of plastics coverings capacitors.
I am not sure that if you cover the clock in silicone in such a way that the silicone is touching both circuit board and the clock will it not actually transmit vibrations?
The idea from Andrea with the thin wires will be partly cancelled or not?
Sand will work on the outside of the aluminium boxes.
Maybe you can of '' fix '' the circuit by surrounding it wirk cork particles
To be continued as the say.
Greetings, Eduard
Maybe with clocks it is no problem but i am sure ( Supersurfer for sure) heard about negative influences of plastics coverings capacitors.
I am not sure that if you cover the clock in silicone in such a way that the silicone is touching both circuit board and the clock will it not actually transmit vibrations?
The idea from Andrea with the thin wires will be partly cancelled or not?
Sand will work on the outside of the aluminium boxes.
Maybe you can of '' fix '' the circuit by surrounding it wirk cork particles
To be continued as the say.
Greetings, Eduard
Hello,
But Andrea decoupled the clock from the board by using thin wires.
And then you will attach your silicon creation to the board . Won't this diminish the effect of the thin wires.
Maybe you can make a little box from your silicon that will just cover the sensitive parts on the board but that will only be connected to the board and not touch the parts at all?
Greetings, Eduard
But Andrea decoupled the clock from the board by using thin wires.
And then you will attach your silicon creation to the board . Won't this diminish the effect of the thin wires.
Maybe you can make a little box from your silicon that will just cover the sensitive parts on the board but that will only be connected to the board and not touch the parts at all?
Greetings, Eduard
i am not saying that what andrea is did/suggested/measured is wrong... is still one of the best solution, but not the only one.
you can surely follow what he did, sure of good results, afaik all the measurements was done with the crystal attached in the very same way he reported on the manuals.
I just wanted to add my experience on top, i have matured quite some insights on that matter and i can't resist to do tests.
Anyway, don't forget that the crystal shell is still attached to the pcb via neoprene "studs" with Andrea's solution.
at high frequency, thoose small cilinders acts like a solid piece of concrete (... i am exagerating..)
they haven't nor the flexibility nor the mass to stops what really matters for such tiny piece of crystal quartz.
for crystals > 5/6 mhz physical size makes them more mechanically resonant with high freq rather than lows.
at any rate, i think that matter is still "marginal" (the whole vibration isolation topic...). What sounds good is the topology and the careful component selection.
As i am not pretending to be right, but just sharing my thoughts :
Pick your poison and live happy
Last edited:
Hello,
The only clock i ever touched is the clock on my DDDAC mainboard.
I forgot about the neoprene Andrea is also using!!.
If you want damping materials to work you should take about its properties . I think if you use a piece of neoprene, 10*10 and 5 mm thickness and you put a clock on top i think it wont work. If you install the schock absorbers from a Fiat Ducato on your Fiat 500 they wont work.
Maybe you could add weight to the clock. You could cast a lead housing . With a cavity inside that will allow you to fix the clock with something that will prevent the lead acting as a heat sink. Then mount this housing onto the board with a single vibration damping element that is optimized for this load.
A Japanese company called Taica makes them . NOT sure if they also make them for very tiny loads. Available at Mouser.
ALMOST sure that if you execute this well it will be worthwhile .
Greetings, Eduard
P.S As you can see in the attachment i have worked with lead before
The only clock i ever touched is the clock on my DDDAC mainboard.
I forgot about the neoprene Andrea is also using!!.
If you want damping materials to work you should take about its properties . I think if you use a piece of neoprene, 10*10 and 5 mm thickness and you put a clock on top i think it wont work. If you install the schock absorbers from a Fiat Ducato on your Fiat 500 they wont work.
Maybe you could add weight to the clock. You could cast a lead housing . With a cavity inside that will allow you to fix the clock with something that will prevent the lead acting as a heat sink. Then mount this housing onto the board with a single vibration damping element that is optimized for this load.
A Japanese company called Taica makes them . NOT sure if they also make them for very tiny loads. Available at Mouser.
ALMOST sure that if you execute this well it will be worthwhile .
Greetings, Eduard
P.S As you can see in the attachment i have worked with lead before
Attachments
Right.
Right.
dapening mass must be calibrated upon the mass of the item you are trying to damp. Past a certain ratio you have no benenefit. So it is marginal the difference between 50grams and 200kg when you are trying to damp 0.05 grams of quartz.
(this is an over semplification, harmonic resonances and many other aspects must be taken into account, but if you speak about lead vs silicone we can limit the scope to mass)
I think it can benefit from that "heatsinking" effect, by increasing the thermal mass you render the quartz far less suspeptible to thermal short term variations, when the assembly came to equilibrium there's more inertia to beat to make a variation.
Conservation of energy reign
Last edited:
Hello Italy,
I remember in the eighties the French published about a lead housing for the Denon DL103 cartridge.
If you use just one '' grommet '' made by Taica you will need a minimum weight of 50 grams and a maximum weight around 180 grams. So i think it can be done!
I presumed the clock would raise in temperature after switching on and then after reaching a certain temperature it will stay there. IF you get a close bonding with the lead the temperature will stay at the same level. I mean it will copy the surrounding temperature and not be changed by the clock working or not.
If the clock doesnt care about the temperature as long as it stays the same then it will be okay.
Greetings, Eduard
I remember in the eighties the French published about a lead housing for the Denon DL103 cartridge.
If you use just one '' grommet '' made by Taica you will need a minimum weight of 50 grams and a maximum weight around 180 grams. So i think it can be done!
I presumed the clock would raise in temperature after switching on and then after reaching a certain temperature it will stay there. IF you get a close bonding with the lead the temperature will stay at the same level. I mean it will copy the surrounding temperature and not be changed by the clock working or not.
If the clock doesnt care about the temperature as long as it stays the same then it will be okay.
Greetings, Eduard
Dear Holland,
absolute temperature value is good only for frequency accuracy.
as already stated many times, we are not interested in it... so temperatura can be kept within reasonable range and nothing "bad" happens.
You can do both ways, or increase the thermal mass with thermal conductive materials, or you can thermally isolate it.
Both has pros and cons.
Coming back to the vibration topic...
At work i have developed a small test jig to measure vibration (for relative measurements, not absolute) that aided the design of a product of us. As soon as i have finalized the prototype of the "silicone shell" i will test it.
In the mean time i have in mind a quick test here at home... if i can i will make a video of it
Italy over and out
Last edited:
Hello,
So the idea with oven some time ago had the main goal to keep the temperature at the same level all the time.
I remember Jean Hiraga writing about thermal distorsion with transistors that could be dealt with by making them work at a higher temperature so the temperature would fluctuate less. But then they said the transistors sounded better at a higher temperature.
If with clocks the only issue is keeping the temperature the same then it can be done two ways. BUT i think i would use some 3 mm foam to fix the clock into the lead housing.
Greetings, Eduard
Someone will have to try. I think if we go the extra mile with so many boards for just the clocks we should invest time in this matter too!
So the idea with oven some time ago had the main goal to keep the temperature at the same level all the time.
I remember Jean Hiraga writing about thermal distorsion with transistors that could be dealt with by making them work at a higher temperature so the temperature would fluctuate less. But then they said the transistors sounded better at a higher temperature.
If with clocks the only issue is keeping the temperature the same then it can be done two ways. BUT i think i would use some 3 mm foam to fix the clock into the lead housing.
Greetings, Eduard
Someone will have to try. I think if we go the extra mile with so many boards for just the clocks we should invest time in this matter too!
there's no proof that higher temps makes clocks sounds better... at least to my knowledge.
Surely a controlled temperature is useful for keeping long term freq stability.
Usually the temperature is kept higher than ambient just to have headroom (so the control loop of the oven can go higher and lower to compensate).
Quite pathetic attempt to evalute with "household" tools
i have used two cellphone, one with high datarate acquisition over accelerometers, while the other was generating vibrations (with the vibration motor).
i will redo all of this with proper sweep test over wide range of frequencies...
but for the moment i'll leave you with this (imho promising) results.
the first is the transmitted vibration with the two devices coupled rigidly together.
the second is with 3 stud of 5mm diameter/5mm height closed cell neoprene.
i also tested by adding a whole 500gr spool of solder between the two devices, at this high frequencies was almost like having a direct connection
the last is the two devices isolated by the silicone dampener.
I think there are some misconceptions about temperature. Laptech publishes a temperature at which each oscillator produces optimum phase noise. In the case of the last GB, my 45/48MHz crystals came with specific temperature spec around 85C that corresponds to their published Q. I believe that Andrea found that temperature was less important for the performance of the 5/6MHz family and implemented without the oven making our lives much simpler. If you were looking for the ultimate performance I guess that temperature stability and a specific temperature would be the answer.
As I explained several time in digital audio we look for short term stability.
Long term stability does not matter unless you have to synchronize other devices.
If you need long term stability you can add a oven by yourself.
We have tested the oscillators with and without the oven and since there was no phase noise difference we decided to avoid the complexity and the cost of the oven.
So there are no plans to implement the oven.
Long term stability does not matter unless you have to synchronize other devices.
If you need long term stability you can add a oven by yourself.
We have tested the oscillators with and without the oven and since there was no phase noise difference we decided to avoid the complexity and the cost of the oven.
So there are no plans to implement the oven.
i am sure of that !
i want only to take a peek at the document of laptech that wlowes are referring to, just to be sure what we are deling with.
In the end, if the measurements was the same... then there's no need to add complexity over it.
but the curiosity remains...
i suspect, and is just a suspect, that the thermal data is referring as measure of how much/what is ideal when laptech's are used for OCXO's.
not every quartz can handle the temps usually needed for an ocxo with ease, and who has the technology that keeps optimal phase noise at ocxo's temps then surely is good marketing material.
Last edited:
The phase noise measurements are related to the oscillator and not to the crystal.
The crystal itself does not oscillate so there is no way to measure its phase noise until it's operating in an oscillator circuit.
The temperature specified by the crystal manufacturer for a SC-Cut crystal is the ideal temperature at which you get the exact oscillating frequency and the best long term stability.
SC-Cut crystals have a specific inflection curve related to the temperature and usually the inflection point is placed in the range 80-100°C, so they need an oven to operate at this temp point.
The operating temperature does not affect the short term stability, there are no phase noise difference at 25°C and 90°C.
As I already said we have measured the phase noise with and without the oven and there was no difference.
The crystal itself does not oscillate so there is no way to measure its phase noise until it's operating in an oscillator circuit.
The temperature specified by the crystal manufacturer for a SC-Cut crystal is the ideal temperature at which you get the exact oscillating frequency and the best long term stability.
SC-Cut crystals have a specific inflection curve related to the temperature and usually the inflection point is placed in the range 80-100°C, so they need an oven to operate at this temp point.
The operating temperature does not affect the short term stability, there are no phase noise difference at 25°C and 90°C.
As I already said we have measured the phase noise with and without the oven and there was no difference.
Attachments
AT the time of the last GB when ovens and temperature were a consideration, we received the test data from Laptech with our crystals.
The Well Tempered Master Clock - Group buy See Post 359.
- Status
- Not open for further replies.