The article is a nostalgic reminder of when I was first interested in electronics and got interested in quartz resonators for projects. Grinding down old war surplus FT243-style crystals and other types so as to resonate at frequencies I needed was a common practice with hobbyists back then.
I used many similar techniques to the same end of removing quartz which raised its frequency. Grinding materials included abrasives such as jeweler's rouge, cerium oxide, commercial polishes such as Brasso and Silvo and even HF solution. I'd place the quartz on a small section of plate glass and slide it through a slurry of the abrasive periodically testing its frequency until I'd reached my target.
There's an art to this that's too long to mention here except to say abrasives were used strategically, course grinding would get me near the desired frequency and I'd finish off with a fine abrasive. Then there was the job of re-aging the crystal after its recent abuse to increase its stability. Other techniques were involved such as not lowering its Q factor, etc. which I'll not cover here.
The most desired crystal cut was from the XT-plane (being the most stable) but it was generally difficult to get as it's only a small section of the quartz crystal (also each cut oscillates only over a limited range of frequencies). I used to have a book that explained these cuts in detail, their frequency ranges and electronic properties along with the basic crystallography which I lost years ago. A quick glance at the book would have shown that a great deal of science, engineering and skill is involved in the selection of quartz and its manufacture into useful resonators.
BTW, the mentioning of HF will likely horrify chem-phobic readers. We were well aware of its dangers and took special precautions never to come in contact with it.
Further to my unforgivable boo-boo. This article provides a pretty good introduction to the subject and there's a diagram of a quartz crystal at the beginning indicating the position of AT cuts in relation to the crystal and other cuts.
I've seen much more detailed diagrams of the crystal and cuts/cutting angles but I can't find one online.
.
Those precision diagrams are required by those who cut the quartz at the start of the manufacturing process.
> BTW, the mentioning of HF will likely horrify chem-phobic readers
I would not say I am chem-phobic, but yes indeed that stood out. HF is nasty stuff, and yes requires some care I suspect.
The other details are fascinating, though - the intersection of mechanical, crystallographic, and RF (?) properties of a crystal that you can adjust through abrasives and selection of the cut.
Working with HF was extremely difficult before WWII, but it has become much easier after the invention of Teflon.
Teflon is not affected by HF, so if you use only vessels of Teflon to hold the HF solution and tweezers made of Teflon for handling anything that you submerge in the solution of HF, it works fine.
Besides using Teflon for anything that is in contact with the HF solution, you must do all work under a hood that evacuates the vapors of HF emitted by the solution, otherwise handling a HF solution would be very dangerous. It is good that gaseous HF is lighter than air, so after being evacuated it will continue to rise in the air, while becoming more and more diluted.
In my case the HF came in small black plastic cylindrical bottles of 50 perhaps 75 ml (could have been 100ml but that seems a bit big). Back then they were commonly available at local hardware stores.
The bottle was flat on top from which protruded its 'neck' — more a threaded spigot of perhaps 8mm in dia. and about 10 to 12mm high with a small hole in it for the HF to exit. A red plastic cap screwed onto the thread to seal the bottle.
If the bottle were ever knocked over (which it never was) very little would have spilled out (picture the threaded top and screw cap on the small bottles of Tabasco sauce and you've pretty much got it).
So the HF only came out in drops which were poured directly into a small plastic beaker of about 50ml containing about 20ml of H2O and the crystal. The crystal rested on some finely corrugated plastic and not lying flat on the bottom to ensure the etching process was reasonably uniform.
The operation was done outside and if I recall the HF was neutralised with NaOH.
The soln was very dilute and the process could easily take over an hour and was used mainly to finish off crystals that had already been through the abrasive process.
I don't know what type of plastic the bottle was made of but it was jet black.
I can't recall exactly but I'm pretty certain most crystal are synthetic these days, at least so for the common cuts. I'd imagine most watch quartz would be synthetic but that's only a guess.
I wish I could but it's s long while ago. More precisely the book wasn't lost but loaned to someone who never returned it and I lost track of him.
The book was a harback with a bright yellow dust jacket and I think its title was printed in red. It was about 300 pages and was no lightweight, full of resonance equations, tables etc.—the sort of book you'd find in the lab of a commercial crystal manufacturer.
Can't remember the price but it was damned expensive. At times over the years I've had need to refer to it and I still get a bit peeved when I think about it.
ok, could it be:
Crystal Oscillator Design and Temperature Compensation :Marvin E. Frerking (1978).
or perhaps
Introduction to Quartz Crystal Unit Design : Virgil E. Bottom (1982)?
Could well be either, it was around the late 70s early 80s when the book went walkabout.
Damn nuisance my memory isn't better. However, I'd recognize it instantly if I saw it. Must check those references out. There may be enough info online for me to figure it out. Thanks.
"Crystal Oscillator Design and Temperature Compensation :Marvin E. Frerking (1978)."
Ah, that has to be the book, just did a search and there's a photo of the cover on Amazon that I recognize. I said it was bright yellow with red writing and I was pretty close with its red lines and bits.
240 pages and not 300, but that's not too bad after close on 50 years. And I was correct about the price too, it is expensive, Amazon has it at $110 for the hardback.
I've just spent the weekend tuning brass reeds from an organ. It sounds like a very similar process, except you can grind both ends of the tongue to raise and lower frequency.
Funny you should mention that another hobby of mine was rescuing old harmoniums and making them playable again—you know, fixing and replacing reeds, renewing bug-eaten felt, sealing rat holes in the bellows, etc.
I grant you it's not in the same league as voicing a diapason though. :-)
I reckon adjusting and tweaking things goes with the territory. I'm pretty much at home tweaking crystals, fixing reeds, aligning IF stages in radio and TV equipment, there's much of a sameness in the way one tackles all of them.
BTW, I've actually repaired reeds by soldering them. Not a good fix though as the solder can fatigue with use. Throws out equal temperament a bit too but most can't hear the difference.
By 'repaired' you mean closing a fracture? I'm interested to hear your experiences! Electronics solder or silver solder?
This one is 60 cents sharp across the board (not uncommmon), but I wanted a social instrument. So I brought them down with solder. The bottom two octaves have worked out well. The next two... we'll see. I now have the fear that I've weakened the brass by heating it. But it still sounds nice and speaks well. Fingers crossed.
Oh that is so cool. I played the one in Liepaja, Latvia for a bit and it was absolutely amazing. It's love/hate for me (like the harpsichord), I love the instruments but I usually do not like the music that is played on them because of the grating effect. I have pretty bad tinnitus which really spoils a lot of music for me, extremely annoying.
Ha-ha, I didn't read this until I'd written the above reply, seems we have something in common.
Incidentally, I'm one of those mad people who'll put on a recording of Helmut Walcha playing Bach's Passacaglia and Fugue in C minor on a Silbermann and turn the volume up until the room shskes.
Adding solder has also been frequently used to correct the resonance frequency of quartz crystals that have been ground too much, and I mean during industrial mass-production, not only in a home-lab setting.
How are people sticking stuff to quartz? I know less than nothing, but the pieces of quartz you find in rock don't look like they'd take a solder bond.
I'd assumed that with piezo crystals etc there was a mechanical connection rather than an electrode bonded to the crystal?
But if you can add solder presumably there is some kind of molecular connection with the metal?
Electrodes are deposited on the crystal in vacuum, e.g. by metal evaporation or sputtering, in the same way as they are deposited on the semiconductor crystals used to make transistors or integrated circuits.
The electrodes may consist of multiple layers, a base layer that adheres strongly to quartz and a top layer that is solderable, e.g. made of nickel or silver.
The pins of the package that hosts the crystal resonator are soldered on the electrodes, in places well chosen so that they will not damp much the oscillations of the crystal.
When the mass of the crystal must be increased to shift the resonance frequency, excess solder may be deposited on the electrodes.
There is already a silver patch bonded to the crystal where the wire connects to. Adding weight to that obviously will not make the load curve any better but if you do it with just enough to drop you back down below where you wanted to be then it can be a saving move. You could also put a trimmer in parallel, but that might not have enough range (and can also end up overloading the crystal so the oscillator won't start).
I remember when I worked at Motorola some decades ago, there was a little section in the factory area in Schaumburg that made quartz crystals. It's all long since been sold off (I think all the equipment went to China) but I remember all the signs for cyanide alarms, presumably due to some step in the manufacturing process.
The quartz crystals themselves were grown with carefully controlled levels of specific impurities (like scandium) in order to reduce their temperature sensitivity.
They may have had baths for galvanic deposition of silver. Those typically used cyanide solutions.
Galvanic deposition of silver has been frequently used for increasing the thickness of thin metal layers that had been deposited in vacuum, in order to adhere to the crystal.
Believe or not, when I explained to many non techies how quartz watches work and how any computers’ hardware clocks work in the same principle, they were all surprised how elegant and how efficient the mechanism is. I was also impressed when I first learned about it. True science and engineering beauty.
You mean could you get it down to a low enough frequency? Hmm
I guess you can get down into audio frequency but maybe the amplitude will be tiny, you probably want a piezo mechanism that will give you more of a rumble.
You can get quartz to resonate down to upper audio frequencies with certain crystal cuts and manufacturing techniques but it's difficult. Typically, the lowest frequency in common use is with its use in watches with a frequency of 32,768Hz (that's about the lower limit where manufacturing and frequency combine to make a useful product).
For electronic circuits such as frequency reference markers where frequency stability is important the lowest practical frequency is 100kHz with 1Mhz preferred, and where frequency tolerances are tight 5 and 10MHz are much preferred with operation in a temperature stabilized oven to minimize frequency drift.
The most frequency-stable crystal cuts are at those frequencies, as frequencies increase (say >10MHz to 100MHz), which at the highest frequencies require the crystal to operate in overtone mode, frequency stability again tends to decrease.
As mentioned in the article there are a lot of common watch crystals that oscillate at 32.768 kHz, but they are tuning fork crystals rather than bulk mode (or modern SAW), which have much higher frequencies. It would be challenging to lower the frequency of these, but perhaps you could evaporate Au or W onto the tips and reseal them to get into the audible frequencies. Much easier would be to get two crystals which beat against each other in the audio range. Even a couple of 6MHz crystals 100ppm off would be ~6kHz (temp sensitive), and you would need some driving circuits, but you might be able to hear the beat by driving an electret microphone or a really tiny tweeter coil speaker (probably need an amp though).
It's much easier to build/buy electrical rather than mechanical LC components to hit audio frequencies ~100Hz-10kHz.
Yes, but those normally just convert the electrical oscillations to sounds, they are not parts of the oscillators that determine the audible frequency.
In the past, audio RC oscillators or LC oscillators were used, with the former being preferred as the latter required too bulky inductors to reach so low frequencies.
Nowadays, it is usually simpler to not use any audio oscillator, but to use some microcontroller that divides the frequency of its clock until reaching the desired audio frequency.
And a similar question: if you took a normal magnet and rotated it very quickly, say 10k rpm, would it emit an RF signal at (10k/60) hz? I'm 95% sure the answer is yes but I've never seen this demonstrated.
The article is a nostalgic reminder of when I was first interested in electronics and got interested in quartz resonators for projects. Grinding down old war surplus FT243-style crystals and other types so as to resonate at frequencies I needed was a common practice with hobbyists back then.
I used many similar techniques to the same end of removing quartz which raised its frequency. Grinding materials included abrasives such as jeweler's rouge, cerium oxide, commercial polishes such as Brasso and Silvo and even HF solution. I'd place the quartz on a small section of plate glass and slide it through a slurry of the abrasive periodically testing its frequency until I'd reached my target.
There's an art to this that's too long to mention here except to say abrasives were used strategically, course grinding would get me near the desired frequency and I'd finish off with a fine abrasive. Then there was the job of re-aging the crystal after its recent abuse to increase its stability. Other techniques were involved such as not lowering its Q factor, etc. which I'll not cover here.
The most desired crystal cut was from the XT-plane (being the most stable) but it was generally difficult to get as it's only a small section of the quartz crystal (also each cut oscillates only over a limited range of frequencies). I used to have a book that explained these cuts in detail, their frequency ranges and electronic properties along with the basic crystallography which I lost years ago. A quick glance at the book would have shown that a great deal of science, engineering and skill is involved in the selection of quartz and its manufacture into useful resonators.
BTW, the mentioning of HF will likely horrify chem-phobic readers. We were well aware of its dangers and took special precautions never to come in contact with it.
Heavens, that should be AT cut not XT. What on earth was I thinking of?
Some of you guys should have picked that up. ;-)
Further to my unforgivable boo-boo. This article provides a pretty good introduction to the subject and there's a diagram of a quartz crystal at the beginning indicating the position of AT cuts in relation to the crystal and other cuts.
I've seen much more detailed diagrams of the crystal and cuts/cutting angles but I can't find one online. . Those precision diagrams are required by those who cut the quartz at the start of the manufacturing process.
https://xoprof.com/2023/09/unleashing-the-mystery-of-crystal...
> BTW, the mentioning of HF will likely horrify chem-phobic readers I would not say I am chem-phobic, but yes indeed that stood out. HF is nasty stuff, and yes requires some care I suspect.
The other details are fascinating, though - the intersection of mechanical, crystallographic, and RF (?) properties of a crystal that you can adjust through abrasives and selection of the cut.
Working with HF was extremely difficult before WWII, but it has become much easier after the invention of Teflon.
Teflon is not affected by HF, so if you use only vessels of Teflon to hold the HF solution and tweezers made of Teflon for handling anything that you submerge in the solution of HF, it works fine.
Besides using Teflon for anything that is in contact with the HF solution, you must do all work under a hood that evacuates the vapors of HF emitted by the solution, otherwise handling a HF solution would be very dangerous. It is good that gaseous HF is lighter than air, so after being evacuated it will continue to rise in the air, while becoming more and more diluted.
In my case the HF came in small black plastic cylindrical bottles of 50 perhaps 75 ml (could have been 100ml but that seems a bit big). Back then they were commonly available at local hardware stores.
The bottle was flat on top from which protruded its 'neck' — more a threaded spigot of perhaps 8mm in dia. and about 10 to 12mm high with a small hole in it for the HF to exit. A red plastic cap screwed onto the thread to seal the bottle.
If the bottle were ever knocked over (which it never was) very little would have spilled out (picture the threaded top and screw cap on the small bottles of Tabasco sauce and you've pretty much got it).
So the HF only came out in drops which were poured directly into a small plastic beaker of about 50ml containing about 20ml of H2O and the crystal. The crystal rested on some finely corrugated plastic and not lying flat on the bottom to ensure the etching process was reasonably uniform.
The operation was done outside and if I recall the HF was neutralised with NaOH.
The soln was very dilute and the process could easily take over an hour and was used mainly to finish off crystals that had already been through the abrasive process.
I don't know what type of plastic the bottle was made of but it was jet black.
So is this art of selecting right crystals is applicable for synthetic crystals also or these can be grown perfect each time to specifications?
I can't recall exactly but I'm pretty certain most crystal are synthetic these days, at least so for the common cuts. I'd imagine most watch quartz would be synthetic but that's only a guess.
This awesome. Could you perhaps remember the title of the book you lost?
I wish I could but it's s long while ago. More precisely the book wasn't lost but loaned to someone who never returned it and I lost track of him.
The book was a harback with a bright yellow dust jacket and I think its title was printed in red. It was about 300 pages and was no lightweight, full of resonance equations, tables etc.—the sort of book you'd find in the lab of a commercial crystal manufacturer.
Can't remember the price but it was damned expensive. At times over the years I've had need to refer to it and I still get a bit peeved when I think about it.
ok, could it be: Crystal Oscillator Design and Temperature Compensation :Marvin E. Frerking (1978). or perhaps Introduction to Quartz Crystal Unit Design : Virgil E. Bottom (1982)?
Could well be either, it was around the late 70s early 80s when the book went walkabout.
Damn nuisance my memory isn't better. However, I'd recognize it instantly if I saw it. Must check those references out. There may be enough info online for me to figure it out. Thanks.
Thank you!
"Crystal Oscillator Design and Temperature Compensation :Marvin E. Frerking (1978)."
Ah, that has to be the book, just did a search and there's a photo of the cover on Amazon that I recognize. I said it was bright yellow with red writing and I was pretty close with its red lines and bits.
240 pages and not 300, but that's not too bad after close on 50 years. And I was correct about the price too, it is expensive, Amazon has it at $110 for the hardback.
Good one!
Fantastic! Also found used on abebooks! Thank you
That sinking feeling of knowing you've ground off just a bit too much...
I've just spent the weekend tuning brass reeds from an organ. It sounds like a very similar process, except you can grind both ends of the tongue to raise and lower frequency.
And, if you're sneaky, you can add solder.
Funny you should mention that another hobby of mine was rescuing old harmoniums and making them playable again—you know, fixing and replacing reeds, renewing bug-eaten felt, sealing rat holes in the bellows, etc.
I grant you it's not in the same league as voicing a diapason though. :-)
I reckon adjusting and tweaking things goes with the territory. I'm pretty much at home tweaking crystals, fixing reeds, aligning IF stages in radio and TV equipment, there's much of a sameness in the way one tackles all of them.
BTW, I've actually repaired reeds by soldering them. Not a good fix though as the solder can fatigue with use. Throws out equal temperament a bit too but most can't hear the difference.
That's exactly what I'm doing! Nothing grandiose.
https://blog.afandian.com/tags/harmonium/
By 'repaired' you mean closing a fracture? I'm interested to hear your experiences! Electronics solder or silver solder?
This one is 60 cents sharp across the board (not uncommmon), but I wanted a social instrument. So I brought them down with solder. The bottom two octaves have worked out well. The next two... we'll see. I now have the fear that I've weakened the brass by heating it. But it still sounds nice and speaks well. Fingers crossed.
Oh that is so cool. I played the one in Liepaja, Latvia for a bit and it was absolutely amazing. It's love/hate for me (like the harpsichord), I love the instruments but I usually do not like the music that is played on them because of the grating effect. I have pretty bad tinnitus which really spoils a lot of music for me, extremely annoying.
Wow!
I'm working on something much, _much_ smaller than that!
https://blog.afandian.com/tags/harmonium/
Ha-ha, I didn't read this until I'd written the above reply, seems we have something in common.
Incidentally, I'm one of those mad people who'll put on a recording of Helmut Walcha playing Bach's Passacaglia and Fugue in C minor on a Silbermann and turn the volume up until the room shskes.
You are going to be leveling up in a whole bunch of skills.
Yes! Most important being calibrating what I should and shouldn't attempt yet.
Adding solder has also been frequently used to correct the resonance frequency of quartz crystals that have been ground too much, and I mean during industrial mass-production, not only in a home-lab setting.
How are people sticking stuff to quartz? I know less than nothing, but the pieces of quartz you find in rock don't look like they'd take a solder bond.
I'd assumed that with piezo crystals etc there was a mechanical connection rather than an electrode bonded to the crystal?
But if you can add solder presumably there is some kind of molecular connection with the metal?
Electrodes are deposited on the crystal in vacuum, e.g. by metal evaporation or sputtering, in the same way as they are deposited on the semiconductor crystals used to make transistors or integrated circuits.
The electrodes may consist of multiple layers, a base layer that adheres strongly to quartz and a top layer that is solderable, e.g. made of nickel or silver.
The pins of the package that hosts the crystal resonator are soldered on the electrodes, in places well chosen so that they will not damp much the oscillations of the crystal.
When the mass of the crystal must be increased to shift the resonance frequency, excess solder may be deposited on the electrodes.
Thanks, that's fascinating. I imagine they use a mask and large spread? Or a stepper?
Fun video of 'metallization' on a coarser scale! https://www.youtube.com/watch?v=e-QcseGvU5o
There is already a silver patch bonded to the crystal where the wire connects to. Adding weight to that obviously will not make the load curve any better but if you do it with just enough to drop you back down below where you wanted to be then it can be a saving move. You could also put a trimmer in parallel, but that might not have enough range (and can also end up overloading the crystal so the oscillator won't start).
> according to [2] a train crashed in 1972 due to a badly designed crystal oscillator spontaneously jumping to its third overtone
New fear unlocked
More interesting: amazing sleuthing to figure out that that was the root cause.
Nice to see this get more love, I had a fun time going through that site, there are lots of gems there.
For instance:
https://www.pa3fwm.nl/projects/sdr/
The longer you read, the more amazed you will be.
Dang that fascinating but way above my head. I need to take a few classes in signal processing I think to parse what's all going on here.
I remember when I worked at Motorola some decades ago, there was a little section in the factory area in Schaumburg that made quartz crystals. It's all long since been sold off (I think all the equipment went to China) but I remember all the signs for cyanide alarms, presumably due to some step in the manufacturing process.
The quartz crystals themselves were grown with carefully controlled levels of specific impurities (like scandium) in order to reduce their temperature sensitivity.
They may have had baths for galvanic deposition of silver. Those typically used cyanide solutions.
Galvanic deposition of silver has been frequently used for increasing the thickness of thin metal layers that had been deposited in vacuum, in order to adhere to the crystal.
Believe or not, when I explained to many non techies how quartz watches work and how any computers’ hardware clocks work in the same principle, they were all surprised how elegant and how efficient the mechanism is. I was also impressed when I first learned about it. True science and engineering beauty.
Have a look at the company Silicon Time, they make the most amazing little devices.
With a crystal of the right dimensions and correct input voltage could one feel the crystal vibrating?
There are piezo actuators you could certainly feel but not sure they are made of quartz, maybe PZT.
You mean could you get it down to a low enough frequency? Hmm I guess you can get down into audio frequency but maybe the amplitude will be tiny, you probably want a piezo mechanism that will give you more of a rumble.
You can get quartz to resonate down to upper audio frequencies with certain crystal cuts and manufacturing techniques but it's difficult. Typically, the lowest frequency in common use is with its use in watches with a frequency of 32,768Hz (that's about the lower limit where manufacturing and frequency combine to make a useful product).
For electronic circuits such as frequency reference markers where frequency stability is important the lowest practical frequency is 100kHz with 1Mhz preferred, and where frequency tolerances are tight 5 and 10MHz are much preferred with operation in a temperature stabilized oven to minimize frequency drift.
The most frequency-stable crystal cuts are at those frequencies, as frequencies increase (say >10MHz to 100MHz), which at the highest frequencies require the crystal to operate in overtone mode, frequency stability again tends to decrease.
As mentioned in the article there are a lot of common watch crystals that oscillate at 32.768 kHz, but they are tuning fork crystals rather than bulk mode (or modern SAW), which have much higher frequencies. It would be challenging to lower the frequency of these, but perhaps you could evaporate Au or W onto the tips and reseal them to get into the audible frequencies. Much easier would be to get two crystals which beat against each other in the audio range. Even a couple of 6MHz crystals 100ppm off would be ~6kHz (temp sensitive), and you would need some driving circuits, but you might be able to hear the beat by driving an electret microphone or a really tiny tweeter coil speaker (probably need an amp though).
It's much easier to build/buy electrical rather than mechanical LC components to hit audio frequencies ~100Hz-10kHz.
There are piezo buzzers and beepers, of course.
Yes, but those normally just convert the electrical oscillations to sounds, they are not parts of the oscillators that determine the audible frequency.
In the past, audio RC oscillators or LC oscillators were used, with the former being preferred as the latter required too bulky inductors to reach so low frequencies.
Nowadays, it is usually simpler to not use any audio oscillator, but to use some microcontroller that divides the frequency of its clock until reaching the desired audio frequency.
Those can make sounds at other frequencies but they are very loud on their resonant frequency. Loud enough to damage your hearing if you're unlucky.
And a similar question: if you took a normal magnet and rotated it very quickly, say 10k rpm, would it emit an RF signal at (10k/60) hz? I'm 95% sure the answer is yes but I've never seen this demonstrated.
Nice. I'm an EE by education. We weren't taught how such a key piece of the puzzle worked.
Could someone take the equivalent circuit of the article and put in in CircuitJs to form an oscillator? Thanks!
The account of the youtube video linked in the article has been terminated apparently. Anyone know why or if there's an alternate link?
https://web.archive.org/web/20160422054642/https://www.youtu...
Obligatory: https://youtu.be/wHenisSTUQY?si=oNOHR9QRCG3GYvf2
Just realised this video is referenced in the article! Didn't read it carefully.
Now that was nice. I liked that a lot thank you for sharing.