XO controlled 480Hz Oscillator
in [Design]
|
Prev: The UT804 bench multimeter
Next: Small genset
From: John Fields on 10 Feb 2010 14:41 On Wed, 10 Feb 2010 07:52:52 -0800 (PST), Chris <christopher.maness(a)gmail.com> wrote: >This is the actual crystal I ordered from digikey: > >http://search.digikey.com/scripts/DkSearch/dksus.dll?Detail&name=X975-ND > >If I get the thing dialed in, the 30ppm is tolerable for my >application. I will try 25pF fixed capacitor for the PI network part >of the circuit. --- In a Pierce oscillator, since the load capacitance is presented in series with the crystal, if the crystal has a specified load capacitance of 17pF the load capacitors should be about 35pF each. --- >If I am paranoid should I still figure a value to >limit the current so that @ 12V the max power is ~2mw? --- No, because in your case the crystal is designed to dissipate a _maximum_ of one milliwatt. Also, you don't use the caps to change the crystal's power dissipation, you change it by varying the value of the series resistor. --- >Am I running >into a greater chance that I will not be driving it enough to >oscillate? Or is the minimum power so far below the max power limit >that it does not matter? --- Somewhere between the maximum rated power it'll dissipate when oscillating and the minimum power required to get it to start is where you should be working. Follow Spehro's advice. --- >I will use a 2Meg resistor for the foldback. --- It's not foldback, it's 'feedback' and it's used to bias the inverter into its linear region so that noise will cause the crystal to start ringing which will generate a voltage which will be fed into the amplifier, amplified, and then sent back into the crystal to generate an even larger voltage and... --- >Maybe a 25pF trimmer in parallel will let me dial the freq >right in. --- If it were me I'd use 33pF to ground on the input side of the inverter and about a 70pF trimmer to ground on the other side so that it'd be around the middle of its range at your operating frequency. --- >The specified crystal frequency is what I need. >I am not trying to shift it off the package frequency. >I just want to put some variability in so that I can compensate >for stray capacitance after I have mounted it in the tape deck. --- Then that's what loading it with the recommended load capacitance will get you, since crystals are ground slightly high and the loading caps are used to tune out strays an pull the crystal down to the desired frequency. Here are some other good reads: http://www.ecsxtal.com/store/pdf/quar_des.pdf http://www.ecsxtal.com/store/pdf/oscir_des.pdf http://www.ecsxtal.com/store/pdf/clock_app.pdf JF
From: Chris on 10 Feb 2010 15:09 On Feb 10, 8:49 am, Spehro Pefhany <speffS...(a)interlogDOTyou.knowwhat> wrote: > On Wed, 10 Feb 2010 07:52:52 -0800 (PST), Chris > > > > <christopher.man...(a)gmail.com> wrote: > >On Feb 10, 6:11 am, Spehro Pefhany <speffS...(a)interlogDOTyou.knowwhat> > >wrote: > >> On Wed, 10 Feb 2010 09:44:26 +0200, the renowned Paul Keinanen > > >> <keina...(a)sci.fi> wrote: > >> >On Tue, 09 Feb 2010 18:01:44 -0500, Spehro Pefhany > >> ><speffS...(a)interlogDOTyou.knowwhat> wrote: > > >> >>But the counter > >> >>timebase might not be that *accurate*. ~4MHz crystals are typically > >> >>accurate to +/-30-100ppm. 100ppm is 50 msec error in about 8 minutes, > >> >>just for comparison. > > >> >100 pm might apply to some cheap microprocessor crystals, but for > >> >decades (before frequency synthesizers) , channel crystals were made > >> >for radio telephones to within a few ppm. A 100 ppm crystal would have > >> >caused the transmission on the wrong channel, a slightly smaller error > >> >would cause bad audio distortion, since the signal would be partially > >> >outside the receiver de modulator bandwidth. > > >> >When ordering a frequency for a specific frequency, specify also the > >> >load capacitance or specify the frequency slightly (a few dozen ppm) > >> >above the desired frequency and pull it down to the desired frequency > >> >with a parallel capacitor. Pulling upwards is harder with a series > >> >inductance. > > >> Or if he really needs ppm-level accuracy/stability or better, replace > >> the crystal with a TCXO or VCTCXO (or even OCXO if warm-up time is not > >> important). > > >> Eg.http://www.ecsxtal.com/store/pdf/vc_txo_35sm.pdf > > >> or http://www.abracon.com/Precisiontiming/AOCJY.pdf > > >> Specs are important, because the 4060 apparently won't work with > >> off-the-shelf frequencies, wheras the PIC or some equivalant > >> combination of HCMOS could. > > >> >Overtone crystals (typically above 20 MHz) have a more limited pulling > >> >range. > > >> Related to the inverse of the square of the overtone, so 1/9 range for > >> a 3rd overtone. > > >> Best regards, > >> Spehro Pefhany > >> -- > >> "it's the network..." "The Journey is the reward" > >> sp...(a)interlog.com Info for manufacturers:http://www.trexon.com > >> Embedded software/hardware/analog Info for designers: http://www..speff.com > > >This is the actual crystal I ordered from digikey: > > >http://search.digikey.com/scripts/DkSearch/dksus.dll?Detail&name=X975-ND > > >If I get the thing dialed in, the 30ppm is tolerable for my > >application. I will try 25pF fixed capacitor for the PI network part > >of the circuit. If I am paranoid should I still figure a value to > >limit the current so that @ 12V the max power is ~2mw? > > Yes, particularly if you care much about drift. This particular xtal > is rated at 1mW maximum. > > >Am I running > >into a greater chance that I will not be driving it enough to > >oscillate? > > Yes. You could empirically increase it until it doesn't start then go > down by a fairly large safety margin (maybe 3:1). Or work it out from > the transconductance of the oscillator amplifier (I think NXP > specifies typical values on their HC datasheet). > > http://www.freescale.com/files/microcontrollers/doc/app_note/AN1783.pdf > > > Or is the minimum power so far below the max power limit > >that it does not matter? I will use a 2Meg resistor for the > >foldback. > > DC feedback = bias. > > > Maybe a 25pF trimmer in parallel will let me dial the freq > >right in. The specified crystal frequency is what I need. I am not > >trying to shift it off the package frequency. I just want to put some > >variability in so that I can compensate for stray capacitance after I > >have mounted it in the tape deck. > > >Thanks Guys, > >Chris Maness Here is what I have so far. 2 Meg to bias. Two 20pF for the Pi network. These have an equivalent of 10pF. I would imagine that any stray capacitance would add to this in parallel. Hopefully the stray capacitance is less than 7pF. That would satisfy the 17pF that the crystal wants to see to be on freq. I will add a 20pF trimmer parallel to the crystal for fine tuning. I would imagine I would need to take any frequency measurements off of the next divider stage so as not to introduce more stray capacitance and throw the freq. measurement. As far as the current limiting resistor, I will use a high value like a 20k pot with a 10k resistor in series. I will decrease the value until the circuit starts oscillating. In the final circuit I will use half that value. Here is a photo of the circuit: http://chrismaness.com/backend/spice_simulation/Photo_3.png Thanks, Chris Maness
From: ehsjr on 16 Feb 2010 22:31 Chris wrote: > I finished the this part of the project. The design with the 3.9Mhz > crystal with a CD4060 works perfectly. My tape deck is running > again. However, there is a small 480Hz hum in the record/play audio > now. This is very slight hum, not noticeable when there is dialog. I > did not notice this when the unit was running from a tuning fork time > reference. However, the tuning fork itself was audible outside the > chassis, and maybe masking my ability to hear the hum through the > headphones. > > There is a voltage divider resistor (1k) in series with the circuit. > This was in place because the tuning fork ran off of 12V and the > supply was 24V. My circuit is using a voltage regulator to drop the > voltage down to 12V. There is now about a 8V drop across the > resistor. If I remove the resistor, the hum gets much louder. If I > shunt the power with a 1000uF cap after the resistor the hum gets > louder. If I shunt before the resistor the hum drops into the Nyquist > noise when I monitoring through the preamp. It is still there, but > buried. Is there a better way to get rid of this hum/ripple? The cap > is rather big and not as effective as I would like. > > Thanks, > Chris Maness Got a schematic of the power supply and the regulator? Ed
From: Chris on 16 Feb 2010 23:20 On Feb 16, 7:31 pm, ehsjr <eh...(a)nospamverizon.net> wrote: > Chris wrote: > > I finished the this part of the project. The design with the 3.9Mhz > > crystal with a CD4060 works perfectly. My tape deck is running > > again. However, there is a small 480Hz hum in the record/play audio > > now. This is very slight hum, not noticeable when there is dialog. I > > did not notice this when the unit was running from a tuning fork time > > reference. However, the tuning fork itself was audible outside the > > chassis, and maybe masking my ability to hear the hum through the > > headphones. > > > There is a voltage divider resistor (1k) in series with the circuit. > > This was in place because the tuning fork ran off of 12V and the > > supply was 24V. My circuit is using a voltage regulator to drop the > > voltage down to 12V. There is now about a 8V drop across the > > resistor. If I remove the resistor, the hum gets much louder. If I > > shunt the power with a 1000uF cap after the resistor the hum gets > > louder. If I shunt before the resistor the hum drops into the Nyquist > > noise when I monitoring through the preamp. It is still there, but > > buried. Is there a better way to get rid of this hum/ripple? The cap > > is rather big and not as effective as I would like. > > > Thanks, > > ChrisManess > > Got a schematic of the power supply and the regulator? > > Ed Yes I will post them shortly. The unit runs off of two separate battery packs. 1 24V pack with a negative ground, and one 12V pack with a positive ground. The schematics I have are separate because I was modeling them in LTSpice. Chris
From: Chris on 16 Feb 2010 11:46
I finished the this part of the project. The design with the 3.9Mhz crystal with a CD4060 works perfectly. My tape deck is running again. However, there is a small 480Hz hum in the record/play audio now. This is very slight hum, not noticeable when there is dialog. I did not notice this when the unit was running from a tuning fork time reference. However, the tuning fork itself was audible outside the chassis, and maybe masking my ability to hear the hum through the headphones. There is a voltage divider resistor (1k) in series with the circuit. This was in place because the tuning fork ran off of 12V and the supply was 24V. My circuit is using a voltage regulator to drop the voltage down to 12V. There is now about a 8V drop across the resistor. If I remove the resistor, the hum gets much louder. If I shunt the power with a 1000uF cap after the resistor the hum gets louder. If I shunt before the resistor the hum drops into the Nyquist noise when I monitoring through the preamp. It is still there, but buried. Is there a better way to get rid of this hum/ripple? The cap is rather big and not as effective as I would like. Thanks, Chris Maness |