Forums

CD4093 oscillator frequency

Started by bitrex December 19, 2015
I'm buidling a little relaxaton oscillator for a project using a
 cd4093 quad nand gate with schmitt trigger inputs. I'm using a
 200 ohm resistor and a 100n cap in the RC network.

If I've done my calculations correctly, I was hoping to have an
 oscillator at around 30kHz. But my frequency counter is only
 showing around 8. Supply voltage is 5V. 

I understand that the frequency is dependent on the hysteresis
 thresholds, which are device dependent, but I didn't expect it to
 be that far off. Anything to check?

Can I drop a crystal or ceramic resonator across the resistor to
 "lock" the frequency?

-- 


----Android NewsGroup Reader----
http://usenet.sinaapp.com/
On Sat, 19 Dec 2015 10:52:07 -0500 (EST), bitrex
<bitrex@de.lete.earthlink.net> wrote:

> >I'm buidling a little relaxaton oscillator for a project using a > cd4093 quad nand gate with schmitt trigger inputs. I'm using a > 200 ohm resistor and a 100n cap in the RC network. > >If I've done my calculations correctly, I was hoping to have an > oscillator at around 30kHz. But my frequency counter is only > showing around 8. Supply voltage is 5V. > >I understand that the frequency is dependent on the hysteresis > thresholds, which are device dependent, but I didn't expect it to > be that far off. Anything to check?
Measure the P-P voltage and recalculate using that?
>Can I drop a crystal or ceramic resonator across the resistor to > "lock" the frequency?
A 200 resistor across a crystal is basically shorting the crystal.
On Sat, 19 Dec 2015 10:52:07 -0500 (EST)
bitrex <bitrex@de.lete.earthlink.net> wrote:

> > I'm buidling a little relaxaton oscillator for a project using a > cd4093 quad nand gate with schmitt trigger inputs. I'm using a > 200 ohm resistor and a 100n cap in the RC network. > > If I've done my calculations correctly, I was hoping to have an > oscillator at around 30kHz. But my frequency counter is only > showing around 8. Supply voltage is 5V.
8 kHz, you mean?
> I understand that the frequency is dependent on the hysteresis > thresholds, which are device dependent, but I didn't expect it to > be that far off. Anything to check?
Your calculations. You could also check if the parameters of the CD4093 are correct, by measurement. I've made a PWM-amplifier once with a quad nand and two BC547s. That was fun. But I don't remember how I arrived at the correct frequency. Probably trial and error.
> Can I drop a crystal or ceramic resonator across the resistor to > "lock" the frequency? >
A 30 kHz ceramic resonator? How large would those physically be? I mean, they are not available for that low frequency. joe
On 2015-12-19 07:52, bitrex wrote:
> > I'm buidling a little relaxaton oscillator for a project using a > cd4093 quad nand gate with schmitt trigger inputs. I'm using a > 200 ohm resistor and a 100n cap in the RC network. > > If I've done my calculations correctly, I was hoping to have an > oscillator at around 30kHz. But my frequency counter is only > showing around 8. Supply voltage is 5V. > > I understand that the frequency is dependent on the hysteresis > thresholds, which are device dependent, but I didn't expect it to > be that far off. Anything to check? >
Yep, that datasheet :-) http://www.ti.com/lit/ds/symlink/cd4093b.pdf At 5V suply CD logic is like molasses. The output impedance easily enters the kOhm range so your 200ohms in reality is much higher.
> Can I drop a crystal or ceramic resonator across the resistor to > "lock" the frequency? >
Not really with this one, I'd use a watch crystal chip for that but a 30kHz crystal is going to be pricey. Use a 74HC series device if you want to operate with 5V supply. Or make the resistor tens of kOhms (or better hundreds) and the cap smaller. But if you want any kind of accuracy a CD4093 isn't going to cut it with a 5V supply. -- Regards, Joerg http://www.analogconsultants.com/
In article <ddldtiFf42hU1@mid.individual.net>,
 Joerg <news@analogconsultants.com> wrote:
> Not really with this one, I'd use a watch crystal chip for that but a > 30kHz crystal is going to be pricey.
I've got 60 and 100 KHz tuning fork (watch, I guess) crystals I bought on a lark which were not terribly expensive (right, sorry, they cost the earth by jeorg standards, where all pennies must scream.) For the common 32.768 KHz onesies look like $0.18 and up if "around 30" is "within 10%", 60 KHz looks like $0.49 (or $1.75 in SMT) and would give you an honest 30 with a divide circuit. 30 KHz crystals I don't see on a quick look. -- Cats, coffee, chocolate...vices to live by Please don't feed the trolls. Killfile and ignore them so they will go away.
Joerg <news@analogconsultants.com> Wrote in message:
> On 2015-12-19 07:52, bitrex wrote: >> >> I'm buidling a little relaxaton oscillator for a project using a >> cd4093 quad nand gate with schmitt trigger inputs. I'm using a >> 200 ohm resistor and a 100n cap in the RC network. >> >> If I've done my calculations correctly, I was hoping to have an >> oscillator at around 30kHz. But my frequency counter is only >> showing around 8. Supply voltage is 5V. >> >> I understand that the frequency is dependent on the hysteresis >> thresholds, which are device dependent, but I didn't expect it to >> be that far off. Anything to check? >> > > Yep, that datasheet :-) > > http://www.ti.com/lit/ds/symlink/cd4093b.pdf > > At 5V suply CD logic is like molasses. The output impedance easily > enters the kOhm range so your 200ohms in reality is much higher. > >
Ah. That's it. I'm so used to using the HC stuff I'd forgotten that. I'll try recalculating for a different RC network... -- ----Android NewsGroup Reader---- http://usenet.sinaapp.com/
Ecnerwal <MyNameForward@ReplaceWithMyVices.Com.invalid> Wrote in
 message:
> In article <ddldtiFf42hU1@mid.individual.net>, > Joerg <news@analogconsultants.com> wrote: >> Not really with this one, I'd use a watch crystal chip for that but a >> 30kHz crystal is going to be pricey. > > I've got 60 and 100 KHz tuning fork (watch, I guess) crystals I bought > on a lark which were not terribly expensive (right, sorry, they cost the > earth by jeorg standards, where all pennies must scream.) > > For the common 32.768 KHz onesies look like $0.18 and up if "around 30" > is "within 10%", 60 KHz looks like $0.49 (or $1.75 in SMT) and would > give you an honest 30 with a divide circuit. > > 30 KHz crystals I don't see on a quick look. > > -- > Cats, coffee, chocolate...vices to live by > Please don't feed the trolls. Killfile and ignore them so they will go away. >
The precise freq. isn't critical, just would like it to be fairly stable and above the audio range. I think a watch crystal would work fine -- ----Android NewsGroup Reader---- http://usenet.sinaapp.com/
On Sat, 19 Dec 2015 10:52:07 -0500 (EST), bitrex
<bitrex@de.lete.earthlink.net> wrote:

> >I'm buidling a little relaxaton oscillator for a project using a > cd4093 quad nand gate with schmitt trigger inputs. I'm using a > 200 ohm resistor and a 100n cap in the RC network.
200 ohms is way too low. Scale up R and scale down C, by 20:1 at least.
bitrex wrote:
> > I'm buidling a little relaxaton oscillator for a project using a > cd4093 quad nand gate with schmitt trigger inputs. I'm using a > 200 ohm resistor and a 100n cap in the RC network. > > If I've done my calculations correctly, I was hoping to have an > oscillator at around 30kHz. But my frequency counter is only > showing around 8. Supply voltage is 5V. > > I understand that the frequency is dependent on the hysteresis > thresholds, which are device dependent, but I didn't expect it to > be that far off. Anything to check? > > Can I drop a crystal or ceramic resonator across the resistor to > "lock" the frequency? >
For practical purposes, the threshold(s) are one-half of the CMOS supply voltage (at room temp). Reasonably dependable.

"John Larkin" <jjlarkin@highlandtechnology.com> wrote in message 
news:556b7blona7gtget1lhm3gksgdi3e3b8m9@4ax.com...
> On Sat, 19 Dec 2015 10:52:07 -0500 (EST), bitrex > <bitrex@de.lete.earthlink.net> wrote: > >> >>I'm buidling a little relaxaton oscillator for a project using a >> cd4093 quad nand gate with schmitt trigger inputs. I'm using a >> 200 ohm resistor and a 100n cap in the RC network. > > 200 ohms is way too low. Scale up R and scale down C, by 20:1 at > least.
Most of the CMOS oscillators I've seen, the resistor is more like 2M. The OP could check out the catalogues from various crystal/resonator manufacturers - many include example circuits.