Forums

Me Again, Math Again

Started by Unknown June 25, 2018
http://www.google.com/url?sa=t&rct=j&q=&esrc=s&source=web&cd=1&cad=rja&uact=8&ved=0ahUKEwiSvsnY9-_bAhVDja0KHXHnALoQFggpMAA&url=http%3A%2F%2Fphysics.oregonstate.edu%2F~giebultt%2FCOURSES%2Fph412%2FReading%2Foscillators-1p.pdf&usg=AOvVaw2ID2iftO5An0jNHiC1lYRm

On page 5 of that they have a phase shift oscillator. I would like to use one in  my, well, whatever.  

They got the formula there and the way I read it, the frequency is :

1/[2pi times R times C times the square root of 6]. 

So I want 2 KHz. 

2,000=1/[6.28*R*C*2.449...]

If that is correct, then the values you get for R and C are for each of the three stages ? As in 60 degree delay per R/C ? 

It seems another way to attack this would be to use ascending values in each subsequent stage. For example only - you need 1K and 0.01. Next stage is 10K and 0.001. Then comes 100K and 0.0001. That doesn't look too good for a bipolar because of current etc., and it might run out of signal and not oscillate. An FET should be alright though. 

This has to do with a design I posted a while back. I want to do away with the two stage oscillator. This, like before has the same function, it goes to a buffer that puts out a short pulse to switch the states of the bistable multivibrator. I believe this would be more stable, as the Vcc is  not regulated. What I've seen, correct me if I am wrong, an astable multivibrator frequency is quite dependent on the Vcc. It can be designed out of course but I don't have a hundred years. 

I have it now so that changes in battery voltage equally affects the output to the DUT as well as the scale of the detector end. It nulls out, and I AM going to use 2 more diodes and remove the Vf of the detectors. 

If the battery gets too low, or I don't get the feedback level right it will clip and that will affect the frequency. I am not putting in a regulator. This damn thing is already going to be 5 times more expensive than I ever wanted it, but it will be cool. 

So, where am I right and wrong here. If necessary I can bring u p the old LT file on it, and a new one with the changes. 

Your thoughts. Call me an asshole, call me stupid, call me anything you want, but state your case.
On 06/25/18 19:47, jurb6006@gmail.com wrote:
> http://www.google.com/url?sa=t&rct=j&q=&esrc=s&source=web&cd=1&cad=rja&uact=8&ved=0ahUKEwiSvsnY9-_bAhVDja0KHXHnALoQFggpMAA&url=http%3A%2F%2Fphysics.oregonstate.edu%2F~giebultt%2FCOURSES%2Fph412%2FReading%2Foscillators-1p.pdf&usg=AOvVaw2ID2iftO5An0jNHiC1lYRm > > On page 5 of that they have a phase shift oscillator. I would like > to use one in my, well, whatever. > > They got the formula there and the way I read it, the frequency is : > > 1/[2pi times R times C times the square root of 6]. > > So I want 2 KHz. > > 2,000=1/[6.28*R*C*2.449...] > > If that is correct, then the values you get for R and C are for each > of the three stages ? As in 60 degree delay per R/C ? > > It seems another way to attack this would be to use ascending values > in each subsequent stage. For example only - you need 1K and 0.01. > Next stage is 10K and 0.001. Then comes 100K and 0.0001. That doesn't > look too good for a bipolar because of current etc., and it might run > out of signal and not oscillate. An FET should be alright though. > > This has to do with a design I posted a while back. I want to do away > with the two stage oscillator. This, like before has the same > function, it goes to a buffer that puts out a short pulse to switch > the states of the bistable multivibrator. I believe this would be > more stable, as the Vcc is not regulated. What I've seen, correct me > if I am wrong, an astable multivibrator frequency is quite dependent > on the Vcc. It can be designed out of course but I don't have a > hundred years. > > I have it now so that changes in battery voltage equally affects the > output to the DUT as well as the scale of the detector end. It nulls > out, and I AM going to use 2 more diodes and remove the Vf of the > detectors. > > If the battery gets too low, or I don't get the feedback level right > it will clip and that will affect the frequency. I am not putting in > a regulator. This damn thing is already going to be 5 times more > expensive than I ever wanted it, but it will be cool. > > So, where am I right and wrong here. If necessary I can bring u p the > old LT file on it, and a new one with the changes. > > Your thoughts. Call me an asshole, call me stupid, call me anything > you want, but state your case. >
IIRC the classical vanilla no-frills phase shift oscillator runs at exactly 29/RC, but that's a memory from long ago. Cheers Phil Hobbs -- Dr Philip C D Hobbs Principal Consultant ElectroOptical Innovations LLC / Hobbs ElectroOptics Optics, Electro-optics, Photonics, Analog Electronics Briarcliff Manor NY 10510 http://electrooptical.net http://hobbs-eo.com
It's 29/RC ? 

The referenced paper says something about 29, feedback gain of the amp beta=1/29.

So that means it is the point where I want my frequency ? 

So the amp must be able to handle the load of the feedback network for a voltage gain of 29 ? 

???
On Mon, 25 Jun 2018 16:47:20 -0700 (PDT), jurb6006@gmail.com wrote:

>http://www.google.com/url?sa=t&rct=j&q=&esrc=s&source=web&cd=1&cad=rja&uact=8&ved=0ahUKEwiSvsnY9-_bAhVDja0KHXHnALoQFggpMAA&url=http%3A%2F%2Fphysics.oregonstate.edu%2F~giebultt%2FCOURSES%2Fph412%2FReading%2Foscillators-1p.pdf&usg=AOvVaw2ID2iftO5An0jNHiC1lYRm > >On page 5 of that they have a phase shift oscillator. I would like to use one in my, well, whatever. > >They got the formula there and the way I read it, the frequency is : > >1/[2pi times R times C times the square root of 6]. > >So I want 2 KHz. > >2,000=1/[6.28*R*C*2.449...] > >If that is correct, then the values you get for R and C are for each of the three stages ? As in 60 degree delay per R/C ? > >It seems another way to attack this would be to use ascending values in each subsequent stage. For example only - you need 1K and 0.01. Next stage is 10K and 0.001. Then comes 100K and 0.0001. That doesn't look too good for a bipolar because of current etc., and it might run out of signal and not oscillate. An FET should be alright though. > >This has to do with a design I posted a while back. I want to do away with the two stage oscillator. This, like before has the same function, it goes to a buffer that puts out a short pulse to switch the states of the bistable multivibrator. I believe this would be more stable, as the Vcc is not regulated. What I've seen, correct me if I am wrong, an astable multivibrator frequency is quite dependent on the Vcc. It can be designed out of course but I don't have a hundred years. > >I have it now so that changes in battery voltage equally affects the output to the DUT as well as the scale of the detector end. It nulls out, and I AM going to use 2 more diodes and remove the Vf of the detectors. > >If the battery gets too low, or I don't get the feedback level right it will clip and that will affect the frequency. I am not putting in a regulator. This damn thing is already going to be 5 times more expensive than I ever wanted it, but it will be cool. > >So, where am I right and wrong here. If necessary I can bring u p the old LT file on it, and a new one with the changes. > >Your thoughts. Call me an asshole, call me stupid, call me anything you want, but state your case.
Be warned, phase shift oscillators usually make distorted sine waves. Or no sine waves. -- John Larkin Highland Technology, Inc picosecond timing precision measurement jlarkin att highlandtechnology dott com http://www.highlandtechnology.com
John Larkin wrote:
>Be warned, phase shift oscillators usually make distorted sine waves. >Or no sine waves.
A question, looking at that circuit: is there a difference between 'high pass' phase shifting networks and 'low pass' phase shifting networks in that respect? Seems to me a high pass, as described, is asking for harmonics to be more amplified, equals distortion. while if you reverse R and C like this: -===---===---===--- | | | === === === | | | /// /// /// then you get a low-pass, and no harmonic can make its way? Same phase shift -. What do you think?
On 06/25/18 20:46, jurb6006@gmail.com wrote:
> It's 29/RC ? > > The referenced paper says something about 29, feedback gain of the > amp beta=1/29. > > So that means it is the point where I want my frequency ? > > So the amp must be able to handle the load of the feedback network > for a voltage gain of 29 ? > > ??? >
That's what I was thinking of, minimum gain is (weirdly) exactly 29. Thanks for the reminder! Cheers Phil Hobbs -- Dr Philip C D Hobbs Principal Consultant ElectroOptical Innovations LLC / Hobbs ElectroOptics Optics, Electro-optics, Photonics, Analog Electronics Briarcliff Manor NY 10510 http://electrooptical.net http://hobbs-eo.com
On Monday, June 25, 2018 at 7:47:25 PM UTC-4, jurb...@gmail.com wrote:

> On page 5 of that they have a phase shift oscillator. I would like to use one in my, well, whatever.
[...]
> Your thoughts.
The problem with the standard phase shift oscillator (besides distortion) is if you want to change the frequency, you need to change three components. Here is another approach that only needs two components to change the frequency. You can easily get a stereo potentiometer from an old high-fi stereo in a recycle bin. The author claims good sine wave output. I haven't confirmed this in LTspice, but I suspect it may be better than a plain phase shift oscillator. It can also drive a low impedance load. The author is Gary ZL1AN, and the url is http://www.qsl.net/zl1an/Downloads/Twostage.pdf Note there is no amplitude control such as needed in a Wein bridge. I think the gain stage saturates, and the low pass filter to the op amp follower filters the harmonics to produce the sine wave output. The opamp should be a RRIO for symmetrical limiting.
On Tuesday, June 26, 2018 at 6:08:57 AM UTC-4, Steve Wilson wrote:

> The problem with the standard phase shift oscillator (besides distortion) is if you want to change the frequency, you need to change three components.
> Here is another approach that only needs two components to change the frequency. You can easily get a stereo potentiometer from an old high-fi stereo in a recycle bin.
> The author claims good sine wave output. I haven't confirmed this in LTspice, but I suspect it may be better than a plain phase shift oscillator. It can also drive a low impedance load. The author is Gary ZL1AN, and the url is
> http://www.qsl.net/zl1an/Downloads/Twostage.pdf
> Note there is no amplitude control such as needed in a Wein bridge. I think the gain stage saturates, and the low pass filter to the op amp follower filters the harmonics to produce the sine wave output. The opamp should be a RRIO for symmetrical limiting.
Here is the ASC file. Worked first time. I set it up to run at 2kHz. As expected, amplitude control is via limiting in the gain stage. There is a bit of distortion in the waveform. All odd harmonics. The third harmonic is 30dB down, the rest are much lower. Not bad for such a simple oscillator. Version 4 SHEET 1 1204 680 WIRE 720 -160 224 -160 WIRE 832 -160 720 -160 WIRE 864 -160 832 -160 WIRE 912 -128 784 -128 WIRE 928 -128 912 -128 WIRE 224 -112 224 -160 WIRE 784 -96 784 -128 WIRE 720 -80 720 -160 WIRE 752 -80 720 -80 WIRE 384 -64 352 -64 WIRE 864 -64 864 -160 WIRE 864 -64 816 -64 WIRE 480 -48 448 -48 WIRE 512 -48 480 -48 WIRE 544 -48 512 -48 WIRE 672 -48 624 -48 WIRE 720 -48 672 -48 WIRE 752 -48 720 -48 WIRE 224 -32 224 -48 WIRE 304 -32 224 -32 WIRE 384 -32 304 -32 WIRE 672 -32 672 -48 WIRE 224 -16 224 -32 WIRE 480 -16 480 -48 WIRE 784 16 784 -32 WIRE 672 48 672 32 WIRE 224 80 224 64 WIRE 352 80 352 -64 WIRE 416 80 352 80 WIRE 480 80 480 64 WIRE 480 80 416 80 WIRE 784 80 784 16 WIRE 928 80 928 -128 WIRE 480 96 480 80 WIRE 784 176 784 160 WIRE 928 176 928 160 WIRE 480 192 480 176 FLAG 672 48 0 FLAG 224 80 0 FLAG 480 192 0 FLAG 928 176 0 FLAG 784 176 0 FLAG 912 -128 +12 FLAG 416 -80 +12 FLAG 784 16 -12 FLAG 416 -16 -12 FLAG 304 -32 U1P FLAG 512 -48 U1O FLAG 416 80 U1N FLAG 720 -48 U2P FLAG 832 -160 U2O SYMBOL res 640 -64 R90 WINDOW 0 0 56 VBottom 2 WINDOW 3 32 56 VTop 2 SYMATTR InstName R1 SYMATTR Value 75k SYMBOL res 240 80 R180 WINDOW 0 36 76 Left 2 WINDOW 3 36 40 Left 2 SYMATTR InstName R2 SYMATTR Value 75k SYMBOL cap 656 -32 R0 SYMATTR InstName C1 SYMATTR Value 1n SYMBOL cap 208 -112 R0 SYMATTR InstName C2 SYMATTR Value 1n SYMBOL res 464 -32 R0 SYMATTR InstName R3 SYMATTR Value 12k SYMBOL res 464 80 R0 SYMATTR InstName R4 SYMATTR Value 10k SYMBOL voltage 784 64 R0 WINDOW 123 0 0 Left 2 WINDOW 39 0 0 Left 2 SYMATTR InstName V1 SYMATTR Value -12v SYMBOL voltage 928 64 R0 WINDOW 123 0 0 Left 2 WINDOW 39 0 0 Left 2 SYMATTR InstName V2 SYMATTR Value +12v SYMBOL opamps\\lt1001a 416 -112 R0 WINDOW 3 -32 -18 Left 2 SYMATTR InstName U1 SYMBOL opamps\\lt1001a 784 -128 R0 SYMATTR InstName U2 TEXT 304 -216 Left 2 !.tran 0 15m 0 1u startup TEXT 384 -248 Left 2 ;'Two Stage RC Oscillator TEXT 840 -216 Left 2 !.options plotwinsize=0 TEXT 840 -240 Left 2 !.options numdgt=15
>"Worked first time. I set it up to run at 2kHz."
Is that technically a phase shift oscillator ? I mean, it looks like a fine oscillator but you've got one RC leading ad one RC lagging. Of course this brings it into phase and you have non-inverting amps. The two RCs look like a tank circuit. I see the gain of U2 is unity, and the gain of U1 is not set very high. That's probably why it has low distortion. That is not really a big issue here because it is going to be turned into a little spike to switch a flip flop. However, if it does clip then the frequency may change with Vcc/dd. I would rather not try to compensate for that so it is better if the frequency is stable. I don't want to use coils and I don't want to even go looking for a 2 KHz crystal, so I figured this was the best option. If I am not mistaken, that should run without U2, no ? Of course the math would not be as simple, then it would have a 37.5K source feeding a 150K load. Probably need the gain set higher. And it looks like U2 can just be a common collector bipolar. More later, this is just my fist impression of it. I don't want to say more until I give it that blank stare for a while and think about it. Maybe U1 could be a bipolar with the emitter fed out of a buffer ? It would probably be less drain, and this is battery operated. And I only have 6 volts. I don't know how much head voltage those OP AMPs need. Maybe I should draw up a version with bipolars and see if I can figure out how to get a simulation to run. But if you know of any problem using bipolars do say so. I will keep an eye on this thread.
>"That's what I was thinking of, minimum gain is (weirdly) exactly 29.
Thanks for the reminder! " I have no idea why that is. It must have something to do with the specific amount of loss in the RC network. By trig, 60 degrees begets .5000 and .86603. So you have (I think) 86.6 % of 86.6 % of 86.6 %. Or is it 50 % of 50 % of 50 % ? Actually that sounds more right, you get .86603 the 30 degree way and .500 the 60 degree way. (voltage) Either way it does not add up (down ?) to 1/29th. Half of a half of a half is an eighth. The other way it would be about 64 % but that wouldn't work anyway. So another question, what's new ?