| Wien Bridge, I HATE YOU - Click HERE for Original Thread |
| mod_evil |
Dear Friends,
For many times i'm searching for a good Function Generator to use in place of the good design with ICL8038, MAX038 and XR2206 because them are IMPOSSIBLE to found in Brazil.
I make the Wien Bridge with a OPA228, OPA227, OPA134, NE5534. I used styroflex capacitors, and a Wave Switch.
The capacitors are 1uf, 0.1uf, 0.01uf, 0.001uf, I tested for the potenciometer a 47k wich I have good results. The waveform choose when be good and when don't oscillate.
I put the 10kohm potentiometer and the waveform was good in these 4 scales. But... It choose when oscillate or when don't oscillate. ( The waveform wasn't so good )
Does anybody have a opinion about this? I need a Function Generator!
A digital switch is a good choice to pulll off the waveswitch ( I don't know the name of this switch in English, but it is a rotative switch )?
Best Regards,
Felipe Navarro |
|
|
| kevinkr |
What are you using to control positive feedback in the oscillator? Pots don't work! Traditionally light bulbs were used because of their positive temperature coefficient. Finding the right one might be an adventure, but seems a bit easier than using a precision rectifier and jfet as I once did. (Bulb worked better.) You might want to use one bulb for the higher frequency ranges and add another for the LF ranges or select one with a longer time constant in order to reduce LF distortion.
Look at the old HP201 oscillator, not to mention the old Thorens TD-125 (MKII) electronics as examples of bulb regulated Weinbridge oscillators.
A well implemented Weinbridge will blow the doors off of any function generator in terms of thd.
Note that the chips you listed weren't that great either. I built a generator around the XR-2206 and despite great attention to detail it never worked well. (I was an eng tech at the time I designed and built it.) Incidentally I don't think any of those types are commonly available now, IIRC most of the listed types were in production over 20yrs ago and have not stood the test of time that well.. :devilr: |
|
|
| wakibaki |
What functions, amplitude and repetition rate do you require?
You can output functions from a soundcard.
You can find some free software for tones, e.g TTG, or you can make a .wav file with any waveshape you want with a sample editor or a bit of programming. You might even be able to create the values and export it from Excel.
w
Obviously the rise time is limited and you will only get a really good square wave at fairly low frequencies, but this is true to an extent with any FG. |
|
|
| Iain McNeill |
The Wein bridge is a thing of beauty but hard to get right.
I assume your computer has a sound-card? Why not just use a software function generator. Google Visual Analyzer, Praxis or search your favorite freeware site.
waki- you type faster than me! |
|
|
| forr |
There was a project of a state variable oscillator with stabilisation by diodes which was described by Ian Hickman in Electronics World. Disto lmuch less than 0.1%, simple and interesting.
About 25 years ago, Elektor made a full project with double sided PCB using an XR2206. It got rid of the spikes on the sines at high frequencies quite well. Useful for applications where stabilisation of the output voltage is more important than distorsion (about 0.5 %), I still use it today. |
|
|
| wakibaki |
| quote: | Originally posted by Ecclesiastes
"The race is not always to the swift, nor the battle to the strong."
|
w
Nice to meet you Iain |
|
|
| mod_evil |
| quote: | Originally posted by forr
There was a project of a state variable oscillator with stabilisation by diodes which was described by Ian Hickman in Electronics World. Disto lmuch less than 0.1%, simple and interesting.
About 25 years ago, Elektor made a full project with double sided PCB using an XR2206. It got rid of the spikes on the sines at high frequencies quite well. Useful for applications where stabilisation of the output voltage is more important than distorsion (about 0.5 %), I still use it today. |
I'm using the stabilisation by two 1n4148 diodes. And a negative feedback of gain is 3.
I need from 10hz to 100khz to test amplifiers. Sine and Square waves.
I tested mine soundcard a two weeks ago. It's horrible the square and sine wave at 10khz.
Best Regards,
Felipe Navarro |
|
|
| mod_evil |
This Wein Bridge oscillator is getting me crazy! It works when it wants to.
Best Regards,
Felipe |
|
|
| Conrad Hoffman |
| Go to the Linear Technology site and track down AN5. On page 8 is a simple wien bridge oscillator. The trick is using small light bulbs for stabilization. It can be done with other circuits, but they are way more trouble. For more on that, see the piece Jim Williams of LT wrote, "Max Wien, Mr. Hewlett, and a Rainy Sunday Afternoon". This is in his book, "Analog Circuit Design". Possibly a local library will have it or can get it. Another interesting route is to use the Analog Devices programmable oscillator chips. You'll need to do some software, but they go from near DC to near daylight, and the waveform quality is better than the chips of old, with their triangle-to-sine converters that always left little points on the waveform. Even my favorite Wavetek 185 sweep gen suffers slightly from that. |
|
|
| sawreyrw |
mod_evil,
You need to reduce the value of R2 slightly, because the circuit gain in less than 3 the way it is. 3 is the ideal value, but it may nned to be higher because of component tolerances.
Rick |
|
|
| mod_evil |
A lamp bulb, can help me in every situations...
Now I tested my schematic with a lamp bulb with 40w. This is the smaller which I can get at this moment.
The Wein Bridge now seens okay.
How can I get a square wave from a senoidal wave?
Best Regards! |
|
|
| kevinkr |
| quote: | Originally posted by mod_evil
A lamp bulb, can help me in every situations...
Now I tested my schematic with a lamp bulb with 40w. This is the smaller which I can get at this moment.
The Wein Bridge now seens okay.
How can I get a square wave from a senoidal wave?
Best Regards! |
A small night light bulb might be a bit better, but if you are getting a decent sinewave at LF then perhaps that 40W bulb is the ticket.
Use a comparator to generate a nice square wave.. Make the slicing level adjustable and you can vary the duty cycle.
The aforementioned article by Jim Williams is highly recommended. I do ATE hardware design and he is the absolute authority on extreme precision dc source A to D and D to A converters (how about 1ppm accuracy) amongst a lot other things. While I am gushing I think Bob Pease over at National is worth more than a casual read as well. |
|
|
| wakibaki |
| quote: | Originally posted by mod_evil
How can I get a square wave from a senoidal wave?
|
You can amplify it to >>0.6V and pass it through a diode gate i.e. diodes in antiparallel or 'back-to-back'.
w |
|
|
| kevinkr |
| quote: | Originally posted by wakibaki
You can amplify it to >>0.6V and pass it through a diode gate i.e. diodes in antiparallel or 'back-to-back'.
w |
Diode clipping makes fairly dreadful looking square waves, ok if you need something quick and dirty, but not recommended if you want to measure slew rate or rise times. |
|
|
| Christer |
| Use a comparator as Kevin suggested. Also ground the input to it when switching the output to sine. Otherwise the square wave may interfer and distort the sine. |
|
|
| wakibaki |
Yeah, but there are no issues with duty cycle and slicing, which is why I mentioned it, and you can always tidy things up with a bit of Schmidt trigger logic...
w
|
|
|
| mod_evil |
| quote: | Originally posted by kevinkr
Diode clipping makes fairly dreadful looking square waves, ok if you need something quick and dirty, but not recommended if you want to measure slew rate or rise times. |
Friends,
I tested a IRF640, with a 120ohm resistor in the Gate to make the square wave from the sine.
In the low frequencies < 1khz the square wave converter is distorting the square wave.
I will put some photos about this.
But... Now it isn't working. It works when it wants. I haven't words to write about this.
The senoidal output seems okay.
Best Regards,
Felipe Navarro |
|
|
| jackinnj |
Linear Tech has some great op-amp oscillator circuits. See the various application notes that Jim Williams has written. They show you how to build an AGC with off-the-shelf parts.
Probably the easiest (relatively speaking) high quality signal generator is the Bob Cordell which appeared in Audio in the early 1980's -- this was the design used for the Tektronix SG505 which is a truely great sig gen. You can follow up the Cordell oscillator with the new buffer from National Semiconductor for a device that you can test low impedance loads with.
Cordell's design is on his website. |
|
|
| mod_evil |
I found the SG505 schematic on Diyaudio.com and I'm thinking to build this.
Does anybody have a PCB for this?
Thanks for the help.
Best Regards,
Felipe |
|
|
| jackinnj |
The PCB which Cordell used for his Audio article can be found on page 11 of this PDF:
http://www.cordellaudio.com/instrum...hd_analyzer.pdf
You can save yourself a lot of headache by purchasing a used Wavetek, Krohn-Hite Oscillator, using the chassis, power supply and putting in new switch assemblies. The power supply should be well insulated from the rest of the oscillator -- you can even use some PCB material or a cut up beverage can to prevent emi and rfi from migrating onto the oscillator PCB. |
|
|
| mod_evil |
Dear jackinnj,
I live in Brazil, here haven't that Oscillators... The unique option which we can have here is some China Function Generators, and the Goods Functions Generators from HP and Tektronix is really expensive.
That PCB at the page 11 is the PCB of SG505? In the page 12 there is a schematic different of SG505.
Are you certain?
Best Regards,
Felipe Navarro |
|
|
| jackinnj |
The SG505 and the Cordell Oscillator use the same topology -- they are both state variable oscillators -- Bruce Hofer -- the inventor of the SG505 went on to co-found Audio Precision. Cordell credits Hofer at the end of the article on page 8 of the PDF. The paper is on the AES website "A Comparison of Low Frequency RC Oscillator Topologies" -- written when Hofer was at TEK.
The Cordell oscillator provides only a few number of frequencies, while the SG505 is continuously variable, but in reality, you don't need to check every little nook and cranny. |
|
|
| tschrama |
If you really want low-tech solution for a low THD oscilator, then the lamp-stabilized wien-bridge is very good...especially if you use a low-wattage, high voltage lamp bulb such as a 240volt/3Watt .. these have very slow thermal reactance and will easely give you low THD...
Second hand is also very nice: I bought a Krohn-hite 4400 for ? 150$ off ebay.. wich give me <100dB THD... |
|
|
| jackinnj |
| quote: | Originally posted by tschrama
Second hand is also very nice: I bought a Krohn-hite 4400 for ? 150$ off ebay.. wich give me <100dB THD... |
You and Martin are the only guys who have purchased one of these.
There was another really, really good oscillator described in Audio Amateur a while back -- you use a CMOS clock (divide down with a crystal) and follow it with a pair of LTC1063 low pass oscillators. There is a tiny bit of clock-through which can be removed with a 2nd order active filter. I suppose that you could use other LP switched cap filters, I just happened to buy a bunch of the Linear chips off ebay a couple years back/ |
|
|
| mod_evil |
Friends,
When you need to use a Wien Bridge to generate a sinewave, YOU MUST HAVE A OPA602 or a OPA627.
The sinewave is WONDERFUL with these op-amps.
I'm generating 130khz with the OPA627 and 126,4khz with the OPA602. I will test other smaller cap value to see the hightest frequency which OPA627 can oscillate with the Wien Bridge.
The problem now is with the square wave.
When i'm using a OPA602, to top frequency which I have a clean sinewave was 7khz.
=/
Best Regards,
Felipe Navarro |
|
|
| jackinnj |
| A square wave is the sum of all the odd harmonics of a sine wave. Most voltage feedback opamps run out of gain... |
|
|
| mod_evil |
I will test a 4011 to generate the square from the sine.
Do you have any idea why I'm having problems with the period of the square wave, when it is bigger than 18khz, and when it's smaller than 400hz.
When the square wave is smaller than 300hz the top of the square have a angular problem.
See the photo which it attached here.
The frist photo is the lower problem. And the second photo is the problem with high frequencies.
This is a circuit using a operational in comparator mode. The negative input is connected to the sinewave output, and the positive input is conected to the ground.
Best Regards,
Felipe Navarro |
|
|
| mod_evil |
Another photo.
Another information. The suppy haven't a regulator. |
|
|
| sidiy |
Hello,
I just noticed that the schematic in post 10 is a bit unusual and has little chances to stabilize the amplitude properly. The diodes switch suddenly and the gain reduction is either too low or too high depending on the pot value.
If you want try this arrangement, it's easy. It may work better, a matched pair of diodes (like this one ) will ensure a decent symmetrical waveform.
If the pot P1 is adjusted to about 1.4k you'll get about +/-3Vpp amplitude. You can get almost any max. amplitude you want by resizing the values of R1, R2, P1 but I'll leave the exercise to you.... ;)
Let us know how it turns out. |
|
|
| kevinkr |
| quote: | Originally posted by mod_evil
Another photo.
Another information. The suppy haven't a regulator. |
How about a schematic of the squarewave generator portion of the circuit?
Tilt in the squarewave is usually due to phase shift somewhere in the circuit, and the tilt here I think implies that there is LF rolloff somewhere before the 4011. |
|
|
| kevinkr |
| quote: | Originally posted by sidiy
Hello,
I just noticed that the schematic in post 10 is a bit unusual and has little chances to stabilize the amplitude properly. The diodes switch suddenly and the gain reduction is either too low or too high depending on the pot value.
If you want try this arrangement, it's easy. It may work better, a matched pair of diodes (like this one ) will ensure a decent symmetrical waveform.
If the pot P1 is adjusted to about 1.4k you'll get about +/-3Vpp amplitude. You can get almost any max. amplitude you want by resizing the values of R1, R2, P1 but I'll leave the exercise to you.... ;)
Let us know how it turns out. |
He's using a lightbulb now as is the case with classical implementations of the weinbridge oscillator. Diode clipping never seems to work well in practice although in simulations it can appear to perform ok. |
|
|
| wakibaki |
Gotta lotta droop there...
It's probably there at high frequency too, you just don't notice it so much, 'cause it doesn't have time to drop so far...
Have you got a decent decoupling capacitor up close to the power and ground pins on the comparator and a reasonably stable supply?
What is the load?
w |
|
|
| mod_evil |
| quote: | Originally posted by kevinkr
How about a schematic of the squarewave generator portion of the circuit?
Tilt in the squarewave is usually due to phase shift somewhere in the circuit, and the tilt here I think implies that there is LF rolloff somewhere before the 4011. |
Dear Friends,
Mine power supply is full wave retified and filtered by two 2200uf's caps, two 100uf caps and a decoupling capacitor in order of 1.5uf.
A friend called me to use to convert the sine to a square wave one 4011, but now i haven't none on the stock. I will buy someone.
The diode arrangement is a desaster! It's very problematic. I think the light bulb is better than the diode.
Best Regards,
Felipe |
|
|
| mod_evil |
| quote: | Originally posted by wakibaki
Drive your 4011 like this...
Look here: - Waveform Conversion.
w |
Dear Friend,
Thanks for this schematic. At the voltage divider I can use two 100k resistors. And to drive the another input of the NAND gate to High Level? What resistor do you think I can use here?
Best Regards,
Felipe Navarro |
|
|
| wakibaki |
10k is a common value for a pullup, I would make R1=R2=R3=10k, this will give you an input impedance of ~5k.
100k is OK.
w |
|
|
| mod_evil |
Dear Friends,
The square wave is at home now!
:D
A special thanks to all which helped me with this function generator.
Best Regards,
Felipe Navarro |
|
|
| jackinnj |
| Felipe -- does your scope have a serial/parallel port/gbip/ ethernet connector on the back? I've found some neat programs to dump the screen to a file. |
|
|
| mod_evil |
jackinnj,
No, In the rear and in the front of mine scope haven't any port do mine computer.
In the rear I have only The Z-Axis input ( I don't know how to use this and a use for this ) and the Trigger Output.
Best Regards,
Felipe |
|
|
|
|
