diyAudio

diyAudio (http://www.diyaudio.com/forums/)
-   Software Tools (http://www.diyaudio.com/forums/software-tools/)
-   -   LT Spice design (http://www.diyaudio.com/forums/software-tools/186449-lt-spice-design.html)

ianmgreen 5th April 2011 11:13 AM

LT Spice design
 
1 Attachment(s)
ok so i would definately appreciate some help,

I am currently studying at uni and for my final project my aim was to design and build a 3 band parametric EQ, on building i came across lots of obstacles and got to the point where i couldnt determine if there was a circuit error or a build mistake.

I therefor decided simulation was a better option, using LTspice i created the attatched design, however now im at a standstill with that aswell.

My aim is to create a model of the unit and show its EQ curve (frequency response) show that altering the pots alters performance and be able to tweak component values to create bands appropriate to my project. Once i have proved concept i can begin a final build.

Any help on the design and LTspice would be apreciated.

macboy 5th April 2011 03:43 PM

The first obvious problem is that your opamps have not been connected to any power supply. That is a show-stopper.

I also see that you have connected an AC voltage source to your output... why? I'd suggest removing it.

edit: Another BIG problem, you have labelled the output of each of the 3 EQ sections as "out". Guess what LTSpice does when it sees different wires with the same label? Connects them! Right-click on any one of those wires, select "highlight net", and you will see all wires that LTSpice has connected together to create the "out" net.

I also see very large (10000 to 22000 uF) AC Coupling caps in your signal path. Sure these will help pass bass without roll-off, but they can take a very long time to reach their steady state DC bias after power is supplied (They need to charge up through whatever series resistance is present). This can cause problems in the simulation unless you first simulate the DC operating point, then use that as a baseline when starting the simulation, an advanced technique you are not ready to do.

Also, please attach your potentiometer_standard.lib file. We can't simulate your circuit without this.

ianmgreen 5th April 2011 04:12 PM

1 Attachment(s)
thanks for the assistance, here is the .lib file let me know what you think

macboy 5th April 2011 05:13 PM

I was able to simulate your circuit. It has a flat(ish) gain up to around 1 MHz. I could not get any change in the response by adjusting the pots, and since you didn't label them with their function I was just guessing.

In addition to the suggestions I made above, I also changed the ".noise" statement into a ".ac" statement with 10 Hz to 1 MHz and 10 steps per octave. Why were you doing a noise analysis? What you want is an AC analysis which gives you magnitude and phase response vs. frequency.

ianmgreen 5th April 2011 05:21 PM

sorry i did originally put through an ac signal but was fiddling with the noise gen as im used to using a pink noise signal with an eq unit, i had the same issue when i simulated, no change when pot values were altered, im starting to think its the design, its not mine originally as im using it as a jumping off point and not really sure what im doing. It was originally just one band, as per the design here:

Spectrum Analyzer and Equalizer Designs

i built a single band of the unit and had similar issues, gain control but no filter control, im starting to think i should start from scratch on the design but honestly wouldnt know where to start and have limited time.

any suggestions on why the design isnt operating/alternate designs i can adapt or where i can learn fairly rapidly the basics of designing my own unit?

jcx 5th April 2011 05:32 PM

don't think you've got a simulated circuit "working" until you do a .tran sim and see proper bias V, good waveforms as expected on all nodes

.AC, .noise are not "real" simulations - they linearize the circuit about the dc operating point and just do the Laplace math - won't find clipping, current limited nodes, nonliear operation

only use the .ac, noise analysis "sims" after the sim is validated in .tran analysis with several test waveforms - just like you would use on the bench with a waveform generator, oscilloscope on the real hardware before hooking up the spectrum analyser

ianmgreen 5th April 2011 06:04 PM

to be honest that last one was a bit over my head, this is my first design and im completely new to LTspice. But unfortunately its something i have to do for my final technical project (I dont study electronics, mine is an audio/broadcast engineering degree). So I apologise if i make silly mistakes and need things dumbing down, but i really do appreciate any help you guys can give

macboy 5th April 2011 06:56 PM

Quote:

Originally Posted by ianmgreen (Post 2528541)
sorry i did originally put through an ac signal but was fiddling with the noise gen as im used to using a pink noise signal with an eq unit, i had the same issue when i simulated, no change when pot values were altered, im starting to think its the design, its not mine originally as im using it as a jumping off point and not really sure what im doing. It was originally just one band, as per the design here:

Spectrum Analyzer and Equalizer Designs

i built a single band of the unit and had similar issues, gain control but no filter control, im starting to think i should start from scratch on the design but honestly wouldnt know where to start and have limited time.

any suggestions on why the design isnt operating/alternate designs i can adapt or where i can learn fairly rapidly the basics of designing my own unit?

In the schematics on the page you referenced, the term "mF" is used for microFarads. Sometimes you just need to be able to know from context that mF means micro- not milli-. In fact it is very rare that anyone uses milliFarads. Only pico (pF), nano (nF) and micro (μF) are commonly used. I guess that is why you have absurdly large caps in your circuit.

ianmgreen 6th April 2011 01:17 PM

2 Attachment(s)
Ok, so i have made the suggested corrections,

The performance has altered however I still do not have functioning pots, I followed the guidelines on the LTspice yahoo group when setting them up but feel i have made a mistake somewhere as they are not working.

I have attatched a single band of my design as it stands now and the .lib folder i have used again, can anybody spot where i am going wrong?


All times are GMT. The time now is 11:08 AM.


vBulletin Optimisation provided by vB Optimise (Pro) - vBulletin Mods & Addons Copyright © 2014 DragonByte Technologies Ltd.
Copyright 1999-2014 diyAudio


Content Relevant URLs by vBSEO 3.3.2