New Posts  All Forums:Forum Nav:

AD 8066 problems - Page 2

post #16 of 36
10/15/09 at 7:31pm
Quote:
Originally Posted by tangent View Post
Yes, I'm aware of the high resistor value problem in the Cranky Op-Amps article. I welcome suggestions for alternate configurations that still meet all the requirements. The main thing is low DC offset with high Iib op-amps even accounting for the variable +IN resistance due to the pot. If you leave the pot out, it's easy to get the behavior you want with low resistor values.
Tbh I never found dc offset to be an issue for such a circuit. But in your main article you state a gain of 11(!!!) is a fine starting point... I would say a gain of 3 is a good starting point.

(Thanks for the effort you put into your site btw, great for starting DIYers )
Reply
post #17 of 36
10/15/09 at 7:33pm
Quote:
Originally Posted by thefallenangelx View Post
so actually i cant get it...should i put the r5 res like 10ohm and will it work?
or should i downsize my feedback resistor (4.7k) to 1k or 2k?
You can definitely try R5. It might very well work.

4k7 is fine for a feedback resistor, we were talking about the 100kohm+ range. 4.7kohm is well within the range of the acceptable.
Reply
post #18 of 36
10/15/09 at 9:35pm
Thread Starter 
Quote:
Originally Posted by Slash47 View Post
You can definitely try R5. It might very well work.

4k7 is fine for a feedback resistor, we were talking about the 100kohm+ range. 4.7kohm is well within the range of the acceptable.
So!to let the 8066 i will:
- put R5=10ohm (i just have as for now 47ohm...can i use those instead...or will it affect something somewhat?)
- remove input caps (and put in res instead like 220 - 470ohm?and what will that affect?)
- can i leave on the circuit the 470uf power caps or should i downgrade to the 220 one?
- if i wanna use the ad8620 back...will it work withe the above modification to my circuit?

Thanx everyone!
Reply
post #19 of 36
10/16/09 at 11:26am
Thread Starter 
Quote:
Originally Posted by thefallenangelx View Post
So!to let the 8066 i will:
- put R5=10ohm (i just have as for now 47ohm...can i use those instead...or will it affect something somewhat?)
- remove input caps (and put in res instead like 220 - 470ohm?and what will that affect?)
- can i leave on the circuit the 470uf power caps or should i downgrade to the 220 one?
- if i wanna use the ad8620 back...will it work withe the above modification to my circuit?

Thanx everyone!
Just:
- put 2x R in series (100ohm+150ohm) in the place of the cap input
- put a 47ohm R5 outside the feedback loop

the 8066 sounded slightly better...but the sound its still low and grainy...should i use different values?
or should i put the R5 in the feedback loop?

Can someone please explain me the difference in using input caps or input res?
Reply
post #20 of 36
10/16/09 at 12:26pm
Quote:
Originally Posted by Slash47 View Post
I would say a gain of 3 is a good starting point.
Yes, lowering the gain would help with the DC offset problem. Although it can negatively impact stability and rules out some op-amps, so I can't go too low. I might take it down to 5.

Any more ideas?
Reply
post #21 of 36
10/16/09 at 1:14pm
Quote:
Originally Posted by thefallenangelx View Post
Just:
- put 2x R in series (100ohm+150ohm) in the place of the cap input
- put a 47ohm R5 outside the feedback loop

the 8066 sounded slightly better...but the sound its still low and grainy...should i use different values?
or should i put the R5 in the feedback loop?

Can someone please explain me the difference in using input caps or input res?
An input coupling cap is used to filter out DC, if there's DC from the source (CD or DAP). I don't know of any source with a lot of DC at it's output and therefore coupling caps aren't needed. An exception is when you have an opamp with high bias currents that has to be balanced out by appropriate resistor values. Most of the DIY amps found here don't have input caps, and I guess it's the same for commercial amps.

Input resistors, or "base resistor" if the input is bipolar and "gate resistors" if the input is FET, are used to prevent parasitic oscillation. They also form an RC-filter with the input capacitance and can be beneficial to block out RF.
Reply
post #22 of 36
10/17/09 at 6:08am
I think it's easier to understand pictures.

The basic schematic:


How a pulse looks with and without the 2.2 pF cap:


...and the frequency response, top is without cap to confuse you a bit:

LL
LL
LL
Reply
post #23 of 36
10/17/09 at 6:22am
To illustrate the effect of a capacitive load (1 nF is exaggerated but illustrative):



The pulse without and with both the 2.2 pF cap and the 22R resistor on the output.


... and the frequency response


So you see how the amp could be unstable if not certain measurements are taken.
LL
LL
LL
Reply
post #24 of 36
10/17/09 at 8:29am
Thread Starter 
Quote:
Originally Posted by diditmyself View Post
To illustrate the effect of a capacitive load (1 nF is exaggerated but illustrative):



The pulse without and with both the 2.2 pF cap and the 22R resistor on the output.


... and the frequency response


So you see how the amp could be unstable if not certain measurements are taken.
Yeah, thats useful...but...
- what bout the C1?What value should be used as ur advice?
if i use the r5 resistor could i omit the c1 you put between out and ground, or is it necessary? I just didnt understand the use of C1...
- What program do you use for those response and bode diagram?
Reply
post #25 of 36
10/17/09 at 8:54am
Quote:
Originally Posted by thefallenangelx View Post
Yeah, thats useful...but...
- what bout the C1?What value should be used as ur advice?
if i use the r5 resistor could i omit the c1 you put between out and ground, or is it necessary? I just didnt understand the use of C1...
- What program do you use for those response and bode diagram?
You shouldn't use C1 at all. I just put it there to illustrate the effect of capacitance on the output. Both cables and headphones have "inborn" capacitance. They also have reactance. That means an amplifier can be stable without load and start oscillating when headphones are connected. I've experienced that several times, and to cure it I've used one of the methodes mentioned above.

To test the stability of an amp one can add a capacitor like in the schematic above, but to be clear it shouldn't be there in the finished amp.

The program I'm using is LTSpice. It's free from Linear Technology Linear Technology - Linear Home Page . Its not very intuitive for a noob like me. I've only learnt to scratch the surface, but it's very useful, and I think it almost always predict the real circuit.
Reply
post #26 of 36
10/17/09 at 9:13am
To give you one more example, a zobel network (resistor and capacitor in parallel with the output), that also makes the amp less susceptible to instability due to capacitive loads. Using a zobel has the advantage of unaltered output impedance as compared to the added 22R resistor in series with the output in the example above.

Some say that a zobel is a band-aid for a crappy design. I think it's an ignorant attitude. A zobel can be very handy.


LL
Reply
post #27 of 36
10/17/09 at 9:23am
I just popped an AD8066 in a circuit as the one above, not the exact values but fairly close, and there's an opamp as ground circuit instead of a TLE2426. It plays loud with no distortion. It's not as good as the AD8599 I'm normally using in this amp, but that's another story.
Reply
post #28 of 36
10/17/09 at 7:58pm
Thread Starter 
Quote:
Originally Posted by diditmyself View Post
I just popped an AD8066 in a circuit as the one above, not the exact values but fairly close, and there's an opamp as ground circuit instead of a TLE2426. It plays loud with no distortion. It's not as good as the AD8599 I'm normally using in this amp, but that's another story.
Thank you!just soldered a new cmoy with ur schematic and...ill test it tomorrow!
Oh, in LTspice...where can i found ad8066 opamp (and other opamp) component to simulate some schematic?
Reply
post #29 of 36
10/18/09 at 6:02am
Quote:
Originally Posted by thefallenangelx View Post
Thank you!just soldered a new cmoy with ur schematic and...ill test it tomorrow!
Oh, in LTspice...where can i found ad8066 opamp (and other opamp) component to simulate some schematic?
You can only find LT opamps in the LTSpice package. To use other brands you have to download a subcircuit file from Analog, TI or National. You also need a symbol for every opamp or a symbol called x-opamp. This is where LTSpice is starting to become messy. I'd recommend the "x-opamp". I haven't learned to create symbols.



Open the zip-file below. Put the AD8066.cir in the same folder as your schematic and put the xopamp.asy in the symbol folder. You'll find the x-opamp when you press F2.

In the schematic you have to press "t" and enter this string, and dont forget to push the "spice direcitve" button.


The other thing you have to do is right click the x-opamp symbol and enter the same name in the "spice model" field.


To learn more about LTSpice there's a Yahoo forum. You'll find it if you Google. Diyaudio.com is also essential to learn enough electronics to design your own amp and use LTSpice.
LL
LL
LL
Reply
post #30 of 36
10/20/09 at 11:15am
Quote:
Originally Posted by tangent View Post
Yes, lowering the gain would help with the DC offset problem. Although it can negatively impact stability and rules out some op-amps, so I can't go too low. I might take it down to 5.

Any more ideas?
You should explicitly mention that the integrating capacitor C4 can only be used with opamps that are unity gain stable.

Adding a couple of further reading links to ti / analog app notes on opamp stability probably wouldn't hurt.
Reply
New Posts  All Forums:Forum Nav:
  Return Home