Tangentsoft CMOY tut - substituting parts?
Dec 11, 2008 at 9:56 AM Post #37 of 57
Hmm, whats some popular, easy swaps? I want to start playing around with the configuration now that I'm getting to know the CMOY design.

How do I order from digikey?

EDIT: I tried ordering parts from digikey, but the shipping was as much as the parts themselves............

I'm actually trying to source parts for a Joshatdot CMOY at the moment as I recently received one of his CMOY boards.
 
Dec 11, 2008 at 10:03 AM Post #38 of 57
Quote:

Originally Posted by millwood /img/forum/go_quote.gif
personally, I don't understand the fascination with dc-coupled amps because as far as I know, no speakers can reproduce DC, no functioning audio sources produce DC, and no humans can hear DC.


Are you being serious? I don't mean to be an a-hole but c'mon, acting all high and mighty about being a know-all, and then coming out with that? Priceless.
 
Dec 11, 2008 at 11:14 AM Post #39 of 57
Quote:

Originally Posted by rogerlike /img/forum/go_quote.gif
Are you being serious? I don't mean to be an a-hole but c'mon, acting all high and mighty about being a know-all, and then coming out with that? Priceless.


Can you guys please leave this thread if you have nothing on topic to add.

Quote:

Originally Posted by nsx_23 /img/forum/go_quote.gif
Hmm, whats some popular, easy swaps? I want to start playing around with the configuration now that I'm getting to know the CMOY design.

snip


Yep... shipping from digikey or mouser is expensive.

You will find some easy swaps here.
 
Dec 11, 2008 at 11:37 AM Post #40 of 57
Quote:

Originally Posted by cobaltmute /img/forum/go_quote.gif
It is a fairly simple concept to understand and figure out.


tough questions are easy to understand, that's why so many people have understood them so many times.

Quote:

What I want to know is, should I put resistors there with the same value as R3?


yes in general and no in certain cases. some opamps (opa227 for example) has a bias cancellation circuitry that does not need bias cancellation.

most other opamps will require bias current cancellation (R2=R3//R4).

the problem with the cmoy is that R3 is so small (1k) that the resulting R2 is small. cmoy opamp has a gain of 11x but R1/R2 attenuated the signal first by 0.5x so the effective gain is 5x, on the high-end of the gain range for a typical headphone amp (3 - 5x typically).

another problem with getting rid of the C2 is that if you have a pot in front of the amp, changing volume will change the effective resistance from the non-inverting pin to the ground, which changes your output dc offset.

the best alternative, in my view, is to live with C2.

if you really want to get rid of it, and can live with a lower gain (maybe your headphone is sensitive for example), set R3=10k, R4=22k (resulting gain = 3x), and R2=10k, and use a <47k pot (if you don't have a pot, R1=1k).

everything else being equal, small resistors give you lower thermal noise and make the amp less prone to interference.
 
Dec 11, 2008 at 2:48 PM Post #41 of 57
Quote:

Originally Posted by Ron.id /img/forum/go_quote.gif
Thanks for your reply
biggrin.gif


Ah, I forgot about Tangent's calculator. Btw, what is the meaning of "||" in R2 = R3 || R4?


"||" is a notation for in parallel with.
 
Dec 11, 2008 at 5:00 PM Post #42 of 57
Quote:

Originally Posted by millwood /img/forum/go_quote.gif
yes in general and no in certain cases. some opamps (opa227 for example) has a bias cancellation circuitry that does not need bias cancellation.


I use AD8397 or AD8066. Don't know about bias cancellation circuitry on these opamps.

Quote:

Originally Posted by millwood
if you really want to get rid of it, and can live with a lower gain (maybe your headphone is sensitive for example), set R3=10k, R4=22k (resulting gain = 3x), and R2=10k, and use a <47k pot.


Yes, my cmoy gain is 3x (not 11x as the original). Why 10K for R2 if pot is <47K? I tought it should be ~10x higher than pot resistance
confused_face(1).gif

My pot is 10K, and i use 100K for R2.

Quote:

Originally Posted by Pars
"||" is a notation for in parallel with.


I see... thanks.
 
Dec 11, 2008 at 5:54 PM Post #43 of 57
Quote:

Don't know about bias cancellation circuitry on these opamps.


in principle, you should always pay attention to offset cancellation.

as anything in life, there are exceptions. the need for offset cancellation depends on the magnitude of the bias current. in your case, the ad8066 has a typical bias current of 1pa (1e-12) and ad8397 200na (200e-9). so even if the grounding resistor is 100k higher than the effective resistance on the inverting end, the DC offset voltage at the inverting and non-inverting ends are 1e-12*100e3=.1uv. through the amp's gain of 10x, that's just 1uv on the output. not a big deal. you will find similar stories with opamps utilizing a FET input stage.

ad8397 is a little dicier. its input bias current of 200na on 100kohm generates an offset voltage of 200e-9*100e3=20mv. through an amp's gain of 10x, that's 200mv on the output - a little bit too much.

if you drop ad8397 right into the cmoy design (R2=100k, R3=1k, R4=10k), the voltage caused by the input bias current is 200na*(1k//10k)=~200uv on the inverting end, and 200na*100k=20mv on the noninverting end. multiple that by 11x you get about 200mv on the output.

if you set R2=10k, R3=10k, R4=22k, the inverting end will see 200na*(10k//22k)=1.38mv. the non-inverting end will see 200na*10k=2mv. the difference of .62mv will be amplified 3x and produce a 2mv offset on the output. not bad.

if you set R2=R3//R4=6.9k, the offset voltage will be precisely zero. But at that point, other factors will kick in.

the above assumes of course that C2 is in place.

if you jumpered C2, you will need to factor the wiper position into the calculation, which causes the problems I talked about earlier as the dc offset changes with wiper position.

Quote:

Why 10K for R2 if pot is <47K? I tought it should be ~10x higher than pot resistance


ideally, you don't want the amp's input impedance (=R2//opamp's input impedance =~R2 since opamp's input impedance is much larger than R2) to overwhelm the pot - but even if that happens it is not the end of the world.

Quote:

My pot is 10K, and i use 100K for R2.


the discussion branches out a little depending on if C2 is shorted.

if C2 is not shorted, the input bias current will go through just R2 so our discussion earlier stands.

if C2 is shorted, the input bias current will flow through R2 and the lower part of the pot (from wiper to the ground). since the pot is much smaller than R2, most of the current will go through the pot so effectively the non-inverting end is terminated by the lower part of the pot. since that resistance is anywhere between 0ohm to 10k, you are not that far off the "ideal" value of 6.9k: the max dc offset, reached when the wiper is at the ground, is just 200na*6.9k=1.4mv at the inputs, and 4mv through the amp's 3x gain. the DC offset is 200na*(10k-6.9k) = 2mv, in the opposite direction.

so by shorting C2, you brought the pot into the dc offset calculation; by using a 10k pot, you have basically taken R2 (=100K) out of the calculation for dc offset.

so here is I suggest:

1) if you are keeping C2, set R2=10k, and live with the modified behavior of the pot;
2) if you are shorting C2, use a large R2 or take it out of the circuit, and use a pot that's approximately 2x the value of R3//R4 - that way, you minimize the dc offset on the output.

of course, if you use a fet opamp, the above discussion is a little bit moot - for those amps, "input offset voltage" is the parameter you pay attention to, but it is difficult to null that.
 
Dec 11, 2008 at 6:42 PM Post #45 of 57
ok - lol - dropped thread is still alive...
well, i'm 95% finished building the circuit part of my first cmoy - one solder burn and still struggling a bit, but at least some success...at this stage no pot or audio jacks - i did the first tests with test leads as outlined in tangent cmoy tut - well first after checking whether things run hot, i just checked the dc out - i used a 2227 which used to be in my purchased bsg cmoy - the dc for both channels was around 4.7...so seemed ok and slightly better than in the bsg cmoy. Then i hooked up some cheap buds - the gain of 11 is high i know but tangent recommends it for a start, so i took the advice - so i pushed up my source to 3/100 volume (and then it's already a little too loud) - it sounded good but a lot of hiss(i spose it is quiet hiss cos at 3%volume the audio is much louder than the hiss) - i was concerned at first but then i remembered that bsg cmoy also gives a lot of hiss on maximum gain (which i presume is the same as having no pot) - i just checked it against bsg cmoy (turn the source down and then put the gain on max) and yep, i think the bsg is even slightly more hiss (his gain might be even higher?)...i suppose the ideal would be to have no hiss at max gain, but i have low impedance phones - my highest being 50ohm for the hd595 - so i presume this is normal with gains in the region of 11 and up?
Usually what i do with the bsg cmoy, is to put my source on 25% and then i set the volume of the amp from about 15-25% - and when no source is playing it's absolutely quiet and volume control is not too touchy - workable and quality seems good in that range...would higher impedance phones have less hiss at higher source volumes?
tx for all input - i'm very new to the diy so thanx for all help so far...k
 
You must log in or register to reply here.

Users who are viewing this thread

Total: 2 (members: 0, guests: 2)
Back
Top