(re)searching for a better preamp

[rudi]

Did you use anything special for the 47uf caps? More oscons or morebasic stuff?

I Used a electrolytic SMD cap, but you can use something basic there if you want. i did not notice a change in performance with this cap

 

[filholder]

Hi Rudi,

Sorry to be a pain,

I couldnt get 16mm 3300uf 35v caps anywhere. I had to buy 18mm, this means i will have to alter the layout a bit to make space. I was wondering if you have the designs you sent me as a Eagle file rather than pdf. I havent used Eagle before but i am guess that way has got to be easier than trying to chop up the design in Photoshop, the only other way i can think of doing it.

Regards

Phil Holder

 

[Panelhead]

Dual AD815 Circuit

I loved my buffer after dialing it in a little. So never to leave it alone, I built this one using two AD815's. The unused channel in each has a 562 ohm feedback resistor and a 1K resistor tying non-inverting input to ground.

George

 

[Panelhead]

Clean top side

Most everything is underneath

 

[Panelhead]

Busy side

The amount of wiring and parts makes this a layout pain. Took several attempts to finish.

 

[Panelhead]

Just needs some heatsinks.

To fit on my modular board, the power supply had to go in the middle. This helped a little.
The heatsink used in the stereo version playing fit not fit here. Going to hacksaw up a couple to allow checking.
Even if it does not sound any better than the stereo version, this was six hours well spent. This in my mind is close to optimized for my system, I can move on to another component in the chain.
Thank you Carlos for sharing this overperforming circuit.

George

 

[jaudio]

George

Have you had the chance to measure the dc offset on the unused channel,to see if the null offset circuit has an affect on it?

On my preamp,I used the unused channel for a headphone amp. I hardly ever used headphone but I didnt want the channel to go unused. I used one null circuit for each ad815 with both channel connected to one circuit. I notice that,when adjusting dc for one channel it changes the offset for the other(causes it to go up).
btw the headphone has the gain of 5x and the preamp is a buffer.
In the end,I will more than likely,disconnect the null circuit from the headphone amp and just add a output capacitor.

If you find the time to test dc for both channels,I would be very interested in you findings


Thanks

 

[Panelhead]

I will check

I have measured the offset when both channels do not have the nulling circuit. It varied between 60 mv and 180 mv depending on volume pot position. This was with a 63K input impedance setting resistor, 3K series resistor, 1800 ohm input transformer, and a 25K pot.
The 1K resistor from NI input to ground was chosen to lower this offset. Would have gone lower, but wanted to keep the current going from input to ground at a minimum.
I suspect the nulling circuit connected to one channel will effect the other. Playing around with the nulling voltage, it did vary a little when diddling the other channel. But this may have been shifting the overall null voltage as the loading changed on it.


George

 

[carlosfm]

On my preamp,I used the unused channel for a headphone amp. I hardly ever used headphone but I didnt want the channel to go unused. I used one null circuit for each ad815 with both channel connected to one circuit. I notice that,when adjusting dc for one channel it changes the offset for the other(causes it to go up).
btw the headphone has the gain of 5x and the preamp is a buffer.

Jaudio, please don't do that
You can't expect to adjust offset with such a different gain between both halfs of the chip.
You must really use independent trimming adjustment.
Also, I don't know how you did that, did you duplicate the 390k resistors?
It won't work that well... the null circuit (components after the 9V reg) was made for one channel alone.

 

[jaudio]

Carlos

This is what I have

 

[Panelhead]

Concept Question

I built this voltage nulling circuit to help control input and output offset. It also alows this circuit to work wonderffully in my system.
Looking at it, there is a voltage divider between the 15K resistor and the 500 ohm pot. This will result in a maximum voltage on the 390K resistor of 0.33 volts above ground reference.
This 390K resistor is functioning as a buffer to prevent AC under it from modulating back into the power supply. This is the function also of a drain resistor or plate resistor in a voltage amp circuit.
With a maximum voltage across this resistor so low, how does this not modulate when say a 2 volt signal is applied at the bottom of the 390K resistor?
I got to say, if it is modulating and screwing with the sonics I cannot hear it. Maybe I have not turned up the input level near high enough.
Carlos, you as the designer, what is the reasoning here for why it operates so well? I must be missing something on how this nulling circuit interacts with the AD815.

All Respects,
George

 

[carlosfm]

Carlos
This is what I have

Then you have independent dc null circuits, after all.
It just hit me that efectively it will not work at all, even this way.
Because you are joining both NI (per channel) inputs on the volume pot.
So, you are indeed affecting both channels.
I would suggest you to use a small input coupling cap for the headamp seccions. You must have 100k resistor to ground on each chip channel (not one for both halfs of the chip). Before this resistor, on the headamp halfs, put a 470nF MKP cap.

 

[carlosfm]

Re: Concept Question

Carlos, you as the designer...

Yes, Sir!

...what is the reasoning here for why it operates so well? I must be missing something on how this nulling circuit interacts with the AD815.

I've chosen a high value resistor (hundreds of k) so that it didn't increase crosstalk between channels.
That's all.

😀

 

[jaudio]

The one a have sounds very good too,both the preamp and the headphone amp.

Im going to adjust the circuit to get it as low as possible for the preamp. The headphone will have an output cap,no matter what.
I have listen to both without output cap. I get an incease of 2mv(preamp section) at power amp output,no matter pot position.

I cant remember what it was for the headphone amp but it was not large

 

[jaudio]

No, don't cascade them.
The pot feeds the two halfs of the AD815 (one channel), and use ~200k instead of 100k for R3.
Don't forget to adjust DC-offset on the headphone amp with the same load as your headphones.

I would suggest you to use a small input coupling cap for the headamp seccions. You must have 100k resistor to ground on each chip channel (not one for both halfs of the chip). Before this resistor, on the headamp halfs, put a 470nF MKP cap.

Excuse the earlier post,I was typing it while you posted you response

Ok change the 200k resistor to 100k?

Put a small cap at headphone input instead of at output?
If I use the small cap at input will i still be able to use the null circuit?

 

[carlosfm]

Ok change the 200k resistor to 100k?

Where do you have 200k resistor?
I mean, look at my schematic.
You need R3 (100k) one each amplifier stage, don't mix this with the two halfs of the chip.
So, you need 4x 100k resistors.

Put a small cap at headphone input instead of at output?

Yes. At the input you will use a small, good quality cap.
At the output, for the headphone seccion, you would need a big electrolythic.

If I use the small cap at input will i still be able to use the null circuit?

Use an input cap on the headphone amp seccions and adjust dc offset on the output with the dc null circuit.
With the input cap, these headphone amp seccions will be 'isolated' from the line preamp seccions, this way you will be able to adjust dc offset independently on the 4 channels.
Cool, no?

 

[jaudio]

Carlos thanks for your response

1. I have 4 200K,I will change them to 100k

2. The input cap goes bewteen the R3 and the NI input pin?


Yes very cool

 

[carlosfm]

Carlos thanks for your response

1. I have 4 200K,I will change them to 100k

Ok, I understood you had 2x200k. Let them be, then.
On a second thought, you will put a lower load on the volume pot, because you have 4 channels.
Let those 4x200k, and use a 10k~20k volume pot.
Only if you can't adjust dc properly, then change those resistors to 100k.

2. The input cap goes bewteen the R3 and the NI input pin?

No, it goes before, between the volume pot and R3, of course...
That's only for the headamp seccions.
You can use 220nF, if you keep those 200k resistors.
Just figure it out, the only thing you have to do is insert a pair of 220nF caps, nothing more.

Yes very cool

Right. 😎

 

[jaudio]

I just found some .27uf caps. (bewteen the wires and the pink and yellow caps)

Now,I will have to leave them standing upright and unsecured or put them on the board with the optional output caps(two outputs one with caps and one without)the green caps will be removed. Will there be a problem running input and output so close.

The layout is giving me h*ll

 

[carlosfm]

Will there be a problem running input and output so close.

The problem is all those unshielded wires running so close to each others.