It looks to me that 2 Teddyregs, instead of IC3 and the VREF circuit suggested by Russ, will do a better job.
Joshua_G said:
Actually probably not bad, but probably not better either.
The teddy reg just does not have low enough output impedance. That not to say it would not work well, but we really need ultra low impedance here to keep distortion down.
Even with a large caps the output impedance will not be as low in the audio band as the on-board supply is.
This why feedback is very good at the AVCC pins.
Cheers!
Russ
Re: Re: Next level VREF
Hi Fred,
Yes, that Linear part is very nice. The new buffalo will utilize something exactly like that. There is another part (LT1027) which looks promising too.
One thing I like about the LEDs is that they are cheap and very easy for a DIY mod, which is why I chose them for this example. They should provide a good low noise reference.
I will see if I can knock something together for you guys quickly.
I will do a simple single sided PCB here shortly that uses the LEDs. Then you folks who want to etch it can.
But the simple P2P that Rolls and avr300 did, are actually nice and would be just as effective.
Cheers!
Russ
fmak said:
Hi Fred,
Yes, that Linear part is very nice. The new buffalo will utilize something exactly like that. There is another part (LT1027) which looks promising too.
One thing I like about the LEDs is that they are cheap and very easy for a DIY mod, which is why I chose them for this example. They should provide a good low noise reference.
I will see if I can knock something together for you guys quickly.
I will do a simple single sided PCB here shortly that uses the LEDs. Then you folks who want to etch it can.
But the simple P2P that Rolls and avr300 did, are actually nice and would be just as effective.
Cheers!
Russ
Single Side PCB PDFs for VREF
Here is the VREF with bipolar CCS on a small single side PCB.
The zip includes everything you should need.
Cheers!
Russ
Here is the VREF with bipolar CCS on a small single side PCB.
The zip includes everything you should need.
Cheers!
Russ
Wow, that was fast russ!
So to get this right (as you can tell my PCB experience is not huge) what I should do is:
when I look at the finished etched PCB, the copper is on the bottom side and looking through the fibreglass the traces should look exactly the same as the bottom PDF?
Then the mirror image would be useful where people use the toner method, right?
Another question: why are the green LEDs important? Is it just the forward voltage? I have some HLMP6000 that have a Vf of 1.6-1.7V would they do instead? I also need to find a BC560 - would others eg 2n5087 do?
Fran
So to get this right (as you can tell my PCB experience is not huge) what I should do is:
when I look at the finished etched PCB, the copper is on the bottom side and looking through the fibreglass the traces should look exactly the same as the bottom PDF?
Then the mirror image would be useful where people use the toner method, right?
Another question: why are the green LEDs important? Is it just the forward voltage? I have some HLMP6000 that have a Vf of 1.6-1.7V would they do instead? I also need to find a BC560 - would others eg 2n5087 do?
Fran
Hi Fran. I aim to please. 😉 It only took about 5 minutes to do.
For toner method use the normal (non-mirrored) bottom layer. It will make more sense once you have done one.
Then use the mirrored top silk screen if you want to transfer the components and RefDes etc.
The LED just should be low noise with a forward voltage drop of 1.7-1.8v at about 5ma. Stay away from blue, white, or high brightness types. Any standard red/green/amber should be just fine. You may need to adjust R3 to get your 3.3-3.4V
Cheers!
Russ
For toner method use the normal (non-mirrored) bottom layer. It will make more sense once you have done one.
Then use the mirrored top silk screen if you want to transfer the components and RefDes etc.
The LED just should be low noise with a forward voltage drop of 1.7-1.8v at about 5ma. Stay away from blue, white, or high brightness types. Any standard red/green/amber should be just fine. You may need to adjust R3 to get your 3.3-3.4V
Cheers!
Russ
Russ is the PNP type important. I think I have some 2n5087 here.... will any pnp do or must it be BC560?
Fran
Fran
woodturner-fran said:
Use the lowest noise type you have.
2N5087 look fine, but the pin out looks like it is opposite of BC560(EBC instead of CBE) so you will need to put it in backward.
BC560 and BC327-25 would be good. But there are many others that will also work very well.
Have fun!
Russ White said:
I have just found a 2SA798 with one side of complemenatry pair only. Noise figure 0.5 dB. Will try it.
What is the current draw for AVCC?
.
fmak said:
I made up the circuit using 2N5087 and 2SA798 and there is virtually no difference in measured noise level which is about 4uV 1MHz and 0 and 4 mA current draw. This is very quiet (4 nV/sqrt Hz) although I don't know what the 1/f noise levels are. Maybe not too important for AVCC? Is ceratiainly be much less than what some 'high end; cost?' people are appearing to put in.
fmak said:
AVCC is dynamic. O depends on the signal out of the DAC. But the idle current is around 20ma if I remember correctly. The idle current of AVCC is depndant on the voltage at the outputs (0V for IVY). When using the DAC in voltage output mode the idle current will be pretty low.
The swing is very big, at around 16ma pp in stereo current output mode. This is why low impedance is key.
The opamp can easily handle it current wise. But adding a follower (in the feedback loop) would not hurt either. 🙂
Cheers!
Russ
Extension of the idea...
Here is a low noise regulator based on the VREF idea above.
The output (pass) device can either be To-220 or TO-92 depending on what you need.
I added the capability to take feedback(Sense) from the load instead of the right at the output. You just extend RF to the further pad and wire the sense wire to the pad marked "Sense"
Here is a low noise regulator based on the VREF idea above.
The output (pass) device can either be To-220 or TO-92 depending on what you need.
I added the capability to take feedback(Sense) from the load instead of the right at the output. You just extend RF to the further pad and wire the sense wire to the pad marked "Sense"
Attachments
Russ White said:
Russ
You are prolific. Is this going to be available?
One question - is the previous vref based on the pnp transistor enough to drive the Buffalo LM4562? Or are you just developing a new feedback regulator?
My experience with the ALW regulator for general use is that a tracking preregulators improves performance significantly. The Teddy series pass prereg actually doesn't do anything extra.
Fred
up and running!!
Hi
No big news, just wanted to say that my baby is up and running (including the new VREF). Everything worked as it should. Soldering and assembling was straight forward.
And what about the sound? All I can say: I'm very happy. To me it just sounds right - and that's fantastic. 🙂
Next thing will be a nice case - but first I'll have to sit down, take a deep breath and have some listenings.
Brian and Russ, thanks a lot for making this possible.
Cheers!
Mario
Hi
No big news, just wanted to say that my baby is up and running (including the new VREF). Everything worked as it should. Soldering and assembling was straight forward.
And what about the sound? All I can say: I'm very happy. To me it just sounds right - and that's fantastic. 🙂
Next thing will be a nice case - but first I'll have to sit down, take a deep breath and have some listenings.
Brian and Russ, thanks a lot for making this possible.
Cheers!
Mario
Attachments
- Status
- Not open for further replies.