Cheers Carlos,
Sorry i didnt mean to sound rude it's just frustration, I seem to have hit my limit. While my soldering skills are pretty damn decent now and my case building skills pretty good my ability to translate complex cuicuits to vero board are quite limited.
Thanks again for sharing your work, one day maybe ill be up to understand enough of this thread to build the circuit myself.
How does the jfet circuit sound in comparison, is it very much worse?
Regards,
Phil
Sorry i didnt mean to sound rude it's just frustration, I seem to have hit my limit. While my soldering skills are pretty damn decent now and my case building skills pretty good my ability to translate complex cuicuits to vero board are quite limited.
Thanks again for sharing your work, one day maybe ill be up to understand enough of this thread to build the circuit myself.
How does the jfet circuit sound in comparison, is it very much worse?
Regards,
Phil
Ideas -
Do not let the circuit drawings scare you. It is very easy to build, I even built one and it worked the first time! I am glad the dip version is 1.50 here, and the SM is 10.00. I do not like building using SM parts.
The voltage into the input is to set the input voltage close to zero. This voltage is very highly buffered, it is really just an adjustable voltage reference.
With the straight resistive feedback loop, the dc offset at the input and output track fairly closely. The input offset varies with input impedance, that is why it changes when the pots are rotated.
My thoughts on clipping off the leads is this chip is not going to like being unloaded. There needs to be a feedback resistor and most likely an input impedance to prevent ringing in the unused side. This will degrade the sonics if this occurs. And maybe make it heat up.
The dip package has several unused pins as it is. I really wanted to clip off the unused channel +/- input, output also. Use a different side in each. Then bolt the two 815's together back to back with a heatsink in the middle. Loading the input and putting on a feedback resistor is not much trouble.
Time will tell on the sonics. With all new resistors, caps, pots, and wiring it is expected to not sound its best. But it really sounded nice stone cold.
George
Do not let the circuit drawings scare you. It is very easy to build, I even built one and it worked the first time! I am glad the dip version is 1.50 here, and the SM is 10.00. I do not like building using SM parts.
The voltage into the input is to set the input voltage close to zero. This voltage is very highly buffered, it is really just an adjustable voltage reference.
With the straight resistive feedback loop, the dc offset at the input and output track fairly closely. The input offset varies with input impedance, that is why it changes when the pots are rotated.
My thoughts on clipping off the leads is this chip is not going to like being unloaded. There needs to be a feedback resistor and most likely an input impedance to prevent ringing in the unused side. This will degrade the sonics if this occurs. And maybe make it heat up.
The dip package has several unused pins as it is. I really wanted to clip off the unused channel +/- input, output also. Use a different side in each. Then bolt the two 815's together back to back with a heatsink in the middle. Loading the input and putting on a feedback resistor is not much trouble.
Time will tell on the sonics. With all new resistors, caps, pots, and wiring it is expected to not sound its best. But it really sounded nice stone cold.
George
JFET PREAMP/ FILHOLDER
Hello Filholder,
You can have your cake and eat it too!
Just use the dual JFET, 2SK389, in cascode as a type of "buffer" prior to the AD815 as posted earlier. You will have the gain, the quality of Carlos' preamp, and VERY low to no DC offset.
🙂 Bruce
P.S. You could match 2 JFETs instead, per channel.
Hello Filholder,
You can have your cake and eat it too!
Just use the dual JFET, 2SK389, in cascode as a type of "buffer" prior to the AD815 as posted earlier. You will have the gain, the quality of Carlos' preamp, and VERY low to no DC offset.
🙂 Bruce
P.S. You could match 2 JFETs instead, per channel.
No i can't have my cake and I am straving. If i cant work out the original circuit how can i cut of the front of it and put on a buffer.
I am going to do some studying of the circuit today and might have to ask some questions later.
Like i said a days or two ago if anyone has pcb layouts for their version of this AD815 based Preamp i would be very grateful if i could see/use then.
regards,
Phil
I am going to do some studying of the circuit today and might have to ask some questions later.
Like i said a days or two ago if anyone has pcb layouts for their version of this AD815 based Preamp i would be very grateful if i could see/use then.
regards,
Phil
Hi everyone,
Just to say Rudi has been a star and sent me his pcb masks so i am going to have a go at his AD815 based layout. I need to work out a few of the finer details but i recon it is do-able for me and am very excited.
Thanks for everyone help,
Phil
Just to say Rudi has been a star and sent me his pcb masks so i am going to have a go at his AD815 based layout. I need to work out a few of the finer details but i recon it is do-able for me and am very excited.
Thanks for everyone help,
Phil
Hi All
One thing am am interested in is why do most of thse circuits only use single stage regulation. I would have thought that with something as precise as a preamp, and with lm317 and lm337 being so cheap that dual regulation would be a must. There must be a reason its just i see it.
As usual answers on a postcard to...
Or rather if you would like to comment i would be grateful.
Regards,
Phil
One thing am am interested in is why do most of thse circuits only use single stage regulation. I would have thought that with something as precise as a preamp, and with lm317 and lm337 being so cheap that dual regulation would be a must. There must be a reason its just i see it.
As usual answers on a postcard to...
Or rather if you would like to comment i would be grateful.
Regards,
Phil
Used dual regs in a NAIM pre...
...And it seemed to make an audible improvement, particularly in the highs, but then I think that the AD815 has much better power supply noise rejection ratio's than the NAIM pre I used (NAC 62)
...And it seemed to make an audible improvement, particularly in the highs, but then I think that the AD815 has much better power supply noise rejection ratio's than the NAIM pre I used (NAC 62)
Cannot comment on this
I have tried dual series 3 pin voltage regs in other circuits and did not like the results. Mine has a C-L-C-L-C filter on the output of the 317/337 regs.
This high impedance filter is much better for class A circuits. The results with the AD815 are excellent. Guess feeding a 100K load it is not pushing much current.
With swinging 2 volts max into 100K the steady state value is around 20 uA. With slewing and capacitance effects I doubt if it exceeds 1 ma even on hard transients.
George
I have tried dual series 3 pin voltage regs in other circuits and did not like the results. Mine has a C-L-C-L-C filter on the output of the 317/337 regs.
This high impedance filter is much better for class A circuits. The results with the AD815 are excellent. Guess feeding a 100K load it is not pushing much current.
With swinging 2 volts max into 100K the steady state value is around 20 uA. With slewing and capacitance effects I doubt if it exceeds 1 ma even on hard transients.
George
Sorry for my ignorance but what is a C-L-C-L-C filter?
I really dont see that just cos the AD815 has better PSU noise rejection can be reason on its own. Surely in a preamp you need the adsolute cleanest PSU you can get.
To be honest the reason i ask is cos i built the preamp buffer/PSU on Nuuk's web site a while ago. The issue is while is sounds good for my setup i need gain, so i was looking at the ad815 setup here and Rudi was kind enough to send me hs setup.
Thing is i now have an expensive dual regulated +/- 15v supply in a nice external case which i had built and am not currently using.To be honest it is actually single regulated in the case to +/- 21v and then needs regulating again down to 15v. Could i not adapt Rudis design and use the single stage of regulation on board as the final stage in my setup?
The issue being that the first PSU used about £70 worth of components, is of a very high standard and would save me a shed load trying out this design.
Any thought, or should i just implement the cicuit as described?
Phil Holder
I really dont see that just cos the AD815 has better PSU noise rejection can be reason on its own. Surely in a preamp you need the adsolute cleanest PSU you can get.
To be honest the reason i ask is cos i built the preamp buffer/PSU on Nuuk's web site a while ago. The issue is while is sounds good for my setup i need gain, so i was looking at the ad815 setup here and Rudi was kind enough to send me hs setup.
Thing is i now have an expensive dual regulated +/- 15v supply in a nice external case which i had built and am not currently using.To be honest it is actually single regulated in the case to +/- 21v and then needs regulating again down to 15v. Could i not adapt Rudis design and use the single stage of regulation on board as the final stage in my setup?
The issue being that the first PSU used about £70 worth of components, is of a very high standard and would save me a shed load trying out this design.
Any thought, or should i just implement the cicuit as described?
Phil Holder
Phil, I'm sure that you can use the dual-regulated PSU you have already.
If you adhere to the recommendations for decoupling the power supply pins of the opamp(s), it should all work well. 😉
If you adhere to the recommendations for decoupling the power supply pins of the opamp(s), it should all work well. 😉
Thanks Nuuk,
As always a great source of information, thanks a lot. How much capacitance pre the on-board second stage reg do you think would be enough per rail. Would it be really small like your buffer or would this setup need more? I was thinking about 200-330uf per rail.
Phil
As always a great source of information, thanks a lot. How much capacitance pre the on-board second stage reg do you think would be enough per rail. Would it be really small like your buffer or would this setup need more? I was thinking about 200-330uf per rail.
Phil
Good god that was a quick reply,
Thansk Nuuk,
gonna try my first peice of pcb design i think. Gonna change Rudi design to take my first stage of regualtion and i might try and change the AD815 board since i cannot get the 5 values of smt resistors and caps with buying 50 of each. Might try just widening the layout a little so i can use through hole types.
Thansk Nuuk,
gonna try my first peice of pcb design i think. Gonna change Rudi design to take my first stage of regualtion and i might try and change the AD815 board since i cannot get the 5 values of smt resistors and caps with buying 50 of each. Might try just widening the layout a little so i can use through hole types.
Re: Cannot comment on this
Yes, I have the same experience.
The sound gets slower, thicker, detail goes kaput.
Cascading linear regs does not work for me, it doesn't sound good.
Linear regs have high(ish) output impedance, specially at high frequencies, and cascading only makes the problem worse.
A snubber can make wonders, though.
But I also think that an RCRC (or RCLC) before the reg, and an L on-board on the final circuit (snubberized, bypassed, etc...) is way better than cascading regs.
Btw if you use the secondaries of a small (circuit type) trafo, you have a nice choke. 😎
Panelhead said:
Yes, I have the same experience.
The sound gets slower, thicker, detail goes kaput.
Cascading linear regs does not work for me, it doesn't sound good.
Linear regs have high(ish) output impedance, specially at high frequencies, and cascading only makes the problem worse.
A snubber can make wonders, though.
But I also think that an RCRC (or RCLC) before the reg, and an L on-board on the final circuit (snubberized, bypassed, etc...) is way better than cascading regs.
Btw if you use the secondaries of a small (circuit type) trafo, you have a nice choke. 😎
Re: Re: Cannot comment on this
Hi Carlos,
I'm assuming you by "before the reg," you mean that the filtering stage goes between the power transformer and the regulator. What is the advantage of having the RCRC (or RCLC) before the reg rather than after the reg? I would think that the having the filtering stage after the reg would allow it to smooth out the artifacts that my be produced before the power gets fed to the circuit.
Of course, I know little about the math and theories behind this, so I'm just curious.
Thanks,
KT
carlosfm said:
Hi Carlos,
I'm assuming you by "before the reg," you mean that the filtering stage goes between the power transformer and the regulator. What is the advantage of having the RCRC (or RCLC) before the reg rather than after the reg? I would think that the having the filtering stage after the reg would allow it to smooth out the artifacts that my be produced before the power gets fed to the circuit.
Of course, I know little about the math and theories behind this, so I'm just curious.
Thanks,
KT
KT said:
The first R after the diodes attenuates overshoot.
The RCRC (or RCLC) before the reg lowers PSU ripple, of course... before the reg.
This implementation also avoids large current pulses charging the big caps.
An extra (smaller) RCLC on the preamp board will remove any noise even further, including any noise picked up by the psu wiring (yes, even if you cross the wires).
This may look extreme (and in fact it is, for an op-amp used as a line preamp, but not for some discrete circuits and/or phono stages), but it works better than cascading regulators.
IME.
Nother version
I built one this week leaving off the 330 ohm resistor running from the inverting input to ground. a buffer so to speak. It is working fine.
Previous one used 480 ohm resistors for feedback and for this shunt. This was an attempt to lower the circuit gain. It seems the one with gain is sounding better.
The buffer model has great bass and punch. The first one seemed a little more rounded and fleshy. I think the buffer will improve with more run in. But the first model sounded better straight out of the box.
This buffer model has much better layout and more compact placement. It should be better due to improved grounding and local decoupling. I need to rebuild the first one and see how it sounding with a little better layout.
Overall, very nice. I think that this buffer sounds about as good as any I have heard up to now. The circuit with gain is a little better still.
I would post pics, but lack software to shrink them down from 2 meg to a postable size.
George
I built one this week leaving off the 330 ohm resistor running from the inverting input to ground. a buffer so to speak. It is working fine.
Previous one used 480 ohm resistors for feedback and for this shunt. This was an attempt to lower the circuit gain. It seems the one with gain is sounding better.
The buffer model has great bass and punch. The first one seemed a little more rounded and fleshy. I think the buffer will improve with more run in. But the first model sounded better straight out of the box.
This buffer model has much better layout and more compact placement. It should be better due to improved grounding and local decoupling. I need to rebuild the first one and see how it sounding with a little better layout.
Overall, very nice. I think that this buffer sounds about as good as any I have heard up to now. The circuit with gain is a little better still.
I would post pics, but lack software to shrink them down from 2 meg to a postable size.
George
- Status
- Not open for further replies.