Salas 'pacific' style RIAA
Thx Lee for the layout. I will probably go with point to point for now until I gel down what sounds best.
Salas- I guess the Salas 'pacific' style RIAA from post 110 is now out of date? I ordered all my parts for that schem already.
How does this new one in post 236 compare?
-Kent
Thx Lee for the layout. I will probably go with point to point for now until I gel down what sounds best.
Salas- I guess the Salas 'pacific' style RIAA from post 110 is now out of date? I ordered all my parts for that schem already.
How does this new one in post 236 compare?
-Kent
Salas,
You beat me to the punch, as I just got around to testing the shunt-reg. Yes it works as designed. So everyone is safe using it as-is. I was also thinking about adding the CCs for Vref and bypass.
I have deveated from your design a little bit (as I'm a tube guy at heart) and replaced the second stage with a CCs 6DJ8 and a Mosfet follower (ala NYAL and Joe Curcio). Yes you can run the 6DJ8 at this low voltage. I was running a whole RIAA tube design on 24 volt battery. I also use a pair of fets and cascode for the first stage (kinda like the pearl design) to lower the noise. When I get it all sorted out, I will post a schematic.
You beat me to the punch, as I just got around to testing the shunt-reg. Yes it works as designed. So everyone is safe using it as-is. I was also thinking about adding the CCs for Vref and bypass.
I have deveated from your design a little bit (as I'm a tube guy at heart) and replaced the second stage with a CCs 6DJ8 and a Mosfet follower (ala NYAL and Joe Curcio). Yes you can run the 6DJ8 at this low voltage. I was running a whole RIAA tube design on 24 volt battery. I also use a pair of fets and cascode for the first stage (kinda like the pearl design) to lower the noise. When I get it all sorted out, I will post a schematic.
Better 4R7 1uF RC Zobel in trying such a low value. The ESR of the big lytic prevents instability possibilities on top of having real juice on tap. Will do some traditional mock around with parts anyway. The good thing is that this PSU really helps the RIAA as intended. Very audibly so! Also the 2SK170 CCS under the Zener proved beneficial and stays.
Salas,
Tried the shunt-reg with CCS and cap bypass-loved it! Really made the RIAA sing. I added a .1uf pp to the diode bypass cap and reduced the output bypass to 100uf and .1uf pp. I think it works better with the 100uf/.1uf pp on the output.
Just for grins I hooked it up to my all tube RIAA (runs on 24V battery). WOW what a difference. If it wasn't for the fact the tubes are to noisy for my Audioquest LO MC (.23 mV), I would just use it. If I still had my Blue Point HO MC I would have never tried this RIAA. Anyone who is using this HAS TO use the shunt-reg.
Not sure if I'm going to keep the tube second stage, but did remove the Mosfet (and Mosfet haze) and used the Pass B1 buffer. That made a big difference. I may go back to the original circuit (with the B1 buffer) and just run a tube line stage. I already have a cool type 340 DHT transformer coupled OPT line stage that would work great with it.
Tried the shunt-reg with CCS and cap bypass-loved it! Really made the RIAA sing. I added a .1uf pp to the diode bypass cap and reduced the output bypass to 100uf and .1uf pp. I think it works better with the 100uf/.1uf pp on the output.
Just for grins I hooked it up to my all tube RIAA (runs on 24V battery). WOW what a difference. If it wasn't for the fact the tubes are to noisy for my Audioquest LO MC (.23 mV), I would just use it. If I still had my Blue Point HO MC I would have never tried this RIAA. Anyone who is using this HAS TO use the shunt-reg.
Not sure if I'm going to keep the tube second stage, but did remove the Mosfet (and Mosfet haze) and used the Pass B1 buffer. That made a big difference. I may go back to the original circuit (with the B1 buffer) and just run a tube line stage. I already have a cool type 340 DHT transformer coupled OPT line stage that would work great with it.
Hi Salas
It seems you did it again !!!
Do you plan to build one point to point ?
While I wait patiently by some dedicated pcb work, could you explain the way the schematic works (in Human language) so all that are reading can grasp the idea ?
Due to my lack of knowledge (I am a civil engineer and an electronics lover but I must study a lot to understand how things work), I can surelly understand a schematic by I have some difficulty just "reading" some of them..
Outstanding work.
Ricardo
It seems you did it again !!!
Do you plan to build one point to point ?
While I wait patiently by some dedicated pcb work, could you explain the way the schematic works (in Human language) so all that are reading can grasp the idea ?
Due to my lack of knowledge (I am a civil engineer and an electronics lover but I must study a lot to understand how things work), I can surelly understand a schematic by I have some difficulty just "reading" some of them..
Outstanding work.
Ricardo
Cygnus x1.
Thanks. Yes it sounds nice! Mission accomplished. We would be looking to experiment with some cap choices of different values and makes for best coherence. Thanx for your info.
RCruz.
How can I put it simpler than in my initial explanation...Hmm. OK. There is a Mosfet on top, and has a transistor in a loop. That little 5R6 sees the current that passes through the Mosfet. When it drops 0.67V (Vbe silicon standard voltage) the transistor opens and the Mosfet gate is driven to open too. This loop keeps the current constant. Vbe/Rset=0.67/5.6=120mA. 3 times higher than we may need for a 2 channel buffered Riaa. Ample.
That constant current runs through our regulator and keeps current related non linearity very low. The Zener, the second transistor, and the final Mosfet are the actual voltage regulator. We call it ''shunt'' because its elements are parallel to the voltage, not in series as in the mega majority of regulators. The Zener sits on the 0.67V Vbe of the transistor and keeps the voltage at Vzener+Vbe. To enhance this voltage reference (which is not at all a perfect part) we use a very silent NJFET in Idss CCS mode on its tail, so to lock the current through it. Also a capacitor filters enough Zener self noise. Now the transistor senses differences of Voltage and counteracts them. It has strong hfe gain so it does it well. Because it does not have high power also, we give it the Mosfet to control. That Mosfet manages the current and drives to the ground the excess that our load does not demand. The more current we use, the less hot the Mosfet becomes. The Final capacitor bypasses residual noise to the ground and is our final reservoir talking to our load. There is a contention area about its quality and value. We need some initial rectified and filtered voltage on the input of our circuit that must be higher than the voltage drop of our first Mosfet. 6-8V more VDCin than our target regulated output value will do.
Choose Zener Voltage, add 0.67V, here is your total output voltage for the RIAA version you are going to make. Keep your input vpltage 6-8V higher and you are done. Easy. Some Zeners are set higher than nominal, so it might work like Vz nominal+2V=Vout reg in practice. Does not matter. +/-2V PSU deviation from nominal I specify in versions its OK for Salas RIAA,
There is one PSU made point to point on operation at a friend's from yesterday and it is being evaluated. I will build more. Needs to break in and see what capacitor parts sound best for it. You see already Cygnus x1 has developed some caps preferences for the circuit he supplies for instance.
And that PSU sounds like helping a lot. This RIAA needed its PSU partner in the same simple concept. So it got it.
All those PSU circuit ideas are truly classic, nothing new to write home about. Its a matter of marrying things in a good manner. Only that. You may find exquisitely performing ideas for shunt regs around the forum, notably from wrenchone and the Serbian mates.
Using far more refined stuff like dif amps, current mirrors, led references etc. I thought I must keep it simple so to complete the ''Simplistic RIAA'' concept though. Use more complex shunts by all means if you feel so. There is a nice one with excellent PCB over at the Pass Threshold NS10 thread. Great work from those guys (the usual suspects).
Thanks. Yes it sounds nice! Mission accomplished. We would be looking to experiment with some cap choices of different values and makes for best coherence. Thanx for your info.
RCruz.
How can I put it simpler than in my initial explanation...Hmm. OK. There is a Mosfet on top, and has a transistor in a loop. That little 5R6 sees the current that passes through the Mosfet. When it drops 0.67V (Vbe silicon standard voltage) the transistor opens and the Mosfet gate is driven to open too. This loop keeps the current constant. Vbe/Rset=0.67/5.6=120mA. 3 times higher than we may need for a 2 channel buffered Riaa. Ample.
That constant current runs through our regulator and keeps current related non linearity very low. The Zener, the second transistor, and the final Mosfet are the actual voltage regulator. We call it ''shunt'' because its elements are parallel to the voltage, not in series as in the mega majority of regulators. The Zener sits on the 0.67V Vbe of the transistor and keeps the voltage at Vzener+Vbe. To enhance this voltage reference (which is not at all a perfect part) we use a very silent NJFET in Idss CCS mode on its tail, so to lock the current through it. Also a capacitor filters enough Zener self noise. Now the transistor senses differences of Voltage and counteracts them. It has strong hfe gain so it does it well. Because it does not have high power also, we give it the Mosfet to control. That Mosfet manages the current and drives to the ground the excess that our load does not demand. The more current we use, the less hot the Mosfet becomes. The Final capacitor bypasses residual noise to the ground and is our final reservoir talking to our load. There is a contention area about its quality and value. We need some initial rectified and filtered voltage on the input of our circuit that must be higher than the voltage drop of our first Mosfet. 6-8V more VDCin than our target regulated output value will do.
Choose Zener Voltage, add 0.67V, here is your total output voltage for the RIAA version you are going to make. Keep your input vpltage 6-8V higher and you are done. Easy. Some Zeners are set higher than nominal, so it might work like Vz nominal+2V=Vout reg in practice. Does not matter. +/-2V PSU deviation from nominal I specify in versions its OK for Salas RIAA,
There is one PSU made point to point on operation at a friend's from yesterday and it is being evaluated. I will build more. Needs to break in and see what capacitor parts sound best for it. You see already Cygnus x1 has developed some caps preferences for the circuit he supplies for instance.
And that PSU sounds like helping a lot. This RIAA needed its PSU partner in the same simple concept. So it got it.
All those PSU circuit ideas are truly classic, nothing new to write home about. Its a matter of marrying things in a good manner. Only that. You may find exquisitely performing ideas for shunt regs around the forum, notably from wrenchone and the Serbian mates.
Using far more refined stuff like dif amps, current mirrors, led references etc. I thought I must keep it simple so to complete the ''Simplistic RIAA'' concept though. Use more complex shunts by all means if you feel so. There is a nice one with excellent PCB over at the Pass Threshold NS10 thread. Great work from those guys (the usual suspects).