Maxamillion
Super Member
DC Coupled RSIIb Equalizer! No Coupling Caps!
I designed and built a new output board for my RSIIb low frequency EQ unit. Since I am single amping, all the signal passes through this unit, so I want to make it as transparent as possible.
There is a previous thread where I outline all the upgrades I had already done to the EQ, but one thing still bugged me - the output coupling caps. Even though I had upgraded these from electrolytics to a cascade of Mundorf Supreme and Silver/oil, I kept thinking that they were limiting the performance (coupling caps = BAD for sound!). My tube preamp is transformer coupled, my source is DC servo coupled, and my amps have no input or output coupling caps either, so the EQ unit was bugging me!
At first I explored using a transformer instead of the caps, and I talked with Kevin at K&K audio and Per Lundahl, who suggested the LL1585 as an option. I almost pulled the trigger on this since the LL1671s I have in my tube preamp are fantastic, until I talked to Mark Gilmore of Gilmore Audio. I contacted him to ask how much DC his amps could take on their inputs (not much it turns out, <250mV) and he told me that I'd be better off designing a new output stage using opamps, to replace the stock single-ended MOSFET source follower output stage. After I got over my audiophile aversion to opamps I decided to give it a try.
So here's what I did. The new output stage (the tan board in picture #1) consists of an OPA827 single opamp on a Brown Dog adapter (talk about honing your soldering skills! tiny, tiny SMT opamp, but sweet sounding!), wired as a unity gain buffer. I installed a DIP socket on the board so that I can swap in different opamps, just in case I want to try different opamps in the circuit. Picture #2 is a shot of the underside of the board. I added quite a bit of local power supply capacitance for PS decoupling, as the stock unit runs the MOSFETs directly off the regulators. The board uses one 220uF Nichicon Muse KZ per rail (shared between channels) followed by a 10uF Black Gate N (one per rail per channel) and a 0.1uF ceramic (Y5U, one per rail per channel), all between the rail and ground. There is a 470R stopper resistor on the input and a 22R output resistor, both Texas Components TX2575. Before people start ragging me about the "boutique parts", know that all these components other than the OPA827 came from my spares box (yes I've bought nice stuff over the years!). I pulled the 51R, 22R, 470R and 100K resistors from around the MOSFET circuit to disable it - I left the MOSFETs in there as they are totally disconnected after pulling the resistors; no sense adding more work to the job!
How does it sound - in a word FANTASTIC! Crystal clear, with slamming bass and a nice mellow midrange. I listened to it for 12 hours yesterday, and I kept shaking my head in disbelief, it was that much better sounding than stock. It's also much quieter than the stock unit, which had a low level hiss when in circuit that could be heard from about a foot from the speakers; the new unit eliminates that almost completely, most likely due to the greatly improved local power supply for the output circuit. This is as big an upgrade to the overall sound as new Graz diaphragms, and in conjunction with them reveal details that I never knew were there. In fact, the system sounds BETTER with the EQ in the mix than without it! I attribute this to the very low output impedance (~40ohms) and increased current drive of the output buffer, which is able to drive my Gilmore Raptor amps (input impedance of 8kohms) without breaking a sweat - my preamp with its ~400ohms output impedance is greatly helped in this regard by the EQ. No coupling caps = transparency! DC offset was measured at <1mV per channel, so no problems there - be sure to use a low DC offset opamp like the OPA827 in this application, or you will have problems with DC offset on your amps and speakers (remember, the amp multiplies the DC going into it by its gain factor, unless it has a DC blocking cap).
I also put the EQ on the scope. Pictures 3 and 4 are 2kHz and 20kHz square waves - nice and flat, with no trace of oscillation.
Next steps? I might add some local capacitance for the other OPA827s, the ones that add the bass boost. Nothing like a nice fat reservoir of current to draw from! Just need to find the appropriate places to stick some caps from +/- to common.
When I get a chance I'll post a schematic too, in case anyone else wants to try this.
Edit: see post # 4 below re: C0G ceramics - essential for this mod.
.
I designed and built a new output board for my RSIIb low frequency EQ unit. Since I am single amping, all the signal passes through this unit, so I want to make it as transparent as possible.
There is a previous thread where I outline all the upgrades I had already done to the EQ, but one thing still bugged me - the output coupling caps. Even though I had upgraded these from electrolytics to a cascade of Mundorf Supreme and Silver/oil, I kept thinking that they were limiting the performance (coupling caps = BAD for sound!). My tube preamp is transformer coupled, my source is DC servo coupled, and my amps have no input or output coupling caps either, so the EQ unit was bugging me!
At first I explored using a transformer instead of the caps, and I talked with Kevin at K&K audio and Per Lundahl, who suggested the LL1585 as an option. I almost pulled the trigger on this since the LL1671s I have in my tube preamp are fantastic, until I talked to Mark Gilmore of Gilmore Audio. I contacted him to ask how much DC his amps could take on their inputs (not much it turns out, <250mV) and he told me that I'd be better off designing a new output stage using opamps, to replace the stock single-ended MOSFET source follower output stage. After I got over my audiophile aversion to opamps I decided to give it a try.
So here's what I did. The new output stage (the tan board in picture #1) consists of an OPA827 single opamp on a Brown Dog adapter (talk about honing your soldering skills! tiny, tiny SMT opamp, but sweet sounding!), wired as a unity gain buffer. I installed a DIP socket on the board so that I can swap in different opamps, just in case I want to try different opamps in the circuit. Picture #2 is a shot of the underside of the board. I added quite a bit of local power supply capacitance for PS decoupling, as the stock unit runs the MOSFETs directly off the regulators. The board uses one 220uF Nichicon Muse KZ per rail (shared between channels) followed by a 10uF Black Gate N (one per rail per channel) and a 0.1uF ceramic (Y5U, one per rail per channel), all between the rail and ground. There is a 470R stopper resistor on the input and a 22R output resistor, both Texas Components TX2575. Before people start ragging me about the "boutique parts", know that all these components other than the OPA827 came from my spares box (yes I've bought nice stuff over the years!). I pulled the 51R, 22R, 470R and 100K resistors from around the MOSFET circuit to disable it - I left the MOSFETs in there as they are totally disconnected after pulling the resistors; no sense adding more work to the job!
How does it sound - in a word FANTASTIC! Crystal clear, with slamming bass and a nice mellow midrange. I listened to it for 12 hours yesterday, and I kept shaking my head in disbelief, it was that much better sounding than stock. It's also much quieter than the stock unit, which had a low level hiss when in circuit that could be heard from about a foot from the speakers; the new unit eliminates that almost completely, most likely due to the greatly improved local power supply for the output circuit. This is as big an upgrade to the overall sound as new Graz diaphragms, and in conjunction with them reveal details that I never knew were there. In fact, the system sounds BETTER with the EQ in the mix than without it! I attribute this to the very low output impedance (~40ohms) and increased current drive of the output buffer, which is able to drive my Gilmore Raptor amps (input impedance of 8kohms) without breaking a sweat - my preamp with its ~400ohms output impedance is greatly helped in this regard by the EQ. No coupling caps = transparency! DC offset was measured at <1mV per channel, so no problems there - be sure to use a low DC offset opamp like the OPA827 in this application, or you will have problems with DC offset on your amps and speakers (remember, the amp multiplies the DC going into it by its gain factor, unless it has a DC blocking cap).
I also put the EQ on the scope. Pictures 3 and 4 are 2kHz and 20kHz square waves - nice and flat, with no trace of oscillation.
Next steps? I might add some local capacitance for the other OPA827s, the ones that add the bass boost. Nothing like a nice fat reservoir of current to draw from! Just need to find the appropriate places to stick some caps from +/- to common.
When I get a chance I'll post a schematic too, in case anyone else wants to try this.
Edit: see post # 4 below re: C0G ceramics - essential for this mod.
.
Attachments
Last edited: