9 thoughts on “Accuphase C-245 preamplifier clone (BRZHIFI C-245 Pro)”
I’ve had the BRZHIFI take on the Accuphase C-245 in my system long enough now to have gone through the full arc with it: unboxing lust, mild sonic disappointment, obsessive lid-off staring, a couple of op-amp swaps, one genuinely annoying design problem, and finally a soldering session that turned it into a preamp I actually don’t want to give up. That’s the story below.
For context, this is not an Accuphase. It’s a Chinese-built homage to the C-245, a preamp Accuphase released as the successor to the long-running C-200 line. MUSMYS builds it as a finished unit — CNC’d aluminum chassis, big VU meters, remote, XLR and RCA in and out — and sells it for a fraction of what a real C-245 costs on the used market. What you’re buying is the circuit idea and a spectacular amount of hardware, not the badge.
The topology, and why it matters
The original C-245 is built around two Accuphase signatures, and the clone follows both reasonably faithfully.
The first is current feedback. Most line stages you’ve owned are voltage feedback designs: the output voltage is divided down and fed back to a high-impedance inverting input. That works, but the feedback network sees a high impedance node, and high impedance plus stray capacitance means phase rotation as frequency climbs. Accuphase instead feeds the signal back as a current into a low-impedance node, then converts it to voltage with a transimpedance stage. Because the feedback input is low impedance, phase shift at high frequencies is tiny, gain-bandwidth stays roughly independent of closed-loop gain, and you can get excellent measured performance with a modest amount of global NFB. That last part is the audible bit, in my experience — designs that hit their numbers without burying the circuit under 40dB of correction usually sound less “squeezed.”
The second is MCS — Multiple Circuit Summing. Rather than one input pair, the C-245 runs three identical amplifier circuits in parallel and sums their outputs. Signal is correlated across all three and adds coherently; noise is uncorrelated and only adds as the square root. You get a meaningful S/N improvement essentially for free, plus some distortion cancellation from device averaging. Accuphase went further and used it in a dual configuration, so the full preamp carries six discrete line amp modules — three per channel — in a fully dual-mono layout with separate supplies per channel.
So the philosophy is: discrete, modular, paralleled, dual-mono, current feedback, Class A biased line stage. It’s a 1990s Japanese engineering flex and it still reads well today.
Specs, briefly
Pure Class A line stage
Frequency response: 20Hz–20kHz (+0, −0.15dB); 3Hz–200kHz (+0, −3dB)
I’m not going to pretend those numbers tell you how it sounds. The one that actually matters in practice is input sensitivity, and I’ll come back to it, because it’s the source of the biggest problem with this preamp.
How it sounds out of the box
Honestly? Nice. Perfectly pleasant. Nothing to write home about.
Stock, it’s a competent, slightly polite line stage. The background is quiet, there’s no grain or glare, tonal balance is essentially neutral with a touch of softness up top. Nothing offends. But nothing grabs you either. Images are placed correctly but not sculpted, dynamics are fine but not startling, and the midrange — which is where I live — is a little flat and anonymous. If I’d bought this, listened for a week, and sold it, I’d have written a shrug of a review and moved on.
The good news is that I opened it first.
The inside is absurd, and I love it
This thing is overbuilt to a degree that makes no rational sense at the price. Pull the top cover and it looks like something that should cost five times as much.
Fully dual-mono amplifier modules — separate discrete line stage boards per channel, laid out symmetrically, with heat coming off them the way a Class A circuit should
A serious power supply section: large transformer, a filter bank with far more capacitance than a line stage will ever need, and a discrete regulator board with its own heatsinked pass devices
Separate relay-switched input selection board, so signal switching happens next to the connectors instead of running long leads to a front-panel switch
A motorized volume control on its own assembly, with the microcontroller and remote receiver isolated on a dedicated control board away from the audio path
A separate VU meter driver board with its own sensitivity trim pots — a detail that becomes very important later
Output muting relays so the amp never sees a turn-on thump
Board count, connector count, wiring dress, the sheer mass of the chassis — none of it is necessary. All of it is delightful. This is the kind of interior I’d happily photograph and stare at. It’s a piece of art, and it’s a big part of why I kept going instead of moving it along.
The weak link: NE5532
All that discrete Class A glory, and then the balanced input stage hands the signal to an NE5532. It’s a fine chip. It’s also, to my ears, one of the dullest sounding op-amps in common use — flat, closed-in, and utterly uninterested in texture. Every complaint I had about the stock sound traces back to that one socket.
Which is fantastic news, because it’s socketed.
Burson V7 Classic — this is the mod
Dropping in a Burson V7 Classic discrete op-amp brought the preamp to life. Not subtly. Not “if you listen carefully on the right track.” Immediately and across the entire range:
Bass gained real weight, definition, and pitch — notes now have a leading edge and a decay instead of just being present
The midrange finally has body and texture. Voices stopped being a location on a soundstage and became a person
Treble opened up with air and extension, but without a hint of the hardness that usually comes with “more detail”
Soundstage grew in every dimension, and more importantly, images inside it got dense and three-dimensional
Microdynamics — the small-scale expressive stuff — improved enormously, which is what makes the whole thing sound alive rather than accurate
With the Burson in, this became one of the best sounding preamps I’ve ever had in my system. It comprehensively outperformed both the Schiit Freya and the Freya+, which I’ve lived with and liked. Not close, not “different flavors” — better.
Burson V6 & V7 Vivids also worked great, I just proffered V7 Classic in this specific system.
I also tried a cheap X52B module, mostly out of curiosity, and even that was a significant step up from the 5532. It’s smoother, more relaxed, more forgiving. Too much so, for me — it rounds off transients and sands down the edges that give music its bite. Good value, wrong direction.
One important caveat: this only applies to the XLR inputs. The unbalanced RCA path doesn’t run through the same module, so if you’re feeding this preamp single-ended, op-amp rolling does nothing for you. Plan accordingly.
The real problem: gain
Here’s where the clone diverges from the original in a way that matters.
The genuine C-245 has tone controls. That tone control circuit sits in the signal path and, as implemented, knocks about 10dB off the overall gain of the preamp. MUSMYS deleted the tone controls — understandable, most of us wouldn’t use them — but did not compensate for the gain they were removing. The result is a preamp that is wildly oversensitive.
In practice, useful volume lived in roughly the first 30% of the control’s rotation. Everything past about 10 o’clock was academic. Fine adjustment at normal listening levels was almost impossible, because a few degrees of knob rotation was the difference between background and loud. My Holo Audio Spring 3 has no volume control of its own, so there was no way to shed gain upstream either. This wasn’t a nitpick — it was a genuine usability failure.
Fixing it
The fix is in the XLR input boards. Getting the intended signal drop proved difficult at first. Several of the modifications—including external XLR pads and altering internal feedback resistors—completely failed to create any attenuation. The preamplifier’s active instrumentation amplifier topology and self-correcting feedback loops kept nullifying those adjustments.
The Solution: I eventually found that the fix required implementing a passive voltage divider directly on the output of the Balance Input Amp Ass’y.
The Final Configuration: A precise -15 dB attenuation was achieved by physically repositioning the shunt resistor (R22) to sit after the series resistor (R21), ensuring the divider was downstream of the discrete transistor output stage. The successful setup used a 1.8 kΩ series resistor and a 390 Ω shunt to ground, which forced the necessary signal drop without interfering with the unit’s internal unity-gain buffer operation.
The difference is night and day operationally. The volume control now uses its full range, fine adjustment is easy, and the preamp simply feels finished in a way it didn’t before. Nothing was lost sonically — I still have far more gain than my system needs.
One knock-on effect: with 15dB less signal, the VU meters barely twitched. Fortunately there are sensitivity trim pots on the meter board, so it’s an internal adjustment rather than a redesign. I had to run them essentially to maximum to get meaningful needle movement again — which works, but leaves no headroom for further tweaking. Worth knowing before you go in.
Final verdict
Stock, this is a nice-sounding preamp inside a jaw-dropping chassis, held back by a $1 op-amp and a gain structure that nobody finished thinking through. If you buy one and leave it alone, you’ll get a decent, slightly dull line stage with beautiful meters, and you’ll wonder what the fuss is about.
Buy one and treat it as what it actually is — a superb hardware platform waiting for two fixes — and it becomes something else entirely. The Burson V7 Classic transforms the sound. The 10dB input attenuation transforms the usability. Together they turned this into one of the finest preamps I’ve had in my system, comfortably ahead of many preamps that I had before, and at a price that’s frankly ridiculous for what’s inside the box.
This is great feedback – I received mine 2 weeks ago. I also found the volume input to be way too high – luckily I can adjust the level out on my streamer to 500mv which makes it acceptable. I’m certainly going to be doing what you suggest with op amp. Is it just one op amp chip that needs replacing and is it in on the preamp part of the board that is missing easily accessible? Thanks again for sharing!
Unfortunately no. But it’s a simple change. Just remove R22. Replace R21 with 1.8 kΩ resistor. Right after that, at the output of R21, add another resistor (390 Ω) One leg to R21, one leg to to the ground (pins 1, 6 or 7 – either would work)
I did your resistor mod and added V7 classics. Big difference!
The chips on my boards were soldered in, but luckily the Bursons come with sockets. I’m not the best at soldering but got it done.
Well worth it.
Thanks again.
I have tried one and did notice the muffled sound, so I will try the opamp replacement.
I am not sure I can follow the instructions for the other change but also didn’t notice the sensitivity in volune that you described. Will have to re-test it again, after seeing this review.
Super useful reviews. Looking forward to seeing more!
A really interesting review and food for thought for me. I’ve just bought the 245 Pro model. I’m using it with a Fidler phono stage that has RCA and XLR out. Power amp is a Rotel rb-1552 mkii.
If I run full xlr it gets to around 85db at 2 metres with the volume at zero!
I’ve had to run preamp to amp via rca with Rothwell 20db attenuators. If I get braver I’ll try your mods.
I’ve had the BRZHIFI take on the Accuphase C-245 in my system long enough now to have gone through the full arc with it: unboxing lust, mild sonic disappointment, obsessive lid-off staring, a couple of op-amp swaps, one genuinely annoying design problem, and finally a soldering session that turned it into a preamp I actually don’t want to give up. That’s the story below.
For context, this is not an Accuphase. It’s a Chinese-built homage to the C-245, a preamp Accuphase released as the successor to the long-running C-200 line. MUSMYS builds it as a finished unit — CNC’d aluminum chassis, big VU meters, remote, XLR and RCA in and out — and sells it for a fraction of what a real C-245 costs on the used market. What you’re buying is the circuit idea and a spectacular amount of hardware, not the badge.
The topology, and why it matters
The original C-245 is built around two Accuphase signatures, and the clone follows both reasonably faithfully.
The first is current feedback. Most line stages you’ve owned are voltage feedback designs: the output voltage is divided down and fed back to a high-impedance inverting input. That works, but the feedback network sees a high impedance node, and high impedance plus stray capacitance means phase rotation as frequency climbs. Accuphase instead feeds the signal back as a current into a low-impedance node, then converts it to voltage with a transimpedance stage. Because the feedback input is low impedance, phase shift at high frequencies is tiny, gain-bandwidth stays roughly independent of closed-loop gain, and you can get excellent measured performance with a modest amount of global NFB. That last part is the audible bit, in my experience — designs that hit their numbers without burying the circuit under 40dB of correction usually sound less “squeezed.”
The second is MCS — Multiple Circuit Summing. Rather than one input pair, the C-245 runs three identical amplifier circuits in parallel and sums their outputs. Signal is correlated across all three and adds coherently; noise is uncorrelated and only adds as the square root. You get a meaningful S/N improvement essentially for free, plus some distortion cancellation from device averaging. Accuphase went further and used it in a dual configuration, so the full preamp carries six discrete line amp modules — three per channel — in a fully dual-mono layout with separate supplies per channel.
So the philosophy is: discrete, modular, paralleled, dual-mono, current feedback, Class A biased line stage. It’s a 1990s Japanese engineering flex and it still reads well today.
Specs, briefly
I’m not going to pretend those numbers tell you how it sounds. The one that actually matters in practice is input sensitivity, and I’ll come back to it, because it’s the source of the biggest problem with this preamp.
How it sounds out of the box
Honestly? Nice. Perfectly pleasant. Nothing to write home about.
Stock, it’s a competent, slightly polite line stage. The background is quiet, there’s no grain or glare, tonal balance is essentially neutral with a touch of softness up top. Nothing offends. But nothing grabs you either. Images are placed correctly but not sculpted, dynamics are fine but not startling, and the midrange — which is where I live — is a little flat and anonymous. If I’d bought this, listened for a week, and sold it, I’d have written a shrug of a review and moved on.
The good news is that I opened it first.
The inside is absurd, and I love it
This thing is overbuilt to a degree that makes no rational sense at the price. Pull the top cover and it looks like something that should cost five times as much.
Board count, connector count, wiring dress, the sheer mass of the chassis — none of it is necessary. All of it is delightful. This is the kind of interior I’d happily photograph and stare at. It’s a piece of art, and it’s a big part of why I kept going instead of moving it along.
The weak link: NE5532
All that discrete Class A glory, and then the balanced input stage hands the signal to an NE5532. It’s a fine chip. It’s also, to my ears, one of the dullest sounding op-amps in common use — flat, closed-in, and utterly uninterested in texture. Every complaint I had about the stock sound traces back to that one socket.
Which is fantastic news, because it’s socketed.
Burson V7 Classic — this is the mod
Dropping in a Burson V7 Classic discrete op-amp brought the preamp to life. Not subtly. Not “if you listen carefully on the right track.” Immediately and across the entire range:
With the Burson in, this became one of the best sounding preamps I’ve ever had in my system. It comprehensively outperformed both the Schiit Freya and the Freya+, which I’ve lived with and liked. Not close, not “different flavors” — better.
Burson V6 & V7 Vivids also worked great, I just proffered V7 Classic in this specific system.
I also tried a cheap X52B module, mostly out of curiosity, and even that was a significant step up from the 5532. It’s smoother, more relaxed, more forgiving. Too much so, for me — it rounds off transients and sands down the edges that give music its bite. Good value, wrong direction.
One important caveat: this only applies to the XLR inputs. The unbalanced RCA path doesn’t run through the same module, so if you’re feeding this preamp single-ended, op-amp rolling does nothing for you. Plan accordingly.
The real problem: gain
Here’s where the clone diverges from the original in a way that matters.
The genuine C-245 has tone controls. That tone control circuit sits in the signal path and, as implemented, knocks about 10dB off the overall gain of the preamp. MUSMYS deleted the tone controls — understandable, most of us wouldn’t use them — but did not compensate for the gain they were removing. The result is a preamp that is wildly oversensitive.
In practice, useful volume lived in roughly the first 30% of the control’s rotation. Everything past about 10 o’clock was academic. Fine adjustment at normal listening levels was almost impossible, because a few degrees of knob rotation was the difference between background and loud. My Holo Audio Spring 3 has no volume control of its own, so there was no way to shed gain upstream either. This wasn’t a nitpick — it was a genuine usability failure.
Fixing it
The fix is in the XLR input boards. Getting the intended signal drop proved difficult at first. Several of the modifications—including external XLR pads and altering internal feedback resistors—completely failed to create any attenuation. The preamplifier’s active instrumentation amplifier topology and self-correcting feedback loops kept nullifying those adjustments.
The Solution: I eventually found that the fix required implementing a passive voltage divider directly on the output of the Balance Input Amp Ass’y.
The Final Configuration: A precise -15 dB attenuation was achieved by physically repositioning the shunt resistor (R22) to sit after the series resistor (R21), ensuring the divider was downstream of the discrete transistor output stage. The successful setup used a 1.8 kΩ series resistor and a 390 Ω shunt to ground, which forced the necessary signal drop without interfering with the unit’s internal unity-gain buffer operation.
The difference is night and day operationally. The volume control now uses its full range, fine adjustment is easy, and the preamp simply feels finished in a way it didn’t before. Nothing was lost sonically — I still have far more gain than my system needs.
One knock-on effect: with 15dB less signal, the VU meters barely twitched. Fortunately there are sensitivity trim pots on the meter board, so it’s an internal adjustment rather than a redesign. I had to run them essentially to maximum to get meaningful needle movement again — which works, but leaves no headroom for further tweaking. Worth knowing before you go in.
Final verdict
Stock, this is a nice-sounding preamp inside a jaw-dropping chassis, held back by a $1 op-amp and a gain structure that nobody finished thinking through. If you buy one and leave it alone, you’ll get a decent, slightly dull line stage with beautiful meters, and you’ll wonder what the fuss is about.
Buy one and treat it as what it actually is — a superb hardware platform waiting for two fixes — and it becomes something else entirely. The Burson V7 Classic transforms the sound. The 10dB input attenuation transforms the usability. Together they turned this into one of the finest preamps I’ve had in my system, comfortably ahead of many preamps that I had before, and at a price that’s frankly ridiculous for what’s inside the box.
This is great feedback – I received mine 2 weeks ago. I also found the volume input to be way too high – luckily I can adjust the level out on my streamer to 500mv which makes it acceptable. I’m certainly going to be doing what you suggest with op amp. Is it just one op amp chip that needs replacing and is it in on the preamp part of the board that is missing easily accessible? Thanks again for sharing!
Awesome, I put together a more technical guide here:
https://tubemaze.biz/accuphase-c-245-clone-modifications-and-opamp-change/
Excellent. Thanks. Any chance of a photo of the modified board?
Unfortunately no. But it’s a simple change. Just remove R22. Replace R21 with 1.8 kΩ resistor. Right after that, at the output of R21, add another resistor (390 Ω) One leg to R21, one leg to to the ground (pins 1, 6 or 7 – either would work)
I did your resistor mod and added V7 classics. Big difference!
The chips on my boards were soldered in, but luckily the Bursons come with sockets. I’m not the best at soldering but got it done.
Well worth it.
Thanks again.
Awesome!
I have tried one and did notice the muffled sound, so I will try the opamp replacement.
I am not sure I can follow the instructions for the other change but also didn’t notice the sensitivity in volune that you described. Will have to re-test it again, after seeing this review.
Super useful reviews. Looking forward to seeing more!
A really interesting review and food for thought for me. I’ve just bought the 245 Pro model. I’m using it with a Fidler phono stage that has RCA and XLR out. Power amp is a Rotel rb-1552 mkii.
If I run full xlr it gets to around 85db at 2 metres with the volume at zero!
I’ve had to run preamp to amp via rca with Rothwell 20db attenuators. If I get braver I’ll try your mods.