Haha good one. Probably a better way to improve the sound of NS10s is to go into another room. Terrible speakers - popular because if you can make a recording sound good through them it will sound good on anything
Only simulations but the results are well beyond the simulators
error margins.
The difference is quite large not to say of orders of magnitude
while the added complexity and cost is negligible.
Attachments
Last edited:
Hi Guys
There are several articles about the NS-10M Studio, which was an update of the hifi NS10. The original units came with grille cloth but the later Studio version has a very open steel mesh over the tweeter. The response amendment provided by the grille cloth was lost, so some studio personnel placed tissue over the tweeter. This was completely unnecessary as the Studio variant already took the lack of grille cloth into account.
The actual response of the NS-10M Studio is quite amazing! It takes advantage of the boundary presented by the mixing console to help with bass response. The off-axis dispersion is quite low as the near-field intended use places the listener on-axis in most situations.
I use these speakers with some subs in a "2.2" system for listening to music and watching movies. It sounds far better than any surround system in a home - those are optimised for the nonoptimal theatre enviroment.
Regarding preamps: measurable accuracy is a fine thing and an easy objective to attain. Where Doug and others concentrate on THD, you rarely if ever see an IM plot. IM makes or breaks the sound even when an order of magnitude lower than THD. Since it is impossible to have feedback without the loop itself generating new distortion artifacts, the distortion spectrum of the circuit changes with feedback applied. Every circuit actually has some feedback mechanism within it, so distortions of some type are unavoidable but can be minimised and their balance controlled to some extent.
Doug's circuits in particular are designed to use off-the-shelf parts with no matching. Circuit design controls sensitivity to parameter spreads. This is part of why he glosses over complimentary-diff inputs and push-pull VAS. In both cases, matched components are required to assure equal gains for the circuit halves, and matching introduces extra cost. On the other hand, the balanced circuitry greatly reduces even-order distortion and provides a 'potentially' smoother distortion profile provided other facets of the design are cared for. To reduce sensitivity to device matching requires emitter degeneration that adds noise - linearity can be equal with or without degeneration despite the usual graphs you see to the contrary.
Opamps have their appeal and their limitations. Discrete circuitry has the advantage that higher voltages can be used along with asymmetrical supplies. The latter is almost universally over-looked these days, with the influence of opamp-based design.
Have fun
Kevin O'Connor
londonpower.com
There are several articles about the NS-10M Studio, which was an update of the hifi NS10. The original units came with grille cloth but the later Studio version has a very open steel mesh over the tweeter. The response amendment provided by the grille cloth was lost, so some studio personnel placed tissue over the tweeter. This was completely unnecessary as the Studio variant already took the lack of grille cloth into account.
The actual response of the NS-10M Studio is quite amazing! It takes advantage of the boundary presented by the mixing console to help with bass response. The off-axis dispersion is quite low as the near-field intended use places the listener on-axis in most situations.
I use these speakers with some subs in a "2.2" system for listening to music and watching movies. It sounds far better than any surround system in a home - those are optimised for the nonoptimal theatre enviroment.
Regarding preamps: measurable accuracy is a fine thing and an easy objective to attain. Where Doug and others concentrate on THD, you rarely if ever see an IM plot. IM makes or breaks the sound even when an order of magnitude lower than THD. Since it is impossible to have feedback without the loop itself generating new distortion artifacts, the distortion spectrum of the circuit changes with feedback applied. Every circuit actually has some feedback mechanism within it, so distortions of some type are unavoidable but can be minimised and their balance controlled to some extent.
Doug's circuits in particular are designed to use off-the-shelf parts with no matching. Circuit design controls sensitivity to parameter spreads. This is part of why he glosses over complimentary-diff inputs and push-pull VAS. In both cases, matched components are required to assure equal gains for the circuit halves, and matching introduces extra cost. On the other hand, the balanced circuitry greatly reduces even-order distortion and provides a 'potentially' smoother distortion profile provided other facets of the design are cared for. To reduce sensitivity to device matching requires emitter degeneration that adds noise - linearity can be equal with or without degeneration despite the usual graphs you see to the contrary.
Opamps have their appeal and their limitations. Discrete circuitry has the advantage that higher voltages can be used along with asymmetrical supplies. The latter is almost universally over-looked these days, with the influence of opamp-based design.
Have fun
Kevin O'Connor
londonpower.com
Hi guys
Asymmetrical supplies have different voltage values, not symmetrical (identical but opposite polarity) like +/-12V, +/-15V etc. There was a Technics pre that had a +136V rail along with lower but not symmetrical split rails.
If you read the last paragraph more closely, there is no implication that asymmetrical supplies are "beneficial" to a circuit that otherwise would use symmetric supplies, rather, that certain discrete circuits optimally run from asymmetric voltages.
My point was partly that with the ubiquitous symmetric split rails, most hobbyists and designers do not think in any other term than 'symmetrical'.
Have fun
Kevin O'Connor
londonpower.com
Asymmetrical supplies have different voltage values, not symmetrical (identical but opposite polarity) like +/-12V, +/-15V etc. There was a Technics pre that had a +136V rail along with lower but not symmetrical split rails.
If you read the last paragraph more closely, there is no implication that asymmetrical supplies are "beneficial" to a circuit that otherwise would use symmetric supplies, rather, that certain discrete circuits optimally run from asymmetric voltages.
My point was partly that with the ubiquitous symmetric split rails, most hobbyists and designers do not think in any other term than 'symmetrical'.
Have fun
Kevin O'Connor
londonpower.com
Hi Guys
Most diamond buffers and drivers live within a symmetric circuit that operates best with symmetric supplies.
The LTP is a good example of a circuit that can be operated from asymmetric supplies without detriment.
Another example is the front-end to old-style amps with totem-pole outputs, either BJT or mosfet. Typically, an NPN diff amp with a low negative tail voltage and high positive voltage drives a pair of PNP BJT VASs. The latter might be loaded resistively or with current sources tied to a high negative rail. The PNP collectors drive the base/gate of their respective output device. The output devices will operate from symmetric rails of adequate value to provide the voltage swing to the speaker. Of course, the whole circuit can be turned upside down depending on the device types available for the voltages required.
In the amp example, the output stage might run from, say, +/-40V, and the front end might run off +/-50V, or from +50V and -40V, or from +50V, -40V and -12V. I've seen quite a few with three rails for the front completely unrelated to the output rails.
If you were designing a single-ended preamp and did not want to use current sources or current mirrors, it becomes beneficial to have high supplies to accommodate high R values to control gain while retaining some semblance of linearity. Assuming an NPN input device, the first stage might operate from +40V and -10V. A second stage might have an even higher positive rail and high negative rail. Just an example.
Have fun
Kevin O'Connor
londonpower.com
Most diamond buffers and drivers live within a symmetric circuit that operates best with symmetric supplies.
The LTP is a good example of a circuit that can be operated from asymmetric supplies without detriment.
Another example is the front-end to old-style amps with totem-pole outputs, either BJT or mosfet. Typically, an NPN diff amp with a low negative tail voltage and high positive voltage drives a pair of PNP BJT VASs. The latter might be loaded resistively or with current sources tied to a high negative rail. The PNP collectors drive the base/gate of their respective output device. The output devices will operate from symmetric rails of adequate value to provide the voltage swing to the speaker. Of course, the whole circuit can be turned upside down depending on the device types available for the voltages required.
In the amp example, the output stage might run from, say, +/-40V, and the front end might run off +/-50V, or from +50V and -40V, or from +50V, -40V and -12V. I've seen quite a few with three rails for the front completely unrelated to the output rails.
If you were designing a single-ended preamp and did not want to use current sources or current mirrors, it becomes beneficial to have high supplies to accommodate high R values to control gain while retaining some semblance of linearity. Assuming an NPN input device, the first stage might operate from +40V and -10V. A second stage might have an even higher positive rail and high negative rail. Just an example.
Have fun
Kevin O'Connor
londonpower.com
It's worse than that. Those who would never think to use a tone control will willingly take a pristine analog signal, A/D it in a receiver, run it through Audyssey, then D/A it, all to correct for room problems that could (probably) have been corrected for in the analog domain.
I will attest that this is untrue.
When he says that he can't disclose something, it is either because the circuit in question was designed for a client, or because he hopes to sell it to a client in the future.
As one of his clients, I have been directly involved in both of those cases, thus my assertion.
Last edited:
from an old lag
Hi, my first post here forgive me if this has been covered already
I'm probably just going to build the main preamp and MC section, my own preamp I built some decades back uses the switches on a stereo jack socket to switch out the MC section for aux inputs, I think a basic unit for home hi-fi could be made much more cheaply.
AS a true DIY-er the power supply is standard and I'd probably make my own version. The complete board set adds up to quite a bit.
I see the main pcb's are double sided, to me thats a shame because they cost arm and a leg to buy, is anyone doing a single sided pcb design? I found DIY double sided was a prob, because you can't through plate holes on the kitchen table. So each through joint has to be soldered both sides, and its not practical with stand-up capacitors and I.C holders..
None of this is a criticism of the design technology, I've been around long enough to know Douglas Self knows his stuff. I just think the system could have been made more flexible in its practical construction, and a strip down version would cost less.
I don't see any easy 'purist' alternative to the expensive cermet 1k pots though.. except my own preamp used standard carbon pots of the old style, I've never noticed any great probs over the years,, but it depends whether you're building audiophile standards, or just a boot and braces version to keep cost minimised.
If anyone could work out the working wattage of the pots it'd be useful.. real life audio power levels are significantly less than when sine wave testing.
Finally, one point of order, I see people pass judgement on the sound quality, I wonder if their system was monitored to check against parasitic oscillations. Yes I know.. circuits are modelled ...but if you get an out of spec I.C or a dry joint on a decoupling capacitor, you can get tiny oscillations at high frequencies, and in my experience those do cause sound degradation. So anyone doing sound tests and passing public opinion should really get it checked on a scope by a techie, that things are as they should be..
But its all exciting and I look forward to the build!
Hi, my first post here forgive me if this has been covered already
I'm probably just going to build the main preamp and MC section, my own preamp I built some decades back uses the switches on a stereo jack socket to switch out the MC section for aux inputs, I think a basic unit for home hi-fi could be made much more cheaply.
AS a true DIY-er the power supply is standard and I'd probably make my own version. The complete board set adds up to quite a bit.
I see the main pcb's are double sided, to me thats a shame because they cost arm and a leg to buy, is anyone doing a single sided pcb design? I found DIY double sided was a prob, because you can't through plate holes on the kitchen table. So each through joint has to be soldered both sides, and its not practical with stand-up capacitors and I.C holders..
None of this is a criticism of the design technology, I've been around long enough to know Douglas Self knows his stuff. I just think the system could have been made more flexible in its practical construction, and a strip down version would cost less.
I don't see any easy 'purist' alternative to the expensive cermet 1k pots though.. except my own preamp used standard carbon pots of the old style, I've never noticed any great probs over the years,, but it depends whether you're building audiophile standards, or just a boot and braces version to keep cost minimised.
If anyone could work out the working wattage of the pots it'd be useful.. real life audio power levels are significantly less than when sine wave testing.
Finally, one point of order, I see people pass judgement on the sound quality, I wonder if their system was monitored to check against parasitic oscillations. Yes I know.. circuits are modelled ...but if you get an out of spec I.C or a dry joint on a decoupling capacitor, you can get tiny oscillations at high frequencies, and in my experience those do cause sound degradation. So anyone doing sound tests and passing public opinion should really get it checked on a scope by a techie, that things are as they should be..
But its all exciting and I look forward to the build!
Hi and welcome to the forum
It's funny how we have a preamp in the solid state amplifier forum but here it is!
You probably realize from the early posts and exchanges in the thread that this preamp is intended as a no-compromise showpiece
of just how low distortion could be achieved with standard opamps. That meant using an unusual, low 1k circuit impedance and
uncommon, expensive controls and switchgear, commensurate with the circuit performance. The early versions (the last version was
developed in 1996, IIRC) used more conventional impedances and parts values but you'd never call them simple and inexpensive.
The '96 version's PCB was to full size rack case dimensions.
The publication of this project was sponsored by ELEKTOR magazine and I understand they had a hand in determining at least
the physical format and scope of the design before it was published.
Judging sound quality? I can't see much room for the usual implication there of a little sweet distortion, when absolute lowest
distortion is the design goal. That suggests zero sound quality to me, much as Owdeo discovered at the conclusion of his build.
It's funny how we have a preamp in the solid state amplifier forum but here it is!
You probably realize from the early posts and exchanges in the thread that this preamp is intended as a no-compromise showpiece
of just how low distortion could be achieved with standard opamps. That meant using an unusual, low 1k circuit impedance and
uncommon, expensive controls and switchgear, commensurate with the circuit performance. The early versions (the last version was
developed in 1996, IIRC) used more conventional impedances and parts values but you'd never call them simple and inexpensive.
The '96 version's PCB was to full size rack case dimensions.
The publication of this project was sponsored by ELEKTOR magazine and I understand they had a hand in determining at least
the physical format and scope of the design before it was published.
Judging sound quality? I can't see much room for the usual implication there of a little sweet distortion, when absolute lowest
distortion is the design goal. That suggests zero sound quality to me, much as Owdeo discovered at the conclusion of his build.
The only thing that I see wrong with this pre-amp design is that everything is manual operation. For a purest, then I would have to agree that manual controls can be less intrusive, but as a product for a modern day market, it lacks remote control, thus a very limited market. This whole design, could have all been done on one PCB, I hate manual wiring, if it can be avoided, in a production sense, it adds costs to the mfg process.
So all the power to you, to get up off your a$$ every time, walk over to change the volume/mute when the phone rings. Just because you do not want to use a PGA2320 or similar part, thinking that it will taint the music experience some how = rediculious.
I like a complete solution, this is why I design my own HW. For example, the small cost to add a radio function, i.e. Si4735 or Si4770, costs very little, but completes the design into a more functional unit. Having a big external box, tuner unit, is rediculous extra cost and in the end costs a lot more money. I could go on with functional ehancements, some add extra costs, some are cheap to add, other a bit more, thus they make sense. For Elektor it would have been real easy to add the radio function, since they already offered this radio design previously, "DSP-Radio". Personally I do not understand Elecktor's logic in product requirements specs, what can I say, they offer toys, not well marketed/researched products!! This is my opinion, others can voice there own & defend the product as it is.
FYI, these days a double-sided PCB in small qty is as cost effective as a single sided board, because most of the cost, is NRE costs. If you are talking large volumes then the incremental savings, do make some sense, if you want a lower quality PCB and are a nickel & dime operation.
So all the power to you, to get up off your a$$ every time, walk over to change the volume/mute when the phone rings. Just because you do not want to use a PGA2320 or similar part, thinking that it will taint the music experience some how = rediculious.
I like a complete solution, this is why I design my own HW. For example, the small cost to add a radio function, i.e. Si4735 or Si4770, costs very little, but completes the design into a more functional unit. Having a big external box, tuner unit, is rediculous extra cost and in the end costs a lot more money. I could go on with functional ehancements, some add extra costs, some are cheap to add, other a bit more, thus they make sense. For Elektor it would have been real easy to add the radio function, since they already offered this radio design previously, "DSP-Radio". Personally I do not understand Elecktor's logic in product requirements specs, what can I say, they offer toys, not well marketed/researched products!! This is my opinion, others can voice there own & defend the product as it is.
FYI, these days a double-sided PCB in small qty is as cost effective as a single sided board, because most of the cost, is NRE costs. If you are talking large volumes then the incremental savings, do make some sense, if you want a lower quality PCB and are a nickel & dime operation.
Last edited:
HI Ian
What I meant about parasitics was, they are naughty things that happen in the real world, when someone recreates the physical item. The specs for this design are superb, and yes I've had the same experience, build something, find it sounds totally flat, then realise its just missing all the added hiccups one's used to.
rsavas..
Regarding remote control, I'm not aware what they use now commercially, but many the solid state auto volume circuits would destroy the distortion spec for this kit, ..use motorised servo's to rotate the pot spindles.
and i promise not to make my own single sided boards.
What I meant about parasitics was, they are naughty things that happen in the real world, when someone recreates the physical item. The specs for this design are superb, and yes I've had the same experience, build something, find it sounds totally flat, then realise its just missing all the added hiccups one's used to.
rsavas..
Regarding remote control, I'm not aware what they use now commercially, but many the solid state auto volume circuits would destroy the distortion spec for this kit, ..use motorised servo's to rotate the pot spindles.
and i promise not to make my own single sided boards.
Absolutely, but you can take it for granted that D. Self wouldn't tender an ultimate performance DIY design like this for commercial manufacture - what would be the point, anyway? Take a look at his Cambridge 840 Azur designs for example. Dated now (2006), but they obviously must have remote controls and typical popular features of the rest of the range to be competitive in the mass market. 'Not a bad system remote control, either.
Last edited:
New Doug Self/Elektor Pre-amp
As for mechanical pots, driven by motors, well they have their issues as well and certainly are more costly to implement.
Lots of options, what can I say and every one has their advantages & dis-advanatges. Something that the buyer has to rationalize.
The Cambridge design:
Cambridge Audio Azur 840A Features
Patent pending proprietary Class XD technology.
Oversized low flux toroidal transformers
120 watts per channel (8 ohms)
200 watts per channel (4 ohms)
Completely separate pre-amp and power amp toroidal transformers and circuits make the 840A effectively a pre and power amplifier combination in one chassis
8 audio inputs - including 1 with balanced connections
Sophisticated resistor ladder and relay based attenuator for volume/balance control
The 840A features Incognito Ready TM / A-BUSTM Ready outputs, allowing multi-room capability.
Read a review finding fault in the stepping of the level, not gradual as one would find with a regular pot, everyone has an opinion.
Yet no radio/tuner, same issue as the Elektor design!!
This statement is open to controversy, so you are implying/saying that a 5532 OPA has less distortion than a PGA2320, that is debatable. Could say that a 5534 is a marginally better performer than a 5532.
As for mechanical pots, driven by motors, well they have their issues as well and certainly are more costly to implement.
Lots of options, what can I say and every one has their advantages & dis-advanatges. Something that the buyer has to rationalize.
The Cambridge design:
Cambridge Audio Azur 840A Features
Patent pending proprietary Class XD technology.
Oversized low flux toroidal transformers
120 watts per channel (8 ohms)
200 watts per channel (4 ohms)
Completely separate pre-amp and power amp toroidal transformers and circuits make the 840A effectively a pre and power amplifier combination in one chassis
8 audio inputs - including 1 with balanced connections
Sophisticated resistor ladder and relay based attenuator for volume/balance control
The 840A features Incognito Ready TM / A-BUSTM Ready outputs, allowing multi-room capability.
Read a review finding fault in the stepping of the level, not gradual as one would find with a regular pot, everyone has an opinion.
Yet no radio/tuner, same issue as the Elektor design!!
I occasionally use a Cambridge 650T with this combination: TESTED: Cambridge Azur 840E Preamplifier and 840W Power Amplifier | The Absolute Sound
'Looks like "tuner" programmed in there at 3 on the 840E pre. display, or were you looking for a Receiver?
'Looks like "tuner" programmed in there at 3 on the 840E pre. display, or were you looking for a Receiver?
I guess that I was looking at the Integrated Amplifier version.
What I see is that you can change/program the display, as to the source selection, rather than the default "Input 2". Just like it was done in The Elektor "High-End Pre-amp" that was done around 10 years ago, using a PIC & a PGA2311, used Takamatsu relays for source selection. Why Elektor can not integrate all its articles into one unit is another puzzle?
Not much has changed in 10 years except for the introduction of the Si4735/Si4770 1-chip tuners. The ease of using these parts is outstanding, why anyone would not use them in a modern day receiver/integrated amp is puzzling to me!! I am working on the Si4770 design, as we speak, almost ready to hit the fab button, curious to compare the Si4735 to Si4770 performance, lucky enough to get a sample of the mixer/Balun from Toko to complete my BOM.
What I see is that you can change/program the display, as to the source selection, rather than the default "Input 2". Just like it was done in The Elektor "High-End Pre-amp" that was done around 10 years ago, using a PIC & a PGA2311, used Takamatsu relays for source selection. Why Elektor can not integrate all its articles into one unit is another puzzle?
Not much has changed in 10 years except for the introduction of the Si4735/Si4770 1-chip tuners. The ease of using these parts is outstanding, why anyone would not use them in a modern day receiver/integrated amp is puzzling to me!! I am working on the Si4770 design, as we speak, almost ready to hit the fab button, curious to compare the Si4735 to Si4770 performance, lucky enough to get a sample of the mixer/Balun from Toko to complete my BOM.