Can I route a DSD signal directly to the switching controller's input of a PWM amp, without any D-A conversion? On what conditions?
instead of being so pleased with yourself, do the diy thing and build something
or just listen to some music, and enjoy your life a bit more
but please, stop wasting our time with crap
Well said
OK, so that explains it - you are just trying to go for the "10.000 postings" award.
Bruno stated that the great big UcD amp is simply a bridging of smaller ones. A result is seen at the minimum load impedance figure. Bridging is for increased output for very little sonic penalty. Such penalty would occur in most of the power band, but as clipping nears the bridged amp would have advantage. Also a speaker with very low impedance in a large portion of the audio band could suffer a little bit from a bridged circuit.
1) Harm: Usually not but it depends. There is no reason to ground the heatsink that I can think of. EDIT: Well in some very sensitive circuits, you might want to avoid antenna effects from the heatsink, which could occur even through the capacitance with an insulated TO220 tab, *possibly*. I would always use an insulator between heatsink and tab, mostly for safety reasons but also to avoid any unlikely but possible odd behavior that could result from the conduction with the tab. Coupling heatsinks is done to keep the gain of complementary devices on the same thermal track, but is of no benefit for other devices like regulators. Even when the worst-case TDP of the TO220 in a given circuit doesn't demand a heatsink, longevity and improved performance are great reasons to use the heatsinks if there is room...there usually isn't. But always be aware that a big heatsink can raise the TDP of a TO220 to tens of watts from just 3-5 watts, depending on the device. I prefer TO225 packages, their design can dissipate more (only when heatsinked) than the TO220 in some devices, but the TO220 package is ubiquitous.
Last edited:
Thanks- my concern was for any residual RF, and I was going to use uninsulated sinks, but heck, an insulator and a 'lil hunk of aluminum would go a long way with how hot these get. I'll add them and report back what I see, if that 85 degrees gets down somewhere more manageable.
Quick build question...
I am deciding on how to terminate the speaker wire leads to the nCore module. My first plan is to use silver plated copper wire, and use it without tinning in order to obtain the best contact in the terminal clamp. I know the wire will creep over time allowing the terminal connection to loosen, so my plan is to check it daily for a few days, and tighten the terminal screw to take up the slack. After a couple tightenings like this, I suspect it will settle down and remain tight for a long time.
The other option is to go ahead and tin the wire, and tighten twice, but the connection with the terminal will not be nearly as good.
What do you folks think?
I am deciding on how to terminate the speaker wire leads to the nCore module. My first plan is to use silver plated copper wire, and use it without tinning in order to obtain the best contact in the terminal clamp. I know the wire will creep over time allowing the terminal connection to loosen, so my plan is to check it daily for a few days, and tighten the terminal screw to take up the slack. After a couple tightenings like this, I suspect it will settle down and remain tight for a long time.
The other option is to go ahead and tin the wire, and tighten twice, but the connection with the terminal will not be nearly as good.
What do you folks think?
Third option is to use small fork (or ring) terminals attached to the speaker cables with a crimping tool and bent 90 degrees from the joint. I think that is the way the speaker terminals are meant to be used especially with thick speaker cables.
Something like this:
http://servcat.com/catalog.asp?product=5261295
(I am not sure if the ones in the picture are correct size tho)
I ordered the crimping tool with my last hypex order, and I must say I am very happy with it: it isreally heavy duty and works like a charm:
https://www.hypexshop.com/DetailServlet?detailID=2113
very good buy for the price!
https://www.hypexshop.com/DetailServlet?detailID=2113
very good buy for the price!
Some questions to Bruno
Dear Bruno, I am new to the DIY Hifi world and read the thread attentively and then ordered the Ncore400 with the dedicated power supplies. Now I have some questions I would like to ask you.
1) Which Preamps do you recommend for the Ncores? I will for the time being use the pre out of my Musical Fidelity x-150 and ordered the Interlink ST RCA-XLR with my ncores to connect them to the balanced input I intend to use for the ncore. Do you recommend passive preamps? I use two sources, digital versus my Metrum Octave and phono versus Emmeline Nighthawk F-117 with adjustable gain settings.
2) You are very critical of Pass First Watt amplifiers, as I understand because they don't have enough power. But I wonder if the DIY First Watt B1 buffer preamp is something that might be worth looking into in your opinion. Or should I rather wait for your new preamp/dac project? Will this be OEM only or also for DIY?
3) Did you actually try and measure the Metrum Octave? Given that they use different kind of chips I don't know if the technical problems of NOS that you indicated are as pronounced as with the old chips
4) I am still looking for a case for the ncores and find the cases you used for demo purposes (like the ones pictured by Marja & Henk at 6moons) very cute. Any chance you might tell us were we can get them from?
All the best and looking forward to your answers (and the ncores of course)
Dear Bruno, I am new to the DIY Hifi world and read the thread attentively and then ordered the Ncore400 with the dedicated power supplies. Now I have some questions I would like to ask you.
1) Which Preamps do you recommend for the Ncores? I will for the time being use the pre out of my Musical Fidelity x-150 and ordered the Interlink ST RCA-XLR with my ncores to connect them to the balanced input I intend to use for the ncore. Do you recommend passive preamps? I use two sources, digital versus my Metrum Octave and phono versus Emmeline Nighthawk F-117 with adjustable gain settings.
2) You are very critical of Pass First Watt amplifiers, as I understand because they don't have enough power. But I wonder if the DIY First Watt B1 buffer preamp is something that might be worth looking into in your opinion. Or should I rather wait for your new preamp/dac project? Will this be OEM only or also for DIY?
3) Did you actually try and measure the Metrum Octave? Given that they use different kind of chips I don't know if the technical problems of NOS that you indicated are as pronounced as with the old chips
4) I am still looking for a case for the ncores and find the cases you used for demo purposes (like the ones pictured by Marja & Henk at 6moons) very cute. Any chance you might tell us were we can get them from?
All the best and looking forward to your answers (and the ncores of course)
As Bruno is probably out somewhere enjoying the sunshine after some very rainy days (I am just back from a wet weekend in Assen). let me take a first cut...
I seem to remember Bruno isn't very fond of passive preamps. I know I am concerned with them - unless you get the impedances right, they can lead to nonlinear volume control behaviour and other oddities.
I guess the main issue with non-oversampling DACs is that you either use a steep brickwall filter (with all associated problems), or don't filter and hope your amplifier and speaker can deal with the hight frequency components that are generated by the conversion process. No chipset can alter those basic premises.
I seem to remember Bruno isn't very fond of passive preamps. I know I am concerned with them - unless you get the impedances right, they can lead to nonlinear volume control behaviour and other oddities.
I guess the main issue with non-oversampling DACs is that you either use a steep brickwall filter (with all associated problems), or don't filter and hope your amplifier and speaker can deal with the hight frequency components that are generated by the conversion process. No chipset can alter those basic premises.
Implementation...
Would be key using any passive. The best approach would likely be to put the attenuators in the nCore amp, with very short leads going to the modules, this way the impedance at the output of the attenuators should not be an issue. Additionally, using series attenuators (constant input impedance) should allow one to retain a relatively high input impedance to the amp, with good resistors there is really no downside to series attenuators. One would wonder if an attenuatior directly adjacent to an nCore module could be an antenna for picking up RF though... Mr Putzeys might have some advice on that account. Additionally, anytime one is using a passive, low output impedance of the source (and high current capabilities) will be desired, of course Mr Putzeys has stated his preference that all source components would be better off if they offered low output impedance anyway...
Would be key using any passive. The best approach would likely be to put the attenuators in the nCore amp, with very short leads going to the modules, this way the impedance at the output of the attenuators should not be an issue. Additionally, using series attenuators (constant input impedance) should allow one to retain a relatively high input impedance to the amp, with good resistors there is really no downside to series attenuators. One would wonder if an attenuatior directly adjacent to an nCore module could be an antenna for picking up RF though... Mr Putzeys might have some advice on that account. Additionally, anytime one is using a passive, low output impedance of the source (and high current capabilities) will be desired, of course Mr Putzeys has stated his preference that all source components would be better off if they offered low output impedance anyway...
Well, there are transformer and autoformer based attenuators that might not have as much of an issue driving the ICs. I'd not want to put the attenuator too close to the module, 6" or so of lead length would seem to be optimal, and without connectors will not do much in terms of FR variation. The enclosure might need to be deepened in many cases for stepped attenuators... since they'll be 4 decks (balanced) or dual mono (2 decks per) with the caveat of needing to turn two knobs.
Agreed...
For me as well... Actually, better than any transformer or resistor base attenuator technically. I use digital attenuation in the ESS chip, it operates at 48 bits, so there is no data loss even if one has 32 bit material (and considering an actual audio system will not be better than 19-20 bits of actual resolution anyway), and ESS claims that any quantisation noise is -170 dB or lower, so I hardly think that could be an issue.
I know there are "purists" who still think a digital volume control has "issues", but I cannot hear any. For me the biggest "issue" is getting a good match between DAC output stage and amp input stage-with the input impedance and the gain of the nCore, a good match is easy for me. Going amp direct beat out my Ayre K5-xeMP easily.
Digital software based attenuation for the win =)
Good enough and very much the most simple solution
For me as well... Actually, better than any transformer or resistor base attenuator technically. I use digital attenuation in the ESS chip, it operates at 48 bits, so there is no data loss even if one has 32 bit material (and considering an actual audio system will not be better than 19-20 bits of actual resolution anyway), and ESS claims that any quantisation noise is -170 dB or lower, so I hardly think that could be an issue.
I know there are "purists" who still think a digital volume control has "issues", but I cannot hear any. For me the biggest "issue" is getting a good match between DAC output stage and amp input stage-with the input impedance and the gain of the nCore, a good match is easy for me. Going amp direct beat out my Ayre K5-xeMP easily.
- Status
- Not open for further replies.