Very interesting first impression. I am quite curious about the hiss. Normally I would think that one could measure the spectrum and amplitude and tune it to a level well below hearing threshold for effecient speakers.
As for the caps, tantalum's really do make the sound a bit funny. Is it physically impossible to design a direct coupled class-D amp? No DC offset allows better selection of caps.
As for the caps, tantalum's really do make the sound a bit funny. Is it physically impossible to design a direct coupled class-D amp? No DC offset allows better selection of caps.
Well some of the hiss has to do with EMI I'd imagine.
Sadly due to Lars' rendition of an input stage... being active integrators as opposed to the typical follower with gain.. I can't use the old trick of wiring the RCA signal ground to the inverting input. Bad move... really, I tried it. That leaves one to resorting to a non earthed chassis. There's still hiss but at least this way it's a clean hiss and not affected by the computer/monitor switching supplies via minor ground loops.
Of course the better alternative is to drive the sucker balanced, which I'll do next month when I get some cash to spend on wires.
The hiss is tolerable though, independant of volume, of course. I can't really try spacing the modules out more because they wouldnt' get adequate cooling then.
I have them connected to my cerwin vega's now, on the other disposable speakers of low efficiency there's little hiss to be heard, so I'm not saying it's a problem.
Someone asked me if I'd do a test at higher power levels.. damn ear bleeding I tell you. It gets a little distorted due to lack o' supply, but it seems well controlled aside from that.
I'm holding way back on the commentary so far, will give a more elaborate judgment on it once I've played with it more. You can already see it won't be too bad though.
Lars, I'm bored, I'm going to hack up your case, move things around, and throw in those BHC-T networks you gave me last year. I hope that's OK with you.
Cheers,
Chris
Sadly due to Lars' rendition of an input stage... being active integrators as opposed to the typical follower with gain.. I can't use the old trick of wiring the RCA signal ground to the inverting input. Bad move... really, I tried it. That leaves one to resorting to a non earthed chassis. There's still hiss but at least this way it's a clean hiss and not affected by the computer/monitor switching supplies via minor ground loops.
Of course the better alternative is to drive the sucker balanced, which I'll do next month when I get some cash to spend on wires.
The hiss is tolerable though, independant of volume, of course. I can't really try spacing the modules out more because they wouldnt' get adequate cooling then.
I have them connected to my cerwin vega's now, on the other disposable speakers of low efficiency there's little hiss to be heard, so I'm not saying it's a problem.
Someone asked me if I'd do a test at higher power levels.. damn ear bleeding I tell you. It gets a little distorted due to lack o' supply, but it seems well controlled aside from that.
I'm holding way back on the commentary so far, will give a more elaborate judgment on it once I've played with it more. You can already see it won't be too bad though.
Lars, I'm bored, I'm going to hack up your case, move things around, and throw in those BHC-T networks you gave me last year. I hope that's OK with you.
Cheers,
Chris
It would really help to make things less complicated if Lars just sent a box that is considered a "typical" configuration rather than something that requires modification like chris is doing. The typical configuration would be what is exected for most DIYers to start out with.
Chris,
Is the chassis not closed? Normally the chassis shields the circuits from EMI, especially if it is grounded.
Chris,
Is the chassis not closed? Normally the chassis shields the circuits from EMI, especially if it is grounded.
soongsc said:
Hi,
Ah, I've scared you. What I have here is your most basic implementation of it. There's really not much you can do to it in order to improve on the module itself unless you're simply hellbent on tailering to your very particular liking, in which case, good luck!! Other than swamping op amps which is easy, I don't think there's much worth doing to it.
Adding additional PSU caps isn't at all unreasonable considering the given PSU has only 10 000uF per rail between two channels and with 4 ohm loads.
Replacing the caps that I did was done out of necessity, to get the job done for now. I've asked Lars to send me his chosen replacements in order to test what everyone will be getting, and of course have two modules that sound the same. Call it a repair as opposed to a modification.
The other caps I asked for are needed in order to test it in post filter feedback mode, for my own curiosity. That as well is an option all the modules will have and so can hardly be considered a modification.
It's not my intention to totally customize the thing and make it completely different from what everyone will be getting, that obviously defeats the purpose as my coments on it would hold no relevance.
So I've started out driving it singled ended with everything purely stock other than the two caps I had to change. I added a mains earth connection to the case, which didn't work very well while driving it single ended. I tried the typical connection of having the signal ground to the inverting input for driving it single ended as opposed to connecting it at signal ground.
This was done against Lars' wish, I guess I wasn't fully satisfied with his explanation as to why it shouldnt' be done, and I'm the type that needs to see to believe, so I tried it anyway. While his explanation was lacked somewhat for my liking, he was of course right that it can't handle it well. I later found the input stage isn't your typical follower with gain, but an active differential integrator which I think can explain why you can't connect it that way.
So I changed it back and went with the last option which was to remove the mains earth from the chassis and that helped somewhat.
I might try some bypass caps from the connectors to the case, and your typical X2 bypass on the mains. Aside from this I won't worry about further improving that until I'm driving it balanced.
These grounding issues are no different from what anyone might be faced with. I'd consider this a typical aspect of setting it up that most anyone is likely to endure.
BTW hiss is for the most part inaudible with lesser efficient speakers, and not too bad with mine.
Thus far Lars' new module sounds very good, for just having slapped it together, which I think is the way it is meant to be used. The sound of it thus far is prompting me to implement it better.. consider this a good sign. That doesn't mean modify. Drive it balanced, add a set of caps to the supply because my speakers are hungry, maybe get a less anaemic transformer for it later on, and a few other piddly little things, such as ensuring all the PSU wires are of the same gauge.. (Lars must have been running short when he assembled it).
I hope I've managed to put aside your fears, but if not and you insist Lars send me two more modules... siiigh.. ok🙄
Cheers,
Chris
Looking at the schematics at the NCD site, it seems like it should be driven differentially. If we want to drive it using common single ended output, probably a single to differential resistor circuit is necessary?
Now that I look at the schematics at the NCD site again, the thing looks like it's using feed forward, or is the polarity of the OPAMP mistyped?
Now that I look at the schematics at the NCD site again, the thing looks like it's using feed forward, or is the polarity of the OPAMP mistyped?
Is the NCD an inverting power amp or non-inverting power amp? Asking because I think absolute polarity is important.
classd4sure said:
An amp that has to have the mains chassis earth lifted to work properly? That is, frankly, both ridiculous and irresponsible.
Please reassure me that this isn't the case.
You can earth it of course, but then you should also connect the enclosure to proper GND. Otherwise you will have a noise injection bridge between the enclosure and the circuitry.
You are absolutely right, it perhaps would be ... The earthing issue is different depending on which country you are living in. In some countries (like i.e. the UK) everything is earthed together, (even the GND wire of your oscilloscope probe - which can lead to some rather dramatic events). In other countries it is against safety regulations to earth some applicances.
This particular apparatus is meant to operate under EN60065 safety class 2 (consumer equipment). And thus should (according to the above safety regulation) NOT be earthed.
If you want to earth it anyway, fine, but then you should just add the necessary GND connections to the enclosure. An Earthing resistor is actually present on the PSU unit for eactly this purpose. (Chris: R105). When this resistor is in place, the apparatus is prepared to earthing of the enclosure.
You are absolutely right, it perhaps would be ... The earthing issue is different depending on which country you are living in. In some countries (like i.e. the UK) everything is earthed together, (even the GND wire of your oscilloscope probe - which can lead to some rather dramatic events). In other countries it is against safety regulations to earth some applicances.
This particular apparatus is meant to operate under EN60065 safety class 2 (consumer equipment). And thus should (according to the above safety regulation) NOT be earthed.
If you want to earth it anyway, fine, but then you should just add the necessary GND connections to the enclosure. An Earthing resistor is actually present on the PSU unit for eactly this purpose. (Chris: R105). When this resistor is in place, the apparatus is prepared to earthing of the enclosure.
I thought EN60065 only applied to complete class II double screened assemblies with no user accessible parts?
Exactly, the mains parts, and transformer are approved and setup for safety Class 2 (Double Insulated) operation.
You can change that of course, by adding the Earth connection, but then it's no longer a Class 2 apparatus. Even though the mains parts still have double insulation, it becomes a single insulation Class 1 apparatus, and thus have to comply with a different set of rules, with regards to connectivity and current capability of the Earth connection, accessibility of connected parts etc.
That is why i as a producer cannot uncritically connect things to earth, GND etc at free will. I have to follow the rules strictly, both in my hand built prototypes, and also in my instructions, regarding mains operated equipment.
But the DIY'er can of course do as he like, as long as he makes sure what he does is safe, and operationally viable.
You can change that of course, by adding the Earth connection, but then it's no longer a Class 2 apparatus. Even though the mains parts still have double insulation, it becomes a single insulation Class 1 apparatus, and thus have to comply with a different set of rules, with regards to connectivity and current capability of the Earth connection, accessibility of connected parts etc.
That is why i as a producer cannot uncritically connect things to earth, GND etc at free will. I have to follow the rules strictly, both in my hand built prototypes, and also in my instructions, regarding mains operated equipment.
But the DIY'er can of course do as he like, as long as he makes sure what he does is safe, and operationally viable.
eleson, about transformers.From what I have read on some Swedish manufacturers sites, manufacturing of stock items are
not done in Sweden, but special items are still manufactured here.
And the transformer we wanted , isn´t a stock item.
(I think you did an incorrect translation).
It may be of importance to get these non-stock items quick, and
transformers are not the best you can send by airmail.
not done in Sweden, but special items are still manufactured here.
And the transformer we wanted , isn´t a stock item.
(I think you did an incorrect translation).
It may be of importance to get these non-stock items quick, and
transformers are not the best you can send by airmail.
Golden Mean: You are right, could send 100kg worth of transformers by UPS, and the rest by ship, then you get the same transformer first and later.
The golden mean said:
I think his translation was correct, but he quoted only part of the info. They also say:
"I Växjö tillverkar vi prototyper och mindre serier med mycket korta leveranstider. Här tillverkas också stora transformatorer, upp till 10kVA."
or in english
"In Växjö [ie. in Sweden] we manufacture prototypes and smaller series with very short delivery times. Here we also manufacture large transformers, up to 10 kVA".
Of courser, I would expect that "smaller series with very short delivery time" comes with a quite different price tag than stock transformers. 😉
I think he translated serial production with "manufacturing".
As manufacturing can be aimed at prototyping , short series or
production of stock items, an error occured. But it doesn´t matter much.🙂
As manufacturing can be aimed at prototyping , short series or
production of stock items, an error occured. But it doesn´t matter much.🙂
Lars Clausen said:
The PSU earth point is already connected to the enclosure through the ground lift resistor (I've located them all). I'm not really a fan of ground lift resistors actually.
Yet the enclosure itself was not earthed, which is why I did earth it. Even with all these ground lift resistors in place it is noisier with the mains earth connected.
However my computer is floating which gives its smps ground a high impedance node, and all the switching noise then gets fed through the shield or signal ground to the amp, which is the noise it picks up with the chassis earthed.
If I ground the computer there's far uglier problems, because it in fact is a component that works better without an earth, or a less than ideal one.
Hey, no offense meant Lars, I was just trying to clarify things. I appreciate the difficulties of trying to get anything through certification, and even more the difficulties of getting such a curmudgeonly lot of diyers that we have here to do the right thing.
Instructions? Pah. We don't need no stinking instructions! 🙂
Instructions? Pah. We don't need no stinking instructions! 🙂
Let's say I short all these ground lift resistors (including the 2R2 in the signal ground)..
Doesn't it stand to reason all the noise picked up on the signal earth would then be simply diverted to earth via the chassis instead of creating an error voltage over the ground lift resistors..
Doesn't it stand to reason all the noise picked up on the signal earth would then be simply diverted to earth via the chassis instead of creating an error voltage over the ground lift resistors..
Chris: The amp won't mind, but you would be inviting all kinds of problems with GND loops ( = humming)
- Status
- Not open for further replies.