Maybe the people who have problems or conflicting data with nerdman/Richard's (same guy?) measurements could contact the guy himself instead of dancing around with bavmike. Imo, bavmike is not responsible for nerdman's nor Boggit's design or measurements, so why even argue with him about it? Unless of course the goal is not to get to the truth, but dance around the subject with bavmike.
Last edited:
Yes they are the same guy. Better yet, for anyone who isn't an OEM with an NC-500 based amp in the works, why would they even care how the buffer is built? There's probably 1001 ways to build an input buffer for the NC-500. Richard's modded Hypex circuit was just one of them. Me personally have found a much better way.
Bavmike - can you please stop posting on this thread, unless you have something genuinely beneficial to add. You clearly have zero engineering experience and are frankly ruining a potentially very good thread.
I am a big fan of Hypex and Sonny's work, but you constantly posting photos of their designs or measurements (posted numerous times already), is ultimately a waste of everyone's time and is of no value, when you can't even comprehend what you are posting.
Stay on Audio Shark, where you can keep on doing the good work you have done, which is promoting and marketing Colin and Sonny's excellent work. This is a DIY forum where knowledge is shared, not where you can stake claim to op-amp rolling and repeatedly demonstrate your abundant knowledge on the nc500.
I am a big fan of Hypex and Sonny's work, but you constantly posting photos of their designs or measurements (posted numerous times already), is ultimately a waste of everyone's time and is of no value, when you can't even comprehend what you are posting.
Stay on Audio Shark, where you can keep on doing the good work you have done, which is promoting and marketing Colin and Sonny's excellent work. This is a DIY forum where knowledge is shared, not where you can stake claim to op-amp rolling and repeatedly demonstrate your abundant knowledge on the nc500.
Who has contributed anything useful on this forum about Hypex input buffers or the NC-500 beside's Richard and I? Please share some posts. I haven't seen any measurement data, or any original ideas come from anywhere else.
I am not going to post again, or I fear I will waste even more time on a pointless tirade.
I will say; this isn't a competition - calm down.
Seems like you are just blowing a pile of hot air about nothing. If it wasn't for my contribution here, there would simply just be another Acoustic imagery Atash 500 clone on the market in a different case. Besides the only reason I'm posting on this thread now is because a huge pile of my posts about the NC-500 that started before Boggit came on the scene were moved over to this thread from the Ncore thread.
bavmike you still fail to see that you have come on here and trashed the 'standard' Hypex buffer, trashed Hypex's work and trashed Colin's product by inference - all with questionable analysis. When people have said the equivalent of "that doesn't seem right to me" you've had absolutely nothing to add. And when in a corner you pop in a random pic of some other product that has nothing to do with the conversation.
You can always just stop.
You can always just stop.
Yes, a discussion by proxy is frustrating and unproductive, but unfortunately this hasn't even been a dialogue by proxy, it has been something even worse - one-way argument by proxy.
The problem is precisely that we would like to get to the truth, but that is pretty hard if one of the participants keeps posting proxy arguments to support his case, and then falls back to "I don't know, they aren't my measurements" when asked for clarifications or when the measurements are questioned.
Bavmike might not be responsible for Richard's measurements - but he is responsible for posting them here and making claims based on them. And if he does that, he should also be prepared to explain and defend them, instead of hiding behind "it's not my measurements". If they aren't his measurements, and he can't explain them, then he should not be posting them in support of his claims.
Did you forget to tell him that I didn't need Jorma or Nordost cables. 😀
Sounds a bit strange to me, considering that the cost of components are normally quite low.
Anyway, that's too bad. I guess you can't complain about lack of blind tests anymore now.
🙂
That was based on Arny's advice. I haven't really looked into suitable relays - if you are OK with automobile grade ones (the usual application for high-current relays) it would probably be reasonably cheap.
You don't need an ABX box for blind tests - only for instantaneous-switching ABX. And aren't the golden ears telling us that ABX testing is "unnatural" and misleading anyway?Anyway, that's too bad. I guess you can't complain about lack of blind tests anymore now.
So you can definitely do blind testing in sessions, with an assistant switching components between sessions.
Buffers
It would help if Nerdman (Richard) posted some comments.
As I have said before I have had no problems with noise with either the Burson, Sparkos and LM4562 in either of the Nord or Hypex boards.
It would seem that it is when people use more efficient speakers that the problem occurs.
I use the digital volume control on my Audiolab 8200 CDQ but if I switch to analogue there is something there but not at normal listening volumes.
It would help if Nerdman (Richard) posted some comments.
As I have said before I have had no problems with noise with either the Burson, Sparkos and LM4562 in either of the Nord or Hypex boards.
It would seem that it is when people use more efficient speakers that the problem occurs.
I use the digital volume control on my Audiolab 8200 CDQ but if I switch to analogue there is something there but not at normal listening volumes.
That is not what happened. Richard shared his measurement of the standard Hypex board. I have absolutely nothing to do with his measurement result of the Hypex board. I even said it sounds dead silent in my system. However I haven't measured it myself with any measurement gear. And I didn't trash Colin's board either, I just gave an honest evaluation of it that there's a slight background hiss with the Sparko and LM 4562. And large hiss with the SIL 994. You have been just been blabbering nonsense for weeks, stirring up crap, and not contributing anything useful what so ever.
Last edited:
Richard is the one who posted his measurements here if you didn't notice. I just re-posted them as it seems like some people forgot about them, and are coming up with BS simulation models of a circuit neither Hypex uses in their OEM board, or Richard in his upgraded circuit design.
However if anyone else has tested the board and can debunk Richard's results, lets see the measurements.
Last edited:
Yeah it would be nice if he would elaborate on his testing procedures. As I've said many times between this thread and the Ncore thread, when I test the Hypex OEM board in my system, it sounds like the amps are completely off when I put my ear up to the speakers. The noise I get from Colin's board with Sparko or LM 4562 in place is very small. I need to put my ear to the speaker to hear it, but I still hear something, where I hear dead silence with the Hypex board. So I can imagine there's a reason for this. Most who are used to normal Class A, A/B, tube amps, or have a noisy preamp/DAC will never even think twice about this tiny amount of noise.
But this is all a moot point now anyways as Colin is aware of this, has tried Richards recommendations with positive results, and is coming out with a "revision C" board. None of which would have ever happened if it wasn't for the work I had done together with Richard testing buffer board circuits for the NC-500.
Last edited:
Indeed. It's a shame many of your posts were deleted. How time has flown by since end-September last year. To think it was early November when you first posted noise measurements of the Hypex LM4562-based input buffer. (BTW I especially loved your response to a serious technical question as to why two/dual op amps were needed.)
I only got involved in early January when I first suggested there might be something wrong with those measurements or something very wrong with the circuit implementation as the LM4562 is much better than those results show. I wasn't alone in this view. We had to wait until mid Feb for Colin to post an excerpt of a conversation with Richard. Questions regarding this tidbit went unanswered.
You then received Colin's board for testing and bounced between comparisons between this and the standard Hypex board or the modified Hypex board - it was never exactly clear which one you meant as you got less specific as time went on and so the inference was Colin's was worse than Hypex's badly measuring board, the measurement plot of which you had by then posted several times.
Of course, there was this theme of the lack of drive current of the LM4562 as well - luckily a few of those posts went missing.
Finally in mid March - more than 4 months after you first posted the measurement plot of the Hypex board which served as a model for Colin's - we get a more useful post from Nerdman/Richard. Isn't it great when we finally get to a bit of technical input? But when a couple of very basic questions are asked about this all hell breaks loose and you start claiming there's nothing in the NC500 data sheet which relates to the input buffer and comparing Hypex's recommended buffer design with recommending ketchup with hot dogs or McDonalds recommending Big Macs. At least you started to back-peddle on the buffer output current requirements.
We slowly start to eek out a possible explanation - like trying to pull blood out of a stone. Given Richard tested the Hypex board (not Colin's) it seems, given his recommendations regarding "Ce/1k", the Hypex board didn't have the anti EMI/ESD measures "omitted for clarity" from the application hints of the NC500 data sheet. Someone with the board can confirm this easily. Colin modelled his on this board, with the assistance of Hypex, and lo' and behold a closer examination of his Rev B board layout (we have not seen a full schematic as Colin has preferred not to disclose this and is well within his rights not to do so) suggests he doesn't have these measures in place either. (Maybe the hiss you hear with Colin's board is RF coming from your source which you don't hear on the modified Hypex board because of the EMI countermeasures put in place by Richard - or was it the unmodified Hypex board that you were referring to? Colmo doesn't hear it because his setup isn't delivering such noise to the Nord amp in the first place...?)
So Rev C can include Richard's suggestion or the one I included in my analysis, mine including DC blocking capacitors which have another convenient benefit of low pass filtering any noise from the '47k input resistors', or some other suitable typology. Progress at last. It took awhile and a lot of blabbering in between.
Still outstanding is the debate with respect to the '47k input resistors'. I invited comment on my analysis which suggests they are not at all the dominant source of noise and certainly not the source of the "hiss". I'm still waiting. You have none. Richard hasn't responded either. Rather it seems the dominant source of noise might have been noise ingress into the unshorted(?) inputs while measuring which of course wasn't filtered away by EMI countermeasures. Plus of course the noise of the resistors in the signal path but they have been a constant across all variants I believe.
Richard reintroduced the issue of the required buffer output current capability - the inference being that any implementation involving the LM4562 would be significantly deficient (including Rec C). This was challenged and also remains outstanding. I'm sure people would like this cleared up once and for all (including those owners of NC1200-based amps which use the 'recommended' LM4562-based buffer design).
Next there's the question of whether the feedback network needs to be 'detuned' with Cfb. Richard seems to think so in respect of the Sonic Imagery op amp and that it's good practice generally. Maybe. 270p is a relatively high value. Self used 27p in his discussion of balanced inputs with LM5532 op amps in "Small Signal Audio Design".
Tests I have done with a LME49740-based balanced input (the LME49740 is a quad LM4562) suggest that they're not needed at all. And normally one would place an identical cap across the other resistor making up the feedback network - R11 in Richard's schematic - to quote Self again "to maintain the balance of the amplifier, and hence its CMRR, at high frequencies." Maybe Richard can comment here also. At any rate it's quite likely that such fine-tuning for stability is very much op amp dependent. (I wonder what exactly Hypex implemented on the NC1200 modules.)
You did post one set of measurements which I believe had all the changes to the Hypex board suggested by Richard done - with the LM4562 in place - and another comparable measurement with the SI op amps (of course you were at pains initially to suggest they were class A discrete op amps designed for you - "my discrete pure class A buffer") replacing the LM4562. Given the very good noise and distortion performance of the LM4562 it's interesting to see such a big difference. Perhaps Richard can clarify much more so here again just exactly what tests he did and the test conditions. BTW is "my discrete pure class A buffer" just the (modified) Hypex recommended design with the op amps changed for the SIL994? It's just that so far I have seen (a) Hypex board unmodified (b) Hypex board with Richard's modifications (c) Hypex board with Richard's modifications and SIL994 op amps in lieu of LM4562and (d) Hypex board presumably with Richard's modifications and "my discrete pure class A buffer"/ "the new discrete class A buffer" - these last two charts, c and d, differing from each other...
Four months...
If Colin is going to spend some time on a revision to the board there's a few other things worth looking at as well. One example, mentioned very early on by Barrows is bypass caps at the supply inputs of the op amp. Good practice to prevent instability. There are others, of course, and it is up to Colin to consider them.
Let's get some of these technical questions answered and share the knowledge. That's what this forum is meant to be about...
I only got involved in early January when I first suggested there might be something wrong with those measurements or something very wrong with the circuit implementation as the LM4562 is much better than those results show. I wasn't alone in this view. We had to wait until mid Feb for Colin to post an excerpt of a conversation with Richard. Questions regarding this tidbit went unanswered.
You then received Colin's board for testing and bounced between comparisons between this and the standard Hypex board or the modified Hypex board - it was never exactly clear which one you meant as you got less specific as time went on and so the inference was Colin's was worse than Hypex's badly measuring board, the measurement plot of which you had by then posted several times.
Of course, there was this theme of the lack of drive current of the LM4562 as well - luckily a few of those posts went missing.
Finally in mid March - more than 4 months after you first posted the measurement plot of the Hypex board which served as a model for Colin's - we get a more useful post from Nerdman/Richard. Isn't it great when we finally get to a bit of technical input? But when a couple of very basic questions are asked about this all hell breaks loose and you start claiming there's nothing in the NC500 data sheet which relates to the input buffer and comparing Hypex's recommended buffer design with recommending ketchup with hot dogs or McDonalds recommending Big Macs. At least you started to back-peddle on the buffer output current requirements.
We slowly start to eek out a possible explanation - like trying to pull blood out of a stone. Given Richard tested the Hypex board (not Colin's) it seems, given his recommendations regarding "Ce/1k", the Hypex board didn't have the anti EMI/ESD measures "omitted for clarity" from the application hints of the NC500 data sheet. Someone with the board can confirm this easily. Colin modelled his on this board, with the assistance of Hypex, and lo' and behold a closer examination of his Rev B board layout (we have not seen a full schematic as Colin has preferred not to disclose this and is well within his rights not to do so) suggests he doesn't have these measures in place either. (Maybe the hiss you hear with Colin's board is RF coming from your source which you don't hear on the modified Hypex board because of the EMI countermeasures put in place by Richard - or was it the unmodified Hypex board that you were referring to? Colmo doesn't hear it because his setup isn't delivering such noise to the Nord amp in the first place...?)
So Rev C can include Richard's suggestion or the one I included in my analysis, mine including DC blocking capacitors which have another convenient benefit of low pass filtering any noise from the '47k input resistors', or some other suitable typology. Progress at last. It took awhile and a lot of blabbering in between.
Still outstanding is the debate with respect to the '47k input resistors'. I invited comment on my analysis which suggests they are not at all the dominant source of noise and certainly not the source of the "hiss". I'm still waiting. You have none. Richard hasn't responded either. Rather it seems the dominant source of noise might have been noise ingress into the unshorted(?) inputs while measuring which of course wasn't filtered away by EMI countermeasures. Plus of course the noise of the resistors in the signal path but they have been a constant across all variants I believe.
Richard reintroduced the issue of the required buffer output current capability - the inference being that any implementation involving the LM4562 would be significantly deficient (including Rec C). This was challenged and also remains outstanding. I'm sure people would like this cleared up once and for all (including those owners of NC1200-based amps which use the 'recommended' LM4562-based buffer design).
Next there's the question of whether the feedback network needs to be 'detuned' with Cfb. Richard seems to think so in respect of the Sonic Imagery op amp and that it's good practice generally. Maybe. 270p is a relatively high value. Self used 27p in his discussion of balanced inputs with LM5532 op amps in "Small Signal Audio Design".
Tests I have done with a LME49740-based balanced input (the LME49740 is a quad LM4562) suggest that they're not needed at all. And normally one would place an identical cap across the other resistor making up the feedback network - R11 in Richard's schematic - to quote Self again "to maintain the balance of the amplifier, and hence its CMRR, at high frequencies." Maybe Richard can comment here also. At any rate it's quite likely that such fine-tuning for stability is very much op amp dependent. (I wonder what exactly Hypex implemented on the NC1200 modules.)
You did post one set of measurements which I believe had all the changes to the Hypex board suggested by Richard done - with the LM4562 in place - and another comparable measurement with the SI op amps (of course you were at pains initially to suggest they were class A discrete op amps designed for you - "my discrete pure class A buffer") replacing the LM4562. Given the very good noise and distortion performance of the LM4562 it's interesting to see such a big difference. Perhaps Richard can clarify much more so here again just exactly what tests he did and the test conditions. BTW is "my discrete pure class A buffer" just the (modified) Hypex recommended design with the op amps changed for the SIL994? It's just that so far I have seen (a) Hypex board unmodified (b) Hypex board with Richard's modifications (c) Hypex board with Richard's modifications and SIL994 op amps in lieu of LM4562and (d) Hypex board presumably with Richard's modifications and "my discrete pure class A buffer"/ "the new discrete class A buffer" - these last two charts, c and d, differing from each other...
Four months...
If Colin is going to spend some time on a revision to the board there's a few other things worth looking at as well. One example, mentioned very early on by Barrows is bypass caps at the supply inputs of the op amp. Good practice to prevent instability. There are others, of course, and it is up to Colin to consider them.
Let's get some of these technical questions answered and share the knowledge. That's what this forum is meant to be about...
Last edited:
I'm too busy to read your post right now, but to sum it up, the NC-500 module is for OEM use only. So for anyone who isn't an OEM, if you want an amp that uses the NC-500 modules, you must choose a commercial, finished product. Any mods anyone makes to that commercial finished product are the only thing people on this forum should be concerned about as that would make it a DIY project. So unless you have an NC-500 amp in the works, or plan on modding a commercial amp that uses them, I wouldn't concern yourself with the buffer design so much.