Let's leave it up to others to rewrite history. I can imagine there was a reason posts were deleted. I'll just stay out of it because I don't really care.
Weiss makes an interesting opamp too, I haven't see any DIY-ers using it for some reason.
OP1-BP | Weiss Engineering Ltd.
OP1-BP | Weiss Engineering Ltd.
Well if you are, instead of measuring, at least able to navigate a website you might find north american suppliers....
OK, so just to summarize the discussion as we move on, my original question was about reconciling this:
with this:
So out of the possible answers I offered:
you chose the third alternative, "I'd rather not answer".
Fair enough - your prerogative. Let's move on, we are clearly done with this topic.
Hi All, Here's a solution to eliminate the "hiss" that 3lviz and bavmike mentioned earlier in this thread. bavmike had purchased some of our opamps and brought this "hiss" noise issue to our attention. A few weeks later we had a client contract with us to find a solution for them to power nearfield monitors in there new mobile broadcast trucks and I thought the Hypex stuff would work out just fine in their application. So from our bench testing here we clearly can hear the broadband noise using any number of opamps, AD797, LM4562, Bursons, Dexa, LME's from TI and not just our SIL994Enh's. The Johnson noise of the 47K and 2.2M is the dominant noise source here. First image is stock outta the box so to speak.
I have attached a plot/schematics which will make everything proper with the NC500 buffer stage. This fix works with any opamp. The values for R9, R13 and R14 are grossly out of line for a balanced input stage of this topology. This is the source of the "hiss." sound.
The 2nd plot/schematic is the total "FIX" if one is to use the buffer stage with any confidence. Changing the values for R9, R13 and R14 to what is shown will drop the noise floor dramatically and essentially fix the "hiss."
Additionally, The Hypex System uses a SMPS for powering the power amp. The PS switching noise gets into the audio path at the input. Adding "Ce" will common mode this noise out without affecting the program material and add additional RFI/EMI protection.
The Hypex Poweramp's input impedance is quite low. So the input of the poweramp isn't so sensitive to EMI/RFI. But since it is Low-Z, it requires a gain/buff stage capable of driving out at least 50-70mA for full scale poweramp output. The 994 easily meets this requirement whereas the LM4562 (also a high current out opamp relative to common types) will start to limit/clamp at 60mA i believe. This in and of itself is a very good reason to use the 994 in this application.
As the 994's GBW is >50Mhz, adding capacitors Cfb across R10 and R12 keep the opamp stable, limits it's BW to about 200kHz, keeps ultrasonics from aliasing/modulating with the poweramps switching frequency and is generally proper practice when applied correctly.
We used COG/NPO type caps but polyprop types will work just as well. 1% resistors are fine. If 0.1% are available for R9 and R14, this will improve the CMRR of the buffer stage a bit but this topology isn’t a stunning performer when it comes to CMRR. Feel free to share this info with whomever needs it.
I have attached a plot/schematics which will make everything proper with the NC500 buffer stage. This fix works with any opamp. The values for R9, R13 and R14 are grossly out of line for a balanced input stage of this topology. This is the source of the "hiss." sound.
The 2nd plot/schematic is the total "FIX" if one is to use the buffer stage with any confidence. Changing the values for R9, R13 and R14 to what is shown will drop the noise floor dramatically and essentially fix the "hiss."
Additionally, The Hypex System uses a SMPS for powering the power amp. The PS switching noise gets into the audio path at the input. Adding "Ce" will common mode this noise out without affecting the program material and add additional RFI/EMI protection.
The Hypex Poweramp's input impedance is quite low. So the input of the poweramp isn't so sensitive to EMI/RFI. But since it is Low-Z, it requires a gain/buff stage capable of driving out at least 50-70mA for full scale poweramp output. The 994 easily meets this requirement whereas the LM4562 (also a high current out opamp relative to common types) will start to limit/clamp at 60mA i believe. This in and of itself is a very good reason to use the 994 in this application.
As the 994's GBW is >50Mhz, adding capacitors Cfb across R10 and R12 keep the opamp stable, limits it's BW to about 200kHz, keeps ultrasonics from aliasing/modulating with the poweramps switching frequency and is generally proper practice when applied correctly.
We used COG/NPO type caps but polyprop types will work just as well. 1% resistors are fine. If 0.1% are available for R9 and R14, this will improve the CMRR of the buffer stage a bit but this topology isn’t a stunning performer when it comes to CMRR. Feel free to share this info with whomever needs it.
I'll have an aPX-555 soon. Won't have to send stuff away to get measured. Hopefully it's good enough for the task.
Hi All, Here's a solution to eliminate the "hiss" that 3lviz and bavmike mentioned earlier in this thread. bavmike had purchased some of our opamps and brought this "hiss" noise issue to our attention. A few weeks later we had a client contract with us to find a solution for them to power nearfield monitors in there new mobile broadcast trucks and I thought the Hypex stuff would work out just fine in their application. So from our bench testing here we clearly can hear the broadband noise using any number of opamps, AD797, LM4562, Bursons, Dexa, LME's from TI and not just our SIL994Enh's. The Johnson noise of the 47K and 2.2M is the dominant noise source here. First image is stock outta the box so to speak.
I have attached a plot/schematics which will make everything proper with the NC500 buffer stage. This fix works with any opamp. The values for R9, R13 and R14 are grossly out of line for a balanced input stage of this topology. This is the source of the "hiss." sound.
The 2nd plot/schematic is the total "FIX" if one is to use the buffer stage with any confidence. Changing the values for R9, R13 and R14 to what is shown will drop the noise floor dramatically and essentially fix the "hiss."
Additionally, The Hypex System uses a SMPS for powering the power amp. The PS switching noise gets into the audio path at the input. Adding "Ce" will common mode this noise out without affecting the program material and add additional RFI/EMI protection.
The Hypex Poweramp's input impedance is quite low. So the input of the poweramp isn't so sensitive to EMI/RFI. But since it is Low-Z, it requires a gain/buff stage capable of driving out at least 50-70mA for full scale poweramp output. The 994 easily meets this requirement whereas the LM4562 (also a high current out opamp relative to common types) will start to limit/clamp at 60mA i believe. This in and of itself is a very good reason to use the 994 in this application.
As the 994's GBW is >50Mhz, adding capacitors Cfb across R10 and R12 keep the opamp stable, limits it's BW to about 200kHz, keeps ultrasonics from aliasing/modulating with the poweramps switching frequency and is generally proper practice when applied correctly.
We used COG/NPO type caps but polyprop types will work just as well. 1% resistors are fine. If 0.1% are available for R9 and R14, this will improve the CMRR of the buffer stage a bit but this topology isn’t a stunning performer when it comes to CMRR. Feel free to share this info with whomever needs it.
Hi Richard,
Looks like you have done a bit of work since we last spoke 🙂
Just so people are clear, you didn't mention how to attach the 994 to the stock Hypex board.
Also, is that measured result of the Hypex NC500 OEM buffer combined with the NC500, when it's powered by the Hypex SMPS1200A700?
For me the stock NC 500 board is dead quiet when I listen to the output of my speaker it sounds like it's off. However I can hear a bit of noise from the Nord board with the LM 4562 in place. Perhaps Colin at Nord can use your new schematic to build a new board. I heard he ordered 25 opamps from you.
The original intention for that board was just for OEM's to test drive the NC-500's easily. So anyone serious about making a serious amp based on the NC-500's will make their own custom boards anyways. But for modders, your suggestions will be welcome I'm sure.
Last edited:
Great work! Much appreciated!!
Excellent illustration of how you don't need exotic audiophile components - you just need to understand the actual circuit. Yes, lowering the resistor values will certainly lower the Johnson (thermal) noise. Only downside is that you shouldn't try to drive that buffer with a wimpy high-impedance-output "domestic hifi" source.
Another illustration of the dangers of "op-amp rolling" unless you understand the parameters.
Should be more than good enough! OK, most of the features (and cost) of the APx555 is really on the automation side, intended for automatic tests in a production/manufacturing context, but the specs are good enough for pretty much any purpose.
Do you know of something for strictly data acquisition purposes that can measure 120dB THD+N? plus the DSIO module can measure I2S jitter very accurately as well. A new master clock and module just came out in August.
Another thing is I can imagine your testing was done with the standard IC voltage regulators. The boards I sent in had the Hypex HxR regs installed. They have much higher PSRR ratings.
Like i mentioned earlier i ordered some monolith op amps. However these opamps are ebayed, so never what you get. The labelings are ok tho. Among them was browndoged 2xAD797AR and 2xAD797BR. The AR version seemed to go unstable, lots of hiss and weird squeeking noises coming from the tweeter.
But the BR version _seems_ to be stable. Can't hear anything, when ear is on the tweeter. I find it very interesting choice for the time being. I does not have so wide sound stage yet bigger than with Burson, but it has big slam, controlled bass and easy to listen to mids.
I also got the Sparkos opamp and i find in better than Burson in this setup. I wish it has just a tad more slam tho. I am quite sensitive to harsh sibilance and hear it everywhere, even in my tv and sometimes even IRL too (my optician much have something wrong with her mouth 🙂 ), which is kinda funny. In this regard Sparkos is definitely better choice than Burson. It also has wider sound stage. But maybe if it wasn't that overly wide, there would be more more slam.
Boggit, would it make sense if you sold people some readymade dip8 extensions that has some decoupling caps installed? Lazy person could buy it from you and put under a badly behaving opamp 😉.
But the BR version _seems_ to be stable. Can't hear anything, when ear is on the tweeter. I find it very interesting choice for the time being. I does not have so wide sound stage yet bigger than with Burson, but it has big slam, controlled bass and easy to listen to mids.
I also got the Sparkos opamp and i find in better than Burson in this setup. I wish it has just a tad more slam tho. I am quite sensitive to harsh sibilance and hear it everywhere, even in my tv and sometimes even IRL too (my optician much have something wrong with her mouth 🙂 ), which is kinda funny. In this regard Sparkos is definitely better choice than Burson. It also has wider sound stage. But maybe if it wasn't that overly wide, there would be more more slam.
Boggit, would it make sense if you sold people some readymade dip8 extensions that has some decoupling caps installed? Lazy person could buy it from you and put under a badly behaving opamp 😉.
Last edited:
Please excuse my ignorance. I'm just trying to understand.
I looked at the Nord nc500 specs and it has the low input impedance of 47 k. From what I have seen it seems to me that it is generally preferable to have as high an input impedance as possible to make matching with source gear easier. The higher the input impedance of the amplifier the smaller the demand on the source output. Allowing a wider range of source gear.
First question. Is low/high input impedance a non issue or something that should always be taken into consideration.
Second question: On the UP version with the new input buffer the same input impedance of 47k is stated. Again, excuse my ignorance. But shouldn't the change in inputbuffer/opamps etc in the new input buffer change the input impedance of the amplifier?
I looked at the Nord nc500 specs and it has the low input impedance of 47 k. From what I have seen it seems to me that it is generally preferable to have as high an input impedance as possible to make matching with source gear easier. The higher the input impedance of the amplifier the smaller the demand on the source output. Allowing a wider range of source gear.
First question. Is low/high input impedance a non issue or something that should always be taken into consideration.
Second question: On the UP version with the new input buffer the same input impedance of 47k is stated. Again, excuse my ignorance. But shouldn't the change in inputbuffer/opamps etc in the new input buffer change the input impedance of the amplifier?
Op amp try-outs continues. LME48960 is definitely worth a try. I like it more than Sparkos in my setup. Smooth, but smooth only in the right places, nice sound staging and good bass and slam. ti33er get these and put them into your input buffer and be happy again 🙂.
http://www.ti.com/product/LME49860
Some promising choices are still incoming:
LME49990
OPA1612A
LT1057ACN8
OPA827AID
Current shortlist of opamps of my liking:
LME48960
OPA2132
Sparkos
Not in particular order.
http://www.ti.com/product/LME49860
Some promising choices are still incoming:
LME49990
OPA1612A
LT1057ACN8
OPA827AID
Current shortlist of opamps of my liking:
LME48960
OPA2132
Sparkos
Not in particular order.
Last edited:
So, Boggit, any thoughts about a modified Up Nord buffer, maybe dedicated for the 994 opamp? Nerdman's measurments was very interesting.
That is not really that low, unless you have some really exotic valve gear or something (or try to run one of those strange "let's move everything out to be somebody else's problem" minimalist DACs).
Higher input impedance helps accommodate more gear, but once you have a high enough impedance to cater for your gear, anything higher is not any better. So there is definitely a "good enough" level, depending on your source gear. If you go too high, Johnson noise (thermal noise that is proportional to the resistance) kicks in, as we have seen.