@grataku
Thanks, I guess that means the 2218 is the safe bet. Kind of a pain though since I test jiged the trafos and snubbed the board accordingly. Got to take off the big heat sinks and change out the resistors. Caps will stay the same.
I would just fire it up but am a little afraid of needlessly burning up my Toshiba JFETs.
Don
Thanks, I guess that means the 2218 is the safe bet. Kind of a pain though since I test jiged the trafos and snubbed the board accordingly. Got to take off the big heat sinks and change out the resistors. Caps will stay the same.
I would just fire it up but am a little afraid of needlessly burning up my Toshiba JFETs.
Don
@grataku
Thanks, I guess that means the 2218 is the safe bet.
Nope, that's the one I bought first, the rail start at 24-0-24 and subsequently drops to 21.5-0-21.5 under load. The sound is just not that great with that low a rail, snappiness in the sound evaporates and muddiness sets in. I had a Variac with the 140 overvoltage setting so I was able to bring the primary up and set the rails back to 24-0-24 and test this situation.
I will be putting that 2218 up for sale soon if you want to try it out. I also have a 3222. That goes to 32-0-32V under zero load and gets down to 28-0-28 under load. Still some crazy no load overvoltages.
The PS is by far the most challenging part of the design. Unless a custom transformer is made with very oversized copper and iron and with low ESR I am afraid the problem will not go away with a simple RCR PS. I am sure NP carefully designs the transformers and has them custom built and he doesn't have the issue. For us commoners I think the only solution is a fully regulated board of some type like the buttah' board that can be set to 25-0-25 under load or no load. I guess one can use MicroCap 12 (now free) to simulate the whole system and figure out exactly what transformer will get the job done...But you'd have to have the transformer in hand, load it and measure the ESR to fit in the software which defeats the point. I have the impression that the antek specs are a little 'terse' and not realistic.
Thanks, I guess that means the 2218 is the safe bet.
Nope, that's the one I bought first, the rail start at 24-0-24 and subsequently drops to 21.5-0-21.5 under load. The sound is just not that great with that low a rail, snappiness in the sound evaporates and muddiness sets in. I had a Variac with the 140 overvoltage setting so I was able to bring the primary up and set the rails back to 24-0-24 and test this situation.
I will be putting that 2218 up for sale soon if you want to try it out. I also have a 3222. That goes to 32-0-32V under zero load and gets down to 28-0-28 under load. Still some crazy no load overvoltages.
The PS is by far the most challenging part of the design. Unless a custom transformer is made with very oversized copper and iron and with low ESR I am afraid the problem will not go away with a simple RCR PS. I am sure NP carefully designs the transformers and has them custom built and he doesn't have the issue. For us commoners I think the only solution is a fully regulated board of some type like the buttah' board that can be set to 25-0-25 under load or no load. I guess one can use MicroCap 12 (now free) to simulate the whole system and figure out exactly what transformer will get the job done...But you'd have to have the transformer in hand, load it and measure the ESR to fit in the software which defeats the point. I have the impression that the antek specs are a little 'terse' and not realistic.
Last edited:
I have 400 VA trafo for each M2X mono block (2 x 18V) and I have about 1V voltage drop from idle to warm-up when bias is stabilized. I have CRCLC so also a small drop over choke but that is not much.....about 62 mV.
If you go with 1000 VA trafo for each mono block you will have almost no voltage drop.
If you go with 1000 VA trafo for each mono block you will have almost no voltage drop.
Hi! I did some experimentation with Focusright Scarlett 2i2 and measured my ACA and M2X. Got some strange results is one case and would be happy if the community could comment on that.
My setup was as follows:
- I used internal generator of RAW, set at -10dBFS, 1000Hz
- that gives about 360mV on the output of 2i2 with max monitoring volume, and it's not enough to get 2.83V on the output of ACA or M2X
- so I routed the signal via Shiit Magni 3+ pre-amp on 5.5x gain, setting the volume at the level that gave 2.83V measured on the output of the amplifier (it required 820mV output from pre-amp for ACA and 535mV for M2X)
- I am still waiting for the 8ohms 100W load resistors to arrive, so I measured with the actual speakers (Overnight Sensations) connected
- 2.83V signal from the output of the amp went (in parallel to the speakers) to the line input of the 2i2
- 2i2 line input can take up to 22dBu (9.75V), so I did not use any voltage divider network
- the gain of the 2i2 line input was set to minimum
I got nice and consistent results on my ACA (see the 3 first pictures: Left channel, Right channel and the direct loopback from pre-amp to check that pre-amp did not introduce any significant noise or distortion). THD measured at 0.22% (much better than oficialy stated 0.7% distortion of ACA), and it is 2nd harmonic dominant, but I see the whole handful of other order harmonics, and mainly odd-order harmonics. Also notice that the phases of harmonics are very consistent between the channels (just 1-2 deg of difference between the channels). The hum from the mains is low and really not audible in reality.
With M2X I got mixed results that I do not fully undestand.
Left channel measured well and as expected. THD 0.022% (very close to "official" 0.025%), 3rd harmonics dominant with nice 2nd and a bit of 5th - very much as described by NP (my M2X is with the Tucson buffer now). Hum from the mains is very visible and this was also expected - neither toroid not the auto-formers are shielded yet in my M2X (waiting for the steel cover for the transformer and mu-metal for Edcors to arrive). This mains hum was actually the main reason of these measurements, as I wanted to see it before I make changes and then compare.
But on the right channel I am confused. Please see the picture (5th picture in the attachments). The overall noise floor rose by 25dBFS (from -130 to -105) hiding in effect much of the mains hum, and the harmonics look much different than on the left channel (and if you look at the phases of harmonics, they are totaly different than on the left channel - though in ACA measurement the phases matched almost perfectly between the channels).
What happens here? Why the increase in noise floor level?
I tried to change the loudspeaker used as a load, but same result. If you feed less V to the input, the noise floor on the right channel drops, but with inceasing levels of input the noise floor rises. But it does not rise on the left channel.
Does that hint to some problems in my M2X right channel? What could that be?
When listening to the music, though, I do not hear anything specific on the right channel, it plays as fine as the left to my ears.
Any ideas, please?
-Alvis
My setup was as follows:
- I used internal generator of RAW, set at -10dBFS, 1000Hz
- that gives about 360mV on the output of 2i2 with max monitoring volume, and it's not enough to get 2.83V on the output of ACA or M2X
- so I routed the signal via Shiit Magni 3+ pre-amp on 5.5x gain, setting the volume at the level that gave 2.83V measured on the output of the amplifier (it required 820mV output from pre-amp for ACA and 535mV for M2X)
- I am still waiting for the 8ohms 100W load resistors to arrive, so I measured with the actual speakers (Overnight Sensations) connected
- 2.83V signal from the output of the amp went (in parallel to the speakers) to the line input of the 2i2
- 2i2 line input can take up to 22dBu (9.75V), so I did not use any voltage divider network
- the gain of the 2i2 line input was set to minimum
I got nice and consistent results on my ACA (see the 3 first pictures: Left channel, Right channel and the direct loopback from pre-amp to check that pre-amp did not introduce any significant noise or distortion). THD measured at 0.22% (much better than oficialy stated 0.7% distortion of ACA), and it is 2nd harmonic dominant, but I see the whole handful of other order harmonics, and mainly odd-order harmonics. Also notice that the phases of harmonics are very consistent between the channels (just 1-2 deg of difference between the channels). The hum from the mains is low and really not audible in reality.
With M2X I got mixed results that I do not fully undestand.
Left channel measured well and as expected. THD 0.022% (very close to "official" 0.025%), 3rd harmonics dominant with nice 2nd and a bit of 5th - very much as described by NP (my M2X is with the Tucson buffer now). Hum from the mains is very visible and this was also expected - neither toroid not the auto-formers are shielded yet in my M2X (waiting for the steel cover for the transformer and mu-metal for Edcors to arrive). This mains hum was actually the main reason of these measurements, as I wanted to see it before I make changes and then compare.
But on the right channel I am confused. Please see the picture (5th picture in the attachments). The overall noise floor rose by 25dBFS (from -130 to -105) hiding in effect much of the mains hum, and the harmonics look much different than on the left channel (and if you look at the phases of harmonics, they are totaly different than on the left channel - though in ACA measurement the phases matched almost perfectly between the channels).
What happens here? Why the increase in noise floor level?
I tried to change the loudspeaker used as a load, but same result. If you feed less V to the input, the noise floor on the right channel drops, but with inceasing levels of input the noise floor rises. But it does not rise on the left channel.
Does that hint to some problems in my M2X right channel? What could that be?
When listening to the music, though, I do not hear anything specific on the right channel, it plays as fine as the left to my ears.
Any ideas, please?
-Alvis
Attachments
@grataku
Thanks for taking the time to go thru this. I decided to switch horses and I have boards and parts coming for the SLB. I will be back when I get that part installed in the build. This is my 3hd major change with the M2x. It's like a hobby endowment project.
Appreciate your help.Don
It's closer than mine & not far enough. The shield on the transformer should help though. If you want to compare I posted pics of mine in earlier post.
BTW with the SBL the speaker GND need to go to the PS board directly or you get a ground loop.
Another BTW: the after ~1 hr operation the transformer in that case will be the hottest part of the setup. I measured 65C in mine. Can't hold my hand on it. Heatsinks are 44C for comparison.
something's not right...
I have tucson too for the moment.Now I have to get the same setup you have and reproduce your measurements on mine. It's time.
Could you please PM me your setup details? I am a little behind the times on this. What software do you use? I have been wanting to get an 2i2 or a MOTU for sometimes anyways.
Last edited:
Hi @grataku! It would be very interesting to see your measurements too!
My setup is explained in details in the post. I could only add that I used REW software (V5.20 beta 61) on MacBook Pro running on batteries. I used TRS (stereo) to RCA adapters for input and output to 2i2, then normal RCA-RCA cables from 2i2 to pre-amp and from pre-amp to the amp, and a specially made banana-plugs to RCA cable from the amp to 2i2. You will not be able to reproduce the setup verbatim, I think, but that does not matter so much.
As for my measurements results, the left channel looks very reasonable - harmonics are like expected, I suppose, and the hum from the mains is what I am going to deal with.
But the right channel is strange. Why the noise floor could go up so much on one channel? This makes me wonder and any suggestions would be very much appreciated.
Thanks
- Alvis
My setup is explained in details in the post. I could only add that I used REW software (V5.20 beta 61) on MacBook Pro running on batteries. I used TRS (stereo) to RCA adapters for input and output to 2i2, then normal RCA-RCA cables from 2i2 to pre-amp and from pre-amp to the amp, and a specially made banana-plugs to RCA cable from the amp to 2i2. You will not be able to reproduce the setup verbatim, I think, but that does not matter so much.
As for my measurements results, the left channel looks very reasonable - harmonics are like expected, I suppose, and the hum from the mains is what I am going to deal with.
But the right channel is strange. Why the noise floor could go up so much on one channel? This makes me wonder and any suggestions would be very much appreciated.
Thanks
- Alvis
Last edited:
I have REW measurements here using 2i2 and a homemade interface box from speaker out to 2i2. The measurements are with 3-4W output into a 6R8 ohm resistor. Depending on mains quality and how much DC are on mains the 50 Hz peak goes a bit up and down. Therefor I have a heavy type DC-blocker under ways and I am also thinking about investing in a mains AC regenerator (not a conditioner).
Attachments
Thanks for your measurements, @MEPER. Looks very reasonable to me and quite similar in harmonics to my left channel. And I notice the phases of harmonics match on both channels very well (as it was on my ACA measurements also).
So what could make my measuments of the right channel of M2X so different in respect of the floor level? (I attach the files here again for easier reference).
So what could make my measuments of the right channel of M2X so different in respect of the floor level? (I attach the files here again for easier reference).
Attachments
You have a bit more averaging therefor you curves seems not as "busy". Right channel shows much more noise.
I output more power so probably why I have a bit more THD if we compare left channel.
Is it same chassis so both channels use same PSU? ....or mono blocks?
My are mono blocks.
Also interesting that you noise level decrease from 13k - 20k......like there are some filtering.....
I output more power so probably why I have a bit more THD if we compare left channel.
Is it same chassis so both channels use same PSU? ....or mono blocks?
My are mono blocks.
Also interesting that you noise level decrease from 13k - 20k......like there are some filtering.....