| Netlist |
Test-driving my new born Mono AlephX High Power Amp,
I kept an eye on the scope while playing music.
I put the scope in XY configuration, connecting the A-channel to the +out and the B-channel to the -out.
The whole monoblock is biased high enough: 750mA with 0.47ohm source resistors. The source is a cd player directly connected to the input, non-balanced.
Power supply is +/- 24V
The output drives two 8ohm speakers in parallel.
I played different cd's till finally I took Tori Amos's StrangeLittleGirls from the shelf.
I could not believe my ears and my eyes!!
The volume was rather high compared with other cd’s. I know that the Tori Amos record is superb when talking dynamics, but I wonder what the cause of this behaviour could be.
Thinking that AlephX should be able to produce them accurately, knowing that my speakers
are not the best, but not bad at all and hoping that the cd is well recorded.
The speakers got mad, the scope too, had to dramatically reduce the volume, thinking the AlephX was about to blow!
I started measuring DC and all other kind of voltages, but everything was normal.
There's no way to take a picture from the scope because dynamics are too fast and I don't have a memory scope, but I made an approximated drawing of the distortion.
Is this just normal clipping, severe clipping, is there something I overlooked in building the amp, I wonder….
Did someone else experience such behaviour, Peter…others?
/Hugo – thinks Tori Amos is a naughty girl. :clown: |
|
|
| Taco |
| With the same CD my BOZ clipped like hell, because the volumepot is at the output. My DAC has a output near 2 Volt. So maby pass doesn't like Tori :D. |
|
|
| till |
Are they recorded superb?
I ask because i never understood the reason for this strange noise right at the beginning of Boys for Pele.
Excellent dynamiks? I just listen to Norrington / Berlioz symphonie fantastique ....... |
|
|
| SY |
| Sounds like (bad pun) a classic case of compression, and up near the MSB to boot. Great for AM airplay, which is perfect for dull poseurs (poseuses?) like TA. Image is everything, and I don't mean soundstage. |
|
|
| rickpt |
I agree with SY
The last Red Hot Chili Peppers album has a maximum dynamic range of 6 Db!!!!!!!!! You can clearly listen to distortion even on the TV set
Regards |
|
|
| Edwin Dorre |
| quote: | I could not believe my ears and my eyes!!
The volume was rather high compared with other cd’s. I know that the Tori Amos record is superb when talking dynamics, but I wonder what the cause of this behaviour could be.
Thinking that AlephX should be able to produce them accurately, knowing that my speakers
are not the best, but not bad at all and hoping that the cd is well recorded.
The speakers got mad, the scope too, had to dramatically reduce the volume, thinking the AlephX was about to blow!
I started measuring DC and all other kind of voltages, but everything was normal. |
I had the same with a Michael Jackson album. I was happily testing my Aleph 4 with one eye on the scope and one on the speakers and the Aleph. Suddenly with the first song I say massive clipping. I mean every bass line clipped. I jumped up the unplug everything in 1 second...
After looking and looking... and grabbing it with cooledit it seemed the CD clips..... brrr... bad Michael! He sure did scare the hell out of me!
Edwin |
|
|
| Netlist |
| quote: | Originally posted by till
Are they recorded superb?
I ask because i never understood the reason for this strange noise right at the beginning of Boys for Pele.
Excellent dynamiks? I just listen to Norrington / Berlioz symphonie fantastique ....... |
I'm not saying this cd is a superb recording, but it definitely has superb dynamics ;) Can you give us the label of the Norrington / Berlioz symphonie fantastique?
To Taco: Glad I'm not the only one. Still want to know what's the problem here. Maybe Nelson could explain something. :cool: :)
First track of the cd is just amazing.
I just tried another one: In Requiem on Yara from Rabih Abou-Khalil (Enja record ENJ-9360 2) http://www.enjarecords.com/c9360.htm
The amp (or my cd-player, or my speakers, or..or..) shows the same strange behaviour with the cello. But here the midrange suffers from heavy crackle. I could compare the sound with the sound of a speaker with a broken voice coil.
/Hugo |
|
|
| Netlist |
OK, I'll put Tori on Cooledit tomorrow and let you know.
If someone happens to have Rabih Abou-Khalil let me know how he performs. I will test him also with Cooledit.
Still, I'm not convinced that this is the reason for what I heard.;)
Thanks for the info, till
/Hugo |
|
|
| moe29 |
SY - In what way is Tori Amos a POSER??? Maybe you should
think twice before slagging off on one of the most talented
musicians making music today. Or keep your opinion to yourself
(you don't here me saying a word about Mr. Jackson do you?)
:bigeyes:
Netlist - I refer to a lot of Tori's CD's as reference material when
evaluating equipment. I don't have Strange Little here to listen
to right now, but had not noticed anything strange about it in
the past.... It wouldn't happen to be a BMG disc or something
like that would it?
Pics of the Aleph-X????!!! |
|
|
| SY |
Yawn.
Go look up "poseur" in a decent dictionary. (edit: I see you did that and edited out your criticism of my spelling)
My kid sister used to say the same thing about Bobby Sherman. |
|
|
| faustian bargin |
um...'poser' is the low-brow way to spell 'poseur'. maybe i'm missing some humor or something.
***edit - simultaneous post with SY. sorry bout that...not that you need defending or anything. |
|
|
| moe29 |
hey i edited my post!!!! you read to fast!!! i was getting ahead
of myself |
|
|
| SY |
| You mean "too fast.";) |
|
|
| moe29 |
Please!!!! Tori Amos is talented, you may not like her music,
that i can understand, but to compare her to a teen-beat
heart throb is crazy |
|
|
| Netlist |
Alright, alright, this thread was not intended to start a discussion on musical preferences.:devily:
In this thread, I don't care if Tori is a poseur or not, I just wanted to say what I said in the first post.
That said, I noticed that when I unplugged the second speaker, which is paralleled, things behave better. The XY-axis is not a bow anymore, except at top-level volume.
I guess it has something to do with impedance.
I’ll post some pictures soon, in another thread, as those are way off topic.
/Hugo ;) ;) |
|
|
| till |
and how does it look with no speaker or a > 8 R resistor instead?
doesn´t this mean to less bias for low impedace load? |
|
|
| Netlist |
| quote: | Originally posted by till
and how does it look with no speaker or a > 8 R resistor instead?
doesn´t this mean to less bias for low impedace load? |
I'll connect my speakers in series tomorrow.
Not sure what you mean with "less bias for low impedance".
Do you mean I should lower the bias to drive low impedance? :confused:
/Hugo - bedtime now. :angel: |
|
|
| AudioFreak |
| It sounds like the bias is too low for the impedance of the load. If the output devices and heatsinks you are using can take it, I'd try increasing the bias. |
|
|
| Netlist |
Did some calculations and here are the results.
The whole calculation is for one monoblock.
As I recall, the supply voltage is +/-24V, which is 48V over the rails.
There are 16 output devices, 8 on each side. Two times 4 for the current sources and two times 4 for the outputs.
I refer to Grey’s original schematic:
http://www.diyaudio.com/forums/atta...s=&postid=32776 (Thanks Grey, Nelson)
Source resistors are 0.47ohm. Voltage over the source resistors is 0.35V. So I have 0.745mA times 8 = 5.96A lets say 6A bias.
R19 and R29 are 47k. Feedback resistors are 120k. Input resistors R18 & R28 are 10k.
DC stability resistors from common diff-pair source to outputs are 2.35k (two paralleled 4.7k).
Both absolute and differential DC are extremely low. Diff-DC can be reduced to almost 0 with V2 and absolute DC is 15-17 mV
R1/R4 and R44/R45 are not used.
No C2 and C4.
R12, R34 are 2.2k.
Sinewave 400Hz unbalanced.
Output power is now 18W RMS into 8ohm. With the 4ohm load 14W RMS.
I added another 2.2k parallel to R12, R34 and the results are a bit better:
24V * 2.2A = 52.8W RMS into 8 ohm
10V * 1.7A = 17W RMS into 4 ohm.
These values are measured just before the XY-axis starts showing distortion. That is become a bow or elipse. I can increase volume further before things get out of hand, but to me it seems the XY-axis should be flat all the way before clipping.
I got rid of the severe clipping at 8ohm and the 4ohm tests are better too.
Still not what it should be when burning that much power.
Sorry if I offended someone (sy, moe29…) yesterday. It was late and I was tired. I just wanted to keep the thread on-topic, that’s all. ;)
Here’s a nice peace about Tori for moe29: http://members.aol.com/AKlink8489/musician.htm :cool:
/Hugo |
|
|
| grataku |
| quote: | Originally posted by Netlist
Did some calculations and here are the results.
The whole calculation is for one monoblock.
As I recall, the supply voltage is +/-24V, which is 48V over the rails.
There are 16 output devices, 8 on each side. Two times 4 for the current sources and two times 4 for the outputs.
I refer to Grey’s original schematic:
http://www.diyaudio.com/forums/atta...mp;postid=32776 (Thanks Grey, Nelson)
Source resistors are 0.47ohm. Voltage over the source resistors is 0.35V. So I have 0.745mA times 8 = 5.96A lets say 6A bias.
R19 and R29 are 47k. Feedback resistors are 120k. Input resistors R18 & R28 are 10k.
DC stability resistors from common diff-pair source to outputs are 2.35k (two paralleled 4.7k).
Both absolute and differential DC are extremely low. Diff-DC can be reduced to almost 0 with V2 and absolute DC is 15-17 mV
R1/R4 and R44/R45 are not used.
No C2 and C4.
R12, R34 are 2.2k.
Sinewave 400Hz unbalanced.
Output power is now 18W RMS into 8ohm. With the 4ohm load 14W RMS.
I added another 2.2k parallel to R12, R34 and the results are a bit better:
24V * 2.2A = 52.8W RMS into 8 ohm
10V * 1.7A = 17W RMS into 4 ohm.
These values are measured just before the XY-axis starts showing distortion. That is become a bow or elipse. I can increase volume further before things get out of hand, but to me it seems the XY-axis should be flat all the way before clipping.
I got rid of the severe clipping at 8ohm and the 4ohm tests are better too.
Still not what it should be when burning that much power.
Sorry if I offended someone (sy, moe29…) yesterday. It was late and I was tired. I just wanted to keep the thread on-topic, that’s all. ;)
Here’s a nice peace about Tori for moe29: http://members.aol.com/AKlink8489/musician.htm :cool:
/Hugo |
How are you doing the measurements? Have you ever listened w/o the scope attached? Are you using a differential probe at the output or just looking at 1 side to gnd?
In all my experimentations I noticed that really bad things happen depending on how instruments are connected. |
|
|
| Netlist |
| quote: | Originally posted by grataku
In all my experimentations I noticed that really bad things happen depending on how instruments are connected. |
I know what you are talking about.
The scope ground is attached to the star ground of the amp.
Channel A goes to +out, channel B to -out.
Bad things happen when there is offset at the outputs and the ground of the probe is connected to - or + out.
There is no difference in sound or power with or without connected instruments.
BTW, the sound is absolutely beautiful when the amp plays in the “safe” area.
I would like to hear from other people what the scope patterns are in XY-mode.
Did anyone ever perform these tests?
/Hugo |
|
|
| Banfi T. |
Well, well...Do we really listen to what we hear? Till, please just look up in the booklet what is that "strange noise" in Boys for Pelé... What makes me curius that you haven't noticed the other, much more relevant effect of the used equipment, namely that it modulates the sound of the Bösendorfer by mechanical means. All over the song...
Or just hit the links above:
http://theatreorgans.com/hammond/fa...ry/mystery.html
http://mixonline.com/ar/audio_tori_amos/
Hmmm.... |
|
|
| grataku |
| quote: | Originally posted by Netlist
I know what you are talking about.
The scope ground is attached to the star ground of the amp.
Channel A goes to +out, channel B to -out.
Bad things happen when there is offset at the outputs and the ground of the probe is connected to - or + out.
There is no difference in sound or power with or without connected instruments.
BTW, the sound is absolutely beautiful when the amp plays in the “safe” area.
I would like to hear from other people what the scope patterns are in XY-mode.
Did anyone ever perform these tests?
/Hugo |
I never used XY-mode. Nelson and maybe Grey or Fred (he never clearly stated whether he got the TexX done). To tell you the thruth, I found your initial post confusing, I never quite understood what the problem was, exactly. Granted I was having breakfast and skimming through the posts...but what is the 'safe' area? |
|
|
| Netlist |
| quote: | Originally posted by grataku
I never used XY-mode. |
XY-mode is used to look at the phase of a signal.
I can't explain you the exact process of how this works, but if you have a scope that can do XY-mode you should try it.
Connect the left output to the A-channel, and the right output to the B-channel. Apply a sinewave to the amp and calibrate the scope for proper display of the XY-axis. When all is well in phase, and both channels of the amp are driven equally with the same signal you should see a nice straight line from the top left to the bottom right of the scope. When this line becomes a bow or an elipse something is wrong. Don't ask me further technical details, other people here are much more qualified to explain what I mean.
Now, play a nice piece of music, in stereo and look what happens. Very nice indeed! :)
With safe area I mean the max volume where the line is still a nice straight line.
To Banfi T. : impressive. The Nexo stuff is indeed superb. I bet they have tons of roadies..and money :clown:
/Hugo |
|
|
| grataku |
| I know what the XY mode is, I just never used it. I normally look at input and output waveforms. So the A and B are on the SAME channel, just on the + and - output, right? I'll try that out tonite and let you know. |
|
|
| Netlist |
| quote: | Originally posted by Netlist
Connect the left (or right) +output of your amp to the A-channel of the scope, which is then your Y-axis, and the right (or left) +output to the B-channel of the scope, which is then your X-axis. Ground of the scope probes is -outputs of the amp. |
Read this again: not to make mistakes.
Don't ask me if you can do this with two monoblocks or with the Alephs.
/Hugo - Hard to write manuals. :clown: |
|
|
| grataku |
quoteA:
Originally posted by Netlist
Connect the left (or right) +output of your amp to the A-channel of the scope, which is then your Y-axis, and the right (or left) +output to the B-channel of the scope, which is then your X-axis. Ground of the scope probes is -outputs of the amp.
Read this again: not to make mistakes.
Don't ask me if you can do this with two monoblocks or with the Alephs.
/Hugo - Hard to write manuals.
quoteB:
I know what you are talking about.
The scope ground is attached to the star ground of the amp.
Channel A goes to +out, channel B to -out.
Bad things happen when there is offset at the outputs and the ground of the probe is connected to - or + out.
Ok, thanks for straightening that out. I got confused by your posts (see quoteA and quoteB). However, I don't agree that in this cased is legit to connect the scope ground to the -output. |
|
|
| Netlist |
| quote: | Originally posted by grataku
However, I don't agree that in this cased is legit to connect the scope ground to the -output. |
In the A-quote, a normal commercial stereo amp is in mind.
In most of these the -output is connected to ground.
The XY-connection to the scope can certanly be done as described.
In quote B I'm talking about the AlephX. First, we have two monoblocks. Not sure if we can connect both of them star points together through the scope ground. Second, in the X-circuits, we have two identical 'channels' connected together like a bridged amp, but with X added. (That's the simple way I see the X-circuit). This means we have four outputs and only two scope probes. That's why I said not to ask me how to do it with two monoblocks or with the Alephs.
/Hugo |
|
|
| grataku |
| quote: | Originally posted by Netlist
In the A-quote, a normal commercial stereo amp is in mind.
In most of these the -output is connected to ground.
The XY-connection to the scope can certanly be done as described.
In quote B I'm talking about the AlephX. First, we have two monoblocks. Not sure if we can connect both of them star points together through the scope ground. Second, in the X-circuits, we have two identical 'channels' connected together like a bridged amp, but with X added. (That's the simple way I see the X-circuit). This means we have four outputs and only two scope probes. That's why I said not to ask me how to do it with two monoblocks or with the Alephs.
/Hugo |
I give up, I have no idea of what you're talking about. Maybe you ought to revise your simple way of thinking about the X. |
|
|
| Helix |
| quote: | Originally posted by grataku
In all my experimentations I noticed that really bad things happen depending on how instruments are connected. |
Like the time I remembered that the scope is connected to earth, as I put the probe and ground across the output of a bridged amp
:bigeyes: |
|
|
| Netlist |
| quote: | Originally posted by grataku
I give up, I have no idea of what you're talking about. Maybe you ought to revise your simple way of thinking about the X. |
Ok, I rarely give up; I make you a drawing of what I want to explain.
Don't blame ME if YOU can't understand what I mean. No offence!!:)
/Hugo |
|
|
| tiroth |
| Netlist, there are not "4 outputs" there are either two outputs referenced to ground, or a single output not referenced to ground, depending on how you want to look at it. |
|
|
| Netlist |
| quote: | Originally posted by tiroth
Netlist, there are not "4 outputs" there are either two outputs referenced to ground, or a single output not referenced to ground, depending on how you want to look at it. |
Thanks tiroth,
could you also explain how to measure the two outputs of two AlephX monoblocks with one scope? Connecting the probes to +out and grounds to scope ground? Or connect the probes to +out and probe grounds to -out?
Is my A-quote correct?
I believe there's a big difference in measuring a normal amp, a bridged amp and an X-amp.
/Hugo |
|
|
| grataku |
"could you also explain how to measure the two outputs of two AlephX monoblocks with one scope?"
with a 2 channel scope, by using two differential probes.
PS I am a little thick-headed, sorry ;) |
|
|
| moamps |
| quote: | Originally posted by Netlist
could you also explain how to measure the two outputs of two AlephX monoblocks with one scope? Connecting the probes to +out and grounds to scope ground? Or connect the probes to +out and probe grounds to -out?
Is my A-quote correct?
I believe there's a big difference in measuring a normal amp, a bridged amp and an X-amp.
/Hugo |
Hi,
why you need measuring two monoblocks with scope in same time? :scratch: If you want measuring phase shift between blocks, just measure phase between two + or - outputs.
You can't connecting ground of probe to any output of bridged (x amp). For measuring one bridged output you need two channel scope, with inputs summed, one input reversed.
If you have diff. probe you can measure two blocks with two channel scope.(grataku answered)
Regards |
|
|
| Netlist |
| quote: | Originally posted by grataku
PS I am a little thick-headed, sorry ;) |
Don't worry, I definitely have the same problem sometimes.. :cheerful:
| quote: | Originally posted by moamps
Hi,
why you need measuring two monoblocks with scope in same time? :scratch: If you want measuring phase shift between blocks, just measure phase between two + or - outputs.
You can't connecting ground of probe to any output of bridged (x amp). For measuring one bridged output you need two channel scope, with inputs summed, one input reversed.
If you have diff. probe you can measure two blocks with two channel scope.(grataku answered)
Regards |
Thanks, that's exactly what I wanted to know. I simply couldn't find it in grataku's answer. I'll give it a try (when I have my two monoblocks)
/Hugo :) |
|
|
| TheMyxin |
[QUOTE]Originally posted by till
[B]Are they recorded superb?
I ask because i never understood the reason for this strange noise right at the beginning of Boys for Pele.
I think this comes from the Instrumentation
the Boersendorfer trough a Leslie Cabinet
as described in the inlay of the CD-Booklet.
:-)) |
|
|
| Netlist |
Back to the first post of this thread:
Did some mods, and the power supply now is +/-22V
The total bias is 7.78A
The AC current gain is 50%.
The number of fets is 16.
This is all for 1 channel, and according to berekening.xls I am supposed to have 100W into 8ohm and 121W into 4ohm.
The reality however, is different.
8ohm: 59W (19V*3.1A)
4ohm: 30W (10V*3A)
measured with a true RMS-DMM just before clipping.
I took pictures of the XY-axis scope result.
The amp is playing extremely well at low volume.
What could be wrong?
The first picture shows the result at 8ohm.
/Hugo |
|
|
| wuffwaff |
Netlist,
seems like you don´t have enough ac current gain and 3A is the max you can get out of the amp.
What you can try is to measure clipping voltage with no or a very high load (the output to ground resistors for example). If you get around 28V (40 volts peak) then at least this part of the amp is ok.
Now you´ll have to see why you don´t get more than 3A . 3A would mean 4.2A peak current and a total of 4.2A bias (both sides) at 50% ac-current gain.
If as you say you´ve got 7.8 A you should have 7.8A peak and 5.5A rms. Since you only have 3A rms your ac-current-gain is only around 10%.
william |
|
|
| vossie |
Something I've noticed (actually picked up from our learned friends on this forums) and might be something for you to look at is your PSU...
Your power supply construction has a LOT to do with the amp's ability to reproduce complex high frequencies (as found with female voices TA and Nora Jones). I noticed with my Aleph 5 that it's high frequency reproduction ability is severely hampered by not having smaller (faster) caps in parallel over the larger, electrolytic reservoirs. Nelson has a very good article about PSU construction that speaks of the finer details around this subject.
All boils down to the fact that the larger caps are not able to deliver the current (due to "internal impedance" and "induction") at the speed which the higher power, high frequencies require. Having a 10uF, a 1uF and a 0.1 uF over your cans makes one huge difference. If you have these check them for faults and play around with them.
Keep in mind that the X amp is basically a double-aleph construction - it is the secondary stage of an Aleph mirrored and shared from the same LTP input stage. This means that double the PSU speed could be required at some stages of music reproduction. You might be better off having these bridging caps applied to your positive side and negative side seperately.
I don't know what your PSU looks like but I bet this is part of your problem - if not the main culprit.
Let us know.
Vossie |
|
|
| Netlist |
Thanks everyone, Highly appreciated.
I live with my AlephX for a few weeks now, and did some more listening test and tweaks. The former TA problems are gone now.
I will come back on the details later, and will try to post a more final version of the mods than is possible now.
I certainly have to change the big PSU caps to a higher voltage rating and I will try the bypass caps too. For the moment I only have four 330nF caps over the diode bridge.
The first five minutes, when the amp is still cold, I hear a small hum. In time I would like to add some nice chokes to the PSU, hoping the hum disappears completely.
You can find picture and more details of the amp here:
http://www.diyaudio.com/forums/show...&threadid=14979
Thanks again for you input.
/Hugo |
|
|
| vossie |
You don't have to go for the best caps (or most expensive ones) on the shelf. Give it a go with the normal 250v /??uF caps used for ripple smoothing over AC loads. Even they make a difference and I have not been able to pick up an audible difference between them and the melalised film caps for example.
The best to do is to give it a "calculated try" and see what happens...:devilr:
Vossie |
|
|
|
