yeah, i think i could copy down the aleph3 service menu from memory at this point...
ill check out the 30 and the 60 for the problem... any reason why this problem exists? and is there a certain type of cap i should use? i think i might have some WIMA's or something...
also, does the value really matter a whole lot, or should i stick with 1nf?
ill check out the 30 and the 60 for the problem... any reason why this problem exists? and is there a certain type of cap i should use? i think i might have some WIMA's or something...
also, does the value really matter a whole lot, or should i stick with 1nf?
WIMA's are just fine. Stick with 1nF
For testing purposes even a ceramic C will do.
Reason for this problem is the insufficient phase margin for the
loop in some implementations of the aleph current source. Phase
margin is influenced by the transconductance of the FET(s).
This will lead to oscillations of the Aleph current source,
sometimes resulting in crackle noise, buzz, excessive hum etc..
The cure is the compensation C
For testing purposes even a ceramic C will do.
Reason for this problem is the insufficient phase margin for the
loop in some implementations of the aleph current source. Phase
margin is influenced by the transconductance of the FET(s).
This will lead to oscillations of the Aleph current source,
sometimes resulting in crackle noise, buzz, excessive hum etc..
The cure is the compensation C
rtirion said:
you lost me at "reason". but if it works, it works.
when i get home from work, ill give it a try if i have caps.
another silly question... should i match input impedance if im bi-amping with another amp with a higher input impedance? i am planning to use this to drive the tweeters on my magnepans. the bass panels are driven by a rotel amp. the rotel's input imped. is 32kohms, and the aleph is 23kohms. the aleph definately sounds much more muted than the rotels. it's sound wasnt really that impressive, but maybe the cap will fix this. i was thinking its becuase the input levels were a bit off, maybe this could be it?
Cowanrg,
Can you post a schematic for your power supply?
From looking at your web page it seems like you
are usting just capacitors after your rectifiers.
A CLC power supply set up may take away some
of the hum/buzz you are hearing. Adding a 2mH
inductor to my ZEN amp's PSU really quieted it
down. (see Return of Zen)
Just a thought.
m.
Can you post a schematic for your power supply?
From looking at your web page it seems like you
are usting just capacitors after your rectifiers.
A CLC power supply set up may take away some
of the hum/buzz you are hearing. Adding a 2mH
inductor to my ZEN amp's PSU really quieted it
down. (see Return of Zen)
Just a thought.
m.
Moe,
yes, im just using a simple power supply.
i am about to add the cap to my amp (after reading through many threads, it looks to be the issue). i also have some other concerns that might make me think its more than a humming issue.
i tested the amp's power output, and at max, im getting around 12.5 watts. thats WAY too low. so it seems like its not getting enough bias. also, when i did a frequency sweep test tone, it really didnt like the 10hz. i dont know exactly what it was doing, but it didnt give as much juice, AND the LED on it was blinking. that was strange. this was all from a 8ohm dummy load too.
im going to throw in the cap, and since im running so cool right now (not even really all that warm to the touch), im going to increase R113 (or remove it), to give more bias.
if all that doesnt work, i guess i could try a filter on the supply.
yes, im just using a simple power supply.
i am about to add the cap to my amp (after reading through many threads, it looks to be the issue). i also have some other concerns that might make me think its more than a humming issue.
i tested the amp's power output, and at max, im getting around 12.5 watts. thats WAY too low. so it seems like its not getting enough bias. also, when i did a frequency sweep test tone, it really didnt like the 10hz. i dont know exactly what it was doing, but it didnt give as much juice, AND the LED on it was blinking. that was strange. this was all from a 8ohm dummy load too.
im going to throw in the cap, and since im running so cool right now (not even really all that warm to the touch), im going to increase R113 (or remove it), to give more bias.
if all that doesnt work, i guess i could try a filter on the supply.
well, more tinkering, and im more confused.
there is still a slight hum. i cant tell if its less than before (it might be), but its still there, and that's unacceptable. im also not able to get more than about 20 watts out of it.
i did what rtirion suggested, by sticking a 1nf cap across B and C of Q105. it appears to operate in the same manner, maybe less buzz/hum.
here is a note i find interesting... if i understand correctly, by increasing the bias, i will get a higher output power, as well as better sound. well, i removed R113, and am now getting a 2.8A draw at the positive rail. this was up from ~1.8A before. however, when i throw it the exact same sine wave i did before, im getting the exact same power output, 12.5 watts. this isnt exactly scientific, but its a signal generator on my computer, that im outputting at full output. i have used this to test amps before, and it typically comes very close to the amp's max power output rating. in this case, it doesnt even come close.
moe has suggested doing a filter at the power supply. is there a way to test is this is the cause? i guess im asking if there is some sort of way to make a conclusion what is going wrong. i saw a thread where rtirion was testing voltages and getting less than desired. he knew what to do to get the right values, of course i dont.
i guess a question to ask would be for nelson himself... can this amp be dead silent when it's running, or will it always have a little bit of hum or hiss to it? the original didnt have a filter on the supply, and i have seen many others do it without and not complain about hum. am i just being picky?
there is still a slight hum. i cant tell if its less than before (it might be), but its still there, and that's unacceptable. im also not able to get more than about 20 watts out of it.
i did what rtirion suggested, by sticking a 1nf cap across B and C of Q105. it appears to operate in the same manner, maybe less buzz/hum.
here is a note i find interesting... if i understand correctly, by increasing the bias, i will get a higher output power, as well as better sound. well, i removed R113, and am now getting a 2.8A draw at the positive rail. this was up from ~1.8A before. however, when i throw it the exact same sine wave i did before, im getting the exact same power output, 12.5 watts. this isnt exactly scientific, but its a signal generator on my computer, that im outputting at full output. i have used this to test amps before, and it typically comes very close to the amp's max power output rating. in this case, it doesnt even come close.
moe has suggested doing a filter at the power supply. is there a way to test is this is the cause? i guess im asking if there is some sort of way to make a conclusion what is going wrong. i saw a thread where rtirion was testing voltages and getting less than desired. he knew what to do to get the right values, of course i dont.
i guess a question to ask would be for nelson himself... can this amp be dead silent when it's running, or will it always have a little bit of hum or hiss to it? the original didnt have a filter on the supply, and i have seen many others do it without and not complain about hum. am i just being picky?
Cowanrg,
Something's not right. You mention that the amp is running cool,
there's no way an Aleph 3 runs cool unless you have 3 foot high
heatsinks! From looking at your pictures, those heatsinks should
be hard to touch for more than 5 seconds or so.
You need to stop and go over the circuit again with a fine tooth
comb. Something's not right. By no means am i an expert at
troubleshooting, but adding small bypass caps or adding a CLC
filter like i mentioned is not the problem.
Start with the power supply, make sure you are getting good
reading for both + and - rails.
You may also think about replacing all the active parts on the PCB.
Replace the power resistors.
It just seems to me something is fundamentally wrong, you need to start there.
Something's not right. You mention that the amp is running cool,
there's no way an Aleph 3 runs cool unless you have 3 foot high
heatsinks! From looking at your pictures, those heatsinks should
be hard to touch for more than 5 seconds or so.
You need to stop and go over the circuit again with a fine tooth
comb. Something's not right. By no means am i an expert at
troubleshooting, but adding small bypass caps or adding a CLC
filter like i mentioned is not the problem.
Start with the power supply, make sure you are getting good
reading for both + and - rails.
You may also think about replacing all the active parts on the PCB.
Replace the power resistors.
It just seems to me something is fundamentally wrong, you need to start there.
well, i dont have a scope, and im not going to get one just yet, unless someone has one for CHEAP.
i would tend to agree about the amp, something isnt running quite right.
a couple of the power resistors are running really hot, i would assume hotter than they should be. i could replace them i guess. it seems too simple of a problem.
as far as how hot they get, when i have them biased at 2.8A, they get HOT. at their 1.8A, they run warm, but not hot.
from the power supply, i am getting about 26.8v unloaded, and around 23v or so loaded. this is from both rails. the current draw seems to be right on as well. (at 1.8A normally biased, i do have an RMS meter, so that seems right).
since there are two monoblocks, i just have a hard time believing that each one is wrong in the exact same way... they are built the same, but each was done at a different time (i built one first, then the second one). the first one has been COMPLETELY rebuilt 3 times (using all new parts each time).
i really wish nelson could get in here and give some advice. every other thread ive seen him reply in with problems, he got them solved immediately. he really knows this circuit
im just at a loss here.
i would tend to agree about the amp, something isnt running quite right.
a couple of the power resistors are running really hot, i would assume hotter than they should be. i could replace them i guess. it seems too simple of a problem.
as far as how hot they get, when i have them biased at 2.8A, they get HOT. at their 1.8A, they run warm, but not hot.
from the power supply, i am getting about 26.8v unloaded, and around 23v or so loaded. this is from both rails. the current draw seems to be right on as well. (at 1.8A normally biased, i do have an RMS meter, so that seems right).
since there are two monoblocks, i just have a hard time believing that each one is wrong in the exact same way... they are built the same, but each was done at a different time (i built one first, then the second one). the first one has been COMPLETELY rebuilt 3 times (using all new parts each time).
i really wish nelson could get in here and give some advice. every other thread ive seen him reply in with problems, he got them solved immediately. he really knows this circuit
im just at a loss here.
...maybe you're trying to do too much at once.
why are you trying different bias levels before the circuit is working?
Use the parts as specified in Kristijan's manual (stock parts).
If your power supply rails are measuring correctly, if you are getting:
good Vgs readings across each FET within .01 or so of each other
(you said they're matched)
and you're getting a .5v reading over your power resistors
and a normal Vgs reading on each of your differential pair FETS
and a DC offset unter 100mV
Then you should get music, and not much hum or buzz.
Is this the status of the amps right now?
Sorry if i'm confused, it just seems like you're so close to finishing,
and i can feel your frustration!
good luck,
m.
why are you trying different bias levels before the circuit is working?
Use the parts as specified in Kristijan's manual (stock parts).
If your power supply rails are measuring correctly, if you are getting:
good Vgs readings across each FET within .01 or so of each other
(you said they're matched)
and you're getting a .5v reading over your power resistors
and a normal Vgs reading on each of your differential pair FETS
and a DC offset unter 100mV
Then you should get music, and not much hum or buzz.
Is this the status of the amps right now?
Sorry if i'm confused, it just seems like you're so close to finishing,
and i can feel your frustration!
good luck,
m.
yeah, im just getting far too frustrated. another couple days of this, and it goes on the shelf, my workshop gets closed up for a few months, and i forget about DIY.
i will get all the measurements i can. i will even try and test the transistors individually and see if they are actually matched. i bought them off a group purchase here on diyaudio about a year or so ago. they were matched for an aleph.
i know i have a dc offset of around 2mv, so that is good.
i will check on the vgs readings of the fets and the differential pair.
i will get all the measurements i can. i will even try and test the transistors individually and see if they are actually matched. i bought them off a group purchase here on diyaudio about a year or so ago. they were matched for an aleph.
i know i have a dc offset of around 2mv, so that is good.
i will check on the vgs readings of the fets and the differential pair.
my listening room is pretty great actually. its about 23' x 20'. the 3 of the walls are exterior walls. the back wall (where you have your back to), is actually open to the rest of the house, which goes back another 30' maybe. the ceiling is vaulted to about 15'. the back wall is about 2 stories high, not just a single story because of the vault.
on the main wall, i have my main rack, which holds my tv, dvd, receiver, and some other components. i have two amp stands that hold my rotel 2-channel amps (currently powering my mains). the alephs will go on top of the rotels (not stacked, but i have another shelf for the amps stands). these stands are RIGHT next to each speaker, so i only need a 3' speaker cable.
the mains are magnepan 1.6QR's, powered by rotel 1070 amps (one per speaker). i use a sonic frontiers SFT-1 as a transport, a denon dvd-2200 for dvd's, dvda, and sacd.
i also have dual subs. i use two velodyne hgs-15x's crossed over at 80hz. each one sits right behind the each front speaker. my rears are magnepan smgc's.
all cables are either audioquest jaguar interconnects or audioquest pike's peak speaker cables. the sub cables are synergistic research resolution reference x2, and the digital cable is a MIT digital reference.
on the main wall, i have my main rack, which holds my tv, dvd, receiver, and some other components. i have two amp stands that hold my rotel 2-channel amps (currently powering my mains). the alephs will go on top of the rotels (not stacked, but i have another shelf for the amps stands). these stands are RIGHT next to each speaker, so i only need a 3' speaker cable.
the mains are magnepan 1.6QR's, powered by rotel 1070 amps (one per speaker). i use a sonic frontiers SFT-1 as a transport, a denon dvd-2200 for dvd's, dvda, and sacd.
i also have dual subs. i use two velodyne hgs-15x's crossed over at 80hz. each one sits right behind the each front speaker. my rears are magnepan smgc's.
all cables are either audioquest jaguar interconnects or audioquest pike's peak speaker cables. the sub cables are synergistic research resolution reference x2, and the digital cable is a MIT digital reference.
Set your meter on VAC (same as measuring your house wiring) and put your leads across the supply capacitors. Positive side to ground and negative side to ground. Same as measuring VDC. You should be seeing levels in the millivolt range when the amp is biased. As you draw more current from your supplies the ripple will increase. If I remember correctly my alephs were about 65mV when biased at 3 amps each rail. What we are looking for is a capacitor that might be bad causing excessive current draw. You wouldn't want to see anything up around a couple hundred mV.
BDP
Don't let it get you down. I had the same power up woes that made me shelf my Aleph 2 for a few months.
Did you try grounding the input and see if the noise was being picked up from interconnects or ground loop issues with other equipment?
Also, lets talk about getting you a scope for that big workshop. Your deep enough into this game to invest a few bucks in a scope. Ebay is the place to be for this one. For a few hundred dollars you can go first class and get a nice Tektronics. Or if you don't want to spend much, do what I did. I got a B&K 2120 for $40.00 that works perfect. Or get what ever falls in between in your budget. But don't waist your money on any PC card based or cheesy hand held LCD Velleman garbage. Get Tek or B&K or maybe HP(I don't really know their scopes, but they made hella fine stuff back in the day)
Keep us posted
o yea
Don't bother removing the output fets to rematch Vgs, just check to see that all of the source resistors have the same (close) voltage drop. And if you had a scope, I'd say use AC coupling to check for noise across as well. and checking for noise at the gates could tell you if the problem was coming from the + or - half.
Did you try grounding the input and see if the noise was being picked up from interconnects or ground loop issues with other equipment?
Also, lets talk about getting you a scope for that big workshop. Your deep enough into this game to invest a few bucks in a scope. Ebay is the place to be for this one. For a few hundred dollars you can go first class and get a nice Tektronics. Or if you don't want to spend much, do what I did. I got a B&K 2120 for $40.00 that works perfect. Or get what ever falls in between in your budget. But don't waist your money on any PC card based or cheesy hand held LCD Velleman garbage. Get Tek or B&K or maybe HP(I don't really know their scopes, but they made hella fine stuff back in the day)
Keep us posted
o yea
Don't bother removing the output fets to rematch Vgs, just check to see that all of the source resistors have the same (close) voltage drop. And if you had a scope, I'd say use AC coupling to check for noise across as well. and checking for noise at the gates could tell you if the problem was coming from the + or - half.
sounds good everyone. at least i have a few things to check out right now.
im going to go down there and do all the measurements all of you requested. i hope to see some strange things. i really hope some measurements are off.
i think i probably should get a scope. if i can get one for around $40, that wold be nice. i really dont have a huge budget for this stuff.
i have tried shorting the input (also just plain disconnecting it alltogether), but it acts the same. i disconnected the earth ground and am now getting a bit less hum, and no loops when i connect an input (however there is a a bit of a thump when i connect something, but only at that second when its connected, and i can deal with that.)
so, im making a list of everything i need to measure, and i will post it ALL here in a few hours.
im going to go down there and do all the measurements all of you requested. i hope to see some strange things. i really hope some measurements are off.
i think i probably should get a scope. if i can get one for around $40, that wold be nice. i really dont have a huge budget for this stuff.
i have tried shorting the input (also just plain disconnecting it alltogether), but it acts the same. i disconnected the earth ground and am now getting a bit less hum, and no loops when i connect an input (however there is a a bit of a thump when i connect something, but only at that second when its connected, and i can deal with that.)
so, im making a list of everything i need to measure, and i will post it ALL here in a few hours.
well, these measurements are seeminingly good. im not sure how close they are supposed to be though. few are actually within their range, most are close though. hopefully this can tell someone something.
so here is every measurement i know how to do:
PSU
Current:
positive rail: 2.9A
negative rail: 2.9A
Voltage (loaded)
positive rail: 29.99V
negative rail: 23.02V
vAC ripple: 0.049V
DC offset: 1.17mV
FETS (Vgs):
Q106: (4.9 - 0.7) = 4.2
Q107: (4.89 - 0.69) = 4.2
Q108: (18.2 - 22.3) = 4.1
Q109: (18.2 - 22.3) = 4.1
Q105: (4.8 - 0.0017) = 4.79
Q103: (14.35 - 17.82) = 3.47
Q101: (3.622 - 0.013) = 3.609
Q102: (3.62 - 0) = 3.620
Power Resistors:
R120: (0.692 - 0.25) = .666 (oooh, scary)
R121: (0.681 - 0.02) = .661
R122: (23.1 - 22.46) = .64
R123: (23.04 - 22.35) = .69
(all other power resistors measured 0.004 or 0.005, but aren't mentioned in service manual, so i wont type it all)
Other:
R108: (18.35 - 23.13) = 4.78
Z103: (23.2 - 14.4) = 8.8
so here is every measurement i know how to do:
PSU
Current:
positive rail: 2.9A
negative rail: 2.9A
Voltage (loaded)
positive rail: 29.99V
negative rail: 23.02V
vAC ripple: 0.049V
DC offset: 1.17mV
FETS (Vgs):
Q106: (4.9 - 0.7) = 4.2
Q107: (4.89 - 0.69) = 4.2
Q108: (18.2 - 22.3) = 4.1
Q109: (18.2 - 22.3) = 4.1
Q105: (4.8 - 0.0017) = 4.79
Q103: (14.35 - 17.82) = 3.47
Q101: (3.622 - 0.013) = 3.609
Q102: (3.62 - 0) = 3.620
Power Resistors:
R120: (0.692 - 0.25) = .666 (oooh, scary)
R121: (0.681 - 0.02) = .661
R122: (23.1 - 22.46) = .64
R123: (23.04 - 22.35) = .69
(all other power resistors measured 0.004 or 0.005, but aren't mentioned in service manual, so i wont type it all)
Other:
R108: (18.35 - 23.13) = 4.78
Z103: (23.2 - 14.4) = 8.8
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.