Netlist said:
This and input caps solved the problem for me.
/Hugo
Where did you put the input caps? Just in serie after the resistor on IN- like in the normal Aleph?
What happened when you did not use the cap? I have put in the 2k3 for R46 and R47 and even then it moves around. It does not go down and then stay put at a resolution of 1mV. It keeps moving up and down a bit all around in a space of 250mV or so not really something to stay awake over but...... I even have put in the 100ohm's...but that does not do much either, nothing even...
Edwin
I assume you connect unbalanced.Edwin Dorre said:
Then the picture is how I did it. The -IN is your cable shielding and is grounded.
DC drift.
In "This is not just another Aleph" Peter gives a good explanation too about the way he did the coupling. He said it was important to use same values for both caps.
Good luck.
/Hugo
Attachments
, but while measuring you have terminated the input? If so ignore me Netlist said:
I assume you connect unbalanced.
Then the picture is how I did it. The -IN is your cable shielding and is grounded.
DC drift.
In "This is not just another Aleph" Peter gives a good explanation too about the way he did the coupling. He said it was important to use same values for both caps.
Good luck.
/Hugo
Hugo,
have you experimented with the values of r19 and 29? I am using 10K or maybe 15K, don't remember. Nelson uses 10K I believe, it keeps the gain low.
I just went over 1k post so my mission here has been accomplished.
Congrats with your 1K posts!!!grataku said:
Hugo,
have you experimented with the values of r19 and 29? I am using 10K or maybe 15K, don't remember. Nelson uses 10K I believe, it keeps the gain low.
I just went over 1k post so my mission here has been accomplished.
I might start a congrats thread for you. 
I never went that low with R19/29. Currently I have 47K in there.
What is the purpose of keeping the gain low? Could a lower R19/29 also affect the offset?
/Hugo – 0.650K posts. 4.72 posts a day. Posts Law: 650/4.72 = 137.71 days on this forum
Netlist said:
I assume you connect unbalanced.
Then the picture is how I did it. The -IN is your cable shielding and is grounded.
DC drift.
In "This is not just another Aleph" Peter gives a good explanation too about the way he did the coupling. He said it was important to use same values for both caps.
Good luck.
/Hugo
Hmm.. I used the Aleph-X as single-ended although when he is finished I will use him with a balanced Aleph-P.
I tried putting in the 2.2uf Elco's but this did not do very much. I put them in and shorted the GND and the IN+ , the IN- is also shorted via the new inserted elco to GND. I only had 2.2uf elco's in house, but I will buy some cheapo test filmcaps today.
But... no go... I absolute DC keeps running down. It moves from 2V to 200mv very quickly, then slowly walks from the 0mV boundary to minus -100mV. Then I turn the trimpot up to 50mV and he keeps walking down... I can keep doing this for more than an hour...
I have 2k2 for the Ian resistors, 47ohm to ground, 2.2uf elco in place... no go...
Ok, time for some serious reading and searching on diyaudio.com !
Edwin
P.S. O, I forgot to tell. I use CRCRC (6x 22000uf 63V BC components 154) with 0.11 ohm per R. I have an ripple of about top-to-top of 0.5V with my full load of 400W ... Is this a source of my offset trouble?
I assume you mean 47k to ground? For R19/29?Edwin Dorre said:
As a result you terminated the input.
Your PSU seems OK to me, I mean, I wouldn't know how this can affect offset.
I re-measured a few things in my amp and at first switch on, when completely cold the offset is 11mV, and within a minute or so goes down to 1 to 3 mV and stays there.
Is the voltage over the 390ohm resistors stable?
/Hugo
My DC offset (at speaker's output) is usually very low and doesn't drift much (less than 10mV). My absolute offset is 4V or so at start up and after an hour or two, goes down to 0 and stays there. It doesn't bother me much as I keep the amps all day on when listening.
The offset at speakers terminals is adjusted by changing series resistor value at CCS of differential pair. The absolut offset is changed by changing the bias at the output (balancing both sides by checking voltage drop on source resistors).
I'm not using 100ohm resistor from output to ground as I didn't notice any advantage when using it, and I wanted to limit parts count. 4.7 ohm Ian resistors are in use. The amp is using 8 pairs of devices per channel and is very stable, just takes time to settle down
The offset at speakers terminals is adjusted by changing series resistor value at CCS of differential pair. The absolut offset is changed by changing the bias at the output (balancing both sides by checking voltage drop on source resistors).
I'm not using 100ohm resistor from output to ground as I didn't notice any advantage when using it, and I wanted to limit parts count. 4.7 ohm Ian resistors are in use. The amp is using 8 pairs of devices per channel and is very stable, just takes time to settle down
Found some interesting posts from Peter in "The AlephX" thread:
"Posted by Peter Daniel on 11-22-2002 02:48 PM:
Let me give you an example:
The first channel readings at start up was 1.8V absolute offset with 4mV difference on both sides. Half an hour later it was 0.6V with 6mV difference between sides. After 2 hours it was -0.6V on both sides with difference of 8mV. When I was observing that I thought I was dreaming.
Now the second channel, on one side the absolute DC offset is about 0.1V at start up and it wonders up maybe to 0.2V after warm up. The other side though is more like 5V in the beginning and then goes down to 1V, it can be adjusted to lower value, but then drifts away again. So the difference is usually around 0.5V eventualy.
I have to find what is causing that kind of behaviour. I don't think it has to do anything with resistors or capacitors, but rather with semis and I have to figure out which ones. "
Posted by Peter Daniel on 11-22-2002 07:55 PM:
Good news everybody. Just before I started desoldering differential pair, I noticed I had a broken solder joint on input to ground resistor. Second channel works equaly well as the first one. "
/Hugo
"Posted by Peter Daniel on 11-22-2002 02:48 PM:
Let me give you an example:
The first channel readings at start up was 1.8V absolute offset with 4mV difference on both sides. Half an hour later it was 0.6V with 6mV difference between sides. After 2 hours it was -0.6V on both sides with difference of 8mV. When I was observing that I thought I was dreaming.
Now the second channel, on one side the absolute DC offset is about 0.1V at start up and it wonders up maybe to 0.2V after warm up. The other side though is more like 5V in the beginning and then goes down to 1V, it can be adjusted to lower value, but then drifts away again. So the difference is usually around 0.5V eventualy.
I have to find what is causing that kind of behaviour. I don't think it has to do anything with resistors or capacitors, but rather with semis and I have to figure out which ones. "
Posted by Peter Daniel on 11-22-2002 07:55 PM:
Good news everybody. Just before I started desoldering differential pair, I noticed I had a broken solder joint on input to ground resistor. Second channel works equaly well as the first one. "
/Hugo
It seems that the drifting of the absolute DC has stopped a bit. It still moves a bit around at 0mV (-50mV t/m +50mV) and does not stay put at a fixed value but I think this willl not hurt that much. It seems it need just 2 hours of running to make up his mind.
The diff DC offset stays put at 18mV at does not move at all!
What is a good RMS AC voltage of the sinus wave running through the amp to measure the active current gain? I tried it with 9V and at max power (31V) but both give difference values for the ACS resistors!
Edwin
The diff DC offset stays put at 18mV at does not move at all!
What is a good RMS AC voltage of the sinus wave running through the amp to measure the active current gain? I tried it with 9V and at max power (31V) but both give difference values for the ACS resistors!
Edwin
Netlist said:
Did you check the solderings?
This looks pretty OK.
I never checked at different power levels. Is there a huge difference?
/Hugo
I am allmost ready to boot up channel 2, so I will check it with this one, to see how he behaves. I do not think I made a bad soldering at the same place in two channels
I re-checked but there is not a big difference. It is 53% to 52% or so... nothing to worry about. I think I more or less measured difference in warmth of the amp than on output power (too much time between measurements during warmup)! My fault!
Edwin
Netlist said:
Did you check the solderings?
/Hugo
I started up channel 2 and this behaves the same! I have about 44mV of diff DC offset and the absolute offset once the amp is hot keeps about 20mV but finds it very funny to move up and down around this 20mV center point...
I dunno... I think I stop worying about it and make one of these facts of live... I can try 1K for the ''Ian' resistors . Channel 1 of running on 2k2, channel 2 is running at 4k7 and I do not see much difference in stability. Both are running with 100ohm to ground.
Edwin
Indeed R12/34 are the resistors. A lower R12/34 increases the percentage of the current source.MikeW said:Are R12 and R34 the resistors to set current gain? The lower the resistor the more current gain? If the negative clips before the positive do you need more or less current gain?
This is the explanation by Mr. Pass himself:
http://www.passdiy.com/projects/zenv2-4.htm
Someone else will be able to give you the exact theory of clipping, but the current source should be set at 50%.
Sorry if you already mentioned, but what is the value of your feedback R. Also, in post#48 you say you have 47ohm to ground. Is that for R19/29 and did you mean 47k?Edwin Dorre said:
Maybe the whole DC stuff will stabilize once the amp is burned-in long enough.
/Hugo
Netlist said:
Sorry if you already mentioned, but what is the value of your feedback R. Also, in post#48 you say you have 47ohm to ground. Is that for R19/29 and did you mean 47k?
Maybe the whole DC stuff will stabilize once the amp is burned-in long enough.
/Hugo
The feedback R (r16, r27) is the default 100K . I have 4x big 47 ohm 10W metalfilm resistors, when I put them in serie I have about 100 ohm and when I use then alone I have 47 ohm to put output to ground.
R46 and R47 are 4k7 in one amp and 2k2 in the other...
R19/R29 are the default 10K also. I build more or less the Hifizen/Carpenter/Grataku design.
Yes, lets hope so...
Edwin
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.