hiMr. Nelson,Wuf Waf, Grey, GL and others 🙂 me as not a electronic engineer mooving a bit closer to good sound 🙂 as sicking audiophile I am thinking that my Black gate N warm in or just my ears get used to aleph x sound.
I have a question regarding aleph x improoving -
there is a resistor from diferential pair to output. usualy 2.2kohm on kristijan schematic, 10kohm on gl and I do not see it on Gray schematic. whren I mooved from 2.2kohm to 10kohm-strange thing happen- sound improoved a litle bit-it sounds more analogue. however abolute dc ofset from 1V or so went to 6V ! I would like to ask how to vary absolute dc ofset? with pots I can vary it but not much also when I am adjuting I get that voltage across source resistors lifts to a 0.60-0.66V firs thing0 replaced 33ohm resitor to ground from output with 100ohm. - has no influence.
I find folowing-
""
There are two types of offset with the Aleph-X to consider.
Absolute Offset: The voltage difference between an output terminal and ground.
Relative Offset: The voltage difference between the the positive and negative output terminals.
The absolute offset will affect your amplifier performance near it's maximum output, as the offset will cause the amplifier to start clipping at lower levels than can be achieved with proper adjustment. The good news is that the absolute offset can be set by adjusting the current in the CCS for the differential input transistors (the IRF9610's). Measuring it is easy using a multimeter, leave the amplifier on but with no input signal, connect the common probe to ground, the test probe to an output terminal, and read the voltage.
how to adjust current in the ccs for diferential input? I run 24 mosfet .channlel, 12A bias. , 22.5V rails
"""o
😕
I have a question regarding aleph x improoving -
there is a resistor from diferential pair to output. usualy 2.2kohm on kristijan schematic, 10kohm on gl and I do not see it on Gray schematic. whren I mooved from 2.2kohm to 10kohm-strange thing happen- sound improoved a litle bit-it sounds more analogue. however abolute dc ofset from 1V or so went to 6V ! I would like to ask how to vary absolute dc ofset? with pots I can vary it but not much also when I am adjuting I get that voltage across source resistors lifts to a 0.60-0.66V firs thing0 replaced 33ohm resitor to ground from output with 100ohm. - has no influence.
I find folowing-
""
There are two types of offset with the Aleph-X to consider.
Absolute Offset: The voltage difference between an output terminal and ground.
Relative Offset: The voltage difference between the the positive and negative output terminals.
The absolute offset will affect your amplifier performance near it's maximum output, as the offset will cause the amplifier to start clipping at lower levels than can be achieved with proper adjustment. The good news is that the absolute offset can be set by adjusting the current in the CCS for the differential input transistors (the IRF9610's). Measuring it is easy using a multimeter, leave the amplifier on but with no input signal, connect the common probe to ground, the test probe to an output terminal, and read the voltage.
how to adjust current in the ccs for diferential input? I run 24 mosfet .channlel, 12A bias. , 22.5V rails
"""o
😕
For differential offset, you adjust the loading resistors coming
off the Drains of the input diff pair, and by this I mean you
vary one of them versus the other. Generally I use a pot of
much higher value than the resistors and put the wiper to v-
and each leg to a resistor.
off the Drains of the input diff pair, and by this I mean you
vary one of them versus the other. Generally I use a pot of
much higher value than the resistors and put the wiper to v-
and each leg to a resistor.
thanks Mr. Nelson, this is 2.2kohm resitors in my atached shematic?
if yes I am surprised because variuos aleph version has very diferent values(in Grolins version I can se it at all) for example one diyers easily variies from 4 to 10kohm. when I went from 2.2khz to 10kh I get 6V absolute ofset. is it normal? as I one of the forum member said theese resistors are evils and adviced to keep them as high as posible..
If I have good realitive and absolute dc ofset does it means my aleph x works great and sounds great? I do not have osciloscope. if theese two values are main I can stay tune with them slightly varying ather values/desired bias,output devices number..
just for records last month I killed 48 irfp240 mached mosfets.. and more or less source resistors.
expensive hobby..
if yes I am surprised because variuos aleph version has very diferent values(in Grolins version I can se it at all) for example one diyers easily variies from 4 to 10kohm. when I went from 2.2khz to 10kh I get 6V absolute ofset. is it normal? as I one of the forum member said theese resistors are evils and adviced to keep them as high as posible..
If I have good realitive and absolute dc ofset does it means my aleph x works great and sounds great? I do not have osciloscope. if theese two values are main I can stay tune with them slightly varying ather values/desired bias,output devices number..
just for records last month I killed 48 irfp240 mached mosfets.. and more or less source resistors.
expensive hobby..
It is my experience that the resistors providing common mode
feedback (the 2.2K to which you refer) have values which must
be optimize for any given version of the design. To determine
the optimal values, you must perform a series of tests;
absolute offset stability, CMRR, distortion vs amplitude and
frequency, and of course listening.
Lower resistance gives greater common mode feedback and
so it gives the greatest absolute DC stability, but it is not also
an absolute that lower values degrade the performance, as
I have occasionally observed better performance at lower values.
When we put a version of the circuit into production we test over
a range of values as a matter of procedure.
In general however, I recommend higher values, CCS
adjustment, thermal compensation, and resistive loading of the
output to ground. The goal is to minimize the absolute DC
(the output referenced to ground) and the standard is a few
hundred millivolts or better.
feedback (the 2.2K to which you refer) have values which must
be optimize for any given version of the design. To determine
the optimal values, you must perform a series of tests;
absolute offset stability, CMRR, distortion vs amplitude and
frequency, and of course listening.
Lower resistance gives greater common mode feedback and
so it gives the greatest absolute DC stability, but it is not also
an absolute that lower values degrade the performance, as
I have occasionally observed better performance at lower values.
When we put a version of the circuit into production we test over
a range of values as a matter of procedure.
In general however, I recommend higher values, CCS
adjustment, thermal compensation, and resistive loading of the
output to ground. The goal is to minimize the absolute DC
(the output referenced to ground) and the standard is a few
hundred millivolts or better.
seems I have only one method from what you mentioned- my ears. when lower value I generaly liked sound better but also with this mod and increasing resistors from output to ground up to 100ohm I noticed somewhat limited power handling. abslolute ofset was 6V amp started clip much more earlier and it from listening it has some aded discortion. but it was very easy acceptable. I will try to play this way- to asemble few boards with all parts mached as well as posible 1%resistor 0.6w , 2x 1% 10w resisitors , 0,01V mached transitors. hope this will help to stabilize amp.
by the way Mr. Nelson You was wright- the bass of aleph x is much more better than aleph5. just few weeks needed to burn in new caps I think. with high quality bass drivers in wright enclosure it is so tight , informative,and holds subtile nuanses that I have never heard with other amps. very subtile. even 1952years records sounds absolutely tought. althought its performance very depends on other componnets.
by the way Mr. Nelson You was wright- the bass of aleph x is much more better than aleph5. just few weeks needed to burn in new caps I think. with high quality bass drivers in wright enclosure it is so tight , informative,and holds subtile nuanses that I have never heard with other amps. very subtile. even 1952years records sounds absolutely tought. althought its performance very depends on other componnets.
Would anyone care to elaborate any more on the "thermal compensation" methods that are both simple and effective?
I am finally getting around to assembling my Aleph-x's (just finished a long speaker building project), and I have read alot of ideas for taming the absolute offset during the warm-up period, but no one seems to have found or posted a simple solution (besides just deal with it)... I would imagine the commercial amplifiers settle down fairly quickly.
In my case, my amplifier will probably see a lot of use during the warm-up / stabilization period, so shorter is better.
Thanks,
Jeff
I am finally getting around to assembling my Aleph-x's (just finished a long speaker building project), and I have read alot of ideas for taming the absolute offset during the warm-up period, but no one seems to have found or posted a simple solution (besides just deal with it)... I would imagine the commercial amplifiers settle down fairly quickly.
In my case, my amplifier will probably see a lot of use during the warm-up / stabilization period, so shorter is better.
Thanks,
Jeff
When you want to be very deliberate about thermal
compensation, you can always buy a range of thermistors
from Digikey and put them in the appropriate spot.
The thermistor decreases in resistance as the temperature
rises, and you generally want to bracket it with fixed values
of resistance in parallel and in series with it. The proportions
of all three values are tailored to the need, and are usually
easy, if tedious, to determine experimentally. Of course you
want the thermistor in some thermal contact with the
circuit.
This 3 component network takes the place of an existing
resistor in the circuit. If your XA amplifier drifts negative with
temperature (which is what they usually do), you will want to
put this network in place of the resistor between the Source
and positive supply on the current source transistor which
feeds the differential pair.
😎
compensation, you can always buy a range of thermistors
from Digikey and put them in the appropriate spot.
The thermistor decreases in resistance as the temperature
rises, and you generally want to bracket it with fixed values
of resistance in parallel and in series with it. The proportions
of all three values are tailored to the need, and are usually
easy, if tedious, to determine experimentally. Of course you
want the thermistor in some thermal contact with the
circuit.
This 3 component network takes the place of an existing
resistor in the circuit. If your XA amplifier drifts negative with
temperature (which is what they usually do), you will want to
put this network in place of the resistor between the Source
and positive supply on the current source transistor which
feeds the differential pair.
😎
It seems possible to add the trimpot to the Hifizen board.
Cutting 2 traces and cleaning a few spots to solder it on to.
Just to be sure, do you mean R23 and R25 on attached schematic?
The 2.2k (and 10k) resistors mentioned is R46 and R47, right?
Just to avoid any confusion😉
Ouch😱
Steen🙂
Thanks Nelson!
I had tossed the idea around in my mind previously, this gives me more confidence. 🙂
By the way, which component(s) should I thermally couple the thermistor to? The current source fet, the main heat sinks, diff pair, or just the ambient air inside the housing? My guess is the current source fet.😕
(the more variables I can eliminate, the more likely I am to get it working more effectively...)
Thanks,
Jeff
I had tossed the idea around in my mind previously, this gives me more confidence. 🙂
By the way, which component(s) should I thermally couple the thermistor to? The current source fet, the main heat sinks, diff pair, or just the ambient air inside the housing? My guess is the current source fet.😕
(the more variables I can eliminate, the more likely I am to get it working more effectively...)
Thanks,
Jeff
steenoe said:
Yes, although I don't see a mention of 10K. BTW, you should
seek to maximize the resistances of R46 and R47, so it's
helpful to use thermal compensation to achieve this.
yup said:
Pick a spot and then adjust your compensation resistances
for the thermistor in that spot. Don't forget to warm the amp
up with the cover on.
Nelson Pass said:
Papa-just one qu(ir)estion :
did you use thermistors in this places in your factory X amps?
(humble me trying to learn if this trick is another one of "good enough for production"...........)
- Status
- Not open for further replies.