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?
I am glad you brought this up. I have started to play with that after seing that edwin had put pots in for r 12 and 34. Considering the kinds of distortion reduction and efficiency increase afforded by the active CS, this is probably the most important part of the circuit to be looked at, once the DC issue are settled.
Clipping should be symmetrical if the ac gain is set to be the same on both sides. If clipping is asymmetrical the current gain is probably off on one side.
I use a differential probe so I can see that directly. I am not sure but it may not be the same if you use a se probe and measure each output side to gnd, I haven't done that in a while and I don't remember what happens. One could always use two se probes and add the waves on the scope, I think I have been through this discussion before.
You may also find that you can maximize the output by playing with the r 12 and 34. I can get 15V rms output at 15V rails with what I believe to be 50% gain setting. This value can be pushed up to 20Vrms, so maybe that is the real 50% current gain?? I don't know. I am still trying to figure out if I did it right.
I hope I finally can help someone,
As far as I believe setting the ACS is very straight forward. The output resistors (the lot in parallel which are also with the normal aleph's) should put out twice the current than the negative resistors. I will explain my case:
I have 8 times 0.56 as output resistors and 4 negative mosfets (as also 4 current source mosfets which totals to 16 mosfets per channel) . So when you measure the AC RMS voltage drop across the output resistors and across one negative 0.56 resistor you can calculate the amount of current running through these stages. I use a 1Khz sin wave with a output of 9V AC RMS as a base to measure the working of the ACS.
I get for instance with a 1Khz 9V AC RMS:
0.080V across the 8 times parallel output resistors and
0.078V across one of the negative output resistor.
This means:
0.080 / (0.56 / 8) = 1.14A
0.078 / (0.56 / 4) = 0.557A
So I have about (1 - (0.557 / 1.14)) * 100 = 52%
The CS puts out 1.14A and the negative side puts out 0.557A so the CS is responsible for about 52% of the output!!!
So the CS can output a bit more than double the output of current. The higher gain the more the CS can multiply the current.
Edwin
As far as I believe setting the ACS is very straight forward. The output resistors (the lot in parallel which are also with the normal aleph's) should put out twice the current than the negative resistors. I will explain my case:
I have 8 times 0.56 as output resistors and 4 negative mosfets (as also 4 current source mosfets which totals to 16 mosfets per channel) . So when you measure the AC RMS voltage drop across the output resistors and across one negative 0.56 resistor you can calculate the amount of current running through these stages. I use a 1Khz sin wave with a output of 9V AC RMS as a base to measure the working of the ACS.
I get for instance with a 1Khz 9V AC RMS:
0.080V across the 8 times parallel output resistors and
0.078V across one of the negative output resistor.
This means:
0.080 / (0.56 / 8) = 1.14A
0.078 / (0.56 / 4) = 0.557A
So I have about (1 - (0.557 / 1.14)) * 100 = 52%
The CS puts out 1.14A and the negative side puts out 0.557A so the CS is responsible for about 52% of the output!!!
So the CS can output a bit more than double the output of current. The higher gain the more the CS can multiply the current.
Edwin
Member
Joined 2002
but yes, pics is what we want. 
Netlist said:You surely did a nice job, writing out the practice of adjusting the ACS.
I admit it took me a long time to figure out how to do it, a few weeks ago.
Now tell us something about the sound...
/Hugo
I only can test drive one channel at a time. Because I missing a part of my second power supply (10W 0.22 resistors). Hopefully I will get them tomorrow. Then she must prove herself for a few days on less expensive speakers. So I will probably test drive her on the big B&W Nautilus 802's this coming weekend if all is going well. I hope to post a piece about the sound the next week, and perhaps a comparison with my 'old' Aleph 4.
The pictures are more or less the same as on page 1 only two channels.
Nothing new. I will post more pictures when my case is ready which wil take at least 2 to 3 weeks more (have to order the alu panels soon)! Have some patience!Edwin
P.S. Ah, what the h*ll, I will post a total show of both channels when they are downstairs in the living room when I have the last parts...
Member
Joined 2002
Thanks Edwin,
I noticed that the pots for the sensing resistors are rather touchy, I had a really hard time with adjustment using a single turn pot but everything is fine now, I have already put in the fixed resistors and everything is stable.
I would add that, in order to get symmetrical clipping as it should be, the absolute dc (output to gnd) should be allowed to settle down to its steady state value, and if not 0, it should be set to 0 with the diff. ccs pot before proceeding to adjust the ac gain.
I noticed that the pots for the sensing resistors are rather touchy, I had a really hard time with adjustment using a single turn pot but everything is fine now, I have already put in the fixed resistors and everything is stable.
I would add that, in order to get symmetrical clipping as it should be, the absolute dc (output to gnd) should be allowed to settle down to its steady state value, and if not 0, it should be set to 0 with the diff. ccs pot before proceeding to adjust the ac gain.
grataku said:
I use 10 turn bournes trimpots, they are not twitsy whatsoever...
Why do you need symmetrical clipping? Is it a fact of live that with 50% you get symmetrical clipping? I think the 1khz sin 9V RMS AC and then measure the current running is a better and more precise method. Just my 2 cents...
Edwin
Edwin Dorre said:
I use 10 turn bournes trimpots, they are not twitsy whatsoever...
Why do you need symmetrical clipping? Is it a fact of live that with 50% you get symmetrical clipping? I think the 1khz sin 9V RMS AC and then measure the current running is a better and more precise method. Just my 2 cents...
Edwin
I don't see why clipping should be anything but. If all the DCs are set to about 0, the current gain is set to the same value on both sides and equal current is flowing in each side of the circuit.
After setting the correct gain at lower voltages I brought the amp up to clipping and that was symmetric. Of course at the clipping point the DC offset goes crazy but that is ok.
hmmm...
I just connected one channel of the Aleph-X with a real Pass Labs Aleph P (not diy...) with a balanced cable. And yet a mystery, I get double the diff DC offset at the outputs.
So this channel has 30mV diff DC offset when there is no balanced cable connected to the Aleph-X and 60mV with the cable connected from the pre-amp to the Aleph-X. Still no amount to be scared off, however it is anoying...
Why o why? Do I need the 2.2uf caps between the + and - of the balanced plug? Strange because an Aleph P has on both the out+ and out- 4 times 10uf caps.
Who can say something clever about what to do... Aleph X is full of mysteries...
Tomorrow I will get my 8x 10W resistor so I can finish the second channel! It does sound promising listening mono though...
Edwin
I just connected one channel of the Aleph-X with a real Pass Labs Aleph P (not diy...) with a balanced cable. And yet a mystery, I get double the diff DC offset at the outputs.
So this channel has 30mV diff DC offset when there is no balanced cable connected to the Aleph-X and 60mV with the cable connected from the pre-amp to the Aleph-X. Still no amount to be scared off, however it is anoying...
Why o why? Do I need the 2.2uf caps between the + and - of the balanced plug? Strange because an Aleph P has on both the out+ and out- 4 times 10uf caps.
Who can say something clever about what to do... Aleph X is full of mysteries...
Tomorrow I will get my 8x 10W resistor so I can finish the second channel! It does sound promising listening mono though...
Edwin
Edwin Dorre said:
2.2µF BETWEEN + and - ?
I guess you mean in series with the input resistors.
I just read an interesting post in the AlehX thread from Carpenter, followed by a reply from Nelson about R19/29
This AlephX thread is still full of surprises and probably the best reference for all our problems.
Hugo – Just found something clever instead of saying it…
Netlist said:
2.2µF BETWEEN + and - ?
I guess you mean in series with the input resistors.
I just read an interesting post in the AlehX thread from Carpenter, followed by a reply from Nelson about R19/29
O yes,
I need to read and understand more of electronics. Everytime when I think I have read enough something is lurking around the corner
!yes, in series with + and the resistor and with - and the resistor...
I will have to re-read this ten times or so, I feel stupid:
"Lowering the value of resistance to ground at the Gates of the input diff pair affects a couple of things:"
I presume this means a resistor with a lower value like the 10K instead of greys 68k? I have 10K in place ...
"It lowers the tendency for common mode offset to be amplified positively."
I do not know what these words mean
. Amplyfing common mode offset does not sound good. So a lower value of R19/29 is a good thing... So 10k it is..."It lowers the open loop gain of the circuit, and thus the amount of feedback."
Ok, the less feedback the better is a common feeling... All the ads say; "we use no feedback". So no feedback is good... So 10K it is...
But... still...
Edwin
Edwin,
don't worry about it. I had the same problem. As I understand it happens because there is a piece of cable with finite lenght connecting the output cap to the input of the AX and this cable has it's own impedance.
The choice is yours try a different cable, lower the input impedance of the ax or add the input caps. I have chosen to live w/o the additional caps.
Oops, beat me to the reply. you have the 10 k already, sorry. Just connect the thing and listen to it! It'll make you happy
don't worry about it. I had the same problem. As I understand it happens because there is a piece of cable with finite lenght connecting the output cap to the input of the AX and this cable has it's own impedance.
The choice is yours try a different cable, lower the input impedance of the ax or add the input caps. I have chosen to live w/o the additional caps.
Oops, beat me to the reply. you have the 10 k already, sorry. Just connect the thing and listen to it! It'll make you happy
Edwin Dorre said:
Let's clarify the terminology:
a stereo AX amp has two channels, each channel has two sides, on each side there are two sets of output mosfets, the active CCS set and the gain set.
taco
On my working prototype I have one mosfet per set and that mosfet is an irfp044. For my high power version I plan to use six mosfets per set so a total of 24 mosfets/channel. For price considerations I will use the irfp240 although I would prefer the higher transconductance model the irfp140 but that's twice the price at digikey.
Edwin
I haven't thought about it much but I was puzzled by the values of source resistors you have chosen. Basically I come up with a much higher value for bias current than what you declared. I 'll try that again later.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.