Hi everybody !
I'm going to build my own Aleph3 but with some modifications.
As a matter of fact I do not need so big power ( 30W on 8 ohm)
so please comment ( correct) things I'm going to to:
1) remove one parallel stage
a) Q107, Q109, R117, R121, R119, R123
2) Change Q106 and Q108(IRF 244) into IRFP 044
3) Change R120 and R122 from 0.47 to 0.33
4) Change Q105 and Q104 to BC550
5) I suppose that R114 will need to change but as usual with
Aleph current source one has to experiment
6) decrease capacitance of power supply filter by half from
8x22000uF to 4x22000uF
I have question about input stage. It is difficult for me to get Q101 and Q102 as a balanced pair. Have You any experience
whether IRF9610 taken "from shelf" have bad influance on sonic impression.
regards
JacekPlacek
I'm going to build my own Aleph3 but with some modifications.
As a matter of fact I do not need so big power ( 30W on 8 ohm)
so please comment ( correct) things I'm going to to:
1) remove one parallel stage
a) Q107, Q109, R117, R121, R119, R123
2) Change Q106 and Q108(IRF 244) into IRFP 044
3) Change R120 and R122 from 0.47 to 0.33
4) Change Q105 and Q104 to BC550
5) I suppose that R114 will need to change but as usual with
Aleph current source one has to experiment
6) decrease capacitance of power supply filter by half from
8x22000uF to 4x22000uF
I have question about input stage. It is difficult for me to get Q101 and Q102 as a balanced pair. Have You any experience
whether IRF9610 taken "from shelf" have bad influance on sonic impression.
regards
JacekPlacek
I can help on the last question. Buy 10 9610's. the price will be almost the same by bulk pricing. Pick 2 closest pairs and save 4 for junk box. The current source doesn't need to match anything, so use the two most out of range ones here. ( I assume you're making 2 channels)
You should go take a look at the Volks Amp 30 at the Pass labs web site.
Look for R-1 which is located from the source to the drain of Q-2. This resistors final value is selected to attain as close to 0 volts as is possible at the output of the amplifier. This sort of arrangement would also allow your input fets to be slightly mis-matched. *IMPORTANT*---- As long as the 9610's are from the same batch code...or cut from the same die you'll probably find they they are plenty close enough to use. In the States Digi-Key is always nice about supplying fomr the same batch code if you remember to ask. The addition of R-1 will allow you to trim further if need be. As an alternative method you can first substitute different 9610's in the Q-2 position for lowest offset and then trim with the addition of R-1.
I highy reccomend installing a small 10 turn trimmer pot in there and adjust for 0 volts at the output, then measure the pot and replace the pot with a fixed resistor. Be sure to allow at least a 1 hour warm up before making offset measurements as these amps can drift around somewhat till operating temperature stabilizes. An hour is pretty much the the standard stabilization time for these amps.
No matter which method you choose to match up the fets you'll find there is little sonic difference. Just try to get your DC offset to around 50mv or less. Both of my 2's stabilize at between 20 to 30 mv.
Once your amp is up and operating you can also substitute different 9610's and measure for lowest distortion. Just remember to keep an eye on that offset and adjust R-1 as necessary!
Hope this isn't too confusing and helps you out a bit!
Mark
Look for R-1 which is located from the source to the drain of Q-2. This resistors final value is selected to attain as close to 0 volts as is possible at the output of the amplifier. This sort of arrangement would also allow your input fets to be slightly mis-matched. *IMPORTANT*---- As long as the 9610's are from the same batch code...or cut from the same die you'll probably find they they are plenty close enough to use. In the States Digi-Key is always nice about supplying fomr the same batch code if you remember to ask. The addition of R-1 will allow you to trim further if need be. As an alternative method you can first substitute different 9610's in the Q-2 position for lowest offset and then trim with the addition of R-1.
I highy reccomend installing a small 10 turn trimmer pot in there and adjust for 0 volts at the output, then measure the pot and replace the pot with a fixed resistor. Be sure to allow at least a 1 hour warm up before making offset measurements as these amps can drift around somewhat till operating temperature stabilizes. An hour is pretty much the the standard stabilization time for these amps.
No matter which method you choose to match up the fets you'll find there is little sonic difference. Just try to get your DC offset to around 50mv or less. Both of my 2's stabilize at between 20 to 30 mv.
Once your amp is up and operating you can also substitute different 9610's and measure for lowest distortion. Just remember to keep an eye on that offset and adjust R-1 as necessary!
Hope this isn't too confusing and helps you out a bit!
Mark
Thanks Mark and Brian. I think I connect both suggestions.
As a matter of fact me and my colegue build own aplifiers so I will buy about twenty IRF9610 and find best 4 pairs. In further step I will trim R-1.
Regards
Jacek
As a matter of fact me and my colegue build own aplifiers so I will buy about twenty IRF9610 and find best 4 pairs. In further step I will trim R-1.
Regards
Jacek
The next price break usually is at 25 pcs. That's one tube for TO247 cases. If you buy a tube, you should get a sealed tube as it comes from the manufacture.
why?
Why do you want to cut the power? Less heat? Smaller transfomer?
Build a mini-A. I have been using 15-0-15 230 VA transformer with a C-R-C, 17 volt DC with bias @ 3amps per channel. About 105 watts total. It sounds very nice and much easier to carry than my A-30.😉
Why do you want to cut the power? Less heat? Smaller transfomer?
Build a mini-A. I have been using 15-0-15 230 VA transformer with a C-R-C, 17 volt DC with bias @ 3amps per channel. About 105 watts total. It sounds very nice and much easier to carry than my A-30.😉
MikeW I ask You too. Why do You need such big power . In typical
home room power about 2x3-:-2x5 Wats creates very loud music.
That is why I'm going to remove two of four output transistors,
change last two transistors into IRF040N and increase bias current
Regards
Jacek
home room power about 2x3-:-2x5 Wats creates very loud music.
That is why I'm going to remove two of four output transistors,
change last two transistors into IRF040N and increase bias current
Regards
Jacek
Forgive Mike I did not catch You correct last time. As a matter of fact "mini-A" it is Your Aleph3 with lowered power supply ?
I suspect You had to correct some resistors values in "Aleph current source ( R114, R120, R121, R122, R123). I simply have wanted remove components which in my opinion are not needed or may be replaced when we will lower needed power. And that is why instead of four IRF 244 I'm going utilize two IRF040N with increased bias current. R120 and R122 I will change to 0.33 and remove two of four output current measure resistors (R126,R127);
Some modification of R114 will be needed too. Capacitors in power supply may be lowered too.
What do You think about it
Regards
Jacek
I suspect You had to correct some resistors values in "Aleph current source ( R114, R120, R121, R122, R123). I simply have wanted remove components which in my opinion are not needed or may be replaced when we will lower needed power. And that is why instead of four IRF 244 I'm going utilize two IRF040N with increased bias current. R120 and R122 I will change to 0.33 and remove two of four output current measure resistors (R126,R127);
Some modification of R114 will be needed too. Capacitors in power supply may be lowered too.
What do You think about it
Regards
Jacek
Decrease power?
You need lower voltage rails. If you are running your Aleph 3 with .47 source resistor with .47 volts across them there is one amp per Mosfet. Two amps total. Now take out one output Mosfet. Change the source resistors to .33 and raise the bias to .66 volts you have 2 amps bias per one Mosfet. Two amps total.
I also asked the same question about decreasing capacitance. After building a few different Alephs with different types of supplies, use whatever you can fit in the box. More bass. 88000uf per rail is a good start. 😉
You need lower voltage rails. If you are running your Aleph 3 with .47 source resistor with .47 volts across them there is one amp per Mosfet. Two amps total. Now take out one output Mosfet. Change the source resistors to .33 and raise the bias to .66 volts you have 2 amps bias per one Mosfet. Two amps total.
I also asked the same question about decreasing capacitance. After building a few different Alephs with different types of supplies, use whatever you can fit in the box. More bass. 88000uf per rail is a good start. 😉
I used a straight 120,000 mfd per rail in my Mini. Thgere is no audible hum with your ear right up to the speaker. Panasonic makes a series of physically small but really large value caps that will allow you to keep your Mini very small. These caps are only 80mm long by about 35mm in diameter!!
Mark
Mark
But what about utilize LC PI filter in power supply which Nelson sucessfully utilized in ZEN ? I realize it can in some circumstances degradate music during very strong pulses ( if it is really true in topology based on current source) but typically we do not need whole power( only small part of it) so maybe nobody will be able to hear it ?
I consider utilize for each supply ( plus and minus )asymethric PI filter. From diode bridge one capacitor 10000uf , further inductance about 2mH and further three capacitors 10000.
It is much cheaper than solution.
What do You think of that ?
Regards
Jacek
I consider utilize for each supply ( plus and minus )asymethric PI filter. From diode bridge one capacitor 10000uf , further inductance about 2mH and further three capacitors 10000.
It is much cheaper than solution.
What do You think of that ?
Regards
Jacek
- Status
- Not open for further replies.