Thanks, going to give it a go! Bad time to travel; skiing whole weekend, home at monday and from tuesday one week vacation in Spain. So this shot have to wait for over a week.
I just got word that the B.B. boards are done and getting packed for the DHL cargo plane. Should be here in a few days. Big 400VA 25v Trafo is here, CRC PSUs are a go, IXYS MOSFETs and PWM SMTs ordered, Vishay Dale CMF resistors ordered, CPU cooler heat pipes are ready, a BB beta build session coming to a workbench near you soon🙂
I think they are 240mm x 130mm.
Looking really good! My casework went over again, going to toss monoblocks away and modify my "tower" suitable for B.B, it needs some welding again.. Only trafos, heatsinks and capMx pcb's ready and waiting.
Version ALPHA 15W optimised for 4R loads
Danny,
I have done the LTSpice analysis for 15W into 4R loads.
The quiescent is 2.1A
First clip occurs on negative waveform; 2V is lost at the voltage amplifier, and the remainder is lost at the Vgs of the pmos output device.
This is a total of just over 6V, so with a stiff 18V rail it clips at 11.4Vpeak.
It means you get extract 11.1Vp into 4R, which is 15.4W.
This design has THREE output pairs, all IRF240/9240s, but this assumes you are using a 480W large complementary mosfet with transconductance of at least 25 Siemens, and hopefully more. The more the transconductance, the higher the clip voltage. Even so, I would not expect clip to be better than -12Vpk. This is 18W.
I have trimmed R17, the AC current dynamic range resistor. By moving this up and down, you equalise the dynamic current variation of the two output devices, more like classic PP AB operation. Even so, none of the outputs drop below about 280mA at any part of the waveform.
As you bring down the load impedance, you depress the H2 and bring up the H3. THD increases, but as H2 drops the ear perceives 'unnatural' profile, and the sound is not so melifluous. The outputs are running slightly lower heat, around 73W the two, compared to the ALPHA 20 which is optimised for 8R loads. A 4R load is less than ideal for this amplifier.
Cheers,
Hugh
Danny,
I have done the LTSpice analysis for 15W into 4R loads.
The quiescent is 2.1A
First clip occurs on negative waveform; 2V is lost at the voltage amplifier, and the remainder is lost at the Vgs of the pmos output device.
This is a total of just over 6V, so with a stiff 18V rail it clips at 11.4Vpeak.
It means you get extract 11.1Vp into 4R, which is 15.4W.
This design has THREE output pairs, all IRF240/9240s, but this assumes you are using a 480W large complementary mosfet with transconductance of at least 25 Siemens, and hopefully more. The more the transconductance, the higher the clip voltage. Even so, I would not expect clip to be better than -12Vpk. This is 18W.
I have trimmed R17, the AC current dynamic range resistor. By moving this up and down, you equalise the dynamic current variation of the two output devices, more like classic PP AB operation. Even so, none of the outputs drop below about 280mA at any part of the waveform.
As you bring down the load impedance, you depress the H2 and bring up the H3. THD increases, but as H2 drops the ear perceives 'unnatural' profile, and the sound is not so melifluous. The outputs are running slightly lower heat, around 73W the two, compared to the ALPHA 20 which is optimised for 8R loads. A 4R load is less than ideal for this amplifier.
Cheers,
Hugh
Unfortunately, I am unable to copy my name to add to the list
Would you PLEASE someone help me
There are some 420 and 421 s I would like to take 8 PC, please
Thank you🙂🙂
Last edited:
Hi Hugh,
Big thanks for the simulation and thorough analysis !
That's very helpful !
I'll also compare the simulation and the FFT profile to the normal Alpha20 8R,
if it is too much of a compromise I'll rather build the Alpha20 8R and go for biamping.
Regards,
Danny
Big thanks for the simulation and thorough analysis !
That's very helpful !
I'll also compare the simulation and the FFT profile to the normal Alpha20 8R,
if it is too much of a compromise I'll rather build the Alpha20 8R and go for biamping.
Regards,
Danny
Perhaps Hugh can recheck with asymmetric PSU rails - because that was what I did and got good results without the profile tending to go to third harmonic. The clipping occurs first on the negative rail, which is why I made it -24v.
Hi X,
Yes, if you increase the negative rail from -18V by 4V to -22V, you can get clean 15.6Vpk output into 4R (28.1W) which is almost doubling the power.
BUT you have to have two power supplies. One for +18V, and another of -22V. To be honest, you could do it with a double toroidal of 2 x 18Vac, and be done with a couple of rails around 22V. More power, but easier to set up.
But you are dead right, you can supply it this way, and you will get more power.
Cheers,
HD
Yes, if you increase the negative rail from -18V by 4V to -22V, you can get clean 15.6Vpk output into 4R (28.1W) which is almost doubling the power.
BUT you have to have two power supplies. One for +18V, and another of -22V. To be honest, you could do it with a double toroidal of 2 x 18Vac, and be done with a couple of rails around 22V. More power, but easier to set up.
But you are dead right, you can supply it this way, and you will get more power.
Cheers,
HD
One way is to drop the voltage on the upper rail through big series diodes - four can give 2.4v drop.
Hi Hugh, X,
Thanks for the suggestions.
I'll make some simulations with different rail voltages and also changing the gain by changing the R113 feedback resistor,
or are there other options to change the gain ?
Regards,
Danny
It's the least I can do and happy to help out my fellow builders!
Nearly all the devices are now claimed, so I have made the doc read only. I will follow-up by PM within a few days for contact info, etc.
BK
Since you mentioned view only google sheet:
If someone can tell me whether Device ID 85 with Vgs 656, hFE 406 and Device 64 with Vgs 655 and hFE 402 are a good enough match to be used I will put my name in the hat for them.
The same applys for Device ID 49 and 11, Vgs = and hFE 441, 443. If a good enough match together.
That should take care of the devices offered. Will await your reply.
Thanks Kokanee