I hand wound some toroids
I wound some toroids on T106-2 cores. They held about 13uH. The sound is much better than stock but falls shy of my earlier filter with 4uH air cores. Now I will try another set of toroids at 6uH, where I can still get a decent filter response, to see if it is the low value and resultant peak at 30k that I like better or if the air core is the key to the better sound.
.
I wound some toroids on T106-2 cores. They held about 13uH. The sound is much better than stock but falls shy of my earlier filter with 4uH air cores. Now I will try another set of toroids at 6uH, where I can still get a decent filter response, to see if it is the low value and resultant peak at 30k that I like better or if the air core is the key to the better sound.
.
Last edited:
Input cap
I have'nt gotten around to A B ing the little smd input cap (.01uf?) but they will probably just get tossed.
I don't hear any hissing from my listening position 10ft away. One thing I'll say for this amp is it's very quiet (w/ all the gain set on OFF anyway).
What I want to know is: what is the correct zobel for 4 ohm speakers? the 10r/.47 is for 8ohm impedance is it not?
Hi Audio1st,
I just wanted to say a big Thank You for your great post and illustrations. I've been reading up in the D class threads and keep coming across your wonderful diagrams and post's. I found them to be meaningful and very helpful.
Please continue this, I think I can speak for other newbies gratitude that your help encourages us to take on some of these projects.
Cheers,
Mr macgee
I wonder what these sound like
Bristol, UK, 5th Oct: Audium Semiconductor has launched an audio power amplifier IC which, at normal listening levels[1], is 20 times more efficient than competing devices, such as Class D[2] amplifiers, without compromising audio quality. The AS1001 operates from a nominal 1.5V power supply and delivers 100W peak power output. The amplifier is so efficient that battery-powered amplified loudspeakers can run for up to 10 months on a set of four ‘C’ batteries, playing for three hours per day, vastly reducing CO2 from power generation and pollution from battery production[3].
http://www.audiumsemi.com/news002.php
Bristol, UK, 5th Oct: Audium Semiconductor has launched an audio power amplifier IC which, at normal listening levels[1], is 20 times more efficient than competing devices, such as Class D[2] amplifiers, without compromising audio quality. The AS1001 operates from a nominal 1.5V power supply and delivers 100W peak power output. The amplifier is so efficient that battery-powered amplified loudspeakers can run for up to 10 months on a set of four ‘C’ batteries, playing for three hours per day, vastly reducing CO2 from power generation and pollution from battery production[3].
http://www.audiumsemi.com/news002.php
Assume T-Amp efficient is around 80%, then 20 times more efficient is about 1600%?
Hum... it's better than a power generate station, it seen that it general power from nothing.
100W peak power output, assume the loading is 4 ohms, then current is 5A, deliver from a 1.5V power supply? if the power rail is step up, then it's still possible, but if deliver over a 1.5V battery then most people will like to know which brand of battery is it.
Hum... it's better than a power generate station, it seen that it general power from nothing.
100W peak power output, assume the loading is 4 ohms, then current is 5A, deliver from a 1.5V power supply? if the power rail is step up, then it's still possible, but if deliver over a 1.5V battery then most people will like to know which brand of battery is it.
Well if something uses a certain amount of energy, and something else uses 1/20th as much to do the same thing, then the second thing might be said to be 20 times as efficient, no?
Not that I necessarily believe their claims, but it doesn't have to start actually producing energy to be 20x as efficient.
Not that I necessarily believe their claims, but it doesn't have to start actually producing energy to be 20x as efficient.
Do you have the option to try this chipset with batteries? This way you could rule out any raunchy converter noise that might be getting through. Or, if possible, try another adapter and see if its better or worse.
6uH toroids vs 13uH
I tried the smaller filter with 6uH T106-2 toroids and it does sound better than the 13uH filter but is still a long way off of the sonics of the 6uH air cores. The jury is still out on the .47uf/7ohm zobel. I tried it with and without. The difference is very slight and might be better without. I might try type 0 cores next to find out if the ultra low AL will start to sound more like an air core or even some kind of wood core toroid to see if the shape alone will still help reduce the EMI enough.
I tried the smaller filter with 6uH T106-2 toroids and it does sound better than the 13uH filter but is still a long way off of the sonics of the 6uH air cores. The jury is still out on the .47uf/7ohm zobel. I tried it with and without. The difference is very slight and might be better without. I might try type 0 cores next to find out if the ultra low AL will start to sound more like an air core or even some kind of wood core toroid to see if the shape alone will still help reduce the EMI enough.
Shielding
.
I am using the passive crossovers in my Usher two ways for the amp trials. The sims I have been running are very different when adding a moderate amount of driver inductance of 200uH to the purely resistive load most people model with. With the inductance, the filter rings like crazy at 70k with no zobel. Even with the zobel there is still a rise at 50k. With the driver inductance, the effects of the zobel don't change much for 4 or 8 ohms or inductances of 50uH and higher so I don't believe the exact zobel values need to tailored to the speaker as much as people think when using these class D amps because they are following a steeper low pass filter than anything you will find in a class AB amp. The steeper filter into the inductive loudspeaker load seems to lock the zobel to the peak regardless of the 4 or 8 ohm speaker impedance so it is just a question of making the zobel work with the given filter. The trade offs seem to be the remaining rise versus the current in the zobel and how far towards the audio band it's power robbing influence intrudes.
I haven't tried shielding the complete amp but from the infuriated responses anyone mentioning air cores gets, I assumed it still wouldn't be enough.
.
I am using the passive crossovers in my Usher two ways for the amp trials. The sims I have been running are very different when adding a moderate amount of driver inductance of 200uH to the purely resistive load most people model with. With the inductance, the filter rings like crazy at 70k with no zobel. Even with the zobel there is still a rise at 50k. With the driver inductance, the effects of the zobel don't change much for 4 or 8 ohms or inductances of 50uH and higher so I don't believe the exact zobel values need to tailored to the speaker as much as people think when using these class D amps because they are following a steeper low pass filter than anything you will find in a class AB amp. The steeper filter into the inductive loudspeaker load seems to lock the zobel to the peak regardless of the 4 or 8 ohm speaker impedance so it is just a question of making the zobel work with the given filter. The trade offs seem to be the remaining rise versus the current in the zobel and how far towards the audio band it's power robbing influence intrudes.
I received my board today. Took about 10 days by UPS Saver to get to You-Taw USA. Overall, the PCB traces don't look hefty enough to handle the currents involved. My positive fan lead has so much insulation peeled off that the negative wire was touching the one end of R-2. Probably won't power it up till tommrrow.
Mark
Mark
Yes, am well aware of that. T.I. does reccomend 2 oz. copper traces for their chips. At the terminals of these chips where DC goes in and sound comes back out the current draw (at least on the high powered TI stuff) might be measured be in amps. Now this is much lower power but the traces still seem very marginal(thin). They've flowed solder along the rails on the bottom already to beef that area up a bit. I;m not really complaining though how could one for 50 bucks. The thing works and sounds passable out of the box.
Mark
Mark
Would these inductors work well to build the new output filter circuit? They are Coilcraft RFB1010 Series Power Inductors Specs are... 10 µH 0.023 DC res. 23 mhz SRF 6.10 Amps
Linky Dinky: Coilcraft RFB1010 Leaded Power Inductors
Thanks in advance for any advice... there is such a plethora of inductors out there I don't know where to start looking...
Mark
Linky Dinky: Coilcraft RFB1010 Leaded Power Inductors
Thanks in advance for any advice... there is such a plethora of inductors out there I don't know where to start looking...
Mark
Why such a large value for the rail caps. I currently have 2000uf. What gains are made by adding more?
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.