Sorry if I was unclear, and thank you for verifying. Yes, I'm speaking strictly of the Alpha20 4R, as I don't really need the extra oomph of the BB, and really would like to avoid active cooling.
Thanks so much for your replies Anand and Hugh. I think you might be right on the heat and compromise. I checked the speakers I'm planning to build and my current ones and they both seem relatively well-behaved 8-ohm speakers. I just worry I might come across some life changing 4-ohm design... However, I doubt it as high-efficiency and 8-ohms seems to be a nearly settled deal these days.
Do you have a shortcut or reading material on the thermal math for this? I went electrical (the really tiny kind) not mechanical, so I don't have much knowledge on the subject of TDP. I hate to keep asking which DiyAudio case is best, I'd like to just figure it out myself.
The 1.96A build looks to be what I'd like to go ahead with. The feedback tap is okay before the sense resistor as the PCB's have it, right? I didn't notice anything weird in simulation moving it from the output.
Thank you very much for the quick replies!
The preamp in active mode is less than a few hundred ohms (ECC99 running full tilt), as it can drive higher-Z headphones with no issues. The passive mode where only the attenuator is in the chain was my worry (I guess I added the active option for a reason...).
However, I have run some Spice sims across the stepped range and see that the passive 20K ladder/series design works just fine with this design including a thevenin equivalent of my various source outputs -- bode seems smooth from 20-20k in AC simulation. I don't think I will worry much about raising the input impedance now. Do you think I was missing anything?
A nice stereo microscope and a steady hand works wonders. I've done 0201 devices, but really don't like it. I use 0603 as minimum just to save time wherever possible, and 0805 minimum when I might modify without magnification.
I agree they have improved quite a bit, but there's still the lingering high dissipation factor of electrolytics, and I'd prefer my capacitors gobble up as few electrons as possible. I've noticed that in critical positions (speaker crossovers especially), even the new ones aren't nearly as good as film.
In tuning C111, is it sufficient to look at the low frequency response on the Bode plot or are there unmodeled real-world effects that this is also compensating for?
I was speaking more of the brief experiment back around post #1717, #1724, etc. And upon finding it again, I think I see your response:
Effectively, I guess the resolution was "It works fine, don't touch it. The amp is always below unity gain at this point anyways." However, I did notice something similar was implemented in the BB, and I was wondering if there really was a resolution for the 20?
I'll let XRK chime in on suggested transformers for his good looking SLB design in conjunction with the 20 8R/4R. The bridge controller is a very trick circuit that keeps the heat down and helps compensate the Cap-Mx voltage drop. All that's really necessary is the expected Vdrop across the SLB from ideal 1.414*AC.
Any further comments or suggestions on what I've said would be appreciated. I'm pretty convinced to pull the trigger on this design. I always like to think about future source/speaker projects additions and how things would work together with what I make.
And of course, Happy Father's Day to all the dads out there!
-Matt
Hi RL,
To calculate the heatsinks, look for their degC/W factor. For example 0.8C/W for say a 5U Dissipante (I am just making up an approximate number). Then calculate your thermal dissipation. Say, +/-27v rails and two devices so 27v x 1.94amp = 53W per device.
Temp rise of heatsink is 0.8C/W x 53w = 40C. Add that to ambient temp of say 22C and you get 62C. That’s a tad over “acceptable” standard of 55C. That’s the temp that you won’t burn your self because you have 2 seconds to retract your hands. The MOSFETs are probably ok.
So maybe look for heatsink with 0.6C/W rating. That gives 54C heatsink temp - perfect range.
As far as trafos go - it’s tricky because the rated voltage is under no load. With 2 to 4 amps it’s sags typically 5v. Now add 3v for cap Mx and you need 8v headroom on top of desired rail voltage under load. If 27v is desired then 27+8 = 36v. Now take 36v divide by 1.41 and that’s the AC voltage rating you need assuming you get a trafo rated at least 2x to 3z the VA rating you need. So a 300VA to 400VA 25v trafo is what you need for 27v DC under Class A load with SLB.
To calculate the heatsinks, look for their degC/W factor. For example 0.8C/W for say a 5U Dissipante (I am just making up an approximate number). Then calculate your thermal dissipation. Say, +/-27v rails and two devices so 27v x 1.94amp = 53W per device.
Temp rise of heatsink is 0.8C/W x 53w = 40C. Add that to ambient temp of say 22C and you get 62C. That’s a tad over “acceptable” standard of 55C. That’s the temp that you won’t burn your self because you have 2 seconds to retract your hands. The MOSFETs are probably ok.
So maybe look for heatsink with 0.6C/W rating. That gives 54C heatsink temp - perfect range.
As far as trafos go - it’s tricky because the rated voltage is under no load. With 2 to 4 amps it’s sags typically 5v. Now add 3v for cap Mx and you need 8v headroom on top of desired rail voltage under load. If 27v is desired then 27+8 = 36v. Now take 36v divide by 1.41 and that’s the AC voltage rating you need assuming you get a trafo rated at least 2x to 3z the VA rating you need. So a 300VA to 400VA 25v trafo is what you need for 27v DC under Class A load with SLB.
Thanks so much, I had no idea it was that simple on the heat dissipation calculations. Of course, converting aluminum and fin volume to C/W is a different ballgame, but I found a link that covers it for the diyAudio cases:
40mm Heatsink Information – diyAudio Store
Looks like an Antek AS-4224/5 would be in the right ballpark for slightly higher than 24V rails depending on sag.
I have seen other amps (for example ESP P101) designed with the polar input cap orientated like the schematic/PCB has it. I really don't think it will make a difference.
I will put the Silmic II in the second channel to see if I can hear dif between that and the film. Maybe it is just me, but it really bothers me when the cap's lead pitch doesn't match the board; the leads will always have to be longer thus more ESL will be present.
I will put the Silmic II in the second channel to see if I can hear dif between that and the film. Maybe it is just me, but it really bothers me when the cap's lead pitch doesn't match the board; the leads will always have to be longer thus more ESL will be present.
Thanks for all the interest and support for the Alpha Big Boy with Buttah (ABBB) amp! We are almost at 50 boards now. I just verified that the built in solid state relay speaker protection and startup delay circuit works. It works amazingly well. I am really looking forward to this new all-integrated Active Bridge/CRC/CapMx/speaker protection/Alpha amp board come to fruition
The Alpha Big Boy with Buttah (ABBB) 52w Class A Amp GB
Here was SSR test in action using an M2X and XKi speaker:
The Alpha Big Boy with Buttah (ABBB) 52w Class A Amp GB
Here was SSR test in action using an M2X and XKi speaker:
