Thermal conductivity is an intrinsic value. So will be the same for same type of alumina.
The unit for thermal conductivity coefficient is Watts per meter per Kelvin. It is normalized by distance of heat traveling through a thickness. So when you double the thickness - the heat flux will be 1/2. You have to use it as part of the heat flux equation known as Fourier’s Law.
Heat flux Q/(delta time) (Watts) = - k x A x (delta Temp) / (delta Distance)
Negative sign simply means heat flows from hot to cold. K is thermal conductivity coefficient. A is cross sectional area (area of thermal spacer pad), delta distance is thickness.
The unit for thermal conductivity coefficient is Watts per meter per Kelvin. It is normalized by distance of heat traveling through a thickness. So when you double the thickness - the heat flux will be 1/2. You have to use it as part of the heat flux equation known as Fourier’s Law.
An externally hosted image should be here but it was not working when we last tested it.
Heat flux Q/(delta time) (Watts) = - k x A x (delta Temp) / (delta Distance)
Negative sign simply means heat flows from hot to cold. K is thermal conductivity coefficient. A is cross sectional area (area of thermal spacer pad), delta distance is thickness.
Hi guys,
Is it not a good idea to braid the 3 flying leads going to each MOSFET - I thought this reduced the potential for noise to get into the circuit?
Likewise with the flying leads in the SLB?
Andy
mosfet leads
Andy,
Twist the Drain and Source wires together, about one full twist every 20mm.
Then twist the Gate wire in the opposite direction around the twisted pair. Twist just enough to hold the three wires together, about 1 full twist over the full length.
The Source and Drain wires can carry a lot of current and will create a magnetic field. Twisting causes these to cancel out and reduce the tendency to induce low level signals into the input stage. By not twisting the Gate lead in with the other two, (almost) nothing gets coupled into the Gate lead.
Keep the wires short and away from the rest of the wiring if possible, especially the input cable.
Paul Bysouth
Andy,
Twist the Drain and Source wires together, about one full twist every 20mm.
Then twist the Gate wire in the opposite direction around the twisted pair. Twist just enough to hold the three wires together, about 1 full twist over the full length.
The Source and Drain wires can carry a lot of current and will create a magnetic field. Twisting causes these to cancel out and reduce the tendency to induce low level signals into the input stage. By not twisting the Gate lead in with the other two, (almost) nothing gets coupled into the Gate lead.
Keep the wires short and away from the rest of the wiring if possible, especially the input cable.
Paul Bysouth
Hi Andy,
I have also used Keratherm but didn't like it for TO220 devices,
the bolt of TO220 puts too much pressure on one side,
result: the keratherm was cut there because of the sharp edge and made a short circuit.
With Alumina I never have to worry about shorts.
I think Keratherm is more used in automated installation when there's no paste used.
I prefer much more the Alumina with paste than Keratherm.
Danny
I have also used Keratherm but didn't like it for TO220 devices,
the bolt of TO220 puts too much pressure on one side,
result: the keratherm was cut there because of the sharp edge and made a short circuit.
With Alumina I never have to worry about shorts.
I think Keratherm is more used in automated installation when there's no paste used.
I prefer much more the Alumina with paste than Keratherm.
Danny
Last edited:
Good point, Danny - but in my case, I'll be compressing the TO220s between the copper pad of a CPU cooler and the front panel ... so the pressure should be even all over the TO220 case. 🙂
Andy
Sorry, X - I understood the FQA40N25 & FQA36P15 used in the AN were TO220 cases. My mistake.
Andy
X
On the Fischer Electronik website the miniature aggregate coolers look
interesting, forgive the word choice, do you think they could dissipate enough heat for the Alpha, Nirvana, Big Boy amps? If the answer is yes it would facilitate building a 1 or 2 or 3 unit rack mount amp.
On the Fischer Electronik website the miniature aggregate coolers look
interesting, forgive the word choice, do you think they could dissipate enough heat for the Alpha, Nirvana, Big Boy amps? If the answer is yes it would facilitate building a 1 or 2 or 3 unit rack mount amp.
Using a 6 unit sub rack it would be possible to have 4 52 watt class A amps, PCB's mounted vertically. Good thing summer is only 3 months long up here.
Do you have a link to the heatsink?
A 6U rack can hold the 4 amp boards and CPU coolers/fans but not sure if enough room for the trafos and SLB. Maybe an extra deep 400mm rack case. Only way to do it is get dimensions and draw it out in a CAD program. Maybe vertical orientation of trafos can help?
You might consider a separate case for the trafo and SLBs?
Last edited:
Not sure how to give direct link. fischerelectronik.de go to heatsinks then aggregate coolers. The quietest fan has a rating of 17 dba.
A 19" sub rack should have about 16.5" of space to use. Depending on the height of the capacitors (4 Big Boys PCB's mounted vertically) there should be several inches on the side for trafos. Will weigh a ton but for a 2 way active speaker could be end game.
Howdy X, Hugh, et all;
I'm considering building the nirvana, I love my 20w alpha, but would like a bit more power. I'm thinking I'll use an SLB power supply, and can hopefully use my existing alpha 20w case. I have a few points to clarify before I hit X's shop and buy some boards 🙂
The heatsinks I have are 8" tall of this profile:
10.080" Wide Extruded Aluminum Heatsink - HeatsinkUSA
Will those work? If not I have some of the Dell heatsinks, that I could strap fans too, but I would strongly prefer to avoid that, if possible.
As for the SLB, I need a single dual rail board, correct? Stereo build, not dual mono.
I have an Antek 5218 from my Alpha, but it seems this might not be enough. I have an Antek 6224 that I was going to use for a F5 Turbo build, that I could use if that's better suited.
I'll use the DIYAudio store soft start board, as I have one built and ready 🙂
The amp will be used to power my lf/mid-bass horns, so the slight lf bump I read about (iirc) is welcomed, as is the damping factor.
I really like the signature of the Alpha 20w, I used Takman metal films, Clarity Cap 10uF input cap (purple power), and Mills MRA5's for the power resistors. I plan to do the same for the nirvana.
I plan on a separate heatsink for the SLB.
Cheers,
Gable
I'm considering building the nirvana, I love my 20w alpha, but would like a bit more power. I'm thinking I'll use an SLB power supply, and can hopefully use my existing alpha 20w case. I have a few points to clarify before I hit X's shop and buy some boards 🙂
The heatsinks I have are 8" tall of this profile:
10.080" Wide Extruded Aluminum Heatsink - HeatsinkUSA
Will those work? If not I have some of the Dell heatsinks, that I could strap fans too, but I would strongly prefer to avoid that, if possible.
As for the SLB, I need a single dual rail board, correct? Stereo build, not dual mono.
I have an Antek 5218 from my Alpha, but it seems this might not be enough. I have an Antek 6224 that I was going to use for a F5 Turbo build, that I could use if that's better suited.
I'll use the DIYAudio store soft start board, as I have one built and ready 🙂
The amp will be used to power my lf/mid-bass horns, so the slight lf bump I read about (iirc) is welcomed, as is the damping factor.
I really like the signature of the Alpha 20w, I used Takman metal films, Clarity Cap 10uF input cap (purple power), and Mills MRA5's for the power resistors. I plan to do the same for the nirvana.
I plan on a separate heatsink for the SLB.
Cheers,
Gable
Last edited:
Hi Gable,
You will dissipate around 100W for each channel using 27V rails and 1.68A quiescent, pro rata for lower quiescent. So you need to pour heat into your heatsink without incurring more than 25C rise over ambient.
My calculation indicate a 350x150mm heatsink, an MF35 Conrad here in Oz, will just do it. If your American Heatsink is 10.8" tall and 12" long it should be enough for EACH channel providing the fins are vertical for good airflow. X will have better information for your US sourcing; he may answer here too.
Good luck with your build and thanks for building the original 20W Alpha!
Hugh
You will dissipate around 100W for each channel using 27V rails and 1.68A quiescent, pro rata for lower quiescent. So you need to pour heat into your heatsink without incurring more than 25C rise over ambient.
My calculation indicate a 350x150mm heatsink, an MF35 Conrad here in Oz, will just do it. If your American Heatsink is 10.8" tall and 12" long it should be enough for EACH channel providing the fins are vertical for good airflow. X will have better information for your US sourcing; he may answer here too.
Good luck with your build and thanks for building the original 20W Alpha!
Hugh
Excellent, thank you Hugh.
Thank you for designing it, the Alpha 20 is a fantastic amplifier, certainly one of my favorites.
Gable,
The idea was to do a simple, easy to build amp with minimum active devices in the chain. The ALPHA met the requirement but fell short of power and efficiency. The Nirvana was the next step, to almost double power, with almost double efficiency and no less of sound quality. When you compare it with the original you will know if this is successful; I hope so..... certainly it's a lot more efficient and the signal chain is almost the same length.
Texas has similar temperatures to southern Australia so we share heatsink issues. Forced air cooling gives about three times more efficiency dissipation, but most people are leery of the fans. After all, good audio is about quiet!
Ciao,
HD
The idea was to do a simple, easy to build amp with minimum active devices in the chain. The ALPHA met the requirement but fell short of power and efficiency. The Nirvana was the next step, to almost double power, with almost double efficiency and no less of sound quality. When you compare it with the original you will know if this is successful; I hope so..... certainly it's a lot more efficient and the signal chain is almost the same length.
Texas has similar temperatures to southern Australia so we share heatsink issues. Forced air cooling gives about three times more efficiency dissipation, but most people are leery of the fans. After all, good audio is about quiet!
Ciao,
HD
Hi Pcgab,
Hugh is right, use one 8in x 10in heatsink per MOSFET. So four of those and a separate heatsink for SLB and you are good. I used the 6224 for the AN and it was perfectly sized to give you 28.5v which is actually a solid 41w I think.
We have a soft start coming soon. I have built gen 1 but found it could be improved so waiting for gen 2 PCB to arrive. It uses a low RDSon SSR to bypass four 100ohm resistors in parallel. Probably 3 weeks away from being available for sale.
Hugh is right, use one 8in x 10in heatsink per MOSFET. So four of those and a separate heatsink for SLB and you are good. I used the 6224 for the AN and it was perfectly sized to give you 28.5v which is actually a solid 41w I think.
We have a soft start coming soon. I have built gen 1 but found it could be improved so waiting for gen 2 PCB to arrive. It uses a low RDSon SSR to bypass four 100ohm resistors in parallel. Probably 3 weeks away from being available for sale.