Heat sinks at high bias
(been offline for too long!)
For that bias, what rating of heat sinks would be needed?
Cheers
(been offline for too long!)
For that bias, what rating of heat sinks would be needed?
Cheers
This is a total dissipation of 72 x 0.75 = 54 watts per channel. (NOTE MY CORRECTION, HELL, CAN'T DO MATHS!!)
Therefore, each heatsink, one for channel, must move 54 watts which at 0.4C/w is an increase of (73 x 0.4) = 22C over ambient.
If you listen in a room at 25C, which is pretty warm in fact, this means a 0.4C/watt heatsink, which is about 10" long and 3" tall, will run at 47C.
This is just hot enough to leave your finger on it for about 20 seconds. At 65C you can touch the heatsink for no more than 2 seconds. So 47C is acceptable.
OTOH, if it is 35C in the room on a hot day with no AC, turn off the FetZilla. You don't want to fry you FetZilla since the junction is at least 25C ABOVE the heatsink temperature..........
Cheers,
Hugh
Therefore, each heatsink, one for channel, must move 54 watts which at 0.4C/w is an increase of (73 x 0.4) = 22C over ambient.
If you listen in a room at 25C, which is pretty warm in fact, this means a 0.4C/watt heatsink, which is about 10" long and 3" tall, will run at 47C.
This is just hot enough to leave your finger on it for about 20 seconds. At 65C you can touch the heatsink for no more than 2 seconds. So 47C is acceptable.
OTOH, if it is 35C in the room on a hot day with no AC, turn off the FetZilla. You don't want to fry you FetZilla since the junction is at least 25C ABOVE the heatsink temperature..........
Cheers,
Hugh
Last edited:
I have been playing around with this "fetzilla" design on & off for about three years now and I am very pleased with how it has turned out. My version is DC linked with a servo & composite o/p pairs plus some more subtle mods. I have not heard either Hugh's version or his Naksa ( which I'm sure must be very good ) but I am not aware of any shortcomings in the imaging or depth of field department with my fetzilla design as it stands. Recordings made with some space like some old Ella big band songs or Ry Cooder's world music collaborations are reproduced quite holographically.
The subtle mods I refer to have been mostly searching out and eliminating any elements of the circuit that have ANY tendency towards resonance. These circuit elements that have a tendency to "ring" at a particular frequency are excited by a broadband music signal ( but probably not by a 1khz test tone ) and really mess up the sound, including imaging, in a big way. Other resonances can be excited by the action of the power supply.
My personal, but objectively unproven, view is that any reasonable topology can be made to "work" very well and play very nice sounding music if the time is taken to fully optimize each approach. My Fetzilla optimization took me 3 years so far and I'm sure there is still room for improvement.
I think there will always be some difference in "sound" between different well implemented topologies but I also think these differences will be small and probably be a matter of personal preference. Certainly every finished design should be able to reproduced Good Dynamics, Natural Musical Tonality and Spacial Coherence.
mike
Last edited:
Or to put it another way . . .
The fact that Hugh designed his Fetzilla remotely from start to finish, having not ever listened to it but using only the subjective feedback from Greg, and that so many think so highly of this design is a testament to his great knowledge & experience. I know he was never intending it to be a "top of the line" design.
All I'm saying is that with some extra attention on this design there is more potential there which I think, among other things, can meet Hugh's "depth of field" criteria and make it a "high end" contender
The fact that Hugh designed his Fetzilla remotely from start to finish, having not ever listened to it but using only the subjective feedback from Greg, and that so many think so highly of this design is a testament to his great knowledge & experience. I know he was never intending it to be a "top of the line" design.
All I'm saying is that with some extra attention on this design there is more potential there which I think, among other things, can meet Hugh's "depth of field" criteria and make it a "high end" contender
Last edited:
I looked at the range at Conrad Heatsinks, and was thinking of their longer, lower M35-50
4" deep fins * 14" long * just 2" high.
It's 0.43 c/ watt
Now looking for a nice case to accommodate the two boards & heat-sinks
My arithmetic gives a minimum internal size of:
250 wide (the width of two PCBs & two heat-sinks) * 350 deep (the length of the heat-sinks) * just 55 high
4" deep fins * 14" long * just 2" high.
It's 0.43 c/ watt
Now looking for a nice case to accommodate the two boards & heat-sinks
My arithmetic gives a minimum internal size of:
250 wide (the width of two PCBs & two heat-sinks) * 350 deep (the length of the heat-sinks) * just 55 high
Fetzilla case options
At Chassis case item
With the boards mounted side to side, and a series of holes drilled over the heatsinks, these look good:
a) the 80 mm high PRAL-47328R
Rounded corners look good too my eyes
But unnecessarily wide heavy & expensive
b) the 1 unit 65 mm high PRAL-1
Finish either natural or champagne
Or the similar but a little smaller PRAL-1a
At Audio Catalog
The AL-421B
Supplied with a power switch, RCA sockets and an IEC socket
In black or "silver"
$90
And the "AL-Dark Knight"
At Chassis case item
With the boards mounted side to side, and a series of holes drilled over the heatsinks, these look good:
a) the 80 mm high PRAL-47328R
Rounded corners look good too my eyes
But unnecessarily wide heavy & expensive
b) the 1 unit 65 mm high PRAL-1
Finish either natural or champagne
Or the similar but a little smaller PRAL-1a
At Audio Catalog
The AL-421B
Supplied with a power switch, RCA sockets and an IEC socket
In black or "silver"
$90
And the "AL-Dark Knight"
Or if the heatsinks are 300 mm long, and the layout is the normal one with heatsinks on the side, the absolute minimim internal space (with Conrad's fin depth of 48 mm) is 300 wide* 234.
The computer motherboard ATX form factor is 305 * 244, so a PC case may work
if the aim were to have a case of minimal size, this would just fit:
S 3207 Aluminum Amplifier Case Enclosure Amplifier BOX PSU Case | eBay
Do the heatsinks sit right up against the boards?
The computer motherboard ATX form factor is 305 * 244, so a PC case may work
if the aim were to have a case of minimal size, this would just fit:
S 3207 Aluminum Amplifier Case Enclosure Amplifier BOX PSU Case | eBay
Do the heatsinks sit right up against the boards?
You are stressing the heatsink, within limits certainly, but it's working hard.
You MUST ensure that the fins can be exposed to the external of the case.
Otherwise, you can halve the size of the heatsink if you use forced air circulation, and leave it inside the case.
Thermal management is crucial on a Class A SS amp. Get it wrong and MTBF drops like a stone. It is related to the Arrhenius Theorem, which related to chemical reactions as well. Increase temperature 11C, and MTBF drops HALF.
Transformer PER CHANNEL should be at least 160VA.
Cheers,
Hugh
You MUST ensure that the fins can be exposed to the external of the case.
Otherwise, you can halve the size of the heatsink if you use forced air circulation, and leave it inside the case.
Thermal management is crucial on a Class A SS amp. Get it wrong and MTBF drops like a stone. It is related to the Arrhenius Theorem, which related to chemical reactions as well. Increase temperature 11C, and MTBF drops HALF.
Transformer PER CHANNEL should be at least 160VA.
Cheers,
Hugh
Last edited:
Hi Hugh
Thanks for the correlation between heat and MTBF
Im intending to drill lots of holes above and below the heatsinks, and mount the case on feet at least 10 mm high, to facilitate convection
My understanding is that if I keep the C/ watt at 45 or lower, it should be ok
I just found pics of the inside of ajst2duk's case. So the heatsinks are joined to the PCB's * underneath? I'd been thinking of them mounted on the * sides . . .
Cheers
Thanks for the correlation between heat and MTBF
Im intending to drill lots of holes above and below the heatsinks, and mount the case on feet at least 10 mm high, to facilitate convection
My understanding is that if I keep the C/ watt at 45 or lower, it should be ok
I just found pics of the inside of ajst2duk's case. So the heatsinks are joined to the PCB's * underneath? I'd been thinking of them mounted on the * sides . . .
Cheers
Last edited:
You might be referring to the "during construction" photo where I had the PCB topside for final wiring etc - the heatsinks are actually on their sides with vents above and below in the case allowing full convection airflow
Here is a photo of the PCB's + Heatsinks as mounted
AJ
FETZILLA availability?
Price for 2-ch?
Power into this load? Flat 16 Ohm above the bass range, 12 Ohm minimum, easy to moderate phase angle
Sensitivity 93 dB actual: I am driving the speaker now to any desired level with amp making 50W into this load (16 Ohm).
I appreciate members saving me the time required to read all threads. Please summarize general consensus, overall sound quality of FETZILLA compared to AKSA 80 and AKSA 100 considering my specific application (less than average power requirements).
In room bass cutoff is high 30 Hz range. (For first octave I employ four piece distributed sub system ala Gedds/Welti/LeJeune, 1kW active, crossed high 30s). The main amp has little to no first octave requirements.
Price for 2-ch?
Power into this load? Flat 16 Ohm above the bass range, 12 Ohm minimum, easy to moderate phase angle
Sensitivity 93 dB actual: I am driving the speaker now to any desired level with amp making 50W into this load (16 Ohm).
I appreciate members saving me the time required to read all threads. Please summarize general consensus, overall sound quality of FETZILLA compared to AKSA 80 and AKSA 100 considering my specific application (less than average power requirements).
In room bass cutoff is high 30 Hz range. (For first octave I employ four piece distributed sub system ala Gedds/Welti/LeJeune, 1kW active, crossed high 30s). The main amp has little to no first octave requirements.
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.