john curl said:
John,
In their defense, I'd like to note that we (meaning all DIY folk) are at the mercy of whatever heatsinks we can find at surplus sites and the like. It's rare to find big ones and they never come with C/W ratings. Granted, with time and experience, you can get a pretty decent gut feeling for what a heatsink can do, just from its heft, but...
On the other hand, Mike's got more heatsinks available to him than are available in the entire state of South Carolina and he's gotten pretty good at sniffing them out.
My biggest heatsinks came from a junkyard. They brought in some sort of large (meaning refigerator-sized) power supply or UPS sort of thing and simply dropped it to the ground. The impact tore all the innards loose and broke the corners of the heatsinks where they were mounted to the chassis...but they're the best I've been able to do in ten or fifteen years of scrounging around here. On the plus side, I was able to match them against an Aavid/Thermalloy profile, so I've got a pretty good idea where I stand in terms of Pd capability.
On the other, other hand, I'm the guy who got so fed up with difficulties in sourcing and paying for heatsinks that I built a water-cooled system that will get rid of all the heat you could ask for. But that's not for everyone.
Grey
Want some of what I am drinking? 
It's Japanese green tea, brewed at lukewarm temperature, and an occasional bit of American Ginseng powder thrown in...
But seriously, I'm only kidding!
As far as using a fan, for a high quality amp in a high quality system, seems like a no-no to me. Fine for testing purposes though...
I'd go with a few dozen or so press fit TO-5s in a milled copper billet myself... dunno about you purists out there... Curl? You?
_-_-bear
It's Japanese green tea, brewed at lukewarm temperature, and an occasional bit of American Ginseng powder thrown in...
But seriously, I'm only kidding!
As far as using a fan, for a high quality amp in a high quality system, seems like a no-no to me. Fine for testing purposes though...
I'd go with a few dozen or so press fit TO-5s in a milled copper billet myself... dunno about you purists out there... Curl? You?
_-_-bear
john curl said:Please don't confuse them, Bear! They have enough problems as is.
Updates
1. The VAS needs additional filtering. It needs to be on the other side of R3 and R20.
2.Next design will also have a spot for a connector for split rails.
3. The ground for C4 needs to be on the input ground plane.
4. Add a 10 ohm resister from input ground to Star ground.
5. Add clamping diodes to the outputs.
6. Don't drink what Bear is drinking.
I hope to have a complete amp built in a week or two for listening tests.
Mike,
How much current are you running in the front end? Do you really need 1W resistors? Not that a larger part hurts, and some might even argue that it sounds better, but you would be able to save some space on the PCB (and money) by using smaller resistors.
For instance, I'm using RN-60 10 Ohm (i.e. rated 1/4W, but actually 1/2W in real terms) resistors for the folded cascode and even with running the second stage at 100mA, they're quite comfy.
Ditto other 1W parts.
Good luck, fella.
I'm tidying up the wiring on mine and should have them operational again in the next week or so. Even I have limits as to how much of a rat's nest I'll put up with. Once they're re-done, my wife will take pictures and I'll start a thread.
Grey
How much current are you running in the front end? Do you really need 1W resistors? Not that a larger part hurts, and some might even argue that it sounds better, but you would be able to save some space on the PCB (and money) by using smaller resistors.
For instance, I'm using RN-60 10 Ohm (i.e. rated 1/4W, but actually 1/2W in real terms) resistors for the folded cascode and even with running the second stage at 100mA, they're quite comfy.
Ditto other 1W parts.
Good luck, fella.
I'm tidying up the wiring on mine and should have them operational again in the next week or so. Even I have limits as to how much of a rat's nest I'll put up with. Once they're re-done, my wife will take pictures and I'll start a thread.
Grey
I have found that 1/2W resistors or more are usually better sonically. This is partly because they have higher thermal capacitance (please think about what that means) and they tend to promote MORE idle current, rather than LESS idle current in the input and driver stages. I don't want you to cheapen this design with compromises, unless it is absolutely necessary. Please remember, MY initials are on the design, not yours.
John,
I began using the term thermal capacitor in the Aleph-X thread several years ago. In that case, I was speaking of a block of metal (not a heatsink--no fins--just a cube of metal) attached to an active device(s) purely to stabilize the temperature. I assume that you're referring to delta-R/degrees C, yes?
What is your position regarding RN60 military resistors? The "official" rating is 1/4W, but they're the same size as conventional 1/2W resistors. Given that heat dissipation is directly related to size, hence surface area available for radiation, I've always regarded them as 1/2W regardless of the MIL-spec rating.
Judging purely from pictures I've seen online, they seem to be the same physical size as the resistors in your designs.
I've never been comfortable with 1/4W resistors (1/8W in "MIL-speak"), partly because I spent a lot of time fiddling with tubes and you never see 1/4W resistors in tube gear, so they just look wrong to me. That's a purely visceral reaction, but I also find them difficult to read, whether by stripes or printed nomenclature. Furthermore, it was worth the extra two cents or so ea. to get the 1/2W so as to be able to use the same resistors in either solid state or tube circuits. That way I don't have to stock two different wattage ratings of each resistance.
For any who feel inclined to pick nits, note that I'm speaking purely in terms of physical dimensions, not chemical composition, configuration of resistive element, or mechanical construction such as attachment of leads, etc.
Justcallmedad,
I'm intrigued by John's bias circuit because it shows symmetrical impedance on both ends and allows you to dispense with the customary electrolytic bypass. I've played with a couple of variations, including an auto-bias version, and it works well.
Grey
I began using the term thermal capacitor in the Aleph-X thread several years ago. In that case, I was speaking of a block of metal (not a heatsink--no fins--just a cube of metal) attached to an active device(s) purely to stabilize the temperature. I assume that you're referring to delta-R/degrees C, yes?
What is your position regarding RN60 military resistors? The "official" rating is 1/4W, but they're the same size as conventional 1/2W resistors. Given that heat dissipation is directly related to size, hence surface area available for radiation, I've always regarded them as 1/2W regardless of the MIL-spec rating.
Judging purely from pictures I've seen online, they seem to be the same physical size as the resistors in your designs.
I've never been comfortable with 1/4W resistors (1/8W in "MIL-speak"), partly because I spent a lot of time fiddling with tubes and you never see 1/4W resistors in tube gear, so they just look wrong to me. That's a purely visceral reaction, but I also find them difficult to read, whether by stripes or printed nomenclature. Furthermore, it was worth the extra two cents or so ea. to get the 1/2W so as to be able to use the same resistors in either solid state or tube circuits. That way I don't have to stock two different wattage ratings of each resistance.
For any who feel inclined to pick nits, note that I'm speaking purely in terms of physical dimensions, not chemical composition, configuration of resistive element, or mechanical construction such as attachment of leads, etc.
Justcallmedad,
I'm intrigued by John's bias circuit because it shows symmetrical impedance on both ends and allows you to dispense with the customary electrolytic bypass. I've played with a couple of variations, including an auto-bias version, and it works well.
Grey
