easy
Hi,
The schematic is simpler than the leach amp, but the thermal concerns are greater.
Since you know the rating of your heatsink it is actually pretty easy to tell if it big enough.
ie If I understand the numbers you give, the heatsink is rated at 0.5c per watt per 3" length. You have 11", so you end up with total capacity of 0.5/(11/3) c / watt, approx. 0.13c/watt, it's a good heatsink.
Then you pick your voltage rails and idle current and multiply the voltage x 2 x idle current, lets say 40 x 2 x 2.5a, a nice 200w per channel. So the temp increase on your sink for one channel will be 200x0.13, or 26c, add in ambient at say 20c for a final temp of 46c. Not a problem at all. You would be able to touch the heatsink without burning yourself. However, 2 channels would give you a final temp of (400x0.13)+20, for a final temp of 72c, too hot to touch, not good for people, cats or reliability. With a decent fan your heatsink would be plenty big enough for 2 channels.
As you can see, the choice of voltage and current is the key to your dissipation levels.
Hope that helps
Stuart
Hi,
The schematic is simpler than the leach amp, but the thermal concerns are greater.
Since you know the rating of your heatsink it is actually pretty easy to tell if it big enough.
ie If I understand the numbers you give, the heatsink is rated at 0.5c per watt per 3" length. You have 11", so you end up with total capacity of 0.5/(11/3) c / watt, approx. 0.13c/watt, it's a good heatsink.
Then you pick your voltage rails and idle current and multiply the voltage x 2 x idle current, lets say 40 x 2 x 2.5a, a nice 200w per channel. So the temp increase on your sink for one channel will be 200x0.13, or 26c, add in ambient at say 20c for a final temp of 46c. Not a problem at all. You would be able to touch the heatsink without burning yourself. However, 2 channels would give you a final temp of (400x0.13)+20, for a final temp of 72c, too hot to touch, not good for people, cats or reliability. With a decent fan your heatsink would be plenty big enough for 2 channels.
As you can see, the choice of voltage and current is the key to your dissipation levels.
Hope that helps
Stuart
Not Worried!
Ok,
I guess at the end of the day, we don't really NEED 200wrms @ class a, however, if it's needed, there's potential we will get it! I thought that when increasing bias (to say 10amps) you would need a bigger power supply? Does anyone know if an 800va 30-0-30 toroid would be big enough for 200wrms @ 4 ohm into class a? (as previously mentioned, it would be dissapating about 450w in heat!) - I thought with class A you had to have double the power output from the transformer/torroid than what you had in heat? Is that right?
Our other problem is that our heatsinks are being made, so i don't have a specific set of info about them, only way we will know is to test them! We were thinking of copper heatsinks
Not sure though yet but we've got a choice of pretty much anything!
Aaron
Ok,
I guess at the end of the day, we don't really NEED 200wrms @ class a, however, if it's needed, there's potential we will get it! I thought that when increasing bias (to say 10amps) you would need a bigger power supply? Does anyone know if an 800va 30-0-30 toroid would be big enough for 200wrms @ 4 ohm into class a? (as previously mentioned, it would be dissapating about 450w in heat!) - I thought with class A you had to have double the power output from the transformer/torroid than what you had in heat? Is that right?
Our other problem is that our heatsinks are being made, so i don't have a specific set of info about them, only way we will know is to test them! We were thinking of copper heatsinks
Not sure though yet but we've got a choice of pretty much anything!Aaron
size does matter...
For reliability you should have a transformer that is overrated relative to the demand you expect. if you are biasing for 400w of heat, I'd go with at least a 600va transformer, probably 800 for serious welding duty. High end manufacturers seem to use 100% over as the 'safe' margin.
But at the end of the day I'd like to be able to plug mine into a normal wall socket, so I'm gonna try not to pull more than about 10amps, for both channels. Here in the US this limits me to about 600va per channel, which coincidentally is exactly what I've got.
Eventually I'll need to consider the startup current which is going to be huge for big toroids backed up with big caps...oh well one thing at a time.
If everything were set up right the efficiency of the amp can approach 50%, at max power into the correct resistive load. It is nearly always a small fraction of that. The quieter the amp plays the less efficient it is. It occours to me, and I havent done the math to check, the amp may actually be cooler if it is played louder...does more power in the load means less in the heatsink?
Stuart
For reliability you should have a transformer that is overrated relative to the demand you expect. if you are biasing for 400w of heat, I'd go with at least a 600va transformer, probably 800 for serious welding duty. High end manufacturers seem to use 100% over as the 'safe' margin.
But at the end of the day I'd like to be able to plug mine into a normal wall socket, so I'm gonna try not to pull more than about 10amps, for both channels. Here in the US this limits me to about 600va per channel, which coincidentally is exactly what I've got.
Eventually I'll need to consider the startup current which is going to be huge for big toroids backed up with big caps...oh well one thing at a time.
If everything were set up right the efficiency of the amp can approach 50%, at max power into the correct resistive load. It is nearly always a small fraction of that. The quieter the amp plays the less efficient it is. It occours to me, and I havent done the math to check, the amp may actually be cooler if it is played louder...does more power in the load means less in the heatsink?
Stuart
Re: size does matter...
Yes, if it is not sent to a speaker the idle power goes to the heatsinks. For Class A anyway.
I know someone else here in Houston with a pair of 100 watt Aleph 2 clones. I think they generate about 300 watts of heat per side. Here where the air conditioning runs 11 months a year that is a load.
But back in my single days I used to keep a pair of Manley 350's running a lot of the time. They must have gave off about the same amount of heat, the lights dimmed all though the house on power up. Each was on a dedicated 20 amp line.
George
Stuart Easson said:
Yes, if it is not sent to a speaker the idle power goes to the heatsinks. For Class A anyway.
I know someone else here in Houston with a pair of 100 watt Aleph 2 clones. I think they generate about 300 watts of heat per side. Here where the air conditioning runs 11 months a year that is a load.
But back in my single days I used to keep a pair of Manley 350's running a lot of the time. They must have gave off about the same amount of heat, the lights dimmed all though the house on power up. Each was on a dedicated 20 amp line.
George
jacco vermeulen said:
Move the the UK!!! £1.80 or so per LITRE!!!! (i'm in australia and ours is $1.19 per litre on a BAD day for 98 octane!)
Oh and i'm thinking 30-0-30 may be a LITTLE high i'm thinking (it's nearly as much as the KSA-100... I'm assuming now that the rest of the circuit (mainboard) can handle this much voltage?
Aaron
NUTTTR said:
Please also look at the posts at the thread: Older Krell Schematics
http://www.diyaudio.com/forums/showthread.php?threadid=45352
and especially at the post no 12 and 13:
http://www.diyaudio.com/forums/showthread.php?postid=520084#post520084
http://www.diyaudio.com/forums/showthread.php?postid=520084#post520084
,where is stated that the KSA 50 used supply rails of +/- 36.5 VDC and the bias was set to 620 mV DC across 0.68 ohm emitter resistors. That +/- 36.5 VDC voltage would be achieved with 26-27V AC transformer secondary.
cheers,
Argo
Thanks Argo!
I emailed him previously (i was the other member) and excellent that he supplied the answers we needed!!!
After all it seems a MUCH smaller amount of bias current!!!!! (than my "10 amps" lol)
Although our power supplies should be fine 30-0-30 setup...
However i think i might be aiming to have the bias just a little lower than 10 amps or so!!!!
Thanks
Aaron
I emailed him previously (i was the other member) and excellent that he supplied the answers we needed!!!
After all it seems a MUCH smaller amount of bias current!!!!! (than my "10 amps" lol)
Although our power supplies should be fine 30-0-30 setup...
However i think i might be aiming to have the bias just a little lower than 10 amps or so!!!!
Thanks
Aaron
If the 620 mV pr. 0.68 ohm resistor is correct, this means the original had slightly less than 2A bias,--- figures all right for a 50W amp, but vastly lower than the 10a being suggested here.... ( Although I do see the idea for kicking up the bias and pulling more power! )
Somebody have an idea for damping factor of the original?
Just wondering if it really does pay to increase the no. of output devices..
Somebody have an idea for damping factor of the original?
Just wondering if it really does pay to increase the no. of output devices..
AuroraB said:
Yes, although the original would have had about 500/600va transformers (KSA100 had 1400va, so 700va a channel and the ksa 50 was half the power!) at most i'd guess... That's why i'm considering kicking the bias right up, obviously because i have sufficent power available to do so and also want class A all the way to peak @ 4 ohm!!
I'm not so much interested in damping factor as i've listened to many an amp and never really heard damping factors change anything from my view, in fact, amps with lower damping factors, i've found, sounded better to my ears! I'm thinking of increasing the outputs (doubling in fact) due to the amount of bias i'm tempted to run therefore the amount of heat each transistor has to dissapate has then been halved... I don't really want each transistor (assuming previous conversations) pushing out 110+W of heat!!!! 55w to me is a MUCH nicer figure per transistor!
Aaron
heatsinks
Nothing in the circuit would be overloaded at 45v, but some adjustment of currents in the front end would be necessary, and more output transistors would make for a more robust amp, whose name would be KSA100...
By choosing a transformer, caps and rectifier large enough for a ksa100, but with secondary voltages appropriate for the ksa50 it would be relatively starightforward to make the 50 into a 100 as more heatsinking was aquired...some toroids have primaries wound for international use, ie with 100, 110, 120, 220, and 240v. Here in the US one of these could be used to select secondary voltage, 26v secondaries would give approx. 31v when the 100v tap is used with 120v mains. Of course, if you are getting custom transformers, just ask for the correct secondary taps.
Aaron, just how big are your heatsinks? 100w RMS into 4 requires 200w peaks, 7.2amp peak current, 3.6amp bias. With 37v rails I get 67 watts per transistor (25% rated power), lower but still not trivial...
Stuart
Nothing in the circuit would be overloaded at 45v, but some adjustment of currents in the front end would be necessary, and more output transistors would make for a more robust amp, whose name would be KSA100...
By choosing a transformer, caps and rectifier large enough for a ksa100, but with secondary voltages appropriate for the ksa50 it would be relatively starightforward to make the 50 into a 100 as more heatsinking was aquired...some toroids have primaries wound for international use, ie with 100, 110, 120, 220, and 240v. Here in the US one of these could be used to select secondary voltage, 26v secondaries would give approx. 31v when the 100v tap is used with 120v mains. Of course, if you are getting custom transformers, just ask for the correct secondary taps.
Aaron, just how big are your heatsinks? 100w RMS into 4 requires 200w peaks, 7.2amp peak current, 3.6amp bias. With 37v rails I get 67 watts per transistor (25% rated power), lower but still not trivial...
Stuart
Hi Stuart,
I agree it is definately not a "trivial" amount of heat thats for sure!!! The pieces of heatsink we will have are 6mm thick, fins are 50mm deep and 2mm apart, the heatsinks themselves are about 400mmx500mm and there's 3 of them per amp, they will actually form the sides of the amp itself, so they are massive amps, we will obviously cut the heatsinks down a little, but how much, i'm not sure yet as i want absolute maximum cooling area....
Would that be enough? I want to space the transistors out a fair bit!!!
Aaron
I agree it is definately not a "trivial" amount of heat thats for sure!!! The pieces of heatsink we will have are 6mm thick, fins are 50mm deep and 2mm apart, the heatsinks themselves are about 400mmx500mm and there's 3 of them per amp, they will actually form the sides of the amp itself, so they are massive amps, we will obviously cut the heatsinks down a little, but how much, i'm not sure yet as i want absolute maximum cooling area....
Would that be enough? I want to space the transistors out a fair bit!!!
Aaron
Gentlemen
Are we each going to try and get hold of parts by our own or is somebody co-ordinating a single point where one can purchase high quality parts (excluding the PSU parts which I can source locally) in one go ?
One can possibly argue that each person can source the components in his own country but in my area it is very difficult to get quality parts and I have often purchased components which appear to be the real thing but are in fact inferior components. Especially with this amp one would like to make a serious go at building a reliable unit.
Any comments in this regard will be appreciated.
In closing, I have a Krell KSA100 Mk1 and it has given excellent service- in fact I consider it a much more dynanamic unit than my Krell KSA 200s which has the sliding bias facility. However I have found that in summer the KSA100 gets "dangerously" hot and that promted me to buy the KSA200s. In my opinion nothing beats these amps on big speakers when it comes to absolute control and dynamic power. I am therefore very much interested in building a KSA50 which can possibly run cooler and give me as much enjoyment as bigger Krell siblings.
Are we each going to try and get hold of parts by our own or is somebody co-ordinating a single point where one can purchase high quality parts (excluding the PSU parts which I can source locally) in one go ?
One can possibly argue that each person can source the components in his own country but in my area it is very difficult to get quality parts and I have often purchased components which appear to be the real thing but are in fact inferior components. Especially with this amp one would like to make a serious go at building a reliable unit.
Any comments in this regard will be appreciated.
In closing, I have a Krell KSA100 Mk1 and it has given excellent service- in fact I consider it a much more dynanamic unit than my Krell KSA 200s which has the sliding bias facility. However I have found that in summer the KSA100 gets "dangerously" hot and that promted me to buy the KSA200s. In my opinion nothing beats these amps on big speakers when it comes to absolute control and dynamic power. I am therefore very much interested in building a KSA50 which can possibly run cooler and give me as much enjoyment as bigger Krell siblings.
sounds big enough
Aaron, your heatsinks sound big enough to me, but without the dissipation figures it's really hard to know. You could check the aavid, thermalloy etc websites to see if any of those vendors have something similar then extrapolate from there. Probably the easiest way to get accurate results would be to get a 50w resistor, mount it on the sink and put a known amount of power into it, measure the temp and you get all the numbers you need. Or you could do what most people would and build it, start it up and slowly increase the bias...
Jozua, if you are talking about transistors etc, there are a couple of places here that are run by forum members that stock the transistors and I'm sure would happily ship them to you, being light the plastic parts might be the way to go. I have used this place and been very happy with speed of service and quality of product:
http://www.tech-diy.com/hexfets.htm.
The small signal transistors, are easier, I have a couple hundred 2sa970, 2sc2240s, if you need some for this amp I can send them to you, email me offline. I buy them from digikey, they have a price break at 100...
I actually have all the parts to make this amp, which is one of the reasons I am keen to do so, so personally, I'm less interested in a group buy for anything but the boards. I might be tempted for cool heatsinks or cases, but not passives, semis etc.
Stuart
Aaron, your heatsinks sound big enough to me, but without the dissipation figures it's really hard to know. You could check the aavid, thermalloy etc websites to see if any of those vendors have something similar then extrapolate from there. Probably the easiest way to get accurate results would be to get a 50w resistor, mount it on the sink and put a known amount of power into it, measure the temp and you get all the numbers you need. Or you could do what most people would and build it, start it up and slowly increase the bias...
Jozua, if you are talking about transistors etc, there are a couple of places here that are run by forum members that stock the transistors and I'm sure would happily ship them to you, being light the plastic parts might be the way to go. I have used this place and been very happy with speed of service and quality of product:
http://www.tech-diy.com/hexfets.htm.
The small signal transistors, are easier, I have a couple hundred 2sa970, 2sc2240s, if you need some for this amp I can send them to you, email me offline. I buy them from digikey, they have a price break at 100...
I actually have all the parts to make this amp, which is one of the reasons I am keen to do so, so personally, I'm less interested in a group buy for anything but the boards. I might be tempted for cool heatsinks or cases, but not passives, semis etc.
Stuart
less heat...
Jozua,
not that I'd want to discourage you from making the ksa50, but if your ksa100 gets too hot, you could lower the dissipation a lot by reducing it's bias to the level of a ksa50, your heat output would drop 40%, probably more than enough in the context of your climate. It's maximum output level would remain the same, just it would be class B...
Stuart
Jozua,
not that I'd want to discourage you from making the ksa50, but if your ksa100 gets too hot, you could lower the dissipation a lot by reducing it's bias to the level of a ksa50, your heat output would drop 40%, probably more than enough in the context of your climate. It's maximum output level would remain the same, just it would be class B...
Stuart
Although I am going to build a KSA 50 early next year I sit here thinking that this is all happenning about 10 years too late!! The Pass Labs stuff is far superior, much easier and less expensive to build. The only drawback is not working as well into very low ohm loads as the Krell but I can also say that the Pass stuff also sounds much, much better....... Having owned a KSA 80B for quite some time........
Mark
Mark