Patrik Floding said:
Yeh, I presume they have their heads very close to the PCBs´ soldering components like a turbo and thinking who will be the first to post pics of the finished crown-jewel
You mention casework/chassis and I plan to buy a custom design chassis from the Italian company Hi-Fi2000. They have all equipments that is needed for to do custom works, and they also have stated it officially.
I have done some drawings for a complete chassis for to fit together with my Australian Conrads heat sink type MF-350-151.5 (2 heat sinks for a mono block, no fan in my cases). I shall order 6 or 7 chassis from them. Anybody interested to join in?
Regards
Flodstroem said:
Yeh, I presume they have their heads very close to the PCBs´ soldering components like a turbo and thinking who will be the first to post pics of the finished crown-jewel
You mention casework/chassis and I plan to buy a custom design chassis from the Italian company Hi-Fi2000. They have all equipments that is needed for to do custom works, and they also have stated it officially.
I have done some drawings for a complete chassis for to fit together with my Australian Conrads heat sink type MF-350-151.5 (2 heat sinks for a mono block, no fan in my cases). I shall order 6 or 7 chassis from them. Anybody interested to join in?
Regards
I too plan on building monoblocks with no fans.
However, I was planning to try to have the transformers in separate chassis. Not sure if it makes sense, but I was hoping doing so might give some audible benefits.
Knowing what you are prepared to spend on transformers, I'm not sure I could afford these caseworks anyway!
Hi,
270W dissipation through a pair of MF35 151.5 will raise Tc quite high. Is MF350 different?
Are you intending to use To3 or To264 packages?
Conrad specify their sinks @ 80Cdegrees above ambient. One must correct the Rth s-a for the lower temperature difference that most designs demand.
Using To264, I would expect Tc to be about 47Cdegrees above ambient. In hot weather this could lead to Tc approaching 80degC and 250W devices must be de-rated to 140W.
A few devices could run slightly hotter due to Ib differences and tolerance spread in the thermal coupling to the sink.
To3 will survive much better but I would still recommend a better pair of sinks. Corrected Rth s-a<0.1C/W
270W dissipation through a pair of MF35 151.5 will raise Tc quite high. Is MF350 different?
Are you intending to use To3 or To264 packages?
Conrad specify their sinks @ 80Cdegrees above ambient. One must correct the Rth s-a for the lower temperature difference that most designs demand.
Using To264, I would expect Tc to be about 47Cdegrees above ambient. In hot weather this could lead to Tc approaching 80degC and 250W devices must be de-rated to 140W.
A few devices could run slightly hotter due to Ib differences and tolerance spread in the thermal coupling to the sink.
To3 will survive much better but I would still recommend a better pair of sinks. Corrected Rth s-a<0.1C/W
AndrewT said:
Andrew, Yes you are right about the dissipation, it will be 270 Watts if running the amp as the originals at idling currents.
My calculation says: correction factor for 37° over ambient will be ca 1.28 and the C/W for the MF35-151.5 is 0.21 (sorry I confused you, but my mistake)
Then each sink could handle 37/(1.28 x 0.21) which means ca 138 Watts and this is one side (one rail side) there is two of those sinks on my mono block so maximum sink temp will be calculated like this: 28° (at hottest summer) + 37° rise, a maximum of 65°C. But consider this: I havnt calculated dissipation help from the well ventilated chassis and there is some small heat sinks on the back of the main sinks that also help to rise the dissipation factor.
Normally our house temp is between 22° to 24° the most of the time here in our location in the Western Norway and I have estimated the fin- temperature to be between 55 and 60°C. Its not that hot, is it?
I Im going to use 6 pair of the metal, 250W transistors so each transistor must handle 270W/12 =22.5 Watts of dissipation, and without looking in the data sheet, it could look like a security factor of 11 not included the running temp correction factor.
Patrik, you have a point here, but in my case, I will spend most of my money on the transformer, PCBs´, Computer graded Electrolytics. If chassis should be an expensive affair then I had to reconsider it seriously.
Hi Flod,
your calculations are OK (except Factor of Safety, see later).
Just add in the 7Cdegrees for Rth c-s and then derate your To3s for 72degC and you have 180W times 6. Well capable in comparison to a KSA100.
Your 6pair (MJ21193/4) with Tc=<77degC and 2ohm load at phase angle of 60degrees with +-52Vdc +-60mF smoothing shows a margin between 22Apk on the 1second SOAR and 17Apk on the 2ohm locus @ 50Vce (Pout~=410W into 2r). That equates to FoS=1.3 not 11.
The output stage will drive 1ohm to 40degrees and stay inside the 1second SOAR. This is better low load ability than a standard KSA100.
The FoS for 2ohm & 6pair MJL4281 @ the same operating conditions in my proposal is down to 0.9, but then I don't have any speakers that go down to those welding currents (yes, I have TIG welded at a lot less than 38A).
your calculations are OK (except Factor of Safety, see later).
Just add in the 7Cdegrees for Rth c-s and then derate your To3s for 72degC and you have 180W times 6. Well capable in comparison to a KSA100.
Your 6pair (MJ21193/4) with Tc=<77degC and 2ohm load at phase angle of 60degrees with +-52Vdc +-60mF smoothing shows a margin between 22Apk on the 1second SOAR and 17Apk on the 2ohm locus @ 50Vce (Pout~=410W into 2r). That equates to FoS=1.3 not 11.
The output stage will drive 1ohm to 40degrees and stay inside the 1second SOAR. This is better low load ability than a standard KSA100.
The FoS for 2ohm & 6pair MJL4281 @ the same operating conditions in my proposal is down to 0.9, but then I don't have any speakers that go down to those welding currents (yes, I have TIG welded at a lot less than 38A).
AndrewT said:
Hi Andrew, thanks for your safety calculations, however Im going to use the slower MJ15003/4 (20A, 250W) for outputs and 6 pair of those types. Also Im going to mount the DRIVERS on the same heat sink as for the outputs (on two output boards) and that will ensure the drivers to run cooler than if those was fitted on a smaller sink on the main board. Smoothing Caps will be 2 x 50 mF= 100 mF for each rail (which means four caps to each mono block). So there will be plenty of Amperes for to handle speakers down to 1-2 ohms. Not necessarily a 1 or 2 ohms rated speaker, but for a 4 ohms rated speaker when played music on. Some time the amp has to deal with lower impedances than the rated 4 ohms, especially in very sharp/fast polarity shifts in transient rich music. And this is more pronounced if driving a speaker directly, without passive filtering (eg. by active x-overs).
Ah-ha, so you are familiar with TIG weldings
. Interesting, do you have use of that knowledge when constructing/building amps ?barryblue, most of my calculations is done using the equations in the data sheets and catalogues (yes there is a lot of that stuff in the data sheets that is very useful) and some of it I have in my head
. I dont know if Andrew has some but you could use the calculating tips from the data sheets to make a spread sheet for your own.Regards
Hi Flod,
I have mounted the drivers on a separate heatsink to ensure they stay much cooler than the outputs.
Running the drivers on a sink that is up at 50 to 60degC seems something I would rather avoid. I'm sure the drivers, Vbe and cascode would thank me if they could talk.
post1235 mentions my proposal.
Barry,
email, if you require a modified version of Bensen's sheet.
I have mounted the drivers on a separate heatsink to ensure they stay much cooler than the outputs.
Running the drivers on a sink that is up at 50 to 60degC seems something I would rather avoid. I'm sure the drivers, Vbe and cascode would thank me if they could talk.
post1235 mentions my proposal.
Barry,
email, if you require a modified version of Bensen's sheet.
Hello folks,
Anyone having resistors (Dale/Vishay RN60 1/2W) left over from your order to this amp. I am in the need of all values (except those 1 ohm/5W), if your´ sitting with any value left Im interested to buy it for at least 6 boards (maybe more).
Im also in the need of the 1N5309
and the feet´s: ZVN/ZVP2110G (SOT-223)
Ok, send me a PM of the totals your´willing to sell to me
including price, shipping, PayPal-fee and your PayPal name.
(have a PayPal-account)
Thanks in Advance.
Regards
Anyone having resistors (Dale/Vishay RN60 1/2W) left over from your order to this amp. I am in the need of all values (except those 1 ohm/5W), if your´ sitting with any value left Im interested to buy it for at least 6 boards (maybe more).
Im also in the need of the 1N5309
and the feet´s: ZVN/ZVP2110G (SOT-223)
Ok, send me a PM of the totals your´willing to sell to me
including price, shipping, PayPal-fee and your PayPal name.
(have a PayPal-account)
Thanks in Advance.
Regards
AndrewT said:
What is considered the best performing solution:
1) mounting the output drivers (MJE15030/31) on a separete heatsink for a relatively stable thermal operating point. Output drivers will likely run a lower Tj temperature. (like the KSA100 )
2) mounting the output drivers on the same heatsink as the output stage transistors for better thermal tracking. Output drivers will likely run at a higher Tj temperature.
LineSource said:
If I dont remember it to wrong I now PWatts mention (in a posts in this tread) when working on an original KSA100 that the drivers was running extremely hot
. Dont know if this has anything to do with this new design, but I was imaging that the drivers should be running cooler if mounted on the outputs heat sink (and Im not talking about the cascode)Thermal tracking is a very good point as you suggested in #2 LineSource. This was my first thoughts when designing my output boards. Wander what was the Tj on those output drivers in the original KSA100? Anyone knows? Mark? Pierre?
Actually Andrew advocates leaving the drivers OFF the main sink. He has alot of reasons for this and I believe these reasons are true. Krell obviously did the same thing. I am using heat sinks a bit larger than Krell's the exact dimensions of which I can post here when I get back home tommrrow night. My sink does run hot but no where near what I would call extremely hot. The OP stage also tracks very well. On my KSA-50 I mounted them to the main sinks.... tracking is also very good this way but I suspect the drovers are actually getting hotter than they normally would. I can honestly say that I've built em both ways.... thermal tracking with them on board stabilizes very quickly. Thermal stability takes alot longer with them mounted on the board using a decent heat sink... I think KRell's was a little too small but it probably stabilized alot quicker than the one I;m using. Still, my KSA-100 fully stabilizes in one hour which I consider normal for any Class A power amp.
Mark
Mark
Thanks Mark for your valuable information. However, Im going to use more drivers than what the board are designed for and thats one reason for to mount them on the output board. A benefit for to doing this besides the thermal tracking issue is the lot easier hard wirings. Not so many cables to solder and wire.
If my idea isnt working well when finished I could always re-wiring and put the drivers up to the main board including a suitable heat sink.
Mark, is it possible for you to post a picture on your drivers at your main board. Interested to see how you has solved the mountings for the drivers heat sink and the mounts on the case chassis.
Regards
If my idea isnt working well when finished I could always re-wiring and put the drivers up to the main board including a suitable heat sink.
Mark, is it possible for you to post a picture on your drivers at your main board. Interested to see how you has solved the mountings for the drivers heat sink and the mounts on the case chassis.
Regards