Only because of impatient people here who cannot wait until December, and already want to order things now.
Of course they may, but then I also don't want to be blamed that the case won't fit their GB components.
So a large case measuring so 500x500x200mm will fit anything.
😉
Patrick
Of course they may, but then I also don't want to be blamed that the case won't fit their GB components.
So a large case measuring so 500x500x200mm will fit anything.
😉
Patrick
I need one more volunteer to cross check my ACAD drawings vs those from bksabath.
Please send me a PM if you have a dwg and Gerber viewer.
Thx,
Patrick
Please send me a PM if you have a dwg and Gerber viewer.
Thx,
Patrick
just for fun I tried to matched some .47 ohm resistors. It looks like measuring current or voltage is perhaps the better way to group the devices, even with a high precision ohmmeter.
I think when it all comes down to it, either you don't bother as Patrick mentioned or you trim the resulting amp, as Zen Mod mentioned
one more question with regard to the heat sink calculations. Are you leaving the option for attaching a regulated PSU to the heat sink, or are these calcs just for the amplifier boards?
If you are intending on using a regulated supply, It will have to be in its own separate case with heatsinks.
Correct.
Making a larger case with large heatsinks to accomdate more dissipation will make is too expensive to make and to ship.
Then it is better to buy commercial ones.
You may consider putting the regulator device on its own free standing heatsink within the case.
There is not a lot of spcae, but you might manage.
If you allow 4V drop & 4A bias, total dissipation is 16W, so the free-standing sink has to be 2W/°C.
Something like
V 5512G Profilkühlkörper - reichelt elektronik - Der Techniksortimenter - OnlineShop für Elektronik, Netbooks, PC-Komponenten, Kabel, Bauteile, Software & Bücher - ISO 9001:2000 Zertifiziert
(75x50x25mm, not small)
Patrick
Making a larger case with large heatsinks to accomdate more dissipation will make is too expensive to make and to ship.
Then it is better to buy commercial ones.
You may consider putting the regulator device on its own free standing heatsink within the case.
There is not a lot of spcae, but you might manage.
If you allow 4V drop & 4A bias, total dissipation is 16W, so the free-standing sink has to be 2W/°C.
Something like
V 5512G Profilkühlkörper - reichelt elektronik - Der Techniksortimenter - OnlineShop für Elektronik, Netbooks, PC-Komponenten, Kabel, Bauteile, Software & Bücher - ISO 9001:2000 Zertifiziert
(75x50x25mm, not small)
Patrick
Last edited:
> It looks like measuring current or voltage is perhaps the better way to group the devices, even with a high precision ohmmeter.
😉
> either you don't bother as Patrick mentioned
I do bother.
> you trim the resulting amp, as Zen Mod mentioned
That depends what you trim. You suggest trimming THD.
My approach is to trim second (and maybe third) harmonics without getting higher 4th, 5th, ....., before soldering.
If THD is the goal, then opamp and high feedback amp (like the Goldmund ?) rules.
No point going to Class A.
Patrick
😉
> either you don't bother as Patrick mentioned
I do bother.
> you trim the resulting amp, as Zen Mod mentioned
That depends what you trim. You suggest trimming THD.
My approach is to trim second (and maybe third) harmonics without getting higher 4th, 5th, ....., before soldering.
If THD is the goal, then opamp and high feedback amp (like the Goldmund ?) rules.
No point going to Class A.
Patrick
Finished testing of the MF35-151 heatsinks.
Air temperature was 22.1 degrees.
The small rectangles represent the position of mosfets across the heatsink.
Applied Voltages and Current were +/- 26.1V and 2.6A (135W).
Obviously between the mosfets the temperature is quite hot but further away the temperature drops off a fair bit. If there was a way to spread the heat across the heatsink better than we might be in busines.
Also, once the heatsinks are conected to the top and bottom plates and front and back plates the temperature will come down further.
Air temperature was 22.1 degrees.
The small rectangles represent the position of mosfets across the heatsink.
Applied Voltages and Current were +/- 26.1V and 2.6A (135W).
Obviously between the mosfets the temperature is quite hot but further away the temperature drops off a fair bit. If there was a way to spread the heat across the heatsink better than we might be in busines.
Also, once the heatsinks are conected to the top and bottom plates and front and back plates the temperature will come down further.
Attachments
Last edited:
Yes 67.4 degrees is far too hot in my opinion also.
However toward the edge of the heatsink the temperature drops below 50. If we can spread the heat uniformly across the heatsink then we shoud be able to get the temperature to below 60 near the mosfets.
However toward the edge of the heatsink the temperature drops below 50. If we can spread the heat uniformly across the heatsink then we shoud be able to get the temperature to below 60 near the mosfets.
I haven't tested the MF35-151, but empirically it seems like it should be limited to 100W.
You might get a little lower temperatures if you were able to line the devices up 1/3 of the way up from the bottom six in a row, but I still don't you'll get comfortable temperatures at 135W without a fan.
You might get a little lower temperatures if you were able to line the devices up 1/3 of the way up from the bottom six in a row, but I still don't you'll get comfortable temperatures at 135W without a fan.
the PCB will determine the device spacing.
Certainly the top devices are far too high up for optimum cooling and that is why we are seeing the top row temps way above the bottom row temps.
Certainly the top devices are far too high up for optimum cooling and that is why we are seeing the top row temps way above the bottom row temps.
What happened to the approx. +30C over ambient "rule of thumb"? The "put your hand on it for at least 5 sec."? Personally I can hold my hand on 65C for minutes. It looks to me like the outer extremidies are about 55C. I agree speading the load is more effectively using the sink. Hence, I prefer the long board (DiyAudio boards) configuration myself.
Are you planing on 3 paralell pair in your F5?
Are you planing on 3 paralell pair in your F5?
Geez, flg, I thought my hands were numb.
With mica isolators you're at around 1.6 C/W junction to sink. At 22.5W per device on a 65C sink (at the device), that's right around 100C at the junctions. That should be OK, but most people will find your sinks way too hot to touch. Silpads will be a little higher thermal resistance and therefore higher junction temperatures.
With mica isolators you're at around 1.6 C/W junction to sink. At 22.5W per device on a 65C sink (at the device), that's right around 100C at the junctions. That should be OK, but most people will find your sinks way too hot to touch. Silpads will be a little higher thermal resistance and therefore higher junction temperatures.
I'm not sure about that. Once its connected together you will also see an increase in ambient temperature, that in my opinion, will negate whatever you gain by this adding more aluminum mass, even with agressive number of cooling holes.
I hate a class A amp running on too small heatsinks.
One will always be tempted to lower the bias current. Or to use fans.
One will always be tempted to lower the bias current. Or to use fans.
Last edited:
Don't get too excited. Look at the measurement in detail.
The average temperature over the heatsink is 57°C. Minus 22.1, then divided by 135W gives 0.26°C/W.
And that is exactly according to Catalog (0.21 x 1.3 correction factor). So don't blame Conrad.
Our task is to spread the heat out evenly so that there is no 20°C over the area of the sink.
How, don't know yet. Give it time.
And you have to look at the entire picture -- what is the final junction temperature.
For those who cannot sleep because of this, you can read through this :
http://www.semicon.toshiba.co.jp/eng/shared/reliability_pdf/bde0128g_chap03.pdf
Or, order yourself a Hifi2000 today !! 😉
Patrick
The average temperature over the heatsink is 57°C. Minus 22.1, then divided by 135W gives 0.26°C/W.
And that is exactly according to Catalog (0.21 x 1.3 correction factor). So don't blame Conrad.
Our task is to spread the heat out evenly so that there is no 20°C over the area of the sink.
How, don't know yet. Give it time.
And you have to look at the entire picture -- what is the final junction temperature.
For those who cannot sleep because of this, you can read through this :
http://www.semicon.toshiba.co.jp/eng/shared/reliability_pdf/bde0128g_chap03.pdf
Or, order yourself a Hifi2000 today !! 😉
Patrick
- Status
- Not open for further replies.