Hi, I went from using the included silver heatsink to the Zalman purple heatsink, only to find the Zalman heatsink gets too hot without a fan.
I'm not using the Zalman heatsink with the Sure (MSI) fan. I'm running the fan at around 4v now, which means it makes less noise, and is still fast enough to keep the heatsink from getting hot. I did think the fan was a little noisy at 5v myself, but I do have the amp modules on the back of my speakers...
Blow
Most installations have the fan blowing outside air onto the heat source and then let it exhaust out of the case where it will.
For sure. The 41Hz implementations of the same chipset hardly even need heatsinking.
I ran one of these board @24V fan-less in an unvented aluminium enclosure for three months but the stock heatsink did get get very hot, I'd estimate 50 -55C.
Coincidentally, I have been doing some tests on the new version of the board. It comes equipped with a very mechanically noisy 12V fan that plugs into a variable speed fan control that kicks in at 45C.
I am running the board @32-33.5V (to take into account the voltage lost in the input diodes) without a fan.
Ambient ~24-25C
Temperature rise:
Approx 2C every 3mins to a maximum of 52C
Temperature measured at base of heatsink:
30V 45C
32V 46-47C
34V 48-49C
Without extra cooling fins:
32V 50-51C
34V 52C
PSU:
LS75 75W 36V 2.2A
All of these measurements were taken using the Frankenstein heatsink pictured. Sorry no tests done with stock HS. I thought I could get away with a giant block of copper (old Pentium 3 server heatsink) but because the heat gain is slow but constant, a larger surface area is required to disperse the heat away from the chips. A quick blast with the fan and temp drops quickly 10C <1min.
The Zalman HS most people have been using would probably have been a better choice as It probably has a larger surface area but is too tall for my enclosure (old radio chassis).
All this suggests that to be safe I should reduce the voltage or find a quiet fan, but I'm not going to because the board sounds much better @ ~32V and I hate fans
I think that I can get the temp to stabilize at about 50C passively in its case, which is the max operating temperature stated in the Sure manual (although Tripath state 70C) but I think a quiet fan is a far more sensible solution if you are worried about heat.
My only other thought was if there was a way to detect whether there was audio input and invoke the mute if none was detected, although this might be quite annoying for PC users (system sounds etc...)
Hope this helps someone,
Mat
Coincidentally, I have been doing some tests on the new version of the board. It comes equipped with a very mechanically noisy 12V fan that plugs into a variable speed fan control that kicks in at 45C.
I am running the board @32-33.5V (to take into account the voltage lost in the input diodes) without a fan.
Ambient ~24-25C
Temperature rise:
Approx 2C every 3mins to a maximum of 52C
Temperature measured at base of heatsink:
30V 45C
32V 46-47C
34V 48-49C
Without extra cooling fins:
32V 50-51C
34V 52C
PSU:
LS75 75W 36V 2.2A
All of these measurements were taken using the Frankenstein heatsink pictured. Sorry no tests done with stock HS. I thought I could get away with a giant block of copper (old Pentium 3 server heatsink) but because the heat gain is slow but constant, a larger surface area is required to disperse the heat away from the chips. A quick blast with the fan and temp drops quickly 10C <1min.
The Zalman HS most people have been using would probably have been a better choice as It probably has a larger surface area but is too tall for my enclosure (old radio chassis).
All this suggests that to be safe I should reduce the voltage or find a quiet fan, but I'm not going to because the board sounds much better @ ~32V and I hate fans
I think that I can get the temp to stabilize at about 50C passively in its case, which is the max operating temperature stated in the Sure manual (although Tripath state 70C) but I think a quiet fan is a far more sensible solution if you are worried about heat.
My only other thought was if there was a way to detect whether there was audio input and invoke the mute if none was detected, although this might be quite annoying for PC users (system sounds etc...)
Hope this helps someone,
Mat
Attachments
I'm very tempted to try building an Amp11, I need to practice my SMD soldering I think.
Maybe adjusting the dead time could change things a little? The heat isn't really an issue for me but it's a shame to have Class-D/T and not have it as efficient as it could be IMO.
Hifimediy
.
SHOP
.
http://www.diyaudio.com/forums/class-d/164274-new-tk2050-board.html
I would try one of these next.
.
SHOP
.
http://www.diyaudio.com/forums/class-d/164274-new-tk2050-board.html
Thanks Scott, these boards to appear to be better then the Sure boards, I am very tempted to try one to see how they compare both in terms of sound and build quality. The fact that there is no fan on the heatsink is a good thing, and it would be quite interesting to see if one board is significantly more efficient than the other.
I think this may sum it up:
Copied from Modding the Sure Electronics TK2050 2x100W Board (Part 1 of 2) The Hi-Fi Watcher
Mods applied:
Replaced the output inductors with Wuerth 4.7uH 9A rated versions
Changed the fan-based heatsink to a passive Northbridge heatsink from Zalman
Removed the on-board input resistors R14 and R34
Removed capacitors C17 and C25
Removed capacitors C16 and C24 and replaced with appropriate shorting links
Removed the input suppressors underneath the board, by the RCA inputs
Volume controlled bay a 100k Alps pot
Used 4.7uF metalised polypropylene (LCR brand) capacitors between the pot and board input
Used the manual control on the 24V SMPS to turn it up to 27.1V
Mods left to do:
Remove the polarity diodes (replacing with a short) from the DC rails on the board
Consider changing the 24V supply for a 36V supply (turned down to 32V)
Add a 0.47uF + 10 Ohm Zobel network
I think these are most of them, I am sure someone may have other mods, but this will keep you busy.
Copied from Modding the Sure Electronics TK2050 2x100W Board (Part 1 of 2) The Hi-Fi Watcher
Mods applied:
Replaced the output inductors with Wuerth 4.7uH 9A rated versions
Changed the fan-based heatsink to a passive Northbridge heatsink from Zalman
Removed the on-board input resistors R14 and R34
Removed capacitors C17 and C25
Removed capacitors C16 and C24 and replaced with appropriate shorting links
Removed the input suppressors underneath the board, by the RCA inputs
Volume controlled bay a 100k Alps pot
Used 4.7uF metalised polypropylene (LCR brand) capacitors between the pot and board input
Used the manual control on the 24V SMPS to turn it up to 27.1V
Mods left to do:
Remove the polarity diodes (replacing with a short) from the DC rails on the board
Consider changing the 24V supply for a 36V supply (turned down to 32V)
Add a 0.47uF + 10 Ohm Zobel network
I think these are most of them, I am sure someone may have other mods, but this will keep you busy.
Last edited:
Thanks! I have some of those inductors from Vishay, will that do the trick:
Vishay - Inductors - IHLP-5050CE-07 - 5 % DCR Tolerance, Low Profile, Power Inductors
Vishay - Inductors - IHLP-5050CE-07 - 5 % DCR Tolerance, Low Profile, Power Inductors
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.