-
WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required. There are many reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!
Silly surface mount dummy load module
- Thread starter mcdn
- Start date
My reasoning is that the way your boards are built, in order for the back of the pcb to reach 100C, the resistors need to be necessarily at a temperature higher than that (note I am not considering the traces, but I think it is safe to do so, for different reasons). You don't seem to have a problem of too much power generated, as your heat sink is barely warm. In my opinion, your bottle neck is in the thermal coupling between the heat sink and the resistors, unless your experiment is to heat up everything first and then start the fans and take the measurements...
Don't know, I am for sure missing data that you might have, just thought you might be using your IR thermometer to measure the heat sinks like that.
Don't know, I am for sure missing data that you might have, just thought you might be using your IR thermometer to measure the heat sinks like that.
- Thread Starter
- #64
Well exactly, which is why I turned it off at that point
The front side of the resistors seems fine. The back side of the resistors (i.e. the side facing the board) could well be getting to 120C or more. Which is not great, and why I want to add heatsinks to the undersides of the boards. In a revision 2 I would add some big thermal vias under each chip amongst other improvements.But if I can do 200W per board with heatsinks at 80C, then chip temps are likely only 100C, and all is good? To remind everyone: the boards cost $5 each, fans are $5 each, heatsinks can be ordered for $1 each (but 16 are needed), 3D printing is optional. That's around $50 for a 2x400W dummy load. Not so silly after all!
DoctorRobert
Member
- Joined
- Jul 22, 2020
- Messages
- 41
- Likes
- 62
Normally, I would call this a 'very cool project' but I am not so sure about the wording now.
- Thread Starter
- #66
While I wait for the heatsinks, I sorted out V2. Key changes are:
- Single sided aluminium instead of FR4 for the board. This should give much better thermals overall.
- The heatsinks can be connected to the exposed aluminium on the back of the board, and also to the resistors on the front
- Use 4R7 2W 100ppm 1% resistors 19 parallel and 8 serial, for a total of 1.97ohms per board using 152 resistors. This should allow a bit of headroom for connector and trace resistance
- Slight reduction in vertical height to better fit a 140x140mm ducted case
- 2mm wide (1mm internal diameter) thermal vias under each resistor to address the heat buildup from the resistor undersides.
- Definitely using these heatsinks: https://www.aliexpress.com/item/32831608617.html as they are cheap and fit nicely
- M4 holes for cable connection - but note these are only on the top side, you must insulate any screws/bolts/washers on the underside due to the use of single sided aluminium board
Attachments
Last edited:
Really cool idea. Given the low cost of the boards, it would seem efficent to try an eight board of even 32 board 4 ohm load to get max power dissipation without needing much in the way of heatsinks.
Last edited:
- Thread Starter
- #68
KiCAD files and gerbers
github.com
GitHub - mattclarkdotnet/dummyload
Contribute to mattclarkdotnet/dummyload development by creating an account on GitHub.
github.com
- Thread Starter
- #69
Maybe? You still need airflow, and if you use more boards the size of the thing grows to be inconvenient. I'm betting with the V2 boards I can do 800W in a 140x140x200mm ducted case. Don't stand too close to the hot end though!
- Thread Starter
- #70
To clarify, that's $20 per assembled board, i.e. it includes all the resistors, nicely soldered on by our friendly robots at JLC. The boards themselves would only be $2 each.
- Thread Starter
- #71
- Thread Starter
- #72
In the FTC thread I was reminded by @SIY to post measurements. Including the power leads, it's 2.022 ohms at 20C and 2.042 ohms at 60C which is where it reaches steady state at 800W. That's 95ppm/C for the 4R7 resistors, which is bang on spec! (LCSC part page).
- Thread Starter
- #73
I have some more accurate temp measurements now I've installed a ds18b20 on the backside of one of the modules. At 800W at 20C ambient with both fans running at 14V the board stabilises at 70.5C. That fits well with the slight R increase seen with temp, so I'm pretty confident the devices aren't running super hot. The fans are 12V rated but 14V is needed to keep the temp delta to 50C.
Very cool, I never bothered check out Al PCBs in jlcpcb thinking they would be super expensive, but now I can't stop thinking what could I do with that!
I would be very curious to see what happens with only the back side heat sink.
By the way, have you thought on other methods to couple the heat sink to the boards?. I am using Indium with my DIY tpa3255. That could work for you too for the back side, but it could get a bit expensive with your surface area + you would need a means of clapping with some pressure, that is the only detail I am missing in your design.
I would be very curious to see what happens with only the back side heat sink.
By the way, have you thought on other methods to couple the heat sink to the boards?. I am using Indium with my DIY tpa3255. That could work for you too for the back side, but it could get a bit expensive with your surface area + you would need a means of clapping with some pressure, that is the only detail I am missing in your design.