I don't know why they opted to squirt the back side of the boards with heat sink grease rather than directly connecting the front side chips thermally to the heat sinks. The heat sink grease they use is slightly electrically conductive, so it can foul up circuits. Also, their components are ghz speed, so that may have something to do with their choice. A little of their grease on one of those chips and it will shut down the blade. .. even short out the power supply. The sinks have quite a few holes already drilled and tapped (20 to be exact) .. If you were mounting single hole components, you could likely skip. the drilling/tapping. Some of these units don't work, and they are selling them for dirt cheap like as low as $12 or even $5. but they are freakin awesome heat sink /fan combos ... the original miners run the fans at multiple speeds depending on temperature. I don't know the electrical details of that, but the fans are easy to remove. I wound up drilling new holes in the frame to get the sinks where I wanted them. There are thousands of these units out there, so in the long run, the DIY audio people have an awesome supply of kick-*** components for next to nothing. The plate side of the sinks is a full 1/4 inch thick. If you wanted to purchase sinks like these they would cost close to $50 each and the fan is high quality. It is rated at 12 volt 1 amp, so they can scream or run slowly. The fans have 4 wires .. I'm guessing two are for speed feedback to the blades. The frame is perfect for mounting and managing. .. Dream material for DIY audio people .. probably the entire DIY audio user group should be made aware of them. These things will keep showing up and up and up, in newer and newer models even eventually like the s2, s3, s4 etc. and antminer is not the only bitcoin miner manufacturer. BTC mining is not going away any time soon, and the old asics tend to be replaced with newer, faster ones.. The S1's are totally out of style now because now they have S7 ... way more efficient. These things won't become more expensive ... they'll get cheaper Good idea for a DIYer to stock up though because eventually they won't be able to build faster asics any more. That will cut the supply down and it will trickle away, but not before supplying the world with hundreds of thousands of giant heat sinks.
Also, you can very easily swap one of the sinks around so the fins point outwards by just loosening the screws, spin around, and put them back in. I have mine already re-worked with new holes.. I wanted the plate side of one sink to come all the way to the bottom of the frame so I can stick it to the big transformer in my sx-1250, so I had to drill holes into the corners of the frame and then grind access to them with a grinder, cause the bent metal gets in the way. After scooting the two heat sinks apart, I have a fabulous empty space between them to mount stuff. It's just enough space to mount 2 iraud200 boards as well as a double pack of capacitors I sawed out of the old iraud350 boards that weren't good enough in my estimation. I just bypassed their bridge rectifiers with a few blobs of solder. When I get ready to snap an image, The DIYers will catch fever over it. I am using cardboard in stead of standoffs .. cheaper and maybe even easier. I'm going to guess that there will be a lot of copies done of my sx-1250 mod. This mod will make the 1250 kick *** all over the 1980 .. and you won't be able to do the same mod for the 1980 because it's voltage is too high, but new mods could be worked out ... also though .. the 1980 seems to be higher voltage, but lower current, and the supply is only slightly stronger than the 1250, so the 1250 mod is basically just as good, only the 1250s are way cheaper.
Also, you can very easily swap one of the sinks around so the fins point outwards by just loosening the screws, spin around, and put them back in. I have mine already re-worked with new holes.. I wanted the plate side of one sink to come all the way to the bottom of the frame so I can stick it to the big transformer in my sx-1250, so I had to drill holes into the corners of the frame and then grind access to them with a grinder, cause the bent metal gets in the way. After scooting the two heat sinks apart, I have a fabulous empty space between them to mount stuff. It's just enough space to mount 2 iraud200 boards as well as a double pack of capacitors I sawed out of the old iraud350 boards that weren't good enough in my estimation. I just bypassed their bridge rectifiers with a few blobs of solder. When I get ready to snap an image, The DIYers will catch fever over it. I am using cardboard in stead of standoffs .. cheaper and maybe even easier. I'm going to guess that there will be a lot of copies done of my sx-1250 mod. This mod will make the 1250 kick *** all over the 1980 .. and you won't be able to do the same mod for the 1980 because it's voltage is too high, but new mods could be worked out ... also though .. the 1980 seems to be higher voltage, but lower current, and the supply is only slightly stronger than the 1250, so the 1250 mod is basically just as good, only the 1250s are way cheaper.
Last edited:
now trying to deal with hot mode
so, at the moment, both boards working, one hot, one cool .. I think the hot one was cool and switched back to hot. What I thinking is happening is that we have a HF parasite that may or may not make it past the tipping point. The reasons I believe that is because all three transistors + inductor + cap had problems. Cross conduction would only heat the transistors. Hot inductor and blown cap means we have a lot of hf that may or may not be there because the thing can also be cool. Since the circuit can run cool, that indicates that the hf probably should not be there, which means that it is probably a parasite. Looking the circuit over, I see some differences between iraud200 and the original circuit that may effect this issue.
1: some small power supply stabilizing caps and absorption resistor are not there. The engineer may have thought that adding large supply buffering caps solves this problem, but a: those kinds of caps don't necessarily respond in the 0.5mhz - 100mhz range and b: there is no resistor to eat up stray vibrations
2: This board has diodes that will allow the output to bleed charge straight back to the power supply if there is tank circuit oscillation. This could potentially trip up the behavior of the output stage. The whole idea of this may have been to protect the output filter cap, but as we can see, that didn't work. Caps can be punished not just by high voltage, but also by large high frequency currents. .. which is what will happen if parasites show up.
3: The feedback loop has adjusted resistors for a higher gain factor as well as for feedback loop hf filtering, and higher input impedance. These higher resistance values I think may be causing some ambient hum to leak into the feedback loop. .. not sure about that because 6k input impedance should probably not allow hum to leak in. The hum may be leaking in some other way. There isn't much of it, but its enough to know there is an issue.
numbers 1 and 3 are my largest concerns, and so I have a plan to approach this step wise. I will quit when hot mode goes away. Step 1: add power supply stabilising caps plus resistor (around 0.1 uf and 1 ohm), see if it helps maybe also try a small dampening cap/resistor pair from output to ground to try to kill ringing. step2:reduce the gain factor as much as 10% step 3: adjust some of the feedback loop resistors to see if that helps step4: reduce the gain factor even more.
Ultimately, killing the gain factor will kill parasites, but we don't want to do that. Bringing it down 10% is no biggie, but taking off more than that may be a big loss. We want to try everything else before taking more than 10% off of the gain factor. .. So, rather than go through this tit for tat in online text, I am just going to scrap with it and then come back with the results.
so, at the moment, both boards working, one hot, one cool .. I think the hot one was cool and switched back to hot. What I thinking is happening is that we have a HF parasite that may or may not make it past the tipping point. The reasons I believe that is because all three transistors + inductor + cap had problems. Cross conduction would only heat the transistors. Hot inductor and blown cap means we have a lot of hf that may or may not be there because the thing can also be cool. Since the circuit can run cool, that indicates that the hf probably should not be there, which means that it is probably a parasite. Looking the circuit over, I see some differences between iraud200 and the original circuit that may effect this issue.
1: some small power supply stabilizing caps and absorption resistor are not there. The engineer may have thought that adding large supply buffering caps solves this problem, but a: those kinds of caps don't necessarily respond in the 0.5mhz - 100mhz range and b: there is no resistor to eat up stray vibrations
2: This board has diodes that will allow the output to bleed charge straight back to the power supply if there is tank circuit oscillation. This could potentially trip up the behavior of the output stage. The whole idea of this may have been to protect the output filter cap, but as we can see, that didn't work. Caps can be punished not just by high voltage, but also by large high frequency currents. .. which is what will happen if parasites show up.
3: The feedback loop has adjusted resistors for a higher gain factor as well as for feedback loop hf filtering, and higher input impedance. These higher resistance values I think may be causing some ambient hum to leak into the feedback loop. .. not sure about that because 6k input impedance should probably not allow hum to leak in. The hum may be leaking in some other way. There isn't much of it, but its enough to know there is an issue.
numbers 1 and 3 are my largest concerns, and so I have a plan to approach this step wise. I will quit when hot mode goes away. Step 1: add power supply stabilising caps plus resistor (around 0.1 uf and 1 ohm), see if it helps maybe also try a small dampening cap/resistor pair from output to ground to try to kill ringing. step2:reduce the gain factor as much as 10% step 3: adjust some of the feedback loop resistors to see if that helps step4: reduce the gain factor even more.
Ultimately, killing the gain factor will kill parasites, but we don't want to do that. Bringing it down 10% is no biggie, but taking off more than that may be a big loss. We want to try everything else before taking more than 10% off of the gain factor. .. So, rather than go through this tit for tat in online text, I am just going to scrap with it and then come back with the results.
Last edited:
I didn't know about the first part of the info...
The board actually has the wholes for two inductors (with extra wholes for different lead spacing).
I'm going to place an order at a shop that sells ring-core chokes.
They have
40µH rated for 20A and
47µH / 5A.
Both might only fit on the bottom side of the board.
Would 2x 5A be enough?
Will heat be a problem with the inductor on the bottom side?
Did you swap the inductor already?
I am still using the original inductor .. however, mine is not the same as yours, and it's primarily the inductor/transistor combination that determines the power rating. Yours is smaller ... so, guess what ... to get 600w 2 ohm, you have to have 17 amps. The inductors they use are pretty much size minimal. I found a really big one on ebay once. 1pcs New Sagami Digital Amplifier Large Current Shielded Inductor 7G31A 220 22UH | eBay I'm afraid even the inductor I have is not really rated for the 17 amps, which is why I added a heat sink. (big one) but heat sink is not the only issue ... inductors can saturate. I think that one I just posted will fit the bill, but an additional heat sink would still be on my list cause if a parasite fires up, then the inductor takes a serious beating along with the output filter cap. You can't add a heat sink to toroidal inductor very well but 2 40uh inductors in parallel would be good. 10 amps would handle a 4 ohm load.. but the inductor you have would probably already do that with a heat sink. I think these board manufacturers tend to expect more out of inductors than they are rated for .. which works better than the same idea for transistors. Especially with heat sink, you can push a little extra on an inductor.. pretty much never ask a transistor to do more than 2/3 its rated power .. though voltage ratings would be more believable for transistors.
As it goes ... I have had some partial success parasite squashing. I got one of the boards to fire up several times in a row without going to hot mode. The other board is so determined to freak on start-up that it fries my dampener resistors in short order. Right now I only know of one difference between the boards and that is that the one behaving the best has 10uf more stabilizing the piggyback power supply. ... just interim notice I don't have final info. on parasite squashing yet, but if the parasite is active, it's very very mean to dampener resistors on the output. I think it could dump 50 watts into them easily. Have to bring the parasite to almost death before output ring dampening resistors are any good.
As it goes ... I have had some partial success parasite squashing. I got one of the boards to fire up several times in a row without going to hot mode. The other board is so determined to freak on start-up that it fries my dampener resistors in short order. Right now I only know of one difference between the boards and that is that the one behaving the best has 10uf more stabilizing the piggyback power supply. ... just interim notice I don't have final info. on parasite squashing yet, but if the parasite is active, it's very very mean to dampener resistors on the output. I think it could dump 50 watts into them easily. Have to bring the parasite to almost death before output ring dampening resistors are any good.
oh ... btw ... don't touch any inductor that isn't rated for 800khz at least ... 500 khz may be ok after doing the feedback loop mod. But I have this sneaking suspicion I may wind up modding the feedback loop again. I'll probably wind up kicking the gain factor down to 30 or lower. MHZ speed is not uncommon for powdered iron, but I would have my doubts about ferrite. I think there are some higher speed ferrites. Trust me ... not just any inductor will do. This thing punishes inductors, so if it aint ready for 800 khz ... it aint ready for this board.
yeah, the board has 8 holes .. one set of four for the 4 ohm rated inductor, and one set of four with larger spacing for the 2 ohm rated inductor. Larger than the 2 ohm rated inductor won't fit in the designated position and will have to be specially mounted. Whichever set of 4 you use, only two of the holes actually conduct current. The other two connect to inductor leads that do nothing. ..That is, unless you bring in your own inductors.. in which case, you could use a pair of 40uh inductors in parallel and use four holes for them.
sorry ... got hung up
I got hung up and haven't been able to fully investigate the parasite problem so far. However, for anyone interested in info. now, I have some. I tried switching the delay cap in the feedback loop to 200pf .. it seemed like that cooled down the transistors and slightly heated up the inductor. Adjusting the 360 ohm delay resistor just between the chip inputs didn't get me anything. After so much playing around, what I think is that the output filter is ringing, and the ring kicks back into the feedback loop. So, for the one board, dampening the ring stopped the parasite ... for the other it didn't ... however, the cap meter doesn't read both caps the same, so I think that there is somehow accidentally a difference between capacitances, which is causing one board to work and the other to heat up almost to overtemp. I thought it might be because I drilled the hole out for one of the caps .. next time I'll try using air and soldering iron at the same time in stead. It's impossible to solder the proper side of the board underneath the cap where it connects to ground, so I think what was turning hot and cold mode off and on was the ground lead of the output filter cap connecting/disconnecting from the ground plane. When I soldered the lead straight to ground through a different path, the board stayed in hot mode permanently ... so, what I think is that for the board that is working, the capacitor is somehow smaller .. like 0.27 or 0.33 uf .. this changes the ring behavior, and makes the filter not work quite as well, but killing the parasite is worth it I think. So, my next attempt will be to swap out the 0.47 for something between 0.22 and 0.33 with an additional 0.1 dampened with 10 ohms to see if that kills the parasite. If I think it needs more filtering, I will shoot for a third or perhaps third and fourth pole to the filter ... like another inductor at 10 uh in series with the output. ... problem is that for certain frequency ranges, that cap acts like almost a short to ground and causes a lot of ringing. Some characteristics of the board probably add to this behavior. So, that's where I'm at, and I'll update after I have tested out my theory.
I got hung up and haven't been able to fully investigate the parasite problem so far. However, for anyone interested in info. now, I have some. I tried switching the delay cap in the feedback loop to 200pf .. it seemed like that cooled down the transistors and slightly heated up the inductor. Adjusting the 360 ohm delay resistor just between the chip inputs didn't get me anything. After so much playing around, what I think is that the output filter is ringing, and the ring kicks back into the feedback loop. So, for the one board, dampening the ring stopped the parasite ... for the other it didn't ... however, the cap meter doesn't read both caps the same, so I think that there is somehow accidentally a difference between capacitances, which is causing one board to work and the other to heat up almost to overtemp. I thought it might be because I drilled the hole out for one of the caps .. next time I'll try using air and soldering iron at the same time in stead. It's impossible to solder the proper side of the board underneath the cap where it connects to ground, so I think what was turning hot and cold mode off and on was the ground lead of the output filter cap connecting/disconnecting from the ground plane. When I soldered the lead straight to ground through a different path, the board stayed in hot mode permanently ... so, what I think is that for the board that is working, the capacitor is somehow smaller .. like 0.27 or 0.33 uf .. this changes the ring behavior, and makes the filter not work quite as well, but killing the parasite is worth it I think. So, my next attempt will be to swap out the 0.47 for something between 0.22 and 0.33 with an additional 0.1 dampened with 10 ohms to see if that kills the parasite. If I think it needs more filtering, I will shoot for a third or perhaps third and fourth pole to the filter ... like another inductor at 10 uh in series with the output. ... problem is that for certain frequency ranges, that cap acts like almost a short to ground and causes a lot of ringing. Some characteristics of the board probably add to this behavior. So, that's where I'm at, and I'll update after I have tested out my theory.
conundrum
I WAS contemplating purchasing a few of these boards to add to my minidsp multi-driver active monstrosity prototype. However, seeing that there are not five (5) pages dedicated to a redesign of the board, I'll pass.
That said, I am interested in knowing if such a problem is common with class d modules offered in Amazon, eBay, or Aliexpress, or even modules sold by more recognized names (connectelectronix is another source I am contemplating, hyphex being too expensive for what I am to attain).
Would purchasing class d modules from connectelectronix ensure a proper working module?
I WAS contemplating purchasing a few of these boards to add to my minidsp multi-driver active monstrosity prototype. However, seeing that there are not five (5) pages dedicated to a redesign of the board, I'll pass.
That said, I am interested in knowing if such a problem is common with class d modules offered in Amazon, eBay, or Aliexpress, or even modules sold by more recognized names (connectelectronix is another source I am contemplating, hyphex being too expensive for what I am to attain).
Would purchasing class d modules from connectelectronix ensure a proper working module?
when I get it all resolved, I think I'll re-post a short version of the story. My experience with class D modules from Ebay so far is that there is usually some kind of dissatisfaction. If you can anticipate that in advance, you may be able to get around it, but it seems they don't have quality control under control at least in China. .. For example ... I think the l25D board may be fine so long as you don't ask it to drive below 8 ohms. There was an interesting one out there coming from the Philippines that is comparable to the iraud200 that I can't find any more. .. it might be more promising. In any case, if what you need is the short story for this board, well, I got in some parts today and am hoping to know the scoop within a week. If I can kill the parasite on the second board (which I think I can) then a small number of steps may convert junk into gold. On the other hand .. Behringer has come out with some genuine consumer grade class D amps available at multiple locations such as mcm that are starting to interest me.
I haven't seen connectelectronics stuff, but sure sells through parts-express, which I am quite confident has more quality control management than ebay. I have a board from them, but don't have the right power supply for it. .. anyways, I doubt that parts-express would sell anything likely to come back as they are a US vendor, and you can easily return your stuff to them. My problem with their boards was finding one that matches a power supply I already had. If you are ready to get a supply to match, then I think they would have something to interest you.. a little more expensive than the ebay stuff, but there is no question that having something that works out of the box would be worth it.
I haven't seen connectelectronics stuff, but sure sells through parts-express, which I am quite confident has more quality control management than ebay. I have a board from them, but don't have the right power supply for it. .. anyways, I doubt that parts-express would sell anything likely to come back as they are a US vendor, and you can easily return your stuff to them. My problem with their boards was finding one that matches a power supply I already had. If you are ready to get a supply to match, then I think they would have something to interest you.. a little more expensive than the ebay stuff, but there is no question that having something that works out of the box would be worth it.
Last edited:
connectelectronix it is then, since I am more concerned with investing the already diminished portion of my time in designing a suitable enclosure than venturing down a rabbit hole of unpredictability.
That said, I will revisit to view the outcome of your venture.
Thank you for your response.
well, at this point you have a new option because I just succeeded in squashing the parasite on the second board. Basically, I replaced the output filter cap with a smaller one of 0.27 uf 275 volts film cap.. that cooled down the heat sink some, but the parasite was still active. Next, I took a 0.22 uf cap .. soldered it in series with 4.1 ohms 4 watts and attached it between the speaker output and the 0 volt output (the two output terminals) of the board. You can actually listen to what happens by turning on an old fm radio tuner and listening to the disturbance of the radio that the class D boards produce. When I attached the capacitor/resistor to the outputs, the hiss on the radio substantially reduced, and the inductor heat sink began to cool down 60-90 minutes later it became difficult to tell the difference between the temperatures of the heat sinks of the two inductors on the separate boards. .. Parasite dead. The class D boards still disturb the radio tuner somewhat .. especially for weak stations, I think this is a given for class D, but I am going to see what happens if I add a third filter pole in ( a 10 uh inductor I think) The second inductor should not have to be highly expensive because the rf will be mostly dead before it gets there .. it's just a clean-up inductor I'm going to take a ferrite core, cut a slot in it about 1/8 inch wide and wind it until my meter says 10 uh or maybe 7-8 uh. this is not a requirement for the system, but I am hoping to stop the class D boards from interfering with the radio tuner. .. In any case ... all fixed now!!!!!!! I'll start writing up a short version of the story tomorrow. .. I'll title it iraud200 short story and it will be a new thread.
I'll start writing up a short version of the story tomorrow. .. I'll title it iraud200 short story and it will be a new thread.well, at this point you have a new option because I just succeeded in squashing the parasite on the second board. Basically, I replaced the output filter cap with a smaller one of 0.27 uf 275 volts film cap.. that cooled down the heat sink some, but the parasite was still active. Next, I took a 0.22 uf cap .. soldered it in series with 4.1 ohms 4 watts and attached it between the speaker output and the 0 volt output (the two output terminals) of the board. You can actually listen to what happens by turning on an old fm radio tuner and listening to the disturbance of the radio that the class D boards produce. When I attached the capacitor/resistor to the outputs, the hiss on the radio substantially reduced, and the inductor heat sink began to cool down 60-90 minutes later it became difficult to tell the difference between the temperatures of the heat sinks of the two inductors on the separate boards. .. Parasite dead. The class D boards still disturb the radio tuner somewhat .. especially for weak stations, I think this is a given for class D, but I am going to see what happens if I add a third filter pole in ( a 10 uh inductor I think) The second inductor should not have to be highly expensive because the rf will be mostly dead before it gets there .. it's just a clean-up inductor I'm going to take a ferrite core, cut a slot in it about 1/8 inch wide and wind it until my meter says 10 uh or maybe 7-8 uh. this is not a requirement for the system, but I am hoping to stop the class D boards from interfering with the radio tuner. .. In any case ... all fixed now!!!!!!!
Impressive. I presume comparison photos will be part of the brief? I will read the brief when available.
@mgshightech
I just recalled an article I had read a while ago concerning thermal transfer efficiency in heatsink design and application. (I do not have the article with me, nor recall the website.) In essence, what the author of the article proved was how efficiency of heat dissipation is impaired when even a small discrepancy is introduced in attaching components to a heatsink (including the application of too much thermal paste). I presume the heatsink mod you introduced is temporary or simply a rough assembly to affect some reduction in heat? I once had a coupler manufactured to transfer heat from the CPU of a mini-itx motherboard to a heatsink to transfer heat sans the use of any fans. From studying the specifications, I elected 6061 aluminum for the coupler. The machinist told me he did not think heat could transfer fast enough through approximately 1.5" of aluminum, but proceeded to create the coupler to spec. Once applied, the coupler did indeed work very efficiently, transferring all of the heat to the heatsink. I plan to affect the same approach with the modules I elect for the final amplifier design, since it will be a hermetically sealed design. Perhaps you should consider having a coupler made to increase the efficiency of heat transfer to the respective heatsink(s). (Of course, if you achieved a significant reduction of heat with the mod, such may not be necessary, and the original heatsink will suffice.)
I just recalled an article I had read a while ago concerning thermal transfer efficiency in heatsink design and application. (I do not have the article with me, nor recall the website.) In essence, what the author of the article proved was how efficiency of heat dissipation is impaired when even a small discrepancy is introduced in attaching components to a heatsink (including the application of too much thermal paste). I presume the heatsink mod you introduced is temporary or simply a rough assembly to affect some reduction in heat? I once had a coupler manufactured to transfer heat from the CPU of a mini-itx motherboard to a heatsink to transfer heat sans the use of any fans. From studying the specifications, I elected 6061 aluminum for the coupler. The machinist told me he did not think heat could transfer fast enough through approximately 1.5" of aluminum, but proceeded to create the coupler to spec. Once applied, the coupler did indeed work very efficiently, transferring all of the heat to the heatsink. I plan to affect the same approach with the modules I elect for the final amplifier design, since it will be a hermetically sealed design. Perhaps you should consider having a coupler made to increase the efficiency of heat transfer to the respective heatsink(s). (Of course, if you achieved a significant reduction of heat with the mod, such may not be necessary, and the original heatsink will suffice.)
here's my problem ... I don't have a proper pair of 600 watt dummy loads .. just my 2 ohm horn loaded 15 inchers .. computers and such stacked on top of the horns .. when I start to run this thing up to power, well ... it's scary. Those drivers have over 2 inch p-p throw, and it's really loud and I don't dare hold it there for several hours.. I'll wind up with neighbors with question marks in their heads and maybe some broken computers or something, and all my little odds and ends fallen to the floor. The boards are now cool at quiescent state. I don't know how hot they'll get at full power, but if the data I have gathered is any good, well,
a: manufacturer already suggested adding heat sink for 2 ohm load
b: my other class D boards didn't have enough heat sink when driven into low impedance
c:I have checked out the specs on comparable inductors to the ones on the boards, and they may handle up to something around 20 amps DC. However, they are rated much lower for 500 mhz-ish operation .. that tells me that they'll be overheating before they saturate, which tells me that for 2 ohm load, the inductors will be needing some extra heat sink, and with that, they'll be good and won't need replacing, which is a daunting task. So, now I have added quite a bit of extra heat sink to the system to make sure stuff works.
I just stacked a heat sink straight on top of the inductor.. admittedly, this is not the only option. In fact, I considered using a piece of rod as you were suggesting to connect from the inductor to a single, monolithic heat sink that would sink for everything, but I realized that tolerances would be tough to get right. For that reason, I stuck with just putting a heat sink straight to the inductor ... which would be impeded if you have purchased one of these boards with extra size storage capacitors. .. so, probably don't get big caps unless you want to cut an aluminum cylinder as you say. Big caps are a bad idea anyways because if something goes wrong with the board, the caps have enough energy to blow trace off of the board, causing permanent damage. You want fuses between the board and any really big caps. In any case, what you say about heat sinks is true. too much heat sink grease will cause a weakness, but a one inch cylinder of aluminum or even better copper will not have much trouble conducting heat away from the inductor. Aluminum and Copper are pretty much the best at that ... very very good. .. I already dropped images of the test and development setup onto this thread for looking at. However, if you really want cool heat sinks for cheap, go to ebay and look up antminer s1.
I'll be using one of those for my final setup. The s1 comes with two huge heat sinks and a big fan to boot. .. and you can get one for around $20. wow I'm going to use one of the big sinks to sink the transformer of this pioneer sx-1250 so it will be good for another 300 watts at least. Then I'll mount the boards so that the top of the heat sink mounts directly to the flat face of the other large s1 heat sink, and grease the two sinks together. My experience with this is that a big heat sink can suck the heat out of one of these little sinks in less than a minute. There won't be any trouble keeping the original sinks cool under these conditions, and the boards will be good for 900 watts or so into 2 ohms as rated. .. if your power supply can supply it ... .. the supply I have will only be good for around 600 watts/channel, and that assumes I am not using the other two hi-fi class AB channels. .. but 600 w/channel is all my subs need for now. Go look up the antminer s1 on ebay and have a DIYer orgasm. It's an old bitcoin miner that is no longer competent, so they sell for far less than their heat sinks are worth.
a: manufacturer already suggested adding heat sink for 2 ohm load
b: my other class D boards didn't have enough heat sink when driven into low impedance
c:I have checked out the specs on comparable inductors to the ones on the boards, and they may handle up to something around 20 amps DC. However, they are rated much lower for 500 mhz-ish operation .. that tells me that they'll be overheating before they saturate, which tells me that for 2 ohm load, the inductors will be needing some extra heat sink, and with that, they'll be good and won't need replacing, which is a daunting task. So, now I have added quite a bit of extra heat sink to the system to make sure stuff works.
I just stacked a heat sink straight on top of the inductor.. admittedly, this is not the only option. In fact, I considered using a piece of rod as you were suggesting to connect from the inductor to a single, monolithic heat sink that would sink for everything, but I realized that tolerances would be tough to get right. For that reason, I stuck with just putting a heat sink straight to the inductor ... which would be impeded if you have purchased one of these boards with extra size storage capacitors. .. so, probably don't get big caps unless you want to cut an aluminum cylinder as you say. Big caps are a bad idea anyways because if something goes wrong with the board, the caps have enough energy to blow trace off of the board, causing permanent damage. You want fuses between the board and any really big caps. In any case, what you say about heat sinks is true. too much heat sink grease will cause a weakness, but a one inch cylinder of aluminum or even better copper will not have much trouble conducting heat away from the inductor. Aluminum and Copper are pretty much the best at that ... very very good. .. I already dropped images of the test and development setup onto this thread for looking at. However, if you really want cool heat sinks for cheap, go to ebay and look up antminer s1.
I'll be using one of those for my final setup. The s1 comes with two huge heat sinks and a big fan to boot. .. and you can get one for around $20. wow I'm going to use one of the big sinks to sink the transformer of this pioneer sx-1250 so it will be good for another 300 watts at least. Then I'll mount the boards so that the top of the heat sink mounts directly to the flat face of the other large s1 heat sink, and grease the two sinks together. My experience with this is that a big heat sink can suck the heat out of one of these little sinks in less than a minute. There won't be any trouble keeping the original sinks cool under these conditions, and the boards will be good for 900 watts or so into 2 ohms as rated. .. if your power supply can supply it ... .. the supply I have will only be good for around 600 watts/channel, and that assumes I am not using the other two hi-fi class AB channels. .. but 600 w/channel is all my subs need for now. Go look up the antminer s1 on ebay and have a DIYer orgasm. It's an old bitcoin miner that is no longer competent, so they sell for far less than their heat sinks are worth.
the new thread is here: http://www.diyaudio.com/forums/class-d/284655-iraud200-short-story.html#post4566776 that thread is not for discussion but reference, so please do discussing here in stead.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.