Announcement

Collapse
No announcement yet.

Why so few higher powered Class D stereo amp boards?

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

  • tomzarbo
    replied
    Originally posted by neildavis View Post
    Any opening will significantly lower the heat at the amplifier heatsink. It's all about "thermal resistance"--if the thermal resistance to ambient air drops to near zero, the temperature will rise quite a bit.

    I've been using heat pipes to draw the heat to an aluminum plate in the back of the chassis. The 4mm flat pipes available at Digikey fit into the heatsink fins of some of these amps. You can buy a tubing bender from Home Depot to help shape the heat pipe, but with the flat pipe you just need to wrap it around some tubing--the flat heat pipes are easy to work with.. I mix up some of the white thermal paste with 5 minute epoxy (about 1 to 1 mix) for DIY thermal epoxy. You can see the heat pipes in this picture, but this was before using the epoxy:



    Next time I'll buy a shorter heap pipe . I also used some thick silicon thermal pads to heatsink the power supply to the mounting plate. The back panel gets warm, so I know the heat sinking is working.
    Neil, that is a cool way to get rid of some heat. (I wasn't trying to be funny there)

    I like how that works, and I can totally understand how it would do the job. I solder a wire holding it 2-3 inches from the iron, and in a few seconds it's getting pretty dang hot. That looks to be a great way to move heat around to where it can go away easier. I'm going to use that in the future I think. I'm glad I have smart friends to help me with all these projects... or I would have probably burned my house down by now.

    TomZ

    Leave a comment:


  • tomzarbo
    replied
    Originally posted by Millstonemike View Post

    TOm,

    I went to your build thread to see the amp's position in the cabinet. Didn't see the vent so that's better than I thought.

    Just to make sure ... when you post 120o, is that Celsius or Fahrenheit? When I quoted the data sheet's auto shutdown temp, 150o, that's in Celsius. Operating temp range goes up to 85o C (185o F).

    And your right. The temp in the chip will be hotter than measured at the heat sink.
    Click image for larger version

Name:	BTRTerminalVent.jpg
Views:	78
Size:	259.7 KB
ID:	1424232
    Hey, we've got pics again!

    Mike, that was
    Fahrenheit. So the chip can get to 185 degrees F before it cuts out? I don't think I'm getting close to that if I had to guess on the chip itself.

    Here's a pic of the vent and you can see the little heat sink behind it. Hopefully it does the job.

    Thanks for the input and all the guidance too.

    TomZ

    Leave a comment:


  • JRT
    replied
    Originally posted by westrock2000 View Post
    I’d like to see more digital amps that are rated, even at 50 watts, but with less than 1% distortion. Many of these digital amps are rated at at 10% for their advertised power output.
    10% is a good point on the curve to indicate that the amplifier is well into clipping, and is good for assessing headroom. At half of that power level, at signal levels -3_dB below the signal level associated with 10% THD, the amplifier should be operating below clipping and producing much less nonlinear distortion.

    To avoid the nonlinearities associated with clipping, I would argue for 0_dBFS signal levels on tweeters not exceeding a threshold of -6_dB below the signal level associated with 10%_THD, and amplifiers for other drivers not exceeding a threshold of -3dB below the signal level associated with 10% THD.

    Although in music the average power at high frequencies is much lower than it is in lower midrange, the brief crests can be higher. For tweeters, signal crests can reach +3_dB above 0_dBFS, for peaks centered between samples at frequencies near 1/2 sample rate. For midrange and below, crests can extend very much longer than high frequency crests, but do not rise significantly above 0_dBFS in level.

    Leave a comment:


  • neildavis
    replied
    Any opening will significantly lower the heat at the amplifier heatsink. It's all about "thermal resistance"--if the thermal resistance to ambient air drops to near zero, the temperature will rise quite a bit.

    I've been using heat pipes to draw the heat to an aluminum plate in the back of the chassis. The 4mm flat pipes available at Digikey fit into the heatsink fins of some of these amps. You can buy a tubing bender from Home Depot to help shape the heat pipe, but with the flat pipe you just need to wrap it around some tubing--the flat heat pipes are easy to work with.. I mix up some of the white thermal paste with 5 minute epoxy (about 1 to 1 mix) for DIY thermal epoxy. You can see the heat pipes in this picture, but this was before using the epoxy:



    Next time I'll buy a shorter heap pipe . I also used some thick silicon thermal pads to heatsink the power supply to the mounting plate. The back panel gets warm, so I know the heat sinking is working.

    Leave a comment:


  • Millstonemike
    replied
    Originally posted by tomzarbo View Post

    Hey Mike,

    I have the amp fastened to the top of the cabinet using 1/4" (roughly) brass standoffs. It is positioned towards the rear of the cabinet as well. It was the best place to locate it to keep it out of the way placement-wise considering the other components in the space. I agree though, on the bottom would be better thermally speaking. Here is a link to my current "Bantam Table Radio" project and how it's mounted. (the radio is upside down in the pic.) http://techtalk.parts-express.com/fo...26#post1423326

    That temperature measurement is with the amp mounted permanently inside the cabinet. The entire bottom of the center portion of the cabinet is open (there is a screen covering it but it has airflow) and I have installed a soffit vent right next to the amp's heat sink fins to allow any hot air to escape.

    I can't post any pics of this at the moment until the forum gets fixed, but you can see the heat sink fins right next to the vent on the rear of the cabinet. After it's been running for awhile at high volume you can feel a bit of warmth coming from the vent area. Hopefully this will be sufficient to keep the amp cool when it's running at warp-factor 10 for long periods.

    I'm thinking that 120 degrees max would lie within it's normal operating temperature? I think you mentioned before that the chip itself would shut down around 150 degrees or so, and the heat sink would not probably get that hot, so maybe at high volume I'm getting closer to maxing it out than I think. Hopefully, adding the soffit vent right next to the heatsink fins will keep things 'cool enough.'

    Have to say though, it really sounds good. If anyone remembers hearing the 'Bantams' from DIY NY earlier this year, this is them -- just closer together. With this preamp, the TDA7492 amp really does the speaker justice at 24v and I couldn't be happier with how it sounds.

    I originally mentioned wanting to do two versions of this radio, this one, and another with a built-in subwoofer, possibly taller and deeper. I think I'll just stick with this setup for now as it just works so well. I think for a radio, this has bass enough to satisfy nearly anyone.

    I ended up using some black aluminum rigid grill-type material to make a covering for the open underside of the enclosure just to keep anything from getting in there and from any wires from peeking out. I covered the edges with a layer of black electrical tape to conceal the sharp edges. I was able to cut it easily with scissors.

    Hopefully soon I can post some pics of the last few detail items.

    TomZ
    TOm,

    I went to your build thread to see the amp's position in the cabinet. Didn't see the vent so that's better than I thought.

    Just to make sure ... when you post 120o, is that Celsius or Fahrenheit? When I quoted the datasheet's auto shutdown temp, 150o, that's in Celsius. Operating temp range goes up to 85o C (185o F).

    And your right. The temp in the chip will be hotter than measured at the heat sink.

    Leave a comment:


  • tomzarbo
    replied
    Originally posted by Millstonemike View Post

    Are you measuring temp with the unit open? There may be a significant difference when it's buttoned up. Did you mount the amp at the top of the table radio - where heat collects?

    I would mount the amp so the heat sink fins and channels are vertical at the bottom of the cabinet. That creates a passive airflow over the heat sink as it heats the surrounding air.

    See my link (since TT won't allow pic uploads - always a "corrupted file error").
    Hey Mike,

    I have the amp fastened to the top of the cabinet using 1/4" (roughly) brass standoffs. It is positioned towards the rear of the cabinet as well. It was the best place to locate it to keep it out of the way placement-wise considering the other components in the space. I agree though, on the bottom would be better thermally speaking. Here is a link to my current "Bantam Table Radio" project and how it's mounted. (the radio is upside down in the pic.) http://techtalk.parts-express.com/fo...26#post1423326

    That temperature measurement is with the amp mounted permanently inside the cabinet. The entire bottom of the center portion of the cabinet is open (there is a screen covering it but it has airflow) and I have installed a soffit vent right next to the amp's heat sink fins to allow any hot air to escape.

    I can't post any pics of this at the moment until the forum gets fixed, but you can see the heat sink fins right next to the vent on the rear of the cabinet. After it's been running for awhile at high volume you can feel a bit of warmth coming from the vent area. Hopefully this will be sufficient to keep the amp cool when it's running at warp-factor 10 for long periods.

    I'm thinking that 120 degrees max would lie within it's normal operating temperature? I think you mentioned before that the chip itself would shut down around 150 degrees or so, and the heat sink would not probably get that hot, so maybe at high volume I'm getting closer to maxing it out than I think. Hopefully, adding the soffit vent right next to the heatsink fins will keep things 'cool enough.'

    Have to say though, it really sounds good. If anyone remembers hearing the 'Bantams' from DIY NY earlier this year, this is them -- just closer together. With this preamp, the TDA7492 amp really does the speaker justice at 24v and I couldn't be happier with how it sounds.

    I originally mentioned wanting to do two versions of this radio, this one, and another with a built-in subwoofer, possibly taller and deeper. I think I'll just stick with this setup for now as it just works so well. I think for a radio, this has bass enough to satisfy nearly anyone.

    I ended up using some black aluminum rigid grill-type material to make a covering for the open underside of the enclosure just to keep anything from getting in there and from any wires from peeking out. I covered the edges with a layer of black electrical tape to conceal the sharp edges. I was able to cut it easily with scissors.

    Hopefully soon I can post some pics of the last few detail items.

    TomZ

    Leave a comment:


  • Millstonemike
    replied
    Originally posted by tomzarbo View Post
    I had been hesitant to go the full 24v because of the 'warnings' on the product page, even though you've even previously tried to asswage my fears Mike... but you're absolutely right... it seems that these amps can take it just fine.

    I've been doing some near full-output temperature testing on the TDA7492 amp: https://www.parts-express.com/tda749...2x50w--320-606 and with the video player preamp faceplate feeding the source material it's peaking at around 120 or so degrees on the heatsink. That's pretty chill for a legit' amplifier producing real output. I like that little amp a lot more now. It's not until the very last three or four clicks on the preamp that it starts sounding harsh, but by then the woofers are going pretty much berserk at that point so it's a wash. Of course, I can get the preamp turned up to full and the ND91's are still alive, so the preamp determines a lot in this equation, but I think the amp is really doing a phenomenal job putting out some real power in a tiny size.

    TomZ
    Are you measuring temp with the unit open? There may be a significant difference when it's buttoned up. Did you mount the amp at the top of the table radio - where heat collects?

    I would mount the amp so the heat sink fins and channels are vertical at the bottom of the cabinet. That creates a passive airflow over the heat sink as it heats the surrounding air.

    See my link (since TT won't allow pic uploads - always a "corrupted file error").

    Leave a comment:


  • stephenmarklay
    replied
    Originally posted by Millstonemike View Post

    As Wogg noted, forget the 10% THD watt ratings from the mfgs - marketing showmanship.

    The major class D chips from TI, STmicro, etc, do a good job. When they're not driven into clipping expect < .3% THD at power. That said, power out is dependent on two things: power supply voltage and the speaker impedance (notwithstanding chip limitations). My rule of thumb is "clean" W rms = (PSV-1)2 / (2*SPR). That is based on physics and semiconductor losses for typical modules based on these chips.

    Some examples:

    24 V PS and 8 ohm speakers yields 33 W rms (66 W rms for 4 ohm speakers).
    18 V PS and 8 ohm speakers yields 18 W rms (36 W rms for 4 ohm speakers).
    12 V PS and 8 ohm speakers yields 7.5 W rms (15 W rms for 4 ohm speakers).

    The "50 W x 2" ratings you may see reflect the chip's thermal limitation. But if you have a 24 V PS and 4 Ohm speakers you'll get 66 W rms (without a problem with music assuming the chip has a good heat sink and airflow). And that's worst case.

    I recommend a 24 V PS for any of these amps. You don't have to use all the power the 24 V PS can give, but it provides a lot of headroom when needed.
    Thank you very much. I am going to try the 2.1 amp. I appreciate the clarification. For sure with a 24V PS, there is more than enough clean power.

    Leave a comment:


  • tomzarbo
    replied
    I had been hesitant to go the full 24v because of the 'warnings' on the product page, even though you've even previously tried to asswage my fears Mike... but you're absolutely right... it seems that these amps can take it just fine.

    I've been doing some near full-output temperature testing on the TDA7492 amp: https://www.parts-express.com/tda749...2x50w--320-606 and with the video player preamp faceplate feeding the source material it's peaking at around 120 or so degrees on the heatsink. That's pretty chill for a legit' amplifier producing real output. I like that little amp a lot more now. It's not until the very last three or four clicks on the preamp that it starts sounding harsh, but by then the woofers are going pretty much berserk at that point so it's a wash. Of course, I can get the preamp turned up to full and the ND91's are still alive, so the preamp determines a lot in this equation, but I think the amp is really doing a phenomenal job putting out some real power in a tiny size.

    TomZ

    Leave a comment:


  • Millstonemike
    replied
    Originally posted by stephenmarklay View Post

    Thank you Wogg. really 20W is probably sufficient anyway. it is just the marketing that is a bit optimistic.

    Much appreciated.
    As Wogg noted, forget the 10% THD watt ratings from the mfgs - marketing showmanship.

    The major class D chips from TI, STmicro, etc, do a good job. When they're not driven into clipping expect < .3% THD at power. That said, power out is dependent on two things: power supply voltage and the speaker impedance (notwithstanding chip limitations). My rule of thumb is "clean" W rms = (PSV-1)2 / (2*SPR). That is based on physics and semiconductor losses for typical modules based on these chips.

    Some examples:

    24 V PS and 8 ohm speakers yields 33 W rms (66 W rms for 4 ohm speakers).
    18 V PS and 8 ohm speakers yields 18 W rms (36 W rms for 4 ohm speakers).
    12 V PS and 8 ohm speakers yields 7.5 W rms (15 W rms for 4 ohm speakers).

    The "50 W x 2" ratings you may see reflect the chip's thermal limitation. But if you have a 24 V PS and 4 Ohm speakers you'll get 66 W rms (without a problem with music assuming the chip has a good heat sink and airflow). And that's worst case.

    I recommend a 24 V PS for any of these amps. You don't have to use all the power the 24 V PS can give, but it provides a lot of headroom when needed.

    Leave a comment:


  • stephenmarklay
    replied
    Originally posted by wogg View Post

    Your best reference will be to lookup a data sheet on the chip used. Here's the TDA7492 used on that Sure board. With a 20V supply at 1kHz it's at 0.05% distortion right up to almost 20W where it increases dramatically (figure 20 page 18). Of course, that will depend primarily on the power supply voltage, and secondarily on the gain and board setup, but you'll get a much better idea of the real capability.
    Thank you Wogg. really 20W is probably sufficient anyway. it is just the marketing that is a bit optimistic.

    Much appreciated.

    Leave a comment:


  • wogg
    replied
    Originally posted by stephenmarklay View Post

    I pulled this one at random

    https://www.parts-express.com/sure-e...board--320-301

    • Power output (w/ 24 VDC power supply): 50W x 2 (6 ohms, THD 10%), 40W x 2 (8 ohms, THD 10%), 25W x 2 (6 ohms, THD 0.05%), 20W x 2 (8 ohms, THD 0.1%)

    I am not impressed with a rating of 50x2 when 20w at 8ohms is 0.1% distortion.

    I have a “normal” class D amp rated at 2x30 rated 0.005 at full power. That is what my expectations are.
    Your best reference will be to lookup a data sheet on the chip used. Here's the TDA7492 used on that Sure board. With a 20V supply at 1kHz it's at 0.05% distortion right up to almost 20W where it increases dramatically (figure 20 page 18). Of course, that will depend primarily on the power supply voltage, and secondarily on the gain and board setup, but you'll get a much better idea of the real capability.

    Leave a comment:


  • a4eaudio
    replied
    That's just how some of the manufacturers list the boards - Sure always states the THD 10% and then THD 1%. I just ignore the 10% number.
    You can find a Sure board that says it is a 50W board and PE may have a little amp using the same chip and call it a 15W amp because Sure is stating it with a THD 10% and PE is using THD .01%.

    Leave a comment:


  • stephenmarklay
    replied
    Originally posted by wogg View Post
    None of them are 10% at normal levels, they just use that to show a higher rated power than what you can actually get cleanly.
    I pulled this one at random

    https://www.parts-express.com/sure-e...board--320-301

    • Power output (w/ 24 VDC power supply): 50W x 2 (6 ohms, THD 10%), 40W x 2 (8 ohms, THD 10%), 25W x 2 (6 ohms, THD 0.05%), 20W x 2 (8 ohms, THD 0.1%)

    I am not impressed with a rating of 50x2 when 20w at 8ohms is 0.1% distortion.

    I have a “normal” class D amp rated at 2x30 rated 0.005 at full power. That is what my expectations are.

    Leave a comment:


  • wogg
    replied
    Originally posted by stephenmarklay View Post
    I am amazed at the distortion figures on some of these. There are a few that look reasonable but 10% THD at rated power is not ideal
    None of them are 10% at normal levels, they just use that to show a higher rated power than what you can actually get cleanly.

    Leave a comment:

Working...
X