8PAKAMP & 12PAKAMP KIT BUILD THREAD
Moderator: Joe Malone
-
- Posts: 11
- Joined: Tue Sep 28, 2010 4:11 pm
4 Zone Amp Build Commenced - Internal Layout
Here is the proposed layout:
Chips on the PCB underside, mounted to a piece of 75x35x9 unequal angle aluminium with two CPU coolers attached. Angle will attach firmly to the chassis, which may help with dissipation a little... though the thermal resistance across the powder coating will be high I'd say so the assistance is likely to be minimal. On that point, the bridge rectifier is similarly mounted to the chassis; will this provide enough cooling?
Input can be seen in the back left corner of the chassis, with the shielded input signal routed under the PCB so that it is away from the fans and the AC. There is plenty of room for metal shields if hum/buzz proves to be a problem.
Coolers exhaust vertically through pre-existing slots in the chassis (hence their centralised location on the angle) and the fans will likely be controlled by the following heat sensing control circuit to extend their life.
[Source: Elliott Sound Productions - Project 42]
This fan controller is designed to run off the positive rail of a power amplifier, up to +30VDC with component value changes. I will have +/-25VDC here.... can I just tap into this from the +V input on the PCB or will the unequal current drawn from each rail cause problems? I could add a 12V supply but I'd rather not if it's avoidable.
Any comments are be welcomed. Does it look like I have sufficient cooling? Anything glaringly wrong with the layout?
Regards,
Tani.
Chips on the PCB underside, mounted to a piece of 75x35x9 unequal angle aluminium with two CPU coolers attached. Angle will attach firmly to the chassis, which may help with dissipation a little... though the thermal resistance across the powder coating will be high I'd say so the assistance is likely to be minimal. On that point, the bridge rectifier is similarly mounted to the chassis; will this provide enough cooling?
Input can be seen in the back left corner of the chassis, with the shielded input signal routed under the PCB so that it is away from the fans and the AC. There is plenty of room for metal shields if hum/buzz proves to be a problem.
Coolers exhaust vertically through pre-existing slots in the chassis (hence their centralised location on the angle) and the fans will likely be controlled by the following heat sensing control circuit to extend their life.
[Source: Elliott Sound Productions - Project 42]
This fan controller is designed to run off the positive rail of a power amplifier, up to +30VDC with component value changes. I will have +/-25VDC here.... can I just tap into this from the +V input on the PCB or will the unequal current drawn from each rail cause problems? I could add a 12V supply but I'd rather not if it's avoidable.
Any comments are be welcomed. Does it look like I have sufficient cooling? Anything glaringly wrong with the layout?
Regards,
Tani.
You do not have the required permissions to view the files attached to this post.
-
- Posts: 11
- Joined: Tue Sep 28, 2010 4:11 pm
4 Zone Amp Build Commenced - PCB Stuffed
Progress:
LM3886 on board underside. Managed to get them nice and straight. I used a small spring clamp to hold the chips in place, which ensured that the upper surface of the package was flush with the PCB during soldering. Due to the gap formed between the chip and the PCB around the mounting hole, failing to do this would reduce the effectiveness of the mounting screws and could stress the PCB and its connection to the chip. No powered test yet. I'll go to a friend's workshop this afternoon to mount the sinks.
In the meantime... on to the input wiring.
Cheers,
Tani.
LM3886 on board underside. Managed to get them nice and straight. I used a small spring clamp to hold the chips in place, which ensured that the upper surface of the package was flush with the PCB during soldering. Due to the gap formed between the chip and the PCB around the mounting hole, failing to do this would reduce the effectiveness of the mounting screws and could stress the PCB and its connection to the chip. No powered test yet. I'll go to a friend's workshop this afternoon to mount the sinks.
In the meantime... on to the input wiring.
Cheers,
Tani.
You do not have the required permissions to view the files attached to this post.
- Joe Malone
- Site Admin
- Posts: 2080
- Joined: Wed Jan 24, 2007 11:35 pm
- Location: Brisbane, Australia
- Contact:
Re: 8PAKAMP KIT BUILD THREAD
Looking good but mounting the LM3886 on a angle not such a good idea but it should work fine.
The LM3886 when mounted underneath should be placed flat on the bottom of PCB and 2 x flat 3mm washer or 3 x 3mm lock washer inserted in the gap between the PCB hole and mounting tab hole to pack the gap.
The LM3886 when mounted underneath should be placed flat on the bottom of PCB and 2 x flat 3mm washer or 3 x 3mm lock washer inserted in the gap between the PCB hole and mounting tab hole to pack the gap.
Joe
JLM Audio
Capturing Audio without Injury
JLM Audio
Capturing Audio without Injury
-
- Posts: 11
- Joined: Tue Sep 28, 2010 4:11 pm
Re: 8PAKAMP KIT BUILD THREAD
Hi Joe,
Progress:
I've decided to give a "mono blend" type implementation a go. 50k pot, rheostat wired, in place of the mono switch. Means that I'll never have full stereo separation but it's close enough. Plus, the optimum sound might not be either full mono or full stereo, so this lets me tune in between. Mono summing has been moved to the volume pot wipers as you suggested Joe... much better!
Here's the new input network schematic. See any issues?
Cheers,
Tani.
I mustn't have explained it clearly, sorry. The chip package is flush with the PCB... no angle. The spring clip was to make sure that the face was flush and touching the PCB, rather than having it end up cocked at an angle. The close-up photo of the chip in my last post shows that it's flat against the PCB.Joe Malone wrote:Looking good but mounting the LM3886 on a angle not such a good idea but it should work fine.
I'll admit I missed this note, but that is what I was planning to do anyway. In hindsight, I should have placed whatever packers I was planning to use in the gap and fastened the chip to the PCB before soldering. This would have set the gap distance to avoid stressing the solder joints when it's all tightened down. As it is I'm sure it'll be fine... I'll keep an eye on things as I do it up.Joe Malone wrote:... 2 x flat 3mm washer or 3 x 3mm lock washer inserted in the gap between the PCB hole and mounting tab hole to pack the gap.
Progress:
I've decided to give a "mono blend" type implementation a go. 50k pot, rheostat wired, in place of the mono switch. Means that I'll never have full stereo separation but it's close enough. Plus, the optimum sound might not be either full mono or full stereo, so this lets me tune in between. Mono summing has been moved to the volume pot wipers as you suggested Joe... much better!
Here's the new input network schematic. See any issues?
Cheers,
Tani.
You do not have the required permissions to view the files attached to this post.
- Joe Malone
- Site Admin
- Posts: 2080
- Joined: Wed Jan 24, 2007 11:35 pm
- Location: Brisbane, Australia
- Contact:
Re: 8PAKAMP KIT BUILD THREAD
OK coolThe spring clip was to make sure that the face was flush and touching the PCB, rather than having it end up cocked at an angle. The close-up photo of the chip in my last post shows that it's flat against the PCB.
Not sure how smooth a transition from stereo to mono it will be as I think it would be a reverse log function. So may be lots of nothing and then sudden mono at the end. Also width will change when turning the level up or down. But if it is mainly a set and forget system it will be fine.I've decided to give a "mono blend" type implementation a go. 50k pot, rheostat wired, in place of the mono switch. Means that I'll never have full stereo separation but it's close enough. Plus, the optimum sound might not be either full mono or full stereo, so this lets me tune in between. Mono summing has been moved to the volume pot wipers as you suggested Joe... much better!
Joe
JLM Audio
Capturing Audio without Injury
JLM Audio
Capturing Audio without Injury
-
- Posts: 11
- Joined: Tue Sep 28, 2010 4:11 pm
Re: 8PAKAMP KIT BUILD THREAD
Yes that sounds right. I noticed in the simulation that the last ~10% of wiper movement has ~90% of the effect, suggesting a log relationship. Having the level pots change the width though is not desirable... I should have seen that. Might look at putting the mono blend upstream of the pots; as the switches were before. More crosstalk, less variation with level changes. I'll experiment with the real thing and come up with something that works.Joe Malone wrote:Not sure how smooth a transition from stereo to mono it will be as I think it would be a reverse log function. So may be lots of nothing and then sudden mono at the end. Also width will change when turning the level up or down. But if it is mainly a set and forget system it will be fine.
Progress:
The afternoon in a friends workshop proved fruitful. Got the CPU coolers and angle aluminium drilled and "lapped" (I use that term very loosely): Mounted them using some thermal compound: Drilled the holes to mount the chips/PCB to the angle. All 8 go through the chassis base to hold the whole shebang together. The larger holes were pre-existing, but are luckily positioned such that only a small area of the adjacent chips are without thermal contact. More to come...
Tani.
You do not have the required permissions to view the files attached to this post.
-
- Posts: 11
- Joined: Tue Sep 28, 2010 4:11 pm
4 Zone Amp Build - Heat Stress Testing
G'day,
Just put the amp through its paces to ensure that it has adequate heat sinking.... this thing's got some jump
For testing I configured the amp with all 8 channels bridged into 4 pairs, with each pair driving an 8ohm speaker. The heat sinking is as shown in previous posts, fans clipped onto the coolers (running or not). Measurements were taken from the middle of the aluminium angle... this reading was confirmed on numerous occasions to be less than 1deg lower than the case temp (as best as I could measure it) so the thermal coupling must be good. I don't know what power I was supplying to the speakers as the test was completed with music (I didn't fancy listening to a 1kHz pure tone at 100W for 2 hours ). If I build dummy loads I might test it more scientifically down the track.
Temperature measurements (degrees C):
T_amb = 16.5
T_idle = 40.2 (took nearly an hour to reach this so could still be climbing)
T_pp ~ 52 (no fans, music at somewhat louder than 'normal' listening - hence pp meaning Part Power. It took a further ~30mins of playing to reach this before I turned it up)
T_fp ~ 63 (no fans, volume on macbook pro headphone output wide open and playing LOUD - fp Full Power, took ~20mins to climb from 52 to 63 before I flicked on the fans, running slower at 10v supply)
With the amp still playing at "Full Power", the fans dropped the temp to ~50 in ~2mins, ~43 in ~4mins, and it was back to ~33 after ~8mins when I stopped the test.
From this I'm satisfied that the coolers can remove more heat that the chips can supply at pretty much any power output they'll ever see. Also, the aluminium angle spreads the heat adequately and its mass means that the changes in temperature are slow enough for the fan controller to operate smoothly... i.e. with enough hysteresis to avoid rapid on-off cycling.
I'm happy
On to the next step...
Tani.
Just put the amp through its paces to ensure that it has adequate heat sinking.... this thing's got some jump
For testing I configured the amp with all 8 channels bridged into 4 pairs, with each pair driving an 8ohm speaker. The heat sinking is as shown in previous posts, fans clipped onto the coolers (running or not). Measurements were taken from the middle of the aluminium angle... this reading was confirmed on numerous occasions to be less than 1deg lower than the case temp (as best as I could measure it) so the thermal coupling must be good. I don't know what power I was supplying to the speakers as the test was completed with music (I didn't fancy listening to a 1kHz pure tone at 100W for 2 hours ). If I build dummy loads I might test it more scientifically down the track.
Temperature measurements (degrees C):
T_amb = 16.5
T_idle = 40.2 (took nearly an hour to reach this so could still be climbing)
T_pp ~ 52 (no fans, music at somewhat louder than 'normal' listening - hence pp meaning Part Power. It took a further ~30mins of playing to reach this before I turned it up)
T_fp ~ 63 (no fans, volume on macbook pro headphone output wide open and playing LOUD - fp Full Power, took ~20mins to climb from 52 to 63 before I flicked on the fans, running slower at 10v supply)
With the amp still playing at "Full Power", the fans dropped the temp to ~50 in ~2mins, ~43 in ~4mins, and it was back to ~33 after ~8mins when I stopped the test.
From this I'm satisfied that the coolers can remove more heat that the chips can supply at pretty much any power output they'll ever see. Also, the aluminium angle spreads the heat adequately and its mass means that the changes in temperature are slow enough for the fan controller to operate smoothly... i.e. with enough hysteresis to avoid rapid on-off cycling.
I'm happy
On to the next step...
Tani.
Last edited by taniknight on Thu Aug 16, 2012 7:56 pm, edited 1 time in total.
- Joe Malone
- Site Admin
- Posts: 2080
- Joined: Wed Jan 24, 2007 11:35 pm
- Location: Brisbane, Australia
- Contact:
Re: 8PAKAMP KIT BUILD THREAD
Sounds like a great setup so far. Serious heatsinking is always a good idea.
Joe
JLM Audio
Capturing Audio without Injury
JLM Audio
Capturing Audio without Injury
-
- Posts: 9
- Joined: Sun Feb 20, 2011 8:43 pm
Re: 8PAKAMP KIT BUILD THREAD
Wow, I didn't notice this design needs such a big heatsink. And I almost bought a PCB during my recent order (3 BAs & 3 FET DIs which I got yesterday) thinking I could to this in 1U.
- Joe Malone
- Site Admin
- Posts: 2080
- Joined: Wed Jan 24, 2007 11:35 pm
- Location: Brisbane, Australia
- Contact:
Re: 8PAKAMP KIT BUILD THREAD
We have fitted this 8 channel amp into 1 unit rack case with heatsink. It just depends on the amount of power you want per channel. The hardest part about 1 unit high rack case is getting a pancake toroid transformer that fits.sonicwarrior wrote:Wow, I didn't notice this design needs such a big heatsink. And I almost bought a PCB during my recent order (3 BAs & 3 FET DIs which I got yesterday) thinking I could to this in 1U.
Joe
JLM Audio
Capturing Audio without Injury
JLM Audio
Capturing Audio without Injury
-
- Posts: 25
- Joined: Fri Jan 20, 2012 1:25 am
Re: 8PAKAMP KIT BUILD THREAD
Hi Joe,
What is the gain of this 8 channel channel setup...and the singular module too while we are at it? I'm guessing in the 28-30dB range.
Cheers
Paul
What is the gain of this 8 channel channel setup...and the singular module too while we are at it? I'm guessing in the 28-30dB range.
Cheers
Paul
- Joe Malone
- Site Admin
- Posts: 2080
- Joined: Wed Jan 24, 2007 11:35 pm
- Location: Brisbane, Australia
- Contact:
Re: 8PAKAMP KIT BUILD THREAD
Both AMP kit and 8PAKAMP kit are the same circuit.wizardofoz wrote:Hi Joe,
What is the gain of this 8 channel channel setup...and the singular module too while we are at it? I'm guessing in the 28-30dB range.
Cheers
Paul
With no trim pot or volume control fitted gain is 21dB. With 10k trim pot or volume fitted gain is 15dB to -90dB. If you want more gain just change 1k resistor on amp pcb to lower value. 470R will add 6dB. 220R will add another 6dB etc
Joe
JLM Audio
Capturing Audio without Injury
JLM Audio
Capturing Audio without Injury
-
- Posts: 25
- Joined: Fri Jan 20, 2012 1:25 am
Re: 8PAKAMP KIT BUILD THREAD
Thanks Joe, thats perfect...27-28dB gain will be just nice.
-
- Posts: 25
- Joined: Fri Jan 20, 2012 1:25 am
Re: 8PAKAMP KIT BUILD THREAD
Bridging 12dB Wattage, 6dB Voltage level. Parallel no change in dB level.Joe, for the gain (I am no expert in such matters) if you are bridging or paralleling what gain changes would the be if any?
Unbalanced -6dB compared to balancedLikewise if one uses balanced or unbalanced inputs does the gain change at all?
okI need to have all the channels at a the same gain ideally. as some of my channels are bridged now and some are single I am nit sure if I am meeting this goal or not.
-
- Posts: 25
- Joined: Fri Jan 20, 2012 1:25 am
Re: 8PAKAMP KIT BUILD THREAD
I ended up using 476ohms (2x238 in series) but im using un balanced so might have to drop that to maybe 238 to get back up to around 27-28db gain. Will have to measure it once I am done.
Much the same as my 5 channel version but with a 25-0-25 500va as no bridging going on here. Using 2 x 8 way Speakons for the speakers to connect to my Linkwitz Orion's
Much the same as my 5 channel version but with a 25-0-25 500va as no bridging going on here. Using 2 x 8 way Speakons for the speakers to connect to my Linkwitz Orion's