Hi all. I'm starting to put an Aleph 5 together and have ran into some questions I figured I'd run by the folks over here.
If you care to know the reasons why, you can read on after the questions, but suffice it to say that my build is going to be 3 boxes: 1 P/S sitting on the ground and 2 mono blocks getting their DC feed from the P/S. Obviously the amps will not have IEC ground as that'll only be on the P/S. The questions:
1. Should I still do an isolated star ground in the amps? Or is it okay to liberally use their chassis for ground?
2. Should I duplicate everything from the P/S after the coil in each amp? Or should I go ahead and finish the P/S outside the amps and just supply the final voltage to each amp?
3. Am I okay running my DC power from the P/S over to one amp and then the other?
Thanks!
The reasons:
So, after reading posts for a few weeks on DIY Aleph 2 I finally decided that I couldn't stomach a 600W heater for 8-9 hours a day, 5 days a week, so I decided to take the plunge and settle for a 300W heater instead. The decision wasn't an easy one, since my NHT 3.3 speakers are only 87db SPL and would have appreciated the Aleph 2, but I decided that I'd instead do passive bi-amp, keeping my ATI AT-1502 for the subs and the double A5 for everything else. That would bring my overall power requirements much lower, since the 3.3 splits the woofer at 100Hz and 60W for everything above that should be fine..
One of the things I wanted to do, since aesthetics don't mean much to me (although I love the stock look of the A2 and A5), was to separate the P/S from the amp, per one of Nelson's preferences. So after reading much I settled on a 30-0-30 1000va transformer (hoping it'll be +/- 37v) and running +/-V over to the amp. Then I thought purchasing a massive chassis with nothing in it seemed like a silly idea, so I decided to do 2 very small chassis. Each comes with 2 heat sinks of 0.36 C/W each, so my total after 4 heat sinks would be 0.09 C/W which I think should be able to handle 300W without any issues. The chassis run 10" wide and would sit side-by-side.
I'm still trying to figure out how to do volume matching between not only these two amps, but also between this set and the ATI that is going to drive the subs, because I'm using a home-made passive preamp with just one volume control and no balance, so I may have to do an op-amp with a volume control when I build the two A5s to feed the existing amp.
If you care to know the reasons why, you can read on after the questions, but suffice it to say that my build is going to be 3 boxes: 1 P/S sitting on the ground and 2 mono blocks getting their DC feed from the P/S. Obviously the amps will not have IEC ground as that'll only be on the P/S. The questions:
1. Should I still do an isolated star ground in the amps? Or is it okay to liberally use their chassis for ground?
2. Should I duplicate everything from the P/S after the coil in each amp? Or should I go ahead and finish the P/S outside the amps and just supply the final voltage to each amp?
3. Am I okay running my DC power from the P/S over to one amp and then the other?
Thanks!
The reasons:
So, after reading posts for a few weeks on DIY Aleph 2 I finally decided that I couldn't stomach a 600W heater for 8-9 hours a day, 5 days a week, so I decided to take the plunge and settle for a 300W heater instead. The decision wasn't an easy one, since my NHT 3.3 speakers are only 87db SPL and would have appreciated the Aleph 2, but I decided that I'd instead do passive bi-amp, keeping my ATI AT-1502 for the subs and the double A5 for everything else. That would bring my overall power requirements much lower, since the 3.3 splits the woofer at 100Hz and 60W for everything above that should be fine..
One of the things I wanted to do, since aesthetics don't mean much to me (although I love the stock look of the A2 and A5), was to separate the P/S from the amp, per one of Nelson's preferences. So after reading much I settled on a 30-0-30 1000va transformer (hoping it'll be +/- 37v) and running +/-V over to the amp. Then I thought purchasing a massive chassis with nothing in it seemed like a silly idea, so I decided to do 2 very small chassis. Each comes with 2 heat sinks of 0.36 C/W each, so my total after 4 heat sinks would be 0.09 C/W which I think should be able to handle 300W without any issues. The chassis run 10" wide and would sit side-by-side.
I'm still trying to figure out how to do volume matching between not only these two amps, but also between this set and the ATI that is going to drive the subs, because I'm using a home-made passive preamp with just one volume control and no balance, so I may have to do an op-amp with a volume control when I build the two A5s to feed the existing amp.
It seems none of the brightest minds have answered, so you'll have to read mine. 😀
1. No comment. Wiser minds should be consulted to discuss your goal of not having your amp connected to mains earth using mains A/C power.
2. Not sure what you're asking. Is the coil your AC step down power transformer? Is coil an inductor in a CLC filter? Something else? Either way, for a CRC, some people like to put the final cap or caps in the amp chassis, noting that the wiring between the PSU chassis and the amp can be an R in a CRC filter.
3. If you mean running them in parallel, yes. Running them as I read it - "to one amp AND THEN the other" would not be great. You'd have the voltage drop from one amp prior to the other.
I'd also check those heatsinks for de-rating. The Modushop heatsinks that are closest to that thermal dissipation rating are posted in the shop. 200 x 40 x 165 => 0.38 C/W Two pieces would be used for a 4U/400 which is generally accepted to be able to run at reasonable temps at ~80W-100W dissipation per side. You're running at least 50% more than that per side.
Modushop's study shows that a 4U / 400 chassis with 150W per side to run with a 34C temp rise at the tips of the fins and 46C rise at the base of the heatsink. That compares to 26 and 32 respectively at 100W.
Not saying yours will be insufficient or that you've missed anything, but you may want to be certain that the 0.36 C/W is not at thermal saturation at some insanely high temp to be certain and calculate your expected operating temp above ambient.
I'm sure I'm wrong about something and/or missed something, but hopefully others will chime in.
Good luck, be safe, have fun. 🙂
1. No comment. Wiser minds should be consulted to discuss your goal of not having your amp connected to mains earth using mains A/C power.
2. Not sure what you're asking. Is the coil your AC step down power transformer? Is coil an inductor in a CLC filter? Something else? Either way, for a CRC, some people like to put the final cap or caps in the amp chassis, noting that the wiring between the PSU chassis and the amp can be an R in a CRC filter.
3. If you mean running them in parallel, yes. Running them as I read it - "to one amp AND THEN the other" would not be great. You'd have the voltage drop from one amp prior to the other.
I'd also check those heatsinks for de-rating. The Modushop heatsinks that are closest to that thermal dissipation rating are posted in the shop. 200 x 40 x 165 => 0.38 C/W Two pieces would be used for a 4U/400 which is generally accepted to be able to run at reasonable temps at ~80W-100W dissipation per side. You're running at least 50% more than that per side.
Modushop's study shows that a 4U / 400 chassis with 150W per side to run with a 34C temp rise at the tips of the fins and 46C rise at the base of the heatsink. That compares to 26 and 32 respectively at 100W.
Not saying yours will be insufficient or that you've missed anything, but you may want to be certain that the 0.36 C/W is not at thermal saturation at some insanely high temp to be certain and calculate your expected operating temp above ambient.
I'm sure I'm wrong about something and/or missed something, but hopefully others will chime in.
Good luck, be safe, have fun. 🙂
Thanks ItsAllInMyHead. Yes. When I said coils, I meant the "L" inductor in CLC. The transformer has two secondaries that I'll need to join to create a center tap.
I can also run the amps parallel from the power supply as opposed to running the P/S to one and the other. I just figured a foot of 14AWG would practically have 0 resistance, but running them parallel from the P/S seems like a better idea anyway.
I've been running the heat sink calculations using the calculator at Thermal Resistance Calculator for Plate Fin Heat Sink since that was a link that I found in one of the posts, but I've checked another calculator on Heat Sink Size Calculator and it doesn't agree with the first site.
When I put in 200 x 40 x 165 (base of 7mm, fins of 2mm with 20 fins) in the first site, it comes up with 0.38 C/W, which matches the number you gave me, so I think mine at 4 x 0.36 should be okay, but I'm surprised 2 x 0.38 would be okay for the A5 with 270W dissipation. I'm making the assumption that mine will be running at 300W since my voltage will be a bit higher.
I'm not new to electronics, but I'm all new to head sinks and power dissipation. This has been a learning experience to say the least.
I can also run the amps parallel from the power supply as opposed to running the P/S to one and the other. I just figured a foot of 14AWG would practically have 0 resistance, but running them parallel from the P/S seems like a better idea anyway.
I've been running the heat sink calculations using the calculator at Thermal Resistance Calculator for Plate Fin Heat Sink since that was a link that I found in one of the posts, but I've checked another calculator on Heat Sink Size Calculator and it doesn't agree with the first site.
When I put in 200 x 40 x 165 (base of 7mm, fins of 2mm with 20 fins) in the first site, it comes up with 0.38 C/W, which matches the number you gave me, so I think mine at 4 x 0.36 should be okay, but I'm surprised 2 x 0.38 would be okay for the A5 with 270W dissipation. I'm making the assumption that mine will be running at 300W since my voltage will be a bit higher.
I'm not new to electronics, but I'm all new to head sinks and power dissipation. This has been a learning experience to say the least.
Going back to the grounding, I was really referring to the this link: IEC and Audio Ground. The question is, if my P/S chassis is physically disconnected from the amps, do I need to extend the IEC ground to the amps via the diode/resistor to the star ground in the amps? Or can I just keep them separate and thus using the amp body instead of the star ground?
Additionally I think I just made the decision to keep everything for the P/S, including the caps after the inductors, in the same box and just extend the +/- voltage out of the P/S box after everything is finished with the power signal.
Additionally I think I just made the decision to keep everything for the P/S, including the caps after the inductors, in the same box and just extend the +/- voltage out of the P/S box after everything is finished with the power signal.
I'll look forward to learning alongside.
I think I have a better understanding of what you're asking. I don't have a background in electronics, so if it's good with you, I'll use two separate terms:
Mains Earth / aka safety earth - Direct connection to mains earth.
Audio GND - whatever is used for signal return. Could be lifted from mains earth. Could be lifted from power rails. Could be mains earth / the chassis / the same GND as the rails etc. etc.
When it comes to safety, it's best that someone qualified chime in vs. my speculation re: connections to mains earth and your amplifier chassis. Everything below is in reference to your lifted (through the resistor / diodes) Audio GND.
Regarding what I'll call audio GND. I am not sure if the person in the thread you linked was building monoblocks or a single-chassis solution. It seems they were chasing a hum and had their wiring a bit off. The responding person was correcting them to be sure:
1. That their mains earth was connected directly to the chassis and not through the ground lift components.
2. That their "lifted ground" was incorporated into a central return (star ground) and used as the "audio GND" vs. the chassis.
The way you've phrased it, "do I need to extend the IEC ground to the amps via the diode/resistor to the star ground in the amps?"...
Perhaps draw up a schematic or wiring diagram or two re: what you're considering.
There seem to be two (or more) things you're considering. I have to think of it in terms of how it would be wired.
a. V+ V- and GND from the PSU to each amp. Audio GND is already "lifted" from mains earth (resistor / diodes) within the PSU chassis. Star ground within each amp back to a central point.
b. V+ V- and GND from the PSU to each amp. GND at this point would = Mains Earth = The PSU chassis. Resistor / diode network within each amp chassis for the ground lift. Star ground within each amp back to a central point.
Option a. seems to be the most common implementation, I believe.
Hopefully others will chime in. My thoughts are speculation at best.
I think I have a better understanding of what you're asking. I don't have a background in electronics, so if it's good with you, I'll use two separate terms:
Mains Earth / aka safety earth - Direct connection to mains earth.
Audio GND - whatever is used for signal return. Could be lifted from mains earth. Could be lifted from power rails. Could be mains earth / the chassis / the same GND as the rails etc. etc.
When it comes to safety, it's best that someone qualified chime in vs. my speculation re: connections to mains earth and your amplifier chassis. Everything below is in reference to your lifted (through the resistor / diodes) Audio GND.
Regarding what I'll call audio GND. I am not sure if the person in the thread you linked was building monoblocks or a single-chassis solution. It seems they were chasing a hum and had their wiring a bit off. The responding person was correcting them to be sure:
1. That their mains earth was connected directly to the chassis and not through the ground lift components.
2. That their "lifted ground" was incorporated into a central return (star ground) and used as the "audio GND" vs. the chassis.
The way you've phrased it, "do I need to extend the IEC ground to the amps via the diode/resistor to the star ground in the amps?"...
Perhaps draw up a schematic or wiring diagram or two re: what you're considering.
There seem to be two (or more) things you're considering. I have to think of it in terms of how it would be wired.
a. V+ V- and GND from the PSU to each amp. Audio GND is already "lifted" from mains earth (resistor / diodes) within the PSU chassis. Star ground within each amp back to a central point.
b. V+ V- and GND from the PSU to each amp. GND at this point would = Mains Earth = The PSU chassis. Resistor / diode network within each amp chassis for the ground lift. Star ground within each amp back to a central point.
Option a. seems to be the most common implementation, I believe.
Hopefully others will chime in. My thoughts are speculation at best.
From a safety standpoint, the purpose of the mains ground is in case something (the hot wire) becomes loose and touches the metal case. If the mains ground is connected to the metal case, then it'll cause a short and the fuse will blow. So, the question becomes whether it's of any use to extend this ground to the amplifiers if the P/S is housed in a separate metal box. I don't think so, as I can't see any reason behind it.
So then, the P/S supplies +/-V in addition to a ground (isolated from the mains ground) to the amplifier. The question of whether the ground in this circuit should be connected via the resistor/diode to the mains ground is what it's all about. Based on what I've seen so far in the threads here, I think the answer to this is a "Yes", with a diode/resistor in parallel. Then each amplifier's case can become this lifted ground and entry/exit grounds (input, speaker, etc) can just connect to the case wherever convenient. This would be your point 2a.
Your 2b would be if the mains ground got carried over to the amplifiers, in which case a lifted ground within each amp is the way to go with star grounding.
I tend to agree with you that option a seems to be the most logical.
I did draw them up, and added an "Extra" scenario, but I still think 2a is what I should do.
So then, the P/S supplies +/-V in addition to a ground (isolated from the mains ground) to the amplifier. The question of whether the ground in this circuit should be connected via the resistor/diode to the mains ground is what it's all about. Based on what I've seen so far in the threads here, I think the answer to this is a "Yes", with a diode/resistor in parallel. Then each amplifier's case can become this lifted ground and entry/exit grounds (input, speaker, etc) can just connect to the case wherever convenient. This would be your point 2a.
Your 2b would be if the mains ground got carried over to the amplifiers, in which case a lifted ground within each amp is the way to go with star grounding.
I tend to agree with you that option a seems to be the most logical.
I did draw them up, and added an "Extra" scenario, but I still think 2a is what I should do.
Finally found an excellent thread that touches on this subject. There's even posts about two amps sharing the same P/S in an external box. 69 pages and lots to read!