the price has been disclosed in the thread 136EU for the group, i'm in talks with Roberto as to how we will have to order, i'm seeing if I can get mine shipped direct as otherwise i'll be stuck with VAT even though I dont owe it, but you guys in EU should probably get shipped together as much as possible to keep work for them to a minimum and maybe keep shipping costs down. i'm in for 2 x 45/55, or 55/65 depending on Roberto's answer to my question
i'll help coordinate as ive already been talking to Roberto about it, but best not waste money on double fees shifting monies around.
what more information is needed?
i'll help coordinate as ive already been talking to Roberto about it, but best not waste money on double fees shifting monies around.
what more information is needed?
Last edited:
yeah i'll stick with the 2 x 45/55 which I believe means all of those are gone. honestly though i'm torn between 45/55 and 55/65 the lower will allow me to push higher bias, the latter will allow more flexibility with future setups. I would like to earmark 2 of the 45/55 though, i'll think on it as doesnt look like the 55/65 are going to run out shortly. I may also pick up just one dps400 for sub duties, but i'll have to see
So i'm not the only one starting to get doubts about 45/55 vs 55/65...
My main issue is that i haven't decided on speakers yet so i don't know how much power i currently need. Second i'm new to the audio world so i don't really know what to expect and i don't want to be limited by the amp when selecting speakers so more power might be a good option even if it feels like a waste of electricity now.
I'm aiming for a 3U 19" rack chassis 400 mm depth with heatsinks on both sides (pesante dissipante from modu). According to the spare parts section this comes with four 200x40x120 mm heatsinks (two per side) each with a temperature coefficient of 0.5 °C/W so 0.25 °C/W per channel.
Lets assume 400 mA bias that was a good compromise if i remember correctly. That means 36 or 44 W per channel and 127 resp 189 W into 8 ohm according to opcs worksheet.
If we assume that the efficiency of the amp is 65 % that means that 35 % of the output power goes into the heatsink. Which means another 45 resp 66 W. This results in a total dissipated power of 80 W resp 110W (ignoring power supply).
Using the temperature coefficient of 0.25 °C/W this results in a temperature difference of 20 °C and 27.5 °C relative the room temperature. Lets include a derating factor of 0.9 for safety (i live at sea level so it should be 1). This results to a final temperature increase over room temp (25 °C) by 25 °C for the 45/55V version and 32 °C for the 55/65V version.
Both of these sound safe to me for being worst case calculations assuming that my calculations are correct. But i have a feeling that something is missing because this sounds to simple in some sense...
Anyone with a wild estimate on how much power (average) will be dissipated during normal run assuming a fairly inefficient speaker and reasonable listening volume (no hearing damage)?
Still the major problem i have is what type of power do speakers usually require to be driven with reasonable volume without clipping? If i can get away with 45/55 V supply thats great but if there is a significant risk that i will run into speakers that require more i don't want to come up short...
So i would really appreciate some guidelines on this from the more experienced people around.
My main issue is that i haven't decided on speakers yet so i don't know how much power i currently need. Second i'm new to the audio world so i don't really know what to expect and i don't want to be limited by the amp when selecting speakers so more power might be a good option even if it feels like a waste of electricity now.
I'm aiming for a 3U 19" rack chassis 400 mm depth with heatsinks on both sides (pesante dissipante from modu). According to the spare parts section this comes with four 200x40x120 mm heatsinks (two per side) each with a temperature coefficient of 0.5 °C/W so 0.25 °C/W per channel.
Lets assume 400 mA bias that was a good compromise if i remember correctly. That means 36 or 44 W per channel and 127 resp 189 W into 8 ohm according to opcs worksheet.
If we assume that the efficiency of the amp is 65 % that means that 35 % of the output power goes into the heatsink. Which means another 45 resp 66 W. This results in a total dissipated power of 80 W resp 110W (ignoring power supply).
Using the temperature coefficient of 0.25 °C/W this results in a temperature difference of 20 °C and 27.5 °C relative the room temperature. Lets include a derating factor of 0.9 for safety (i live at sea level so it should be 1). This results to a final temperature increase over room temp (25 °C) by 25 °C for the 45/55V version and 32 °C for the 55/65V version.
Both of these sound safe to me for being worst case calculations assuming that my calculations are correct. But i have a feeling that something is missing because this sounds to simple in some sense...
Anyone with a wild estimate on how much power (average) will be dissipated during normal run assuming a fairly inefficient speaker and reasonable listening volume (no hearing damage)?
Still the major problem i have is what type of power do speakers usually require to be driven with reasonable volume without clipping? If i can get away with 45/55 V supply thats great but if there is a significant risk that i will run into speakers that require more i don't want to come up short...
So i would really appreciate some guidelines on this from the more experienced people around.
youve already bought the modushop? the heatsink fins on the pessante are a bit pissant 
for their size and i'de want to make sure they really are rated for that passively.
pretty new area for me too really, need to work it out, probably i'll change to 55/65 but I want to make sure I dont end up in a position I dont have a choice. that being said though, unless I do change I dont think there are any 45/55 left for you to choose
for their size and i'de want to make sure they really are rated for that passively.pretty new area for me too really, need to work it out, probably i'll change to 55/65 but I want to make sure I dont end up in a position I dont have a choice. that being said though, unless I do change I dont think there are any 45/55 left for you to choose
Last edited:
Guys, get 55/65V version. If they'll run hot, you can still cool them actively
An externally hosted image should be here but it was not working when we last tested it.
i'm using conrad 2 x MF35-151 for 4 channel (0.21c/w each) with a fairly modest linear regulated supply for the 2 for the tweeters, so I dont think i'll have any problem. but the problem is it'll be my first time running a digital crossover and the needed power will be much reduced, so there will probably be quite a lot of heat to burn in the output devices, but I dont yet know exactly how much
I do have some fans if need be, but they arent exactly set up for fan forced
I do have some fans if need be, but they arent exactly set up for fan forced
I was first on the 45/55 version so i should be good there, the list look like this at the moment:
QRikard 2x DPS-600-45/55V (Reg LME)
BuildMeSomething 2x DPS-600-45/55V (Reg LME)
lolo 6x 2x DPS-600-45/55V (Reg LME)
qusp 2x 2x DPS-600-45/55V (Reg LME)
So thats a total of 12, unless someone have "stolen" some trough PM...
igor0203: thanks for digging up those posts then i don't have to search later. Then i will stay at maximum 400 mA probably 300, still i will keep calculating on 400 mA just to be on the safe side so i don't end up burning stuff...
qusp if you are using two of the MF35-151 for four channels that means basically 0.42 °C/W per channel and you plan to use 600 mA. The modushop sinks are of poorer quality for sure, but i plan to run only 400 mA bias (can go down to 300 in need) so that should even things out and then they have still got better dissipation at 0.25 °C/W per channel.
Total measurements for the modushop sinks are 400x120x40 mm and the conrad sinks are 350x150x48 mm. So the overall area is similar which should mean that the 0.25 °C/W per channel isn't to far off in reality (i hope) and with added safety margin it should work. I don't want to end up on the boarder line either way and a fan is definitely a last way out which i want to avoid.
If my calculations are correct (please verify my assumptions) i will end up at around 60 °C on the heatsinks in the worst case with the 55/65V supply at full power. 60 does sound a bit high but i think it's within reason otherwise please tell me.
I haven't bought the chassis so i'm open to suggestions. I made a search a while ago and this was the only chassis within reasonable price, the others were a lot more expensive for similar look. I bet they were of better quality but i thought that it wouldn't make such a huge difference if i have a larger safety margin and avoid abusing the chassis physically...
If anyone have suggestions on where to by a 3U (around 3-400 mm depth) chassis with heatsinks for reasonable price (and shipping to sweden) please let me know.
QRikard 2x DPS-600-45/55V (Reg LME)
BuildMeSomething 2x DPS-600-45/55V (Reg LME)
lolo 6x 2x DPS-600-45/55V (Reg LME)
qusp 2x 2x DPS-600-45/55V (Reg LME)
So thats a total of 12, unless someone have "stolen" some trough PM...
igor0203: thanks for digging up those posts then i don't have to search later. Then i will stay at maximum 400 mA probably 300, still i will keep calculating on 400 mA just to be on the safe side so i don't end up burning stuff...
qusp if you are using two of the MF35-151 for four channels that means basically 0.42 °C/W per channel and you plan to use 600 mA. The modushop sinks are of poorer quality for sure, but i plan to run only 400 mA bias (can go down to 300 in need) so that should even things out and then they have still got better dissipation at 0.25 °C/W per channel.
Total measurements for the modushop sinks are 400x120x40 mm and the conrad sinks are 350x150x48 mm. So the overall area is similar which should mean that the 0.25 °C/W per channel isn't to far off in reality (i hope) and with added safety margin it should work. I don't want to end up on the boarder line either way and a fan is definitely a last way out which i want to avoid.
If my calculations are correct (please verify my assumptions) i will end up at around 60 °C on the heatsinks in the worst case with the 55/65V supply at full power. 60 does sound a bit high but i think it's within reason otherwise please tell me.
I haven't bought the chassis so i'm open to suggestions. I made a search a while ago and this was the only chassis within reasonable price, the others were a lot more expensive for similar look. I bet they were of better quality but i thought that it wouldn't make such a huge difference if i have a larger safety margin and avoid abusing the chassis physically...
If anyone have suggestions on where to by a 3U (around 3-400 mm depth) chassis with heatsinks for reasonable price (and shipping to sweden) please let me know.
Did you consider bigger(higher) chassis?
I bought Modushop 5U chassis to be on safe side since this is first time I'm building such powerful amplifier.
I bought Modushop 5U chassis to be on safe side since this is first time I'm building such powerful amplifier.
I have considered higher chassis, but i would prefer to keep the 3U height if possible, because i think that the tiny boards will get lonely in a huge 5U chassis
Seriously it's just about space and looks but i don't want to risk burning stuff just so i can put them into a small chassis hence the discussion. This is my first proper amplifier so i plan to have it last for quite some time so i want to make it proper from the beginning, that should explain some of the obvious questions from my side...
In my case the backpanel is clean (except for I/Os) so i have enough space there to put another 200 or maybe 300 mm long heatsink of the same height (another 0.4 °C/W dissipation). This can either used for power supply or coupled to the other two heatsinks for improved performance. Still i see this as an backup solution to avoid the dreaded fan...
My calculations lead to a total heatsink temperature of 60 °C, what is the normal maximum operating heatsink temperature from a commercial amplifier? 60 sounds a little high in my book but manageable, still i don't have a reference point so i might be completely off.
Seriously it's just about space and looks but i don't want to risk burning stuff just so i can put them into a small chassis hence the discussion. This is my first proper amplifier so i plan to have it last for quite some time so i want to make it proper from the beginning, that should explain some of the obvious questions from my side...
In my case the backpanel is clean (except for I/Os) so i have enough space there to put another 200 or maybe 300 mm long heatsink of the same height (another 0.4 °C/W dissipation). This can either used for power supply or coupled to the other two heatsinks for improved performance. Still i see this as an backup solution to avoid the dreaded fan...
My calculations lead to a total heatsink temperature of 60 °C, what is the normal maximum operating heatsink temperature from a commercial amplifier? 60 sounds a little high in my book but manageable, still i don't have a reference point so i might be completely off.
I am planning on using the 5U from modu, the biggest one 500mm deep, no other choice for 10 channels really.. I haven't worked out yet if I will need active cooling, hopefully not.
I wouldn't mind taking the higher voltage units if people are missing out on the 45/55 version, extra power is always welcomed and I can always bias the woofers for 100mA only, but I think 45/55 is a bit more on the safe side. Roberto, how long to order a new batch of 45/55? I may be happy to step aside and wait if need be, or take the 55/65 but really need to work out the dissipation factors carefully then, I will need help guys. Remember I have full active speakers so the power is NOT an issue, I will rarely go over 10w, but the current demand from the woofers in the 80-100hz range is massive, hence the need for a fast, powerful supply.
I wouldn't mind taking the higher voltage units if people are missing out on the 45/55 version, extra power is always welcomed and I can always bias the woofers for 100mA only, but I think 45/55 is a bit more on the safe side. Roberto, how long to order a new batch of 45/55? I may be happy to step aside and wait if need be, or take the 55/65 but really need to work out the dissipation factors carefully then, I will need help guys. Remember I have full active speakers so the power is NOT an issue, I will rarely go over 10w, but the current demand from the woofers in the 80-100hz range is massive, hence the need for a fast, powerful supply.
Last edited:
The quiescent power dissipation is the dominant heat source in a domestic wide bandwidth amplifier that has a high bias current.
Heat dissipation proportions are quite different in low bias amplifiers, eg. opamps and chipamps and CFP BJT output stage discrete amps.
This Lat FET output stage amplifier falls into the high bias amplifier category.
When used for normal wide bandwidth domestic duty, where average power outputs are usually -40dB to -20dB below clipping, the added heat dissipation due to delivering output power is quite low in relation to the Pq mentioned at the opening sentence.
Let's look at an example.
With a regulated +-50Vdc feeding an FET output stage biased to 300mA, the Pq=30W.
The maximum output power into 8r0 is ~120W, (if the LME is fed with ~+_55Vdc).
-20dB for average output power results in ~1.2W delivered to the load and this at worst will dissipate an extra 2W to 5W in the output stage. This would increase the total output stage dissipation to something of the oder of 32W to 35W. Average listening levels of -30dB to -40dB would have lower dissipation needs.
Do remember, we are not talking about a bass only amplifier, nor are we talking about replaying disco levels where average levels could be -6dB to -10dB below clipping. They are a different animal and deserve a different heatsink design.
Back to domestic duty heatsink design.
If the heatsink is maintained at some arbitrary temperature (Ts) and the output devices have been selected to have a safe temperature de-rated SOAR that matches that arbitrary Ts, then the devices will perform reliably while the heatsink holds close to that arbitrary Ts.
What will that Ts be?
Ts = Ta + [Pd * Rth s-a * De-rating Factor (DF)].
Tc = Ts + [device Pd * Rth c-s].
When a heatsink is tested and specified, it is done with the whole backplate at a common temperature (isothermal) and with the delta T s-a at some specified value, usually around 70C to 80C degrees.
If the operating Delta T s-a is different from that specified, then a correction must be applied. The sink manufacturer usually tells you what that correction is and how to apply it.
For delta T s-a around 20 to 30Cdegrees expect the DF to be 1.2 to 1.5
Lets' assume DF=1.3 and Ta=28°C and Pd=35W and Rth s-a=0.4C/W
Ts = 28 + [35 + 0.4 * 1.3] = 46.2°C
Tc = 46.2 + [35/2 * 0.4] = 53.2°C
The output stage should be checked for Tc=53.2°C. Here I usually add in a 5C Safety Factor and actually do my SOAR modeling at Tc=58°C, for this particular example.
That Safety Factor includes for different Rth c-s, Rth s-a, DF and Ta. These all affect the modeling and thus the prediction of what is acceptable for a de-rated SOAR.
I hope this helps with understanding what should be allowed for in designing the heatsinks and designing the supply voltages. Then informed choices can be made to ensure reliable amplifiers.
I do suggest very strongly that everyone who intends adopting some other design than that prescribed, that they accurately model the SOAR. Use Bensen's spreadsheet or the FET version of my modified spreadsheet based on Bensen's original.
Heat dissipation proportions are quite different in low bias amplifiers, eg. opamps and chipamps and CFP BJT output stage discrete amps.
This Lat FET output stage amplifier falls into the high bias amplifier category.
When used for normal wide bandwidth domestic duty, where average power outputs are usually -40dB to -20dB below clipping, the added heat dissipation due to delivering output power is quite low in relation to the Pq mentioned at the opening sentence.
Let's look at an example.
With a regulated +-50Vdc feeding an FET output stage biased to 300mA, the Pq=30W.
The maximum output power into 8r0 is ~120W, (if the LME is fed with ~+_55Vdc).
-20dB for average output power results in ~1.2W delivered to the load and this at worst will dissipate an extra 2W to 5W in the output stage. This would increase the total output stage dissipation to something of the oder of 32W to 35W. Average listening levels of -30dB to -40dB would have lower dissipation needs.
Do remember, we are not talking about a bass only amplifier, nor are we talking about replaying disco levels where average levels could be -6dB to -10dB below clipping. They are a different animal and deserve a different heatsink design.
Back to domestic duty heatsink design.
If the heatsink is maintained at some arbitrary temperature (Ts) and the output devices have been selected to have a safe temperature de-rated SOAR that matches that arbitrary Ts, then the devices will perform reliably while the heatsink holds close to that arbitrary Ts.
What will that Ts be?
Ts = Ta + [Pd * Rth s-a * De-rating Factor (DF)].
Tc = Ts + [device Pd * Rth c-s].
When a heatsink is tested and specified, it is done with the whole backplate at a common temperature (isothermal) and with the delta T s-a at some specified value, usually around 70C to 80C degrees.
If the operating Delta T s-a is different from that specified, then a correction must be applied. The sink manufacturer usually tells you what that correction is and how to apply it.
For delta T s-a around 20 to 30Cdegrees expect the DF to be 1.2 to 1.5
Lets' assume DF=1.3 and Ta=28°C and Pd=35W and Rth s-a=0.4C/W
Ts = 28 + [35 + 0.4 * 1.3] = 46.2°C
Tc = 46.2 + [35/2 * 0.4] = 53.2°C
The output stage should be checked for Tc=53.2°C. Here I usually add in a 5C Safety Factor and actually do my SOAR modeling at Tc=58°C, for this particular example.
That Safety Factor includes for different Rth c-s, Rth s-a, DF and Ta. These all affect the modeling and thus the prediction of what is acceptable for a de-rated SOAR.
I hope this helps with understanding what should be allowed for in designing the heatsinks and designing the supply voltages. Then informed choices can be made to ensure reliable amplifiers.
I do suggest very strongly that everyone who intends adopting some other design than that prescribed, that they accurately model the SOAR. Use Bensen's spreadsheet or the FET version of my modified spreadsheet based on Bensen's original.
Last edited: