I got an LRS-600N2-48. The fan doesn't run at all, it is supposed to start at 50°C. Until then it is absolut noiseless. Only the green LED reminds you it is powered on.
I get the 600 Watt for 70€, the LRS-350N2-xx is 40€. The point is the "N2" in the code, which is the 200% overload version. So at 80€ for a 700 Watt constant capable supply, with a few seconds of 1400W, it looks quite attractive.
Mean Well seems to meet all European and US/Canada safety and emission regulations? I prefer a SMPS in a metal box from open versions.
I get the 600 Watt for 70€, the LRS-350N2-xx is 40€. The point is the "N2" in the code, which is the 200% overload version. So at 80€ for a 700 Watt constant capable supply, with a few seconds of 1400W, it looks quite attractive.
Mean Well seems to meet all European and US/Canada safety and emission regulations? I prefer a SMPS in a metal box from open versions.
Bryson uses a 120 mm Noctua fan in their top-of-the-line headphone amp. That fan emits single-digit SPL at full speed and less at lower speed. I own several Apple products with inaudible fans. So it is possible to make fan-cooled gear that isn't noisy. But at least from my experience with the Mean Well SE-600, the Mean Well supplies are not optimized for quiet. Those two supplies are louder than most of the test equipment I own.
Some have installed Noctua fans in power supplies to quiet them. So that is an option. I'd still want some sort of linear fan control, though. I built a fan controller for the power supply fan in my PC back in the early 90s. Used an LED as the temperature sensor. Red LEDs like many other PN junctions have a negative temperature coefficient of about -2 mV/ºC, which can be useful in such applications.
Tom
Some have installed Noctua fans in power supplies to quiet them. So that is an option. I'd still want some sort of linear fan control, though. I built a fan controller for the power supply fan in my PC back in the early 90s. Used an LED as the temperature sensor. Red LEDs like many other PN junctions have a negative temperature coefficient of about -2 mV/ºC, which can be useful in such applications.
Tom
Do you have a way of loading it such that it will turn on the fan? I think we all expected the supply to be quiet when the fan is not running. 🙂
Tom
The amp passes the torture test. I ran it at 40 W into 4 Ω on two channels. That may sound like peanuts for a nearly 400 W rated amp, but that's fairly close to the worst case power dissipation. It's also the FTC amp test standard. The heat sinks stabilized at 62 ºC and the heat sink in the power supply reached about that temperature too.
I also ran it with a 32-tone test signal with the signal peaks at 2x 300+ W into 4 Ω. The test signal has a crest factor of 11.3 dB, so it's pretty representative of music. The amp handled that without issue as well.
So far it looks like the Mean Well LOP-600-36 is a good fit for a stereo Modulus-686 amp with 4-8 Ω load ... and a good fit for this 4-channel amp (4 x 8 Ω load). I did connect the channels pairwise out of phase. I.e., I inverted the phase on the input and output of the channels mounted on the left side in the chassis. This balances the load across the two power supplies.
Tom
I also ran it with a 32-tone test signal with the signal peaks at 2x 300+ W into 4 Ω. The test signal has a crest factor of 11.3 dB, so it's pretty representative of music. The amp handled that without issue as well.
So far it looks like the Mean Well LOP-600-36 is a good fit for a stereo Modulus-686 amp with 4-8 Ω load ... and a good fit for this 4-channel amp (4 x 8 Ω load). I did connect the channels pairwise out of phase. I.e., I inverted the phase on the input and output of the channels mounted on the left side in the chassis. This balances the load across the two power supplies.
Tom
Sorry, I use it to power 6 channels of my AVR with 3 D-amps. Even after hours of use it is simply as cold as the room. The most I can pull from it may be 30-40 Watt in normal use, most of it idle current of the amps. Class D is not that effective when large amps run on lowl load.
I'm quite sure the Mean Well fan will be audible when it kicks in, but if needed there should be a SPL in the room which makes it inaudible.
It would need a class A to make it run permanently.
In fact I don't care about fan noise too much. If it is too loud, there are many options to silence it. I care much more about the combination of such a PS with an amp and the audible result.
I have spend most of my life with heavy audio transformers and slowly have to realize that this is not the only way to power an amp any more. Heavy and reliable. Had only one capacitor blowing up ever. In of one of my builds, a conventional supply, a very expensive 10.000mF Roederstein. It was 28 years old.
Switch mode power is not going to disapear in the near future. A rectifier and a few caps are easy to understand for me, these "new" SMPS's are a little more complicated. So I'm still a little suspicious.
It may be stupid from a professional standpoint, I trust switch mode with D-amps, but not so much with A/B.
So your experience is very valueable for me.
The only snags I can think of with SMPSes in audio circuits is that they tend not to like light loads and also that their current limits tend to be hard. It is massively annoying if the supply enters hiccup mode while the music is playing. And it's annoying if the supply whines at light load.
It's pretty easy to calculate the peak load current, but whether the supply whines at light load will often need to be determined by buying a bunch of them and testing them. With the RPS-400-36, about half of them were whiners. Interestingly, none of the ten I bought whined. Go figure...
Some may argue that SMPSes will always have lower reliability than a traditional mains transformer-based supply simply because SMPSes contain more components. That's possible, I suppose. But electronics are pretty darn reliable these days. I have SMPSes in test equipment from the early/mid 80s that still fire up just fine when I push the power button.
I certainly appreciate the lower weight when I go to ship amps with SMPSes. The 4-channel MOD686 amp weighs about 10 kg. Most of that is the heat sinks.
Tom
It's pretty easy to calculate the peak load current, but whether the supply whines at light load will often need to be determined by buying a bunch of them and testing them. With the RPS-400-36, about half of them were whiners. Interestingly, none of the ten I bought whined. Go figure...
Some may argue that SMPSes will always have lower reliability than a traditional mains transformer-based supply simply because SMPSes contain more components. That's possible, I suppose. But electronics are pretty darn reliable these days. I have SMPSes in test equipment from the early/mid 80s that still fire up just fine when I push the power button.
I certainly appreciate the lower weight when I go to ship amps with SMPSes. The 4-channel MOD686 amp weighs about 10 kg. Most of that is the heat sinks.
Tom
And pictures of the completed beast.
Tom
Tom
Hi Tom
Since you are using XLR connectors (locking) and Speakons (locking) did you not consider a Powercon mains connector?
IEC's drive me nuts - the various fancy distribution boards I use have them, and they often just pull out. I put Powercons on my amps, so at least with them there is zero chance of the mains pulling out of those.
Craig
Since you are using XLR connectors (locking) and Speakons (locking) did you not consider a Powercon mains connector?
IEC's drive me nuts - the various fancy distribution boards I use have them, and they often just pull out. I put Powercons on my amps, so at least with them there is zero chance of the mains pulling out of those.
Craig
I think Tom has mentioned that customers are all over the place on IEC versus Powercon, as well as on SpeakOn versus binding posts. As I recall, majority of customers use and want IEC, and I suspect that this is the case because of the ubiquity of IEC power cords.
With all that in mind, I've just bought four sets of powercons to replace the IEC's on the distribution boards.
Bulgin used to make wire clamps to hold their IEC's in.
Well I never - they still do https://www.bulgin.com/products/pub/media/import/attachments/IEC_connectors.pdf page 21.
Of course they only work with a Bulgin IEC cable connector.
Craig
Bulgin used to make wire clamps to hold their IEC's in.
Well I never - they still do https://www.bulgin.com/products/pub/media/import/attachments/IEC_connectors.pdf page 21.
Of course they only work with a Bulgin IEC cable connector.
Craig
Tom,
What do you use to make the stick-on labels I see on the back of your latest Mod-686 above? Looks like a nice, hopefully fairly inexpensive alternative to the methods you would use on the front.
Thanks,
Skip
What do you use to make the stick-on labels I see on the back of your latest Mod-686 above? Looks like a nice, hopefully fairly inexpensive alternative to the methods you would use on the front.
Thanks,
Skip
My client did not request a PowerCon and I didn't ask. So it became an IEC. I think I'll ask next time. It's a heck of a lot easier to make a round hole in those steel panels than a rectangular one.
ModuShop does not offer customization of the steel panels, sadly, so I machined those. I used a carbide tipped hole saw for the 24 mm holes for the Neutrik connectors. $10 from Bezos.
I use a label maker. Mine is a Brother P-Touch P600, but a less fancy model will work too. Just get one that uses the Brother TZe tape.
I de-greased the surface with 99% isopropyl alcohol before I applied the labels. I'm pretty confident they'll stay.
The front panel was digitally printed by ModuShop. The logo is in two colours. It's pretty obvious in person, but somehow the blue and black look nearly the same in the picture. Here's a better picture of the previous build:
Tom
From the post above, the Mean Well LOP-600-36 is a good choice of power supply for a 2 channel modulus 686. Of course, that means one for the +36v supply and one for the -36v supply, a total of two required for a stereo application.
By extension, (for a monoblock build) it would appear a Mean Well LOP-300-36 would be adequate. Of course, for a stereo application that would mean a total of four LOP-300-36's. Two for the required +36v rails, and two for the -36v rails.
I suspect the caveat is that the LOP-300-36 has not been tested to verify that it does not whine.
Comments?
By extension, (for a monoblock build) it would appear a Mean Well LOP-300-36 would be adequate. Of course, for a stereo application that would mean a total of four LOP-300-36's. Two for the required +36v rails, and two for the -36v rails.
I suspect the caveat is that the LOP-300-36 has not been tested to verify that it does not whine.
Comments?
OFF TOPIC, but it looks like a pair of the LOP-300-30's would be a good option for the 86 build, (28.5 ~31.5V) sitting between 27 and 36.
LOP-300-30
LOP-300-30
Not off topic in my view. The LOP series is great. Especially the 150-200% overload "forgiveness" is very, very handy for audio. I recently built a 4-channel MOD686 for a client that has two LOP-600-36 in it. It's intended for four channels of 8 Ω. The LOP-600-36 is the lowest powered option I'd use for a stereo amp intended for 2x 4 Ω operation. You could likely squeeze by with the LOP-400-36, but it's only $36 more for two of the 600 W version, so just go with that. The supplies get pretty toasty with music operation near clipping levels with a 4 Ω load.
The LOP-xxx-27 is a nice option for the Safe-n-Sane build. The LOP-300-30 is the only 30 V option.
I would say that the LOP-300-30 is the absolute smallest you can get away with in a stereo amp. If you provide enough air flow around them for convection and run the two channels in the stereo amp out of phase (so invert the input and the output on one channel) those supplies should provide enough power. I would probably go for the LOP-400-27. Dial the voltage up a bit if you really, really want that extra few watt of output power.
Or... Use the LOP-300-30 and a couple of Noctua fans to move some air over the supplies.
Tom
The LOP-xxx-27 is a nice option for the Safe-n-Sane build. The LOP-300-30 is the only 30 V option.
I would say that the LOP-300-30 is the absolute smallest you can get away with in a stereo amp. If you provide enough air flow around them for convection and run the two channels in the stereo amp out of phase (so invert the input and the output on one channel) those supplies should provide enough power. I would probably go for the LOP-400-27. Dial the voltage up a bit if you really, really want that extra few watt of output power.
Or... Use the LOP-300-30 and a couple of Noctua fans to move some air over the supplies.
Tom
That 4-channel build is marvelous! The connections are looking sweet, and so are those PSU's.
I'm making some drawings for a 5U 400mm chassis to run as cool as possible. There's a 500mm version too but the extra 10cm depth is a major inconvenience. I would take bigger finned heat sinks any day, but that would mess with a whole lot of typical hifi racks out there.
The 5U 400mm chassis should be pretty close to the thermal resistance required for ±36 V sine wave operation at 4 ohm. I can't tell for sure because I've never seen any exact figures for the thermal resistance of the different Modushop heat sinks.
Since I'm already close to the thermal requirements, why not approach the power requirements as well? It would be cool to have for making measurements, and maybe running the PSU's a little cooler during music reproduction too. I have a pair of Connex SMPS800RE ready to go, but they will run out of steam with sine waves eventually. Looking at the LOP-600-36, I find the idea of going dual mono with them very tempting somehow. It would be the opposite end of the 4-channel build; 2 channels with 4 PSU's. Another datapoint for the LOP's would be good to have for the community, it'll be fun to try, and maybe such an amp could be the beauty and the beast all rolled into one. If it doesn't work out, I can always swap an inner plate or bottom plate and insert the Connex PSU's or something else entirely.
@tomchr Do you see any obvious objections to trying a dual mono build with LOP-600-36's? I'd rather prevent unsafe-n-insane, but safe-n-insane is fine with me. And care to share the part number for the terminal blocks you used in that beauty of a 4-channel amp?
I'm making some drawings for a 5U 400mm chassis to run as cool as possible. There's a 500mm version too but the extra 10cm depth is a major inconvenience. I would take bigger finned heat sinks any day, but that would mess with a whole lot of typical hifi racks out there.
The 5U 400mm chassis should be pretty close to the thermal resistance required for ±36 V sine wave operation at 4 ohm. I can't tell for sure because I've never seen any exact figures for the thermal resistance of the different Modushop heat sinks.
Since I'm already close to the thermal requirements, why not approach the power requirements as well? It would be cool to have for making measurements, and maybe running the PSU's a little cooler during music reproduction too. I have a pair of Connex SMPS800RE ready to go, but they will run out of steam with sine waves eventually. Looking at the LOP-600-36, I find the idea of going dual mono with them very tempting somehow. It would be the opposite end of the 4-channel build; 2 channels with 4 PSU's. Another datapoint for the LOP's would be good to have for the community, it'll be fun to try, and maybe such an amp could be the beauty and the beast all rolled into one. If it doesn't work out, I can always swap an inner plate or bottom plate and insert the Connex PSU's or something else entirely.
@tomchr Do you see any obvious objections to trying a dual mono build with LOP-600-36's? I'd rather prevent unsafe-n-insane, but safe-n-insane is fine with me. And care to share the part number for the terminal blocks you used in that beauty of a 4-channel amp?
Last edited:
You can find thermal info here: https://hifi2000.shop/site/user/pages/05.docs/PESANTE _ DISSIPANTE Thermal info.pdf
I don't know about the 5Ux400, but all the other 400 mm deep chassis have heat sinks made from two chunks that are 200 mm wide. I have heard from one customer who had a bit of an extrusion 'burp' on one of the heat sinks that prevented the heat sinks from coming together tightly. He filed off that burp and all was well, but it was annoying. So I'd stay with the 300 mm deep if you can. Then again, don't judge by a one-off experience.
With 5U in height there's no shortage of volume, so you can always have a vertical plate holding stuff or start stacking. The power supplies are easy to stack with long standoffs. You can also put circuitry on the rear panel. I put the ISS there in the 4-channel build. You could even put the power supplies on the heat sinks(!)
I don't see any issues with a dual mono build. I don't see a huge advantage, but no issues. The terminal blocks I used are called barrier terminal blocks. I forget the part number.
Tom
I don't know about the 5Ux400, but all the other 400 mm deep chassis have heat sinks made from two chunks that are 200 mm wide. I have heard from one customer who had a bit of an extrusion 'burp' on one of the heat sinks that prevented the heat sinks from coming together tightly. He filed off that burp and all was well, but it was annoying. So I'd stay with the 300 mm deep if you can. Then again, don't judge by a one-off experience.
With 5U in height there's no shortage of volume, so you can always have a vertical plate holding stuff or start stacking. The power supplies are easy to stack with long standoffs. You can also put circuitry on the rear panel. I put the ISS there in the 4-channel build. You could even put the power supplies on the heat sinks(!)
I don't see any issues with a dual mono build. I don't see a huge advantage, but no issues. The terminal blocks I used are called barrier terminal blocks. I forget the part number.
Tom
Thanks Tom! I'll take my chances with the extrusion, I'd like to have the extra bit of cooling. I'm counting on having to file/sand anyway in case I don't get proper grounding conductivity between chassis parts. My last Dissipante 2U 300mm needed it.
In theory, the extrusion could bite you even with a single heat sink per side, it could cause the front panel to not fit nicely.
I found some nice looking barrier terminal blocks by Eaton: TB100/TB200/TB300, still need to determine the proper size. Their datasheet is good, listing optional parts like which bridges you can order to place between poles of different rows.
Pretty creative idea to mount PSU's on heat sinks. The 5U chassis has a lot of height, 210mm available internally to be exact. A 4-PSU stack is still cutting it pretty close if you want to respect datasheet tolerances for them. You'd have only ~150-250mm to spare depending on if you use the inner plate. It would be fine from a technical perspective I guess, but such a power tower would have a bit of a typical look. I've drawn a t-slot profile rig for them with QCAD. I'm tilting the PSU's backward 90 degrees, kind of like an old file cabinet. So I will have a PSU cabinet instead of a power tower.
In theory, the extrusion could bite you even with a single heat sink per side, it could cause the front panel to not fit nicely.
I found some nice looking barrier terminal blocks by Eaton: TB100/TB200/TB300, still need to determine the proper size. Their datasheet is good, listing optional parts like which bridges you can order to place between poles of different rows.
Pretty creative idea to mount PSU's on heat sinks. The 5U chassis has a lot of height, 210mm available internally to be exact. A 4-PSU stack is still cutting it pretty close if you want to respect datasheet tolerances for them. You'd have only ~150-250mm to spare depending on if you use the inner plate. It would be fine from a technical perspective I guess, but such a power tower would have a bit of a typical look. I've drawn a t-slot profile rig for them with QCAD. I'm tilting the PSU's backward 90 degrees, kind of like an old file cabinet. So I will have a PSU cabinet instead of a power tower.