About the Mean Well PSU options and the whine problem, something in the introduction piece for the LOP 400/500/600 (https://www.meanwell.com/newsInfo.aspx?c=1&i=1303) caught my eye:
In the row where they mention the voltage models for the older RPS series, the 36 V version is not included in the main list. It is mentioned right below that, under "partial wattages" (looks to me like they meant "partial voltages"). It's not clear to me yet what this partial-bit refers to, but I wonder if this could be related to the whine problem. Because the 27 V version is not a "partial" model but included in the main list, and no builds with that version have the whine problem as far as I know. The 36 V version is a "partial" model, so I wonder if that ones operates differently and the problem could be in the parts that are different?
Another thing I was wondering about: some folks have solved the problem by turning a 36 V version all the way up to 40 V. According to the RPS-400 datasheet (https://www.meanwell-web.com/content/files/pdfs/productPdfs/MW/RPS-400/RPS-400-spec.pdf), 40 V is slightly in the over voltage range (39.6 ~ 46.8V) for the 36 V version.
The voltage adjustment range for the 36 V version is listed as 34.2 ~ 37.8V, but apparently they go up to 40 V? Could you maybe run the 27 V versions up to 34 V? So that it would be almost full power for a Mod-686 but without the whine? It would be higher in the over voltage range for the 27 V version though (29.7 ~ 35.1V).
In the row where they mention the voltage models for the older RPS series, the 36 V version is not included in the main list. It is mentioned right below that, under "partial wattages" (looks to me like they meant "partial voltages"). It's not clear to me yet what this partial-bit refers to, but I wonder if this could be related to the whine problem. Because the 27 V version is not a "partial" model but included in the main list, and no builds with that version have the whine problem as far as I know. The 36 V version is a "partial" model, so I wonder if that ones operates differently and the problem could be in the parts that are different?
Another thing I was wondering about: some folks have solved the problem by turning a 36 V version all the way up to 40 V. According to the RPS-400 datasheet (https://www.meanwell-web.com/content/files/pdfs/productPdfs/MW/RPS-400/RPS-400-spec.pdf), 40 V is slightly in the over voltage range (39.6 ~ 46.8V) for the 36 V version.
The voltage adjustment range for the 36 V version is listed as 34.2 ~ 37.8V, but apparently they go up to 40 V? Could you maybe run the 27 V versions up to 34 V? So that it would be almost full power for a Mod-686 but without the whine? It would be higher in the over voltage range for the 27 V version though (29.7 ~ 35.1V).
I contacted Mean Well about the differences between the RPS and the LOP series. They said the LOP was based on the same resonant converter technology as the RPS, so they couldn't rule out that the 36 V LOP would whine as well. That said, it looks like the LOP is quite an improvement on the RPS series, so I'd be tempted to try. Mouser does take returns, so if you treat the power supply gently you do have an out if it turns out to be a whiny supply. All you'd likely have to do is to plug it into the mains.
Tom
Tom
Interesting about the Mouser returns. I'm about to order a Dissipante 4U 400mm with an inner plate for my Mod-686. I'm waiting on a pair of SMS800RE for that build already, but the spacious chassis with the inner plate will allow me to experiment later. I could even fit retrofit a toroid. I'm definitely interested in using Mean Well over Connex, but I don't know if I'll be the first to try a LOP-600-36. Maybe one could even fit those in the safe-and-sane format (making it safe-and-insane? still better than unsafe-and-insane). But you might choose a longer version to get the same thermal resistance as the regular Dissipante 400mm a lot of folks have used.
I'll probably build a second Mod-686 later on too. If I can finish the first one successfully, it's not leaving my living room anymore. I could use another one to demo at friends, hoping they don't want to give it back.
I'll probably build a second Mod-686 later on too. If I can finish the first one successfully, it's not leaving my living room anymore. I could use another one to demo at friends, hoping they don't want to give it back.
Gotta say that I love the 'new' side brackets from ModuShop. Granted, they're not exactly new. I think they started offering them about two years ago, but they're new to me. They're nicely made and the captive nuts make assembly a breeze. Nicely done @Gianluca.
This is the 5U x 300 mm Dissipante with steel panels. I'm building a rack mountable amp for a client. 4x Modulus-686 powered by two Mean Well LOP-600-36.
Tom
This is the 5U x 300 mm Dissipante with steel panels. I'm building a rack mountable amp for a client. 4x Modulus-686 powered by two Mean Well LOP-600-36.
Tom
That is sweet stuff Tom, more Mod-686 is more better. Even with 5U though, with two of those boards on the 300mm heat sinks and running at ±36 V, you wouldn't run them at full power for a longer time right? Of course with sane SPL in a room you'd probably be pretty comfortable thermally.
Please let us know about your experiences with those LOP-600-36's. Those Mean Wells would be a very nice option for more Mod-686's if they run right.
Please let us know about your experiences with those LOP-600-36's. Those Mean Wells would be a very nice option for more Mod-686's if they run right.
It's a 4-channel amp intended for driving 8 Ω speakers. The heat sinks wouldn't be enough for 4 Ω operation, and the power supplies would be marginal too.
I'll test the amp thoroughly, including with 4 Ω load on two of the channels to see if the supplies behave. I hope to be able to recommend them.
Tom
I'll test the amp thoroughly, including with 4 Ω load on two of the channels to see if the supplies behave. I hope to be able to recommend them.
Tom
When I have multiple pieces of metal making up an enclosure or panel, I like using these fasteners in between wherever possible. It gives me the warm and fuzzies that there aren't going to be any gaps in bonding.
https://www.mcmaster.com/products/s...rade-5-steel-serrated-flange-hex-head-screws/
Soon I will get to my Mod-686 build. Finally settled on making monoblocks. Each will have a Power-86 and a 300VA 2x22 VDC secondary transformers. Going to use a 4U x 300 heat sink from modushop. I could always make it 5U in the future, but not sure I need that much power from the amp. Will be powering an 8 Ohm driver as part of a two way active speaker. I currently have 4 of the 4U heat sinks, going to use them here and for my wolverine build. My monoblocks will also have a 5U heat sink on the other side, for any other amp I would like to use such as Pass clones. Each have their own transformer and power supply. Built a tower 230mm wide x 300 mm deep.
Those are neat. In this build I'll run grounding wires between panels to ensure that they're all grounded. I wouldn't want to rip through the powder coating on the steel panels.
Tom
I searched this thread and didn't see this one mentioned. SMPS300RS. 36V at 9.5A peak.
Adequate for one channel?
Adequate for one channel?
The SMPS300RS is a Single supply, hence, the S in the part number. You'd need two of them in series - assuming they can be connected in series - to form ±36 V. A better option would be the SMPS800RE in a custom ±36 V voltage option. I've used that in the past. But, frankly, Mean Well provides a better product with better and much, much more detailed specs and better service, so I'd rather use a pair of those.
And 9.5 A peak is not enough. The Modulus-686 is a bridged amp, so the load can see a total of 72 V when the amp is powered by ±36 V. 72/4 = 18 A, peak for 4 Ω operation. Granted, the output can't go all the way to the rail, so maybe you're looking at more like 67/4 = 16.75 A, peak, and even then that peak won't last long (and yada, yada) so the 11 A, peak of the RPS-400-36 has shown to be enough. And if you invert input and output on one channel you can drive a stereo pair from a pair of RPS-400-36. Unfortunately, however, about half of builders who used that supply had issues with audible whine. In all cases I know of that whine could be removed by turning up the supply voltage of the RPS-400 to 40 V, but I can't guarantee that will be the case for everyone, which is why I no longer recommend the RPS-400-36 for the MOD686. The RPS-400-27 that I use in the Safe-n-Sane build is rock solid.
Tom
And 9.5 A peak is not enough. The Modulus-686 is a bridged amp, so the load can see a total of 72 V when the amp is powered by ±36 V. 72/4 = 18 A, peak for 4 Ω operation. Granted, the output can't go all the way to the rail, so maybe you're looking at more like 67/4 = 16.75 A, peak, and even then that peak won't last long (and yada, yada) so the 11 A, peak of the RPS-400-36 has shown to be enough. And if you invert input and output on one channel you can drive a stereo pair from a pair of RPS-400-36. Unfortunately, however, about half of builders who used that supply had issues with audible whine. In all cases I know of that whine could be removed by turning up the supply voltage of the RPS-400 to 40 V, but I can't guarantee that will be the case for everyone, which is why I no longer recommend the RPS-400-36 for the MOD686. The RPS-400-27 that I use in the Safe-n-Sane build is rock solid.
Tom
Last edited:
Power supply setup:
And inside view of the completed amp.
I've powered it up with two channels plugged in and with four channels plugged into the power supply. No whine that I can hear. So far, so good.
Tom
And inside view of the completed amp.
I've powered it up with two channels plugged in and with four channels plugged into the power supply. No whine that I can hear. So far, so good.
Tom
That’s a beast, but hafta say that the unused cubic space in the power supply section is just crying out for a monster linear PS. 😉
What’s the highest channel count you’re aware of in a single chassis with either the 286 or 686 modules?
What’s the highest channel count you’re aware of in a single chassis with either the 286 or 686 modules?
My 30 year old electronics repair tech memories, Nodding in agreement. This is the way it's done well.
Five, I think. I'm pretty sure someone put a pair of Modulus-686 (one on each heat sink) with two Modulus-286 on one side and one Modulus-286 on the other for an LX521.4. I forget if it was in a 4U or a 5U chassis.
One could do that. You'd probably still have to stack the transformer and the power supply, though, but there's certainly plenty of volume.
Tom
I'm not sure other than I know it's four 8 Ω speakers. Same client ordered one of these amps a few years back. That one I built with two Connex SMPS800RE (custom ±36 V version), but for this build we agreed to go with the Mean Well supplies.
Tom
The new LRS-xxx "N" supplys from Mean Well can do 200% overload for 5 seconds and have a temperature switch for the build in fan. Could a pair of these be an option? Very low priced, too.
The LOP-series allows for 150% overload for a few seconds too. Plenty for audio.
I have a pair of SE-600 that I use for testing on occasion. They are unbelievably loud. If the LRS-series is as loud I'd say they're useless for audio except for audio testing where the fan noise can be tolerated.
Tom
I have a pair of SE-600 that I use for testing on occasion. They are unbelievably loud. If the LRS-series is as loud I'd say they're useless for audio except for audio testing where the fan noise can be tolerated.
Tom
I've owned two fan cooled amps. The first was the good old Krell KSA100II. 100W class A into 8 ohms. The fans were under the power transistor chimneys, and Krell ran them slowly, so fan noise was not an issue. The second one was an Audio Research D125. That was a hybrid, but with four 6550's per channel, which were fan cooled. I had to replace the fans on that because the originals were knackered and noisy. New ones were, like with the Kell, completely unobtrusive.
Back in the day, when I was designing stuff for the company I first worked for, I used a temperature sensor, and adjusted fan speed to suit. Provided you start the fan first, you can reduce the speed to be above stall speed, then increase if the sensor warmed up enough. Same idea on my PC (Dell Optiplex 7090). On start up, whoosh - fan full speed for a second - then it cuts right back. Unless you are doing something needing heavyweight processing, when the fan noise increases.
A bang-bang on/off is very crude.
Craig
Back in the day, when I was designing stuff for the company I first worked for, I used a temperature sensor, and adjusted fan speed to suit. Provided you start the fan first, you can reduce the speed to be above stall speed, then increase if the sensor warmed up enough. Same idea on my PC (Dell Optiplex 7090). On start up, whoosh - fan full speed for a second - then it cuts right back. Unless you are doing something needing heavyweight processing, when the fan noise increases.
A bang-bang on/off is very crude.
Craig
Last edited: