Because many will build this amp for use with consumer sources, including DACs running on 3.3 V rails and providing 900 mV RMS out on a good day.
If you need 20 dB gain, just toss me an email when you order your boards. As I mentioned earlier in this thread, I'll be happy to pluck the gain resistor off the board for you.
If I start receiving a lot of requests for 20 dB gain, I'll be happy to change the default gain for the next batch of boards. It's just a BOM change.
Tom
If you need 20 dB gain, just toss me an email when you order your boards. As I mentioned earlier in this thread, I'll be happy to pluck the gain resistor off the board for you.
If I start receiving a lot of requests for 20 dB gain, I'll be happy to change the default gain for the next batch of boards. It's just a BOM change.
Tom
questions on ICL's in the power circuit
Hi,
I am thinking of which way to go in planning a Modulus-686 build.
I want to listen to music, not generate pure sine waves.
Speakers are relatively easy 8 ohm load, 400 watt max, 95 db sensitive.
Source is differential 4 volts in from a DAC/preamp.
I was planning a build in the Dissipante 300mm 4U chassis.
If I build with the Power 86/ and an AN-5225 transformer, is it reasonable to use ICL's like the CL-60, ala First Watt power supplies, to decrease the inrush, or better to build with two Mean Well RPS-400-36 supplies?
Thanks,
Matt
Hi,
I am thinking of which way to go in planning a Modulus-686 build.
I want to listen to music, not generate pure sine waves.
Speakers are relatively easy 8 ohm load, 400 watt max, 95 db sensitive.
Source is differential 4 volts in from a DAC/preamp.
I was planning a build in the Dissipante 300mm 4U chassis.
If I build with the Power 86/ and an AN-5225 transformer, is it reasonable to use ICL's like the CL-60, ala First Watt power supplies, to decrease the inrush, or better to build with two Mean Well RPS-400-36 supplies?
Thanks,
Matt
I'm happy to help.
It's refreshing to know that your kind of people exist in the world...
So 95 dB @ 1 W, 1 m means 121 dB @ 400 W, 1 m. Watch your hearing...
With 4 V in, you'll want the 20 dB gain option. Also with 95 dB efficient speakers, the lower noise of the 20 dB option would be appreciated, so that's the way to go. Some may argue that a little extra gain is nice, but unless your listening space is huge you'll be fine given the 95 dB sensitivity of your speakers.
The 4U Dissipante chassis has heatsinks large enough to support the MOD686 with ±35-36 V rails. Good choice there. Given your speakers are so efficient and 8 Ω, you could consider going with the 3U chassis. With 4 Ω speakers it would be a bit marginal and I'd encourage you to add a thermal switch, but for 8 Ω speakers it will be fine. I'm just throwing that out there as a possibility. If you want the amp to be virtually bomb proof, go with the 4U chassis.
You have to be careful with the inrush limiters. They can only handle so much energy without exploding, so keep an eye on the amount of energy storage in your supply caps. The other thing to keep in mind is that a typical audio amp doesn't draw enough current to heat up the inrush limiter, so once the initial transient is over, the inrush limiter presents a sizeable resistance in series with the transformer primary. And if the amp does draw enough power, the inrush limiter gets hot and, thus, does not provide inrush limiting on a brief power glitch. Thus, I recommend shorting out the inrush limiter with a relay once the initial transient is over. Then, of course, you have to ensure that the relay disengages early enough when the rail voltage collapses that you get sufficient inrush limiting should you turn the amp off and then on again within a few seconds.
Some argue that power resistors should be used for the inrush limiting as they tend to fail less catastrophically than the NTC inrush limiters. Even with power resistors, you have to watch the transient power when the supply is turned on. Most power resistors only allow for around 2.5x their rated power on transients.
Anyway. Long story. I see a Power-686 board in my future. One that features inrush limiting. That's probably an early spring project.
Another board I see in my future is a speaker protection board. I've had quite a few requests for one of those.
Anyway. If you want inrush limiting, I suggest either doing the math or waiting until I have the Power-686 (or whatever I name it) ready. Alternatively, two of the Mean Well RPS-400-36 would be a good choice.
Tom
Depends on the size of the transformer (VA rating), how paranoid you are about fusing, and how long you want the transformer to last. I seem to recall covering this earlier in this thread...
Regardless of need, many want a soft start. If the market wants, it would make good business sense to cater to this need whether it is a real or perceived need.
There are certainly good technical reasons why one would want a soft start with larger transformers (say 400-500 VA and above). Some prefer to use soft start circuits even with smaller transformers.
I'll also include surge suppression in the Power-686. Some have asked for that feature as well.
Tom
Regardless of need, many want a soft start. If the market wants, it would make good business sense to cater to this need whether it is a real or perceived need.
There are certainly good technical reasons why one would want a soft start with larger transformers (say 400-500 VA and above). Some prefer to use soft start circuits even with smaller transformers.
I'll also include surge suppression in the Power-686. Some have asked for that feature as well.
Tom
Last edited:
Thanks, Tom!
I will stick to my tentative plan of a stereo Modulus-686 in the 4U dissipante chassis as I have built quite a few Class A amps in that box. And, I will likely build with the 500VA Antek transformer / Power 86 power supply, no ICLs, and later, if the power 686 comes to fruition, decide if an update/upgrade is necessary. I should have plenty of real estate in that chassis to do whatever.
Matt
I will stick to my tentative plan of a stereo Modulus-686 in the 4U dissipante chassis as I have built quite a few Class A amps in that box. And, I will likely build with the 500VA Antek transformer / Power 86 power supply, no ICLs, and later, if the power 686 comes to fruition, decide if an update/upgrade is necessary. I should have plenty of real estate in that chassis to do whatever.
Matt
Last edited:
Depends on the size of the transformer (VA rating), how paranoid you are about fusing, and how long you want the transformer to last. I seem to recall covering this earlier in this thread...
I'll also include surge suppression in the Power-686. Some have asked for that feature as well.
I read all thread. In general, under 400VA transformers don't need a soft start. What's the recommended size for one channel (default power from your specs is fine for me)?
24V or 25V output (IIRC), or anything else as I have plenty of choice here in Europe.
Dunno if your power-686 module will be ready by the time I will assemble the amps
If you add the feature make it bypassable, though. It will be a mono module I hope
Sounds like a plan.
350-500 VA per channel is reasonable. I provided a bunch of transformer models here: Post #63.
Math says 328 VA per channel for music (14 dB CF) into a 4 Ω load. 1108 VA per channel for continuous sine wave into 4 Ω. Those numbers are pretty conservative (i.e. err on the larger side) and assume no voltage droop. I'd say somewhere in between is probably fine, so 350-500 VA per channel is my recommendation. Your 400 VA should be a good happy medium.
2x25 VAC would be my first choice. 2x24 VAC is fine as well, but you'll get a little lower output power due to the lower voltage.
Dunno if your power-686 module will be ready by the time I will assemble the amps
I'll keep the Power-86 going, so you can always choose that if you don't want the extra features of the Power-686. You will be able to use the Power-686 to power a mono amp or a stereo amp (or quad MOD86, etc.).
Tom
I remembered correctly then. Need single or double output for a channel?
Anyway, these are the ones I wanna get:
TTSAS0400 - SUPREME AUDIO grade transformer TSAS400VA - voltage to 50V - Shop Toroidy.pl
Potted, shielded and easy mounting
TBH I want more capacitance in the PSU, the p-86 looks too skimpy. Making a CRC firstwatt-style psu until i see your p-686.
Curious what you base that claim on. Also note that if performance is what you want, CRC is not your best choice of topology. The R raises the supply impedance considerably, which is not what you want. You would be better off putting the two Cs in parallel (or choose a larger cap).
In a power amp, the R is kept low to avoid a large voltage across it on the signal peaks. This drastically reduces its effectiveness for filtering - especially when the ESR of the C is taken into account. Also note that the Modulus amps really do not care about the RC filtering in the supply, so you are much better off optimizing for low supply impedance, which gets you back to the regular supply cap used in the Power-86.
For fun, I compared the Modulus-686 performance with the stock Power-86 and with a Power-86 where I had doubled the supply capacitance by soldering two caps in parallel on the bottom of the board. I saw no meaningful difference in performance.
The ripple voltage across the ±35 V supply measures 1.2 V at 300 W into 4 Ω with the stock Power-86. With twice the capacitance, the ripple voltage predictably drops by half to 600 mV. With twice the capacitance, you'll get 2-3 W more out at clipping. That's a difference of about 0.04 dB.
Tom
Modulus-686: 380W (4Ω); 220W (8Ω) Balanced Composite Power Amp with extremely low THD
I can’t help but smile at this. We had our first AES Taiwan section meeting, and there were some talk about engineering not knowing much about music and vise versa, talking with a gentlemen from B&K in Taiwan, I shared my experience about the concept of letting people whom know music just nit pick the sound of systems then let engineering figure out a better way, a concept we used during development of flight controls. We get to really know how difficult it is to satisfy everyone.
Recently, helping a friend sort out some of his old gear, I was listening to a Yamaha receiver through a pair of speakers with Jordan JX92 drivers in direct comparison with my default system, the striking difference yet appreciating each in a different way gave me a better perception of how normal customers might like either.
I can’t help but smile at this. We had our first AES Taiwan section meeting, and there were some talk about engineering not knowing much about music and vise versa, talking with a gentlemen from B&K in Taiwan, I shared my experience about the concept of letting people whom know music just nit pick the sound of systems then let engineering figure out a better way, a concept we used during development of flight controls. We get to really know how difficult it is to satisfy everyone.
Recently, helping a friend sort out some of his old gear, I was listening to a Yamaha receiver through a pair of speakers with Jordan JX92 drivers in direct comparison with my default system, the striking difference yet appreciating each in a different way gave me a better perception of how normal customers might like either.
Last edited:
TBH I want more capacitance in the PSU, the p-86 looks too skimpy. Making a CRC first
TBH I want more capacitance in the PSU, the p-86 looks too skimpy. Making a CRC firstwatt-style psu until i see your p-686.
I must admit I have moments where I feel I'd like to try a capacitance multiplier as part of the power supply.
t
TBH I want more capacitance in the PSU, the p-86 looks too skimpy. Making a CRC firstwatt-style psu until i see your p-686.
I must admit I have moments where I feel I'd like to try a capacitance multiplier as part of the power supply.
t
Check out Bob Cordell's book on power amplifier design. He has a section on 'linear' power supplies. His comment on the 'R' in the CRC is that a very small resistor (~0.1 Ohm IIRC) doesn't add significantly to the drop already inherent in the transformer losses and adds a bit of low-pass filtering. As Tom points out, there is going to be droop under load with a linear supply that you don't get with a switching supply. On the flip side, the switcher will have a hard stop at its current limit.
Geeze Louise, with the ME600-36 regulated supply mentioned in Tom's info page on this model costing only $85 US (PartsExpress), why would anyone want to spend the time & $ on a linear supply? A decent chassis with sufficient heat sinking would likely be more than twice that, and I'd tend to saving $ where ever transparent in the final results.
I've recently blasted together a few classD kits recently using various models of SMPS, including a Hypex with their dedicated supply, and there's more than one nice thing to be said about their simplicity and efficacy - certainly as compared to the requirements for higher rail voltages and current draws in vacuum tube amps.
If a Laptop brick type of SMPS is good enough for Nelson Pass ( ACA kit) ....?
I've recently blasted together a few classD kits recently using various models of SMPS, including a Hypex with their dedicated supply, and there's more than one nice thing to be said about their simplicity and efficacy - certainly as compared to the requirements for higher rail voltages and current draws in vacuum tube amps.
If a Laptop brick type of SMPS is good enough for Nelson Pass ( ACA kit) ....?
Still...
Half the ripple is meaningful to me. The added power was never my ratio.
Anyway Im sure that you will size the new psu-686 for the requirements of their amp.
Unfortunately, that's another one of those ideas that looks great in theory (and works reasonably well in line stages) but doesn't work that well in reality (at least in power amps).
A power glitch on the input will cause the output voltage to drop, which will discharge the cap in the cap multiplier. When the input voltage recovers, the cap will now have to charge through a long time constant. This means on every power glitch, you'll have a few seconds where the cap multiplier does not perform any filtering. If this results in audible hum from the speakers, it's really, really annoying. Every time a heavy load in the neighbourhood turns on, your amp goes "hummmm..."
In a line stage, you can design the cap multiplier to have significantly lower output voltage than its input voltage. That may prevent this issue from occurring. Try that in a power amp and you'll burn gobs of power in the multiplier. Good luck with that.
Yeah. But not very much LP filtering. You'd need supply caps on the order of 0.05-0.1 F (yes, Farad!) to get any meaningful attenuation at 120 Hz. Even then, the attenuation is limited by the ESR of the cap, so you're not likely to get much better than ~10 dB of attenuation that way. I remain under-impressed.
That's certainly true.
Well... You'll need a pair of them and the SE-600 is quite loud as it has a fan in it. But yeah. Point duly noted. You can't build much power supply for the cost of a pair of those Mean Well switchers.
I'll always welcome a fact/science based discussion. A discussion on any topic, be it SMPS vs 'linear' or "how much cap is necessary", can be taken to the extreme. As long as the discussion is fact-based and related to the MOD686 and what is necessary to make it perform well, I'm cool with the discussion.
Tom
Last edited: