Q: When using a linear power supply, the toroidal transformer is typically mounted to the bottom of the chassis via the centre bolt.
Is this absolutely necessary, or could the transformer be mounted such that it does not connect to the chassis itself but instead to a sub-chassis not grounded to the main chassis? I understand that the centre bolt would still be required as it affects the flux of the transformer.
Is this absolutely necessary, or could the transformer be mounted such that it does not connect to the chassis itself but instead to a sub-chassis not grounded to the main chassis? I understand that the centre bolt would still be required as it affects the flux of the transformer.
Yep. That provides a nice solution to the thermal interface as well. As you can imagine, a 200+ W Class AB amp dissipates a bit of power, so getting the thermal system right is a must.
Tom
Yeah. I'll start a thread in the Vendor's Bazar over the weekend.
I'm using a 0.4 K/W heat sink that I bought as scrap aluminum (at scrap aluminum cost) many years ago in Copenhagen.It's about 300x150x50 mm and weighs in at 2.5 kg if I recall correctly.
I need to look at the Dissipante heat sink specs. 4U sounds about right.
If you lower the supply voltage to ±28 V, you can probably get away with the 3U. You'd get around 150 W in that case. That'd be a nice safe and sane build for those who'd like more power than the MOD86 delivers but don't need the 200+ W.
Absolutely. I'll build the boards with the default 26 dB gain. Those who want the slightly lower noise floor can pluck a resistor off the board and reduce the gain to 20 dB.
The gain is adjusted by changing one resistor. This can be done without impacting amplifier stability. It does impact the amplifier noise floor, though, so as the gain is increased, the THD+N gets a bit worse. Figure a 3 dB change in gain -> 3 dB change in THD+N.
The MOD686 has differential (balanced/XLR) input. It can be run single-ended (unbalanced/RCA) as well. I took the lessons learned from the MOD86 and applied them to the '686. Thus, the '686 will start up cleanly (no plop) even with an unbalanced input, unlike the MOD86 where there is a small plop from the input bias current of the THAT1200.
Tom
Hi Tom
This is the link to the Modulus Heat sink specs which I posted a while back in the 286 thread.
http://www.modu.it/CARATTERISICHE_TERMICHE_DISS.pdf
It says 0.41K/W for the 3U. Would that suffice for the 686?
Would be cool though. This way I could upgrade from my 286 to 686.
Thx
SH
That'll work. If you run the amp for hours at the worst case dissipation (150-200 W, 4 Ω, sine wave) expect to see heat sink temperatures around 65 ºC, which is quite hot. The amp will survive just fine though.
For music reproduction, your heat sink will be more than lukewarm but not by much.
Curious: Did you build your MOD286 as mono modules or stereo?
Tom
Ahhh... Should I bite or should I not.
If you agree with Bruno's assumptions about listening distance and SPL, you will find his conclusion valid as well. Change any of the input variables and the result will change. For example, what if you prefer 91 dB SPL at the listening position? That'll bump the 25 W up to 100 W. Or have a large room and sit further away from the speakers? I give you the math here, so you can easily play with the numbers yourself.
As Bruno concludes in the last line, "the occasional clipping event isn't going to spoil the fun." Do you agree? Or would you rather have a system that doesn't clip?
There are also perfectly valid use cases where you do need the power. For example, the woofers in the Linkwitz LX521.4 and Orion speakers require significant power to be available if you want to avoid clipping and rude noises emitted by the speakers. That's common with active speakers where significant boost is added in the filter chain. In fact, I recently spoke with a guy who makes sound systems for recording studios. He needed 1 kW on the woofer channel to be able to support the bass boost of his active speakers ... and this was for near field listening ... and he'd done the math. Now, he was aiming for rather high SPLs, so residential users can likely dial that back some, but still...
Also, don't forget that needs and wants aren't always connected.
Tom
I'll add a smiley next time - as I should have here. (Not that there's anything wrong with having the extra details you added.)
I'd much prefer never to clip. But, in fairness to BP, the full sentence shows he's no longer talking about 25 watts:
Bruno Putzeys said:
Either way, I think we can agree that occasional clipping will have the same effect on the two amps, only it will occur at power levels 10 dB lower on the 25 W amp than on a 250 W amp.
I do suspect that the harshness that comes on many amps when the volume is cranked hard is due to the occasional clipping. At clipping, the distortion spectrum becomes a rather tall forest of harmonics of all orders. In a multi-tone environment that would create all sorts of fun IMD products, which seem to be rather strongly correlated to the listening experience. If you get fewer clipping events, the harshness should go away.
I will also note that a local buddy of mine runs into rude woofer noises with the Parallel-86 on his LX521.4s. I'm suspecting the recording is to blame as the rude noises are only present briefly in one channel on one track. He isn't convinced and suspects that the amp is clipping, which could be the case. As I pointed out earlier: Once you start adding boost to a woofer, you will need to have some power available to drive the cone excursion.
For less demanding speakers, such as the LXmini, I still think the Modulus-86 or Modulus-286 MONO (if you need a little extra headroom) are good candidates.
Tom
I do suspect that the harshness that comes on many amps when the volume is cranked hard is due to the occasional clipping. At clipping, the distortion spectrum becomes a rather tall forest of harmonics of all orders. In a multi-tone environment that would create all sorts of fun IMD products, which seem to be rather strongly correlated to the listening experience. If you get fewer clipping events, the harshness should go away.
I will also note that a local buddy of mine runs into rude woofer noises with the Parallel-86 on his LX521.4s. I'm suspecting the recording is to blame as the rude noises are only present briefly in one channel on one track. He isn't convinced and suspects that the amp is clipping, which could be the case. As I pointed out earlier: Once you start adding boost to a woofer, you will need to have some power available to drive the cone excursion.
For less demanding speakers, such as the LXmini, I still think the Modulus-86 or Modulus-286 MONO (if you need a little extra headroom) are good candidates.
Tom
I think you may be referring to "soft clipping".
An amp like the MOD86 will produce a sine wave cleanly until clipping. It goes from "soft clipping" (assuming that term can be applied to the MOD86) to hard clipping in maybe 100 mV of output swing. That, I think, is why the MOD86 sounds clean until it clips.
In other amps, the THD starts to rise well before the amp clips hard. Some call this soft clipping. Many tube amps behave like this. Nelson Pass' First Watt is apparently another example of this. I generally find such amps to sound harsh and, at least in my own experience, I connect this harshness with the rise in THD at HF. It could also be related to the rise in THD at higher output power.
Now, I don't know if Pass' First Watt sounds harsh to me as I have no experience with it, but based on the measurements that's what I'd expect.
Tom
As I mentioned a couple of pages back I have a new hobby of collecting records with 'realistic' dynamics as final tests if I have enough peak power. But TBH the only time I would get serious power angst is when using a Linkwitz transform on the woofers, where you can easily end up needing an extra 10dB. One of the possible evolutions of my speakers would need that, and I would have the drivers that could handle the power. But that is for the future.
Keep in mind that with the Linkwitz speakers or other systems with electronic crossover and an amp-per-driver, that the power demands on any given amp would be less than that of a single amp driving a traditional speaker with passive crossover after the amp.
But, as Tom said, significant EQ, especially for the woofer, can create large demands for power. A dipole woofer with EQ attempting to reproduce low-frequency effects from a movie will strain to the max of both amp output and driver excursion.
But, as Tom said, significant EQ, especially for the woofer, can create large demands for power. A dipole woofer with EQ attempting to reproduce low-frequency effects from a movie will strain to the max of both amp output and driver excursion.
Cool that will work. Thx.
I built 2 monoblocks of the 286. see here for details
Mod-286 build thread
Wonderful the news keep getting better. Currently for each monoblock I have a 350VA, 2x25VAC installed. How output power would that give me for starters at 8ohms?
Thx