I have yet to find a condition in the lab that supports that. That's why I recommend using a Thiele Network.
45º PM is "textbook stable". I've always designed for 60º PM for typical conditions and 45º PM for the worst case process corner. Due to the quirks in the LM3886 model, I often end up with 80-90º PM in the simulator.
Anyway. I agree with your notion that your questions are probably better served in a different thread. I'll be happy to participate there to the extent my time allows.
BTW: We just crossed 5000 posts and are rapidly approaching 800k page views for this thread. That's pretty awesome. Thanks to Bill for starting this thread. Maybe 2020 will be the year when he completes his amp...
Also thank you to everybody who has supported the project so far and who have made this thread what it is. I really appreciate it.
Cheers!
Tom
I need some help with something very basic.
There's one thing standing in the way of having my amp completely finished.... and that's plastic self adhesive PCB mounts.
Going through endless google results yields so many results but they don't tell me dimension of hole which they fit (to fit tom's boards)
I can't find them on mouser either, so I ended up buying some off ebay and they are too small.
If any one could link me to some it would be greatly appreciated, preferably from a retailer in the UK.
Many thanks guys
Tom
There's one thing standing in the way of having my amp completely finished.... and that's plastic self adhesive PCB mounts.
Going through endless google results yields so many results but they don't tell me dimension of hole which they fit (to fit tom's boards)
I can't find them on mouser either, so I ended up buying some off ebay and they are too small.
If any one could link me to some it would be greatly appreciated, preferably from a retailer in the UK.
Many thanks guys
Tom
Shudder! Please don't rely on anything self-adhesive for mounting the board. I understand the convenience during assembly, but the adhesive tends to fail over time.
Rather, I recommend using screws and standoffs. One way to get the screw holes in the right spots is to use a transfer punch to make a centre punch mark on the bottom panel centred below the hole in the PCB. If I had a set of transfer punches, that's what I'd do. But since I don't I just use a scribe to mark the hole location, centre punch, and drill the hole a little oversized (3.3-3.5 mm for a 3 mm screw).
Tom
HUMmmmmmmmmm
TomC,
I wrote not along telling you about a HUMMMMM I had on my PAR86. It went away (completely) the moment I hooked up a pre/dac with fully balanced outputs. When I first cranked up the PAR86 I was using the only thing I had on hand at the time, a little RCA-connected Nuforce dac with a volume control.
Now, just dead silence.
TomC,
I wrote not along telling you about a HUMMMMM I had on my PAR86. It went away (completely) the moment I hooked up a pre/dac with fully balanced outputs. When I first cranked up the PAR86 I was using the only thing I had on hand at the time, a little RCA-connected Nuforce dac with a volume control.
Now, just dead silence.
I have a couple questions about the Modulus 86, and a comment. I have been looking at modern amplifier kits, and this one really raised my interest with the fact that it gets such perfect specs from an amp which is not discrete.
Comment:
Great to see something like this being done in Alberta!
Questions:
1. Does the board come with a schematic?
2. What was the idea behind limiting bandwidth to exactly the audio band? I see lots of amps with a higher bandwidth.
3. Why the Overture series, and not the TDA series chips?
Comment:
Great to see something like this being done in Alberta!
Questions:
1. Does the board come with a schematic?
2. What was the idea behind limiting bandwidth to exactly the audio band? I see lots of amps with a higher bandwidth.
3. Why the Overture series, and not the TDA series chips?
Is pin 1 of your XLR connected to the chassis? I forgot to make that connection on my 686, and had some hum with no source connected. Once I grounded Pin 1, it went away completely.
TomC,
I wrote not along telling you about a HUMMMMM I had on my PAR86. It went away (completely) the moment I hooked up a pre/dac with fully balanced outputs. When I first cranked up the PAR86 I was using the only thing I had on hand at the time, a little RCA-connected Nuforce dac with a volume control.
Now, just dead silence.
Hi Tom,
I wondered if you had tried running the Mod86 on any of the connex QR power supplies?
I was thinking of a small pair of monoblocks using the SMPS240QR in 30V...
I know you've mentioned in the past that the Mod86 performs just as well on the Bench supply, linear supply and SMPS and you've tested the SMPS300RE and I seem to recall you questioning the point of an unregulated SMPS but the size (and price) advantage for a pair of small mono amps is quite attractive - but only if there are no issues that come with it.
Thoughts?
John
I wondered if you had tried running the Mod86 on any of the connex QR power supplies?
I was thinking of a small pair of monoblocks using the SMPS240QR in 30V...
I know you've mentioned in the past that the Mod86 performs just as well on the Bench supply, linear supply and SMPS and you've tested the SMPS300RE and I seem to recall you questioning the point of an unregulated SMPS but the size (and price) advantage for a pair of small mono amps is quite attractive - but only if there are no issues that come with it.
Thoughts?
John
TomC,
I wrote not along telling you about a HUMMMMM I had on my PAR86. It went away (completely) the moment I hooked up a pre/dac with fully balanced outputs. When I first cranked up the PAR86 I was using the only thing I had on hand at the time, a little RCA-connected Nuforce dac with a volume control.
Now, just dead silence.
I'm glad to hear you got it solved. I've been playing around with my Modulus-686 and grounding lately. I've never personally had issues with any of my amps, but I also use differential sources.
The Modulus-686 is quiet even with its input floating. If I add an XLR-RCA adapter (so turn the connection pseudo-differential as I show in the design doc), I do get faint hum with the input floating. As soon as I plug in a source (or 50 Ω terminator), the amp is completely quiet.
It's not a surprise that you get a little hum with the input floating when using it single-ended. What you're hearing is whatever voltage is induced across the 48 kΩ input impedance. A differential input will pick up roughly equal hum on IN+ and IN-. As the differential receiver senses the difference between IN+ and IN-, the end result will be zero hum.
That's my recommendation, and is the case in all my amps.
Thank you.
Thanks. Yeah... We're not all working in the Oil & Gas sector. Even in Calgary.
Nope.
The bandwidth of the Modulus-86 is < 0.1 Hz to 66 kHz. That's quite a bit wider than the 20 Hz to 20 kHz audio band.
The bandwidth is limited in the low end by the DC servo and in the high end by the RFI filter and the response of the composite amp.
I'm not seeing any advantage of the TDA chips, honestly. If you just skim the headline specs, they look amazing. I mean, who wouldn't want a 150 W amp in a small package. Once you read into the data sheet, you realize that it's 150 W "music power" into a silly low impedance, such as 1 Ω. The LM3886 data sheet is an honest data sheet. I prefer honesty...
Tom
Last edited:
I have. I recommend the SMPS300RE (±30 V version) for use with the Modulus-86. One SMPS300RE can power a stereo Modulus-86.
I think that'd be a sweet setup.
Connex does sometimes use those really cheap Chinese terminal blocks that are made from blue plastic. Other times, they use some green ones. The green ones are much better. I've had several of the blue ones crack when I tightened the screws. They're easy to replace and you can find substitutes at Mouser.
Tom
Hi Danny!
Just let me tell some of my experiences. I’ve built a composite LM3886 amp very similar to Tom’s Modulus, what gave me the basic idea. In the wintertime I’ve simulated-built-measured-listened, simulated-built-measured-listened and so on, made a lot of period and have my conclusions so far…
The amp is really good, but there are some things to consider. A big disappointment was that my favorite Wima FKP capacitors doesn’t perform well as HF filter. A lot of manufacturer use them, but they simply sound dull, without fine details. I’ve tried e.g. Murata-RDE tht C0G caps, they are definitive better, but a teeny cold. My bests are polystyrene types.
The slew-rate of the amp is about 5V/us. It is not too “fast”, but fast enough for >20 kHz. As Douglas Self writes, it is only one of the marketing slogans, that faster is necessarily better!
Phase margin in the simulator > 90 degrees, THD < 0.0001%. I was wondering anyway how good the simulator is! I can model the transient response nearly identical to my measurements.
Other discovery form me, that the capacitor (typ. 220pF) between the + / - inputs of the LM3886 doesn’t help too much, causes only trouble. The usable values here are about 1.5kOhm - 1.5nF RC combination, but didn’t tested it with my ears yet.
Another thing: I do not use a compensating cap in the global feedback like Tom. Tried it with 10-22-33nF values but sonically makes it only worse, the amp stays in any case stable.
And so on…
I would say, keep it as simple as you can, and be aware of component quality if you build it!
Just let me tell some of my experiences. I’ve built a composite LM3886 amp very similar to Tom’s Modulus, what gave me the basic idea. In the wintertime I’ve simulated-built-measured-listened, simulated-built-measured-listened and so on, made a lot of period and have my conclusions so far…
The amp is really good, but there are some things to consider. A big disappointment was that my favorite Wima FKP capacitors doesn’t perform well as HF filter. A lot of manufacturer use them, but they simply sound dull, without fine details. I’ve tried e.g. Murata-RDE tht C0G caps, they are definitive better, but a teeny cold. My bests are polystyrene types.
The slew-rate of the amp is about 5V/us. It is not too “fast”, but fast enough for >20 kHz. As Douglas Self writes, it is only one of the marketing slogans, that faster is necessarily better!
Phase margin in the simulator > 90 degrees, THD < 0.0001%. I was wondering anyway how good the simulator is! I can model the transient response nearly identical to my measurements.
Other discovery form me, that the capacitor (typ. 220pF) between the + / - inputs of the LM3886 doesn’t help too much, causes only trouble. The usable values here are about 1.5kOhm - 1.5nF RC combination, but didn’t tested it with my ears yet.
Another thing: I do not use a compensating cap in the global feedback like Tom. Tried it with 10-22-33nF values but sonically makes it only worse, the amp stays in any case stable.
And so on…
I would say, keep it as simple as you can, and be aware of component quality if you build it!
Attachments
Nice purple!
The Modulus-86, -186, -286, and -686 sound the same below clipping. I.e. they don't make any sound. They just amplify.
I did find that my LXmini speakers would sound a bit harsh when I really cranked them on the Modulus-86. I attributed that to lack of room treatment, though I had no data to back that up.
Once I had developed the Modulus-686, I swapped it in instead of the Modulus-86 fully expecting it to make no difference, but I found that harshness at higher volume levels to now be gone. So I'm wondering if I played loud enough to cause the Modulus-86 to just reach clipping on the peaks. With all the EQ in the LXmini XO, that's not unthinkable.
I have subsequently swapped in the Modulus-286 Kit (LE) instead of the Modulus-686 and found no difference.
Last year I found a pair of KEF R700 on closeout at a good price. I've never owned a pair of reference level speakers, and I figured now's the time. Everybody on my dad's side of the family wears hearing aids, so getting a pair of good speakers while I can still appreciate them seemed like a good idea.
I gotta say that the R700 represent a night-and-day improvement on the LXmini. Holy cow. Apparently that's what happens when you throw science, engineering, and measurements (gasp!) at speaker design on a large scale!
It would be interesting to put the MOD86 on the KEFs and see if I still hear the harshness.
The main differences between the various members in the Modulus family are:
- Modulus-86 has slightly higher noise, thus worse THD+N.
- Modulus-186, 286, 686 have no DC servo. I designed it out to save cost and was able to get good DC specs without it.
- Modulus-86 and 186 provide the same output power (65-70 W, 4 Ω). Modulus-286 provides 120 W (4 Ω), and Modulus-686 provides 360 W (4 Ω).
Tom
Last edited:
I can always lump the 86 on my speakers (Mordaunt Short Performance 880) and crank it up.Nice purple!
The Modulus-86, -186, -286, and -686 sound the same below clipping. I.e. they don't make any sound. They just amplify.
I did find that my LXmini speakers would sound a bit harsh when I really cranked them on the Modulus-86. I attributed that to lack of room treatment, though I had no data to back that up.
Once I had developed the Modulus-686, I swapped it in instead of the Modulus-86 fully expecting it to make no difference, but I found that harshness at higher volume levels to now be gone. So I'm wondering if I played loud enough to cause the Modulus-86 to just reach clipping on the peaks. With all the EQ in the LXmini XO, that's not unthinkable.
I have subsequently swapped in the Modulus-286 Kit (LE) instead of the Modulus-686 and found no difference.
Last year I found a pair of KEF R700 on closeout at a good price. I've never owned a pair of reference level speakers, and I figured now's the time. Everybody on my dad's side of the family wears hearing aids, so getting a pair of good speakers while I can still appreciate them seemed like a good idea.
I gotta say that the R700 represent a night-and-day improvement on the LXmini. Holy cow. Apparently that's what happens when you throw science, engineering, and measurements (gasp!) at speaker design on a large scale!
It would be interesting to put the MOD86 on the KEFs and see if I still hear the harshness.
The main differences between the various members in the Modulus family are:
- Modulus-86 has slightly higher noise, thus worse THD+N.
- Modulus-186, 286, 686 have no DC servo. I designed it out to save cost and was able to get good DC specs without it.
- Modulus-86 and 186 provide the same output power (65-70 W, 4 Ω). Modulus-286 provides 120 W (4 Ω), and Modulus-686 provides 360 W (4 Ω).
I love my "safe-n-sane" Modulus-686 build. 3U x 330 x 300 Mini Dissipante, powered by two Mean Well RPS-400-27-C controlled by my Mean Well Controller. 130 W (8 Ω), 200 W (4 Ω). Perfect!
Tom
See what happens compared to the 686.
Personally, I just can't see a reason not to have the 686 other than size, which isnt even nearly a problem in my smallish room
Congratulations on the new speakers. I think there's a lot of sense in buying well, whilst ears are young enough to appreciate it. At 39, I have a few years left to enjoy this system to the fullest, and I encourage that behaviour [emoji6]
As for DIY speakers, I'm afraid I have never heard 1 pair that were "right" so I'm not surprised by your findings. Enjoy them