In these days of the Internet, pictures also mean nothing - if someone wants to tell a porky, etc, it's pretty easy to fake it - and as for what it sounds like, well ...
I'm one of the few people who is not obsessed, especially these days, with taking 5 million pictures of what's going in their lives, each and every day - I just don't do pic's, unless there is a real use for doing so - and from when I started this audio stuff I have never taken a single snapshot of anything I've done - to me, this is a pointless thing to do, it never occurs to me as something that's important ...
I'm one of the few people who is not obsessed, especially these days, with taking 5 million pictures of what's going in their lives, each and every day - I just don't do pic's, unless there is a real use for doing so - and from when I started this audio stuff I have never taken a single snapshot of anything I've done - to me, this is a pointless thing to do, it never occurs to me as something that's important ...
yet you feel it important to lower the S/N ratio of this thread. No pictures, no schematics, no measurements, no relevance. This is a build thread.
Let me see now ... someone posed this question:
My response was to supply an answer, one answer, to that question, framing some of the relevant parameters - that's why they call this little part of the world "a forum" ...
My response was to supply an answer, one answer, to that question, framing some of the relevant parameters - that's why they call this little part of the world "a forum" ...
If someone posts up the distortion figures of the mod-86 when its driving a non-ohmic load (i.e. a loudspeaker), we might get a little closer to understanding the difference between it and a vanilla LM3886 in terms of real-world performance.
Hi,
Yes, and sorry for little offtopic..
Forgotting datasheet, maybe LM3876 sound different than 3886, like 3875 comparing 3886. For my taste 75 sound better in mid, more natural and smoother..
If sounding 76 is close to 75, I buy it today the Modulus76 !!
Phil.
Yes, and sorry for little offtopic..
Forgotting datasheet, maybe LM3876 sound different than 3886, like 3875 comparing 3886. For my taste 75 sound better in mid, more natural and smoother..
If sounding 76 is close to 75, I buy it today the Modulus76 !!
Phil.
In a composite amplifier, such as the Modulus-86, the sound quality is determined mostly by the error-correction op-amp (LME49710).
When listening on my Alpair 6P full-range speakers, the only sonic signature I hear and those of the Alpair 6P (and its enclosure). The amp is transparent, both with the LM3886 and the LM3876. I really like it.
~Tom
When listening on my Alpair 6P full-range speakers, the only sonic signature I hear and those of the Alpair 6P (and its enclosure). The amp is transparent, both with the LM3886 and the LM3876. I really like it.
~Tom
A quick question before I start building my mod-86: do you need to implement any speaker protection (e.g. an external board using a relay) or is this not necessary?
/path
/path
Is there a turn-on or turn-off "pop" when using the power-86 solution or is an external solution to create a delayed power sequence needed?
Thanks, /path
Thanks, /path
The 3886 has automute to prevent pops on turn on/off and has fairly robust protection. There will be fault modes no o one has thought of, but those will be very unlikely given the millions of these chips in the field.
If the mute resistor is chosen to draw just enough current from the mute pin (~0.2mA) to fully activate the unmute, then when the PSU charge depletes, the mute activates early for a silent shutdown.
This also works at start up by delaying the full unmute until the current comes up to near the final level.
Whereas, using a higher unmute current say around 5mA brings on the unmute very early and delays mute to much later after shutdown.
This also works at start up by delaying the full unmute until the current comes up to near the final level.
Whereas, using a higher unmute current say around 5mA brings on the unmute very early and delays mute to much later after shutdown.
A couple of Caps short, which I need to order but otherwise all the passives are in place. One thing to note for those of us who dont' keep up to date with labelling trends. R13-R15 are 1% normally or 0.1% if you intend to bridge/parallel. Tom's mouser BOM orders the 0.1% parts.
I normally expect high tolerance parts to be black. And 2 of them are, but R13 is banded (but with a purple tolerance band). First banded 0.1% I had ever seen which threw me completely.
The board itself is a joy to solder tho and the mouser delivery has the locations on each bag making it quick even when doing 4 at once.
I normally expect high tolerance parts to be black. And 2 of them are, but R13 is banded (but with a purple tolerance band). First banded 0.1% I had ever seen which threw me completely.
The board itself is a joy to solder tho and the mouser delivery has the locations on each bag making it quick even when doing 4 at once.
Actually, you need Imute > 500 uA to cover the worst case (according to the LM3886 spec sheet). I size the mute resistor to have Imute > 500 uA at the worst case supply voltage (low mains, high output power). I have no clicks or pops in the speakers on turn on/off.
~Tom
I looked up the datasheet.
I have not tried to measure whether the Imute actually required matches the datasheet numbers.
To my eyes 0.2mA (200uA) is on the flat part of the curve, i.e. full unmuting.
0.5mA (500uA) is at least double the minimum required for full unmuting.
It appears, again to me, that <0.14mA is into the muting region.
I would consider 10mA as > 500uA. That would activate and deactivate the mute/unmute function at very different start up/shut down voltages compared to using 0.2mA.
Using 1mA (>500uA) would give a different muting/umuting effect at start up/shut down.
Do you disagree?
I have not tried to measure whether the Imute actually required matches the datasheet numbers.
To my eyes 0.2mA (200uA) is on the flat part of the curve, i.e. full unmuting.
0.5mA (500uA) is at least double the minimum required for full unmuting.
It appears, again to me, that <0.14mA is into the muting region.
I would consider 10mA as > 500uA. That would activate and deactivate the mute/unmute function at very different start up/shut down voltages compared to using 0.2mA.
Using 1mA (>500uA) would give a different muting/umuting effect at start up/shut down.
Do you disagree?
Last edited:
Page 4 in the LM3886 data sheet (above the spec table):
Emphasis mine.
Page 17:
The equation above ignores the voltage drop across the 1 kΩ resistors in the path, by the way... See the equivalent circuit in the data sheet.
Sure. You can look at Fig. 44, 45 and claim that less than 500 uA could be used. I have three issues with this approach:
1) The graphs are intended to show the mute attenuation, i.e. the performance at Imute = 0.
2) Due to the log scale of the graphs (attenuation in dB), it is very hard to tell the attenuation at 200 uA < Imute < 500 uA.
3) The graphs show typical performance. It does not guarantee that all LM3886es that leave the fab meet this spec. Only the parameters listed in the spec table are contractually binding (including the test conditions).
From my experience, there is no benefit of designing for Imute < 500 uA, unless you want to activate the mute function. I seem to get fine results and thump-free startup and shutdown when Imute reaches 500 uA at 3/4 of the nominal supply voltage.
If you run Imute too low (say 200 uA < Imute < 500 uA), you risk mixing some of the voltage present on the GND pin of the LM3886 into the bias current for the input stage in the LM3886. I don't think that's such a hot idea...
As you correctly point out, if you design for Imute >> 500 uA, the startup and shutdown performance will be impacted. Exactly what the impact is and at which Imute it happens, depends on many variables. I would prefer the mute current to have fallen below 100 uA when the rail voltages hit the minimum supply voltage required for reliable operation (not necessarily operation within the specs, but not full collapse of the chip bias circuits).
The bottom line is that as long as you don't have thumps/clicks/pops at startup and shutdown, and the mute function does not activate during normal operation, life is good. Designing for Imute = 500 uA at 3/4 of the nominal supply voltage works for me...
~Tom
Emphasis mine.
Page 17:
The equation above ignores the voltage drop across the 1 kΩ resistors in the path, by the way... See the equivalent circuit in the data sheet.
Sure. You can look at Fig. 44, 45 and claim that less than 500 uA could be used. I have three issues with this approach:
1) The graphs are intended to show the mute attenuation, i.e. the performance at Imute = 0.
2) Due to the log scale of the graphs (attenuation in dB), it is very hard to tell the attenuation at 200 uA < Imute < 500 uA.
3) The graphs show typical performance. It does not guarantee that all LM3886es that leave the fab meet this spec. Only the parameters listed in the spec table are contractually binding (including the test conditions).
From my experience, there is no benefit of designing for Imute < 500 uA, unless you want to activate the mute function. I seem to get fine results and thump-free startup and shutdown when Imute reaches 500 uA at 3/4 of the nominal supply voltage.
If you run Imute too low (say 200 uA < Imute < 500 uA), you risk mixing some of the voltage present on the GND pin of the LM3886 into the bias current for the input stage in the LM3886. I don't think that's such a hot idea...
As you correctly point out, if you design for Imute >> 500 uA, the startup and shutdown performance will be impacted. Exactly what the impact is and at which Imute it happens, depends on many variables. I would prefer the mute current to have fallen below 100 uA when the rail voltages hit the minimum supply voltage required for reliable operation (not necessarily operation within the specs, but not full collapse of the chip bias circuits).
The bottom line is that as long as you don't have thumps/clicks/pops at startup and shutdown, and the mute function does not activate during normal operation, life is good. Designing for Imute = 500 uA at 3/4 of the nominal supply voltage works for me...
~Tom
Last edited:
I am also progressing and I should be ready to start the final check next week-end. It is true that the quality of Mouser packing and labelling is helping a lot!
Dominique
Dominique
I am close to do the final check. The manual says:
Connect the negative lead of a DC voltmeter to ground (J3, pin 2) and touch pin 4 of U4 with the positive lead of the voltmeter. The voltmeter should read -15 V (within a few percent). Touch pin 7 of U4 with the positive lead of the voltmeter. The meter should read +15 V (within a few percent
How do you identify pin 4 and 7 of U4?
Cheers,
Dominique
Connect the negative lead of a DC voltmeter to ground (J3, pin 2) and touch pin 4 of U4 with the positive lead of the voltmeter. The voltmeter should read -15 V (within a few percent). Touch pin 7 of U4 with the positive lead of the voltmeter. The meter should read +15 V (within a few percent
How do you identify pin 4 and 7 of U4?
Cheers,
Dominique
the datasheet shows the pin orientation relative to the manufacturer's ident.
This ident can be a dot depression, or a notch, or a printed label, or a metal tab, or....
LOOK at the datasheet.
This ident can be a dot depression, or a notch, or a printed label, or a metal tab, or....
LOOK at the datasheet.
Hi Dominique
How did you decide which way round to put the ICs in? If you have the 3886 away from you as per the pic http://www.neurochrome.com/audio/wp-content/uploads/2014/09/MOD86_R1p0_Assembled.jpg you can see the dots for pin 1 on all the ICs left hand side furthest away from you.
8 pins are numbered
1 8
2 7
3 6
4 5
Note it wont hurt if you try the meter on all pins until you find the voltages you are looking for.
How did you decide which way round to put the ICs in? If you have the 3886 away from you as per the pic http://www.neurochrome.com/audio/wp-content/uploads/2014/09/MOD86_R1p0_Assembled.jpg you can see the dots for pin 1 on all the ICs left hand side furthest away from you.
8 pins are numbered
1 8
2 7
3 6
4 5
Note it wont hurt if you try the meter on all pins until you find the voltages you are looking for.
Thank you!
I will do the check next week-end, and if everything OK, the first listening test...
Dominique
I will do the check next week-end, and if everything OK, the first listening test...
Dominique