Awesome, thank you, Tom. I think I’ll probably order a pair of LM3886TFs from Mouser. Even though they’re out of stock they at least have an expected date (that lines of with some of the other components I ordered) which seems better to me!
@BrianL oof, I totally didn’t make the connection. Just because I’m off the road system doesn't mean I’m completely disconnected from the outside world.
@BrianL oof, I totally didn’t make the connection. Just because I’m off the road system doesn't mean I’m completely disconnected from the outside world.
It's not that I'm afraid; it's a combination of not knowing what tools I need and how to use them properly; my poor measurement technique; and my inherent cheapness that prevented me from investing in needed metalworking tools. Just as I was going to start buying stuff, I found out that Enco is no longer around. They had good prices for high quality tools. An earlier blow was Amazon buying and shutting down Small Parts Inc.; they had fine tools and very interesting materials.
Yeah, Bill, the electronics industry, and especially the semiconductor portion is a 'cyclical' business, but the current situation is unprecedented. That said, it doesn't necessarily mean that there is a run on LM3886 chips. More likely TI is devoting manufacturing capacity to whatever are the hot demands of the moment -- obviously automotive demand is up. According to the news media, game consoles, computer monitors, etc., are also in high demand.
It's not just the electronics sector that's hit. We're probably all familiar with the sky high prices on lumber. Aluminum is at an all-time high, so the various chassis vendors have had to raise their prices. And ... get this ... on my walk yesterday I bumped into an acquaintance who works in the hotel industry. She reports there's a shortage of cotton, so the hotels can't get sheets. Mind blown.
Tom
Tom
Online Metals is a good place for materials. When I lived in Seattle, I'd just drive to pick up the goods. I miss having that resource so readily available. A metal recycler will often have good materials too. It's obviously more hit/miss, but I've fished quite a few chassis front plates (i.e., aluminum flat bar or thick sheet) out of a scrap bin. One time I got lucky and found an entire tub skid full of heat sinks! I still have some of them.
As for the tools, you don't actually need much. If you already have a corded power drill, I suggest looking into getting a drill stand for it. If you have a mitre saw, getting a metal cutting blade would get you far. A vice can be convenient. If you don't want to permanently mount it, you can do what I did and just put it on a 2x4 that you can then clamp to the work table when you need the vice. Taps, tap handles, tapping blocks are relatively cheap.
As for how to use the tools, perhaps I should put together a video on that. I'm planning to get into video. That seems to be "the thing" these days.
Tom
Last edited:
Hello,
makes it sense to try another OPA as 49720? Maybe a Burson or SparkosLabs. Of course it is expensive, but......
Regards
Ralf
makes it sense to try another OPA as 49720? Maybe a Burson or SparkosLabs. Of course it is expensive, but......
Regards
Ralf
Expensive. Worse performance. Would make the Modulus-86 oscillate every time. Yeah. Not a fan.
I've measured a few discrete opamps as I have an OEM client who likes them. To date I've had exactly one discrete opamp that measured better than an LME49720/LM4562. I think it was the Sparkos. It was slightly lower noise. I don't think it beat the OPA1612, though, and the OPA1612 - while still expensive - is less than 1/10th the cost of the Sparkos.
If you do decide to play with discrete opamps, beware that some of them are marketed as "unity gain stable" even though they exhibit a -40 dB/dec rolloff throughout most of their open loop frequency response. This means these opamps are only stable at unity gain. If you deviate much from unity gain, you will need to add external frequency compensation. This could be tricky in filters (including RIAA front-ends).
There are some (or at least one) that is not that way. I forget which one, but you can check the AVOL curve in the various data sheets. That one did have a -20 dB/dec rolloff with a pole at the unity gain frequency - just as most unity-gain stable opamps do. BTW: If the manufacturer does not give you the AVOL curve, run.
There are some opamps that are marketed as unity-gain stable, have a -20 dB/dec rolloff, and have an internal compensation that gives them a phase wobble around the unity gain frequency. The OPA1612 is like that. That's why the OPA1612 can't be used in the Modulus-86 either. Not without some severe hacking anyway. I think I played with that once and wasn't able to get it stable.
Tom
I've measured a few discrete opamps as I have an OEM client who likes them. To date I've had exactly one discrete opamp that measured better than an LME49720/LM4562. I think it was the Sparkos. It was slightly lower noise. I don't think it beat the OPA1612, though, and the OPA1612 - while still expensive - is less than 1/10th the cost of the Sparkos.
If you do decide to play with discrete opamps, beware that some of them are marketed as "unity gain stable" even though they exhibit a -40 dB/dec rolloff throughout most of their open loop frequency response. This means these opamps are only stable at unity gain. If you deviate much from unity gain, you will need to add external frequency compensation. This could be tricky in filters (including RIAA front-ends).
There are some (or at least one) that is not that way. I forget which one, but you can check the AVOL curve in the various data sheets. That one did have a -20 dB/dec rolloff with a pole at the unity gain frequency - just as most unity-gain stable opamps do. BTW: If the manufacturer does not give you the AVOL curve, run.
There are some opamps that are marketed as unity-gain stable, have a -20 dB/dec rolloff, and have an internal compensation that gives them a phase wobble around the unity gain frequency. The OPA1612 is like that. That's why the OPA1612 can't be used in the Modulus-86 either. Not without some severe hacking anyway. I think I played with that once and wasn't able to get it stable.
Tom
Last edited:
That is a super clean build. Very nicely done. I'm looking forward to hearing your impressions.
Tom
Tom
Hello Tom,
thanks a lot for explaining.
Do you know if it was the smaller Sparkos SS3602 or the bigger SS2590?
By price I would say the smaller because you mentioned 1/10th.
Ralf
PS Very nice 5 Channel amp.
thanks a lot for explaining.
Do you know if it was the smaller Sparkos SS3602 or the bigger SS2590?
By price I would say the smaller because you mentioned 1/10th.
Ralf
PS Very nice 5 Channel amp.
Is there anything at all on the Mod86 that can be improved, maybe better regulators or so, that can improve the sound? Not having an eye on costs.
Ralf
Ralf
I really don't remember which Sparkos it was. The board was maybe 25x25 mm if that tells you anything. Discrete opamps are so far down on my Don't Care List that I don't bother remember which is which.
I highly doubt fancier regulators would improve anything. I'm not able to measure any negative effect of the ones I currently use.
You could lower the noise floor measurably by getting rid of the THAT1200. I've done that in the Modulus-186, -286, and -686.
Tom
I highly doubt fancier regulators would improve anything. I'm not able to measure any negative effect of the ones I currently use.
You could lower the noise floor measurably by getting rid of the THAT1200. I've done that in the Modulus-186, -286, and -686.
Tom
How? Connecting the signal source directly to the lme49710( I have an older version). Any coupling cap needed, and if so , what would be the value of that cap?
Thanks
I personally don’t need the balanced input, this is my computer amp , pcie Creative sound card. Is a capacitor needed ?
In that case you can pluck the THAT1200 and connect from pin 3 to pin 6 on the (now) empty socket. I do recommend taking that connection to ground through something like 47 kΩ to ensure that the LME49710 has a stable DC operating point even if the source is unplugged.
Tom
Tom