Because I was doing a survey of the various solder types available so I could
1) satisfy my own curiosity
2) answer the question that was asked on the topic in an intelligent fashion.
It's how I roll. I need to buy some solder here pretty soon so I figured I'd educate myself on the topic a bit. I prefer to do that by surveying the market and reading about the different alloys and tradeoffs than by following Internet Gospel.
It's entirely possible that both the alloy (Multicore) and the flux (Kester) have to be formulated so they don't eat the tips of your soldering iron. It is also possible that Multicore and Kester solved the problem in different ways.
Tom
+1 on Kester 44 in 0.5mm- More tidy than large diameter and seems to wet the joint quickly which improves heat transfer efficiency. Super easy to use and get nice results regardless of chemical composition.
Tomchr,
Great post and follow-up. I'm ready to wire the Power & Mod86 boards and wonder if you tin the wires before securing them in the screw connections.
I wasn't planning to tin them because, its less work and, I thought crimp/compression terminations might be a little better as electrical conduits. However, I will be soldering to XLRs and binding posts so, your info is helpful.
Great post and follow-up. I'm ready to wire the Power & Mod86 boards and wonder if you tin the wires before securing them in the screw connections.
I wasn't planning to tin them because, its less work and, I thought crimp/compression terminations might be a little better as electrical conduits. However, I will be soldering to XLRs and binding posts so, your info is helpful.
No. You actually do not want to tin the wires before you put them into the screw terminals. The reason is that they tend to rattle loose over time if you tin the ends.
What happens is that the rising cage clamp of the screw terminals crushes the tin a little. That causes the wire to become thinner so over time it will start to develop an intermittent connection. I've occasionally tinned the tips of the wires that I use for bench testing to prevent them from fraying with repeated connect/disconnects. It's pretty common that I have to tighten the screw terminals 3-4 times over the course of, say, a 1-2 month prototyping cycle to maintain a good connection.
Tom
For speaker wire ends I solder on about a 1-inch length of 14 or 16 gauge solid copper wire (like in Romex in-wall wiring), put heat-shrink tubing over the solder joint, and use the copper wire (the end with no solder on it) to make the mechanical connection. It's soft enough to give a little when tightened down, thus giving more connection area, but tough enough to survive many connections and disconnections. It works well for both the clamp-type connectors on many speakers and the "five-way" banana connectors on the back of many amplifiers.
Of course, if you're going to use a "real" connector on the end of the cable (such as XLR, Speak-On, or the dreaded Banana or 1/4" phone), just solder or crimp the cable wire directly to it.
Of course, if you're going to use a "real" connector on the end of the cable (such as XLR, Speak-On, or the dreaded Banana or 1/4" phone), just solder or crimp the cable wire directly to it.
Funny you should mention it. I had that exact same thought just after I hit Submit on my previous post. What can I say? Great minds think alike...
Tom
Just thought of an optional connection to the binding posts. The solder section is half round 1/4" Dia. X 3/8" long with a hole in it. The hole could be tapped and the bare 12AWG wire secured with a copper screw.
On second thought a copper pop rivet would probably provide great compression and no tapping or screws. Either way, no solder and a very tight, compression, termination. Oh, the posts are Cardas bare OFHC copper. The only soldering will be for connecting input wires to the Neutrik XLRs.
Wait a minuet, I could drill and tap the XLR solder cups to 2-56 or just drill through and use a very small copper pop rivet for a compression termination. No solder anywhere. I'm likin this variation. Am I way of base with these variations?
Thanks,
henrylrjr
On second thought a copper pop rivet would probably provide great compression and no tapping or screws. Either way, no solder and a very tight, compression, termination. Oh, the posts are Cardas bare OFHC copper. The only soldering will be for connecting input wires to the Neutrik XLRs.
Wait a minuet, I could drill and tap the XLR solder cups to 2-56 or just drill through and use a very small copper pop rivet for a compression termination. No solder anywhere. I'm likin this variation. Am I way of base with these variations?
Thanks,
henrylrjr
Instead of binding posts, I recommend these. speakON / HPC Speaker Connectors in the Connectors & Adapters Department at Parts Express | 1510
I won't use anything else.
Mike
In my opinion, yes.
The terminal blocks were designed to have a wire stabbed into them and a screw turned. The XLRs were designed to have wires soldered to them. The respective manufacturers went great lengths to ensure that this could be done reliably. There's no reason to be afraid of solder.
Tom
Braindump post... it's been building up since last time I posted here.
I made active crossovers running inside Mod86 circuit by inserting a PLLXO filter between the THAT1200 and the LME49710. So I have a 2 way active crossover without all the fuss of building a separate active XO board. All analog, no computers, and infinite resolution, low price, active xo performance. I'd like a 3 way, but I think I can't do bandpass filter passively. But adding another buffer is not a problem. +/-15V is already available. The space between THAT1200 and LME49710 is fertile ground for filters, volume control, etc.
I used Rod Elliott's sallen key calculator to find the R1, C1 parts values, then built the filters according to PLLXO article on t-linespeaker.org. Dale CMF55 and Vishay 1813s are small, inexpensive and sound very good indeed. The passive filter parts just stick up in the air for the "true DIY" look.
I lifted the appropriate pins of THAT1200 and LME49710 from their sockets and soldered the input of the filter to THAT1200 output, and the filter output to 49710 input pin, with the ground leg of the filter going to the mod86 signal input ground hole on PCB. I connected the amps' signal inputs in parallel to one XLR jack, this makes 24kOhm balanced input impedance.
I built my 2 pairs of mod86s with different gain for this purpose. The stock 20dB amp pair powers the 91.5dB sensitivity ported woofers, while 26dB gain mod86s power the 85.5dB sensitivity bookshelf speakers with their own passive speaker level crossovers. I got lucky that the claimed sensitivity specs were fairly accurate and they blend together pretty well on first listen. I can change the feedback resistors on Mod86 to adjust gain if necessary.
I was surprised how easily I could hear the time alignment adjustment of the bookshelf speaker on top of the woofer box. 1/4" adjustment was easily audible how it affected the bass tone. Tenor sax and everything sounded thin and weird at first. Moving the main speakers 3/4 inch forward made it sound too fat, back 1/4" perfect, 2 minutes, no measurement. That 1/4" distance is a change of only 1.5 degrees phase at 200Hz, but it was easily audible in the music content, not needing to use test tones showing me computer EQ charts. Great drivers, great amplifiers and active crossovers working together to give me extreme bass resolution.
One piece of advice, the TLS.org article mentions you can sometimes use the amplifier resistance as the R2 in the 2nd order hipass filter. Not when the amplifier is LME49710. It's impedance is 250MOhms so it doesn't work. Without a real R2 resistor in the filter, LME49710 outputs DC and 3886 obeys the opamp, putting 29VDC on the speaker terminal. Use a real resistor for R2.
Going back in the thread to suggestions for balanced wire, I wanted to put in a recco for Gotham Cable 10201 GAC2 install cable. It is only one pair, not quad, but it has a solid foil shield, with only 7 strands per conductor, less is more, imo. I know it's taboo to describe wire sonics on this forum, but it sounds very good, compared to other well reviewed quad mic cables I've used, like ProCo. Best part, it's only .20/ft cut to length from redco...
My engineer/amp designer friend tested mod86 this past weekend. He's been playing with 3886 for years, is well aware of their idiosyncracies. He saw the specs and heard it play in my system, had been eager to get it on his scope to see how well Tom had tamed the beast. He saw the 3886 quirks, showed me the ugliness still there when pushed to clipping but well compensated. He was very impressed, he said it is one of the best performing amps that has ever come across his bench.
I still find Mod86 addictive and fun to listen. I go back to my big comfy tube amp to catch my breath sometimes, mod86 is pretty intense. But when I come back it is always such a thrill. But it's not gear thrill, like some magic tube that colors all sounds with exquisite beauty cream, or a subwoofer that beats the crap out of you. Mod86 is dry and uncolored. It's the sound of the instruments in all their glory that is so exciting, and the extreme microdynamics that come through to make the music feel alive. If you love the true sound of the instruments, not necessarily the sound of beautiful electronic distortions, then nothing in my experience touches this. Well, nothing in DIY price range anyway.
Tom, I'm glad you've redesigned the Mod86 to keep it alive despite the TI catalog changes.
Rich
I made active crossovers running inside Mod86 circuit by inserting a PLLXO filter between the THAT1200 and the LME49710. So I have a 2 way active crossover without all the fuss of building a separate active XO board. All analog, no computers, and infinite resolution, low price, active xo performance.
I'd like a 3 way, but I think I can't do bandpass filter passively. But adding another buffer is not a problem. +/-15V is already available. The space between THAT1200 and LME49710 is fertile ground for filters, volume control, etc. I used Rod Elliott's sallen key calculator to find the R1, C1 parts values, then built the filters according to PLLXO article on t-linespeaker.org. Dale CMF55 and Vishay 1813s are small, inexpensive and sound very good indeed. The passive filter parts just stick up in the air for the "true DIY" look.
I lifted the appropriate pins of THAT1200 and LME49710 from their sockets and soldered the input of the filter to THAT1200 output, and the filter output to 49710 input pin, with the ground leg of the filter going to the mod86 signal input ground hole on PCB. I connected the amps' signal inputs in parallel to one XLR jack, this makes 24kOhm balanced input impedance.
I built my 2 pairs of mod86s with different gain for this purpose. The stock 20dB amp pair powers the 91.5dB sensitivity ported woofers, while 26dB gain mod86s power the 85.5dB sensitivity bookshelf speakers with their own passive speaker level crossovers. I got lucky that the claimed sensitivity specs were fairly accurate and they blend together pretty well on first listen. I can change the feedback resistors on Mod86 to adjust gain if necessary.
I was surprised how easily I could hear the time alignment adjustment of the bookshelf speaker on top of the woofer box. 1/4" adjustment was easily audible how it affected the bass tone. Tenor sax and everything sounded thin and weird at first. Moving the main speakers 3/4 inch forward made it sound too fat, back 1/4" perfect, 2 minutes, no measurement. That 1/4" distance is a change of only 1.5 degrees phase at 200Hz, but it was easily audible in the music content, not needing to use test tones showing me computer EQ charts. Great drivers, great amplifiers and active crossovers working together to give me extreme bass resolution.
One piece of advice, the TLS.org article mentions you can sometimes use the amplifier resistance as the R2 in the 2nd order hipass filter. Not when the amplifier is LME49710. It's impedance is 250MOhms so it doesn't work. Without a real R2 resistor in the filter, LME49710 outputs DC and 3886 obeys the opamp, putting 29VDC on the speaker terminal. Use a real resistor for R2.
Going back in the thread to suggestions for balanced wire, I wanted to put in a recco for Gotham Cable 10201 GAC2 install cable. It is only one pair, not quad, but it has a solid foil shield, with only 7 strands per conductor, less is more, imo. I know it's taboo to describe wire sonics on this forum, but it sounds very good, compared to other well reviewed quad mic cables I've used, like ProCo. Best part, it's only .20/ft cut to length from redco...
My engineer/amp designer friend tested mod86 this past weekend. He's been playing with 3886 for years, is well aware of their idiosyncracies. He saw the specs and heard it play in my system, had been eager to get it on his scope to see how well Tom had tamed the beast. He saw the 3886 quirks, showed me the ugliness still there when pushed to clipping but well compensated. He was very impressed, he said it is one of the best performing amps that has ever come across his bench.
I still find Mod86 addictive and fun to listen. I go back to my big comfy tube amp to catch my breath sometimes, mod86 is pretty intense. But when I come back it is always such a thrill. But it's not gear thrill, like some magic tube that colors all sounds with exquisite beauty cream, or a subwoofer that beats the crap out of you. Mod86 is dry and uncolored. It's the sound of the instruments in all their glory that is so exciting, and the extreme microdynamics that come through to make the music feel alive. If you love the true sound of the instruments, not necessarily the sound of beautiful electronic distortions, then nothing in my experience touches this. Well, nothing in DIY price range anyway.
Tom, I'm glad you've redesigned the Mod86 to keep it alive despite the TI catalog changes.
Rich
Just a comment on solder. In the past couple years, based on work-related activities, I've become quite the fan of water-soluble flux. Rosin core flux leaves a messy PCB, even if innocuous for most applications, but requires lots of alcohol or flux cleaner to clean up. Also, a less-than pristine clean job is dangerous as the stuff last removed is the most aggressive part. Low-residue fluxes (aka 'no-clean') fluxes are OK but they really are "can't-clean" -- they don't come off without special flux remover designed for them. Water-soluble flux is removed with warm tap water, and then the board rinsed in distilled water (or DI water). Running under the faucet, I just use a toothbrush to scrub. I make sure to do everything but pots and switches, which I add afterwards with rosin-core or no-clean solder. The only downside to the water-soluble flux is that you must clean it. If I'm working on something over a weekend and am not done, I'll clean it as after a week before cleaning you can discern some less-than-shiny finish.
Also I find that eutectic solder assures fewer cold joints; I gave up on 60/40 years ago.
Also I find that eutectic solder assures fewer cold joints; I gave up on 60/40 years ago.
I use ferrules around the stranded wire.
I have packs of 0.5, 0.75, 1, 1,5 and 2.5mm² (CPC stocks them).
The ferrule should be the smallest that just fits over the stranded wire. When it is crushed down, the internal area becomes smaller than the copper area of the stranded and they too suffer deformation to create gas tight joints.
I have packs of 0.5, 0.75, 1, 1,5 and 2.5mm² (CPC stocks them).
The ferrule should be the smallest that just fits over the stranded wire. When it is crushed down, the internal area becomes smaller than the copper area of the stranded and they too suffer deformation to create gas tight joints.
And can be confused with 40/60.
Whereas 63/37 cannot be confused with 37/63 which is not available.
B..
My engineer/amp designer friend tested mod86 this past weekend. He's been playing with 3886 for years, is well aware of their idiosyncracies. He saw the specs and heard it play in my system, had been eager to get it on his scope to see how well Tom had tamed the beast. He saw the 3886 quirks, showed me the ugliness still there when pushed to clipping but well compensated. He was very impressed, he said it is one of the best performing amps that has ever come across his bench.
If he has tested a better one we would be interested to know what it is. Most I have seen only get within about 10dB, albeit still below any sane threshold of audibility.
The clipping point is an interested one, and why I may have to drop the mod on my bass units when I finally upgrade them*. In current form I have enough head room and gain setup such that I cannot clip them. but clipping recovery is something rarely discussed, but considered by many to be core to why some amplifiers have a 'sound'
* OT but plan is 4 6.5" drivers on the 12" wide baffle. If I put in a linkwitz transform to extend the F3 to 80Hz would need about 18dB more power at 90Hz. So 4 small amps or one big one. Funds did not allow stock piling of Par-86 boards