OK the test went exceptionally well! Very unspectacular, no smoke, no bangs.
I get the correct voltage on both boards on pin 4 and pin 8 of U4 plus minus a max of 0.54V.
For the heatsink I hope it will be fine for the moment, I don't plan to run the amp on high power for longer than a few seconds for testing. Apart from this it will mainly run on rather low to medium levels. One of the build reports in this thread mentioned a similar method and in this case the temperatures are fine.
However, if there are any issues I'll only have to replace this one panel with a heat sink or attach a proper heat sink to the aluminium side panel.
Oh and by the way, is there a reason why you don't offer an unpopulated board for the universial buffer? That doesn't look so hard to solder or am I missing something?
Cheers
Stefan
Very nice.
A few years ago I wrote quite a bit on why I no longer support DIY SMD builds. You can find my thoughts in Post #4579.
I would like to add to the thoughts in that post: SMD components drift in/out of stock much, much faster than leaded parts. That was the case even before the current supply crunch. So in addition to the issues mentioned in Post #4579, supporting DIY SMD builds also means I get to spend most of my days helping people find substitute parts. Now let me be clear. I do enjoy helping people. But I would much, much rather help them choose which of my products to buy than help them find a substitute capacitor that was in stock yesterday, but now isn't. There was a lot of that in the couple of DIY SMD projects I offered before I came to the conclusion that I just couldn't support these builds any longer. There's only one of me and only 24 days in my day.
Tom
Hello Tom,
although I'd really like to try this build, I can't blame you for this decision. With a bit of practice smd builds with 0805 or larger are easier for me because of the better handling of the soldering iron for example. But from my first attempts I know that it is a different cup of tea for a beginner. I have destroyed quite a few boards and components before I had my first working atmega boards. Flux, solder, solder tip, tip cleaning method and temperature had to be adjusted.
I hope diy audio won't be reduced to case building and sticking components together eventually. From your comments i think you may be worried some important parts won't be available forever and unpopulated smd boards are not an option for the reasons you posted. What a dilemma!
Regards
Stefan
although I'd really like to try this build, I can't blame you for this decision. With a bit of practice smd builds with 0805 or larger are easier for me because of the better handling of the soldering iron for example. But from my first attempts I know that it is a different cup of tea for a beginner. I have destroyed quite a few boards and components before I had my first working atmega boards. Flux, solder, solder tip, tip cleaning method and temperature had to be adjusted.
I hope diy audio won't be reduced to case building and sticking components together eventually. From your comments i think you may be worried some important parts won't be available forever and unpopulated smd boards are not an option for the reasons you posted. What a dilemma!
Regards
Stefan
Somewhat ironically it was actually the experienced builders who ran into trouble with the MOD286 Rev. 1.1. Overconfidence?
Meanwhile I think only one person ran into trouble with the original HP-1 headphone amp even though it contained 500+ SMD parts. Go figure. Were these builders more careful having just spent $600 on boards and parts? Were they a different demographic (maybe younger with better eyesight and steadier hands)?
Either way. I would hate supporting those builds now. I'd do nothing all day but respond to email about obsolete parts. I will always support my customers, but I'd rather design circuits than source parts.
Tom
Meanwhile I think only one person ran into trouble with the original HP-1 headphone amp even though it contained 500+ SMD parts. Go figure. Were these builders more careful having just spent $600 on boards and parts? Were they a different demographic (maybe younger with better eyesight and steadier hands)?
Either way. I would hate supporting those builds now. I'd do nothing all day but respond to email about obsolete parts. I will always support my customers, but I'd rather design circuits than source parts.
Tom
I get the sense that I responded to this last night via email, but I'll respond here for completeness.
For the two 1000 uF caps, any 1000 uF (or higher), 50 V (or higher) electrolytic cap with a diameter of 16 mm (or lower) and a pin spacing of 7.5 mm will work. In fact, to find a substitute I type those criteria into Mouser's search tool, check the "in stock" box, and sort the results by price.
That's how I found these: UPW1H152MHD (1500 uF, 50 V) and these: UBY1H102MHL1TN (1000 uF, 50 V, 135ºC (!))
Substitutes can often be found by searching for less "popular" values. Everybody loves 10, 100, 1000, but 90% of the time 11, 110, and 1100 will work just as well. I ordered 110 Ω resistors yesterday for example. 100 Ω was out of stock for the SMD resistor array I wanted. The resistors will be series resistors for LEDs, so if the difference in resistance is inconsequential.
Tom
For the two 1000 uF caps, any 1000 uF (or higher), 50 V (or higher) electrolytic cap with a diameter of 16 mm (or lower) and a pin spacing of 7.5 mm will work. In fact, to find a substitute I type those criteria into Mouser's search tool, check the "in stock" box, and sort the results by price.
That's how I found these: UPW1H152MHD (1500 uF, 50 V) and these: UBY1H102MHL1TN (1000 uF, 50 V, 135ºC (!))
Substitutes can often be found by searching for less "popular" values. Everybody loves 10, 100, 1000, but 90% of the time 11, 110, and 1100 will work just as well. I ordered 110 Ω resistors yesterday for example. 100 Ω was out of stock for the SMD resistor array I wanted. The resistors will be series resistors for LEDs, so if the difference in resistance is inconsequential.
Tom
Last edited:
I found myself mulling over this. I actually don't find it to be a dilemma. At least not a dilemma on my end.
I'm also not worried about any important parts being discontinued. Of course that's always a risk but I don't see any indication that it'll happen any time soon. I'm sure TI and others will keep selling popular products as long as they remain popular. Look at the LM317, for example. It's been around since 1976 (AFAIR) and just refuses to die.
That said, it's pretty clear (and has been for quite a while) that leaded parts are slowly going away. All the interesting new opamps, voltage regulators, MOSFETs, etc. come exclusively in SMD packages. So to further push the performance envelope I have to use surface mounted components.
Since I cannot support all-DIY SMD builds for the reasons given, this means that my all-DIY circuits will offer the best performance that I can wring out of the leaded parts. It's a good thing that the LM4562/LME49720 comes in a DIP package. Granted, the OPA1612 is better (it has a smidge lower distortion and noise), but the OPA1612 doesn't work well in the Modulus-86 (it causes the amp to be unstable with certain loads), so the LM4562/LME49720 is still the best opamp for the Modulus-86. That's how I can keep offering the Modulus-86 as an all-DIY project.
The DIP version of the LM4562/LME49720 survived TI's culling of overlap with the National Semiconductor chips, so I would expect that IC in that package option to stick around for a while.
Some might say that having a few SMD components on a board is OK, especially if the devices are large. Take the MOSFETs in the Guardian-86/686 for example. The trouble with mixing SMD and leaded (aka PTH/pin-through-hole) parts is that some will immediately shy away from the project because it contains SMD. And let's face it; soldering the tab on the MOSFETs on the Guardian boards does push most soldering irons to (or past) their limit. Also, SMD MOSFETs go in/out of stock so fast (especially these days) that the substitute I recommended yesterday is out of stock today.
With these SMD+PTH circuits I also get requests for boards with the SMD soldered, which is fine as long as the production volume is low, but it doesn't scale well. Last night I spent 5-6 hours soldering up Guardian boards and my neck and back are cranky today as result. #gettingOldSucks.
Getting boards with many PTH components assembled by my assembly house is possible but very expensive. All PTH parts are stuffed manually. My assembly guys did look at buying a machine for it, but found it wasn't up to snuff. If I want the boards assembled at a price that's low enough for me to sell the boards and make a profit, I have to go with as much SMD as possible. That's the only solution that I'm aware of that allows for stellar performance and that scales well.
Another option would be to have an all-PTH version and an all-SMD version of the same circuit. That's sorta similar to the Modulus-86 vs Modulus-186. In reality, that's probably only realistic for the Modulus-86/186 because I can get SOTA performance either way. It would not be a solution for the Guardian, for example, as I can't get PTH MOSFETs with the same low RDS(on) as the SMD ones I use. Also, maintaining two versions of the same circuit means twice the work for less than twice the benefit, so that doesn't seem to be a wise use of my time from a business standpoint.
Anyway. Those are my thoughts this frosty and sunny Sunday morning.
Tom
Thank you for offering your business perspective
I am always very interested in the decision-making analysis of a private company. The need to survive and make a profit is wonderful for sharpening Occam's Razor. Thank you so much for sharing your thoughts on the manufacturing dilemma you face through your many posts over the course of years.
I am always very interested in the decision-making analysis of a private company. The need to survive and make a profit is wonderful for sharpening Occam's Razor. Thank you so much for sharing your thoughts on the manufacturing dilemma you face through your many posts over the course of years.
Speaking of the Guardian mosfet, it is out of stock until the end of the year.
Using solder paste was an experience on it. Actually it was kind of fun. I also did my my Norwood cards for the M2X. Getting the solder paste on the opamp pads was challenging. I had a few solder bridges to fix between pins with the soldering iron.
Using solder paste was an experience on it. Actually it was kind of fun. I also did my my Norwood cards for the M2X. Getting the solder paste on the opamp pads was challenging. I had a few solder bridges to fix between pins with the soldering iron.
You're quite welcome. I find that typing out my thoughts helps me see things at a different level of clarity as well.
Speaking of the Guardian mosfet, it is out of stock until the end of the year.
That's not bad. I've seen a few lead times in early 2023 now. The 500 micro controllers I ordered from Mouser in April originally had a lead time in October (this year). That became November. Then March 2022. Then "will advise"... This for a part that used to be in stock by the thousands.
If you apply the solder paste with a stencil and use a reflow oven, the MOSFETs are easy to solder. I bought a relatively cheap Black & Decker toaster oven and a thermocouple thermometer for such things. I've soldered QFN packages that way as well. Those are fun....
For the Guardian MOSFETs I just use a 6.3 mm wide chisel tip on a METCAL MX500. That works every time.
Tom
The mosfet was my first try with a SMD. I didn't use a stencil. I just put paste on all the solder pads. I used a bit too much paste. The mosfet floated around a bit. I used a pan on a hot plate instead of a toaster oven. I was concerned I had a solder bridge, so I checked the mosfet for continuity between the pins after I installed it.
Yeah. Hot plate reflow works too. I'd be a bit concerned about damaging the solder mask with that approach, though. I suppose if you keep a close eye on the temperature you'd be OK.
If I didn't have the toaster oven setup, I'd use a heat gun. Again, just be careful with the temperature.
Tom
If I didn't have the toaster oven setup, I'd use a heat gun. Again, just be careful with the temperature.
Tom
Hello Tom,
thanks for your thoughts, I think I looked at the whole thing from the wrong angle without considering the production and business side.
I've been trying out hot air and solder paste but went back to hand soldering pretty soon. Paste without stencils is just not for me. The solder paste I had was behaving very funny and didn't melt the same way as my usual solder with lead. I resisted the impuls to build a reflow oven from a toaster grill.
On the build I'm in the final stages, Just have to finish the wiring, test the power supply, drill a few holes for the connectors and then put some paint on.
Stefan
thanks for your thoughts, I think I looked at the whole thing from the wrong angle without considering the production and business side.
I've been trying out hot air and solder paste but went back to hand soldering pretty soon. Paste without stencils is just not for me. The solder paste I had was behaving very funny and didn't melt the same way as my usual solder with lead. I resisted the impuls to build a reflow oven from a toaster grill.
On the build I'm in the final stages, Just have to finish the wiring, test the power supply, drill a few holes for the connectors and then put some paint on.
Stefan
What is the temperature limit for the LM3886? The working temp range goes up to 70, but where does it get dangerous?
I'm asking because instead of the ISS I'm using a cheap soft start module for now. It came with 2 temperature probes which are labelled 75 degrees C. My first impulse would be not to use them because by the time they get to 75 C the LM3886 must be quite a bit warmer unless I could mount the probe on top of the3886 somehow.
Life goes on the first wire of the primary side of the transformer, neutral on the 4th with 2 and 3 bridged for 230V, right? Apologies for repeating this, but how critical is this, considering that a (symmetrical) German plug can be plugged in 2 ways with life and neutral exchanged? I was reading up on PAT testing and a requirement seems to be that life and neutral are in the correct order which isn't a problem with a UK plug because it only fits into the socket with life and neutral in fixed positions. We have a PAT tester at work so I'll get our guy to test it for me. Really sorry for the basic questions.....
Regards
Stefan
I'm asking because instead of the ISS I'm using a cheap soft start module for now. It came with 2 temperature probes which are labelled 75 degrees C. My first impulse would be not to use them because by the time they get to 75 C the LM3886 must be quite a bit warmer unless I could mount the probe on top of the3886 somehow.
Life goes on the first wire of the primary side of the transformer, neutral on the 4th with 2 and 3 bridged for 230V, right? Apologies for repeating this, but how critical is this, considering that a (symmetrical) German plug can be plugged in 2 ways with life and neutral exchanged? I was reading up on PAT testing and a requirement seems to be that life and neutral are in the correct order which isn't a problem with a UK plug because it only fits into the socket with life and neutral in fixed positions. We have a PAT tester at work so I'll get our guy to test it for me. Really sorry for the basic questions.....
Regards
Stefan
No problem. It took me a few years to make that adjustment as well.
Not for me either. Paste + heat gun (or better yet an actual hot air tool) makes more sense. But you can get similar results with normal wire solder and flux + heat gun. So for me the paste really only makes sense when I'm 'baking' the boards.
If you're comparing leaded solder like 60/40, 63/37, or 62/36/2 to lead-free solder, you're probably in for a surprise. The cheapest lead-free solder paste is usually SAC305 (Sn/Ag/Cu @ 96.5/3.0/0.5 %) which is pretty awful to work with. It doesn't flow well, doesn't wet easily, and has a large plastic region. It results in grainy solder joints which look exactly like a cold solder joint made with 60/40 (Sn/Pb).
If you're going lead-free, I recommend 99.3/0.7 (Sn/Cu) or the germanium/bismuth doped version. I've been using 99.3/0.7 (Sn/Cu, sans germanium) for the past few days and it behaves a lot like 62/36/2 (Sn/Pb/Ag) leaded solder, albeit at a melting point of 227 ºC versus 183 ºC for the leaded solder. Like 63/37 (Sn/Pb), 99.3/0.7 (Sn/Cu) is an eutectic alloy, i.e., it has no plastic region.
I have yet to try the version with germanium (or bismuth) doping. The Ge/Bi is supposed to make the solder flow a bit better and avoid solder bridging.
For more information, have a look at my article on choosing solder: Choosing Solder - Navigate the Different Alloys, Types of Flux, etc. – Neurochrome
150 ºC die temperature if I recall correctly. It says in the Absolute Maximum table of the data sheet, which is usually the first table in the data sheet.
Tsk, tsk, tsk...
That's pretty hot for an external heat sink. I'd stay below 65 ºC.
I'm not quite sure what you're referring to there. If your transformer has two 115 or 120 VAC windings for the primary, you need to connect them in series and connect one end to live and the other to neutral. And you're right. You can swap live and neutral.
Tom
Tsk, tsk, tsk...
I know, I really feel bad about spoiling a Neurochrome amp with a china board. But I've done the first step now and admitted my bad decision making in public!
Yes, that! I referred to the connection of lead #2 and #3 on the primary side of the transformer to get both primary windings in series for 230V mode.
Still a bit fuzzy about the ability to swap life and neutral. I always thought you can just swap the 2 but with what I was reading about PAT testing, the power-86 and some other schematics and "stuff" I wasn't so sure anymore. So thanks for the clarification about this particular case. Last time I asked pretty much the same question you mentioned that it can be necessary to have life and neutral on particular pins for safety reasons. I know what I need to finish the amp now but I've already looked up some courses in our local community college. It's driving me nuts to get held up by these holes in the basic knowledge.
Thanks
Stefan