I have one Rev. 1.01 boards left in my Modulus project box. I must have sold an odd number of them and ended up with a spare. Toss me an email if you're interested in buying.
I'm down to seven (7) of the Rev. 2.2 boards. Those who are sitting on a stash of LME49710s they'd like to use should buy now as Rev. 2.3 will use the LME49720. Aside from that, Rev. 2.3 will not feature any significant changes.
Tom
I'm down to seven (7) of the Rev. 2.2 boards. Those who are sitting on a stash of LME49710s they'd like to use should buy now as Rev. 2.3 will use the LME49720. Aside from that, Rev. 2.3 will not feature any significant changes.
Tom
Last edited:
What would happen if V+ and V- from Power86 were accidentally reversed on the J connection on Mod86? Would it blow the LM3886TF?
Today I was finishing another Mod86 Amp, which I started last year.
I clicked the power switch, heard a snap, saw a spark, and a puff of smoke. Powered off right away, but the smoke.... oh dang. The spark was in the area of the lm3886. I then noticed the lm3886 plastic seems to be cracked near the legs.
I'm unsure if I had stressed the legs when mounting the heatsink, and this caused a short, or if I might have had V+ and V- backward. Didn't think about it until after it was removed for inspection, but I think those wires from Power 86 may have been wrong (I should have turned the amp around on the workbench before hooking them up!)
In the instructions I'm on the step just before populating the Ux Sockets with LMExxxx and THATxxxx (so at least those weren't damaged)
In current state, doing the test for +\- 17v reads 0. (Or maybe 1 Volt,.. can't remember)
Sound like I blew up LM3886 ? I'm wondering if I damaged anything else.
I've double checked for solder bridges, especially around diodes and other high-pin-density areas. Looks okay.
I've ordered another LM3886TF/NOPB. Hoping that replacing it will take care of it.
Today I was finishing another Mod86 Amp, which I started last year.
I clicked the power switch, heard a snap, saw a spark, and a puff of smoke. Powered off right away, but the smoke.... oh dang. The spark was in the area of the lm3886. I then noticed the lm3886 plastic seems to be cracked near the legs.
I'm unsure if I had stressed the legs when mounting the heatsink, and this caused a short, or if I might have had V+ and V- backward. Didn't think about it until after it was removed for inspection, but I think those wires from Power 86 may have been wrong (I should have turned the amp around on the workbench before hooking them up!)
In the instructions I'm on the step just before populating the Ux Sockets with LMExxxx and THATxxxx (so at least those weren't damaged)
In current state, doing the test for +\- 17v reads 0. (Or maybe 1 Volt,.. can't remember)
Sound like I blew up LM3886 ? I'm wondering if I damaged anything else.
I've double checked for solder bridges, especially around diodes and other high-pin-density areas. Looks okay.
I've ordered another LM3886TF/NOPB. Hoping that replacing it will take care of it.
As a start I'd remove the dead LM3886. You might not have done the electrolytic capacitors any favours either. I don't have my boards with me at the moment so Tom would need to advise which are purely in the audio path and wouldn't have seen any reverse voltage, but you'd probably do well to replace any that saw reverse voltage. Once the dead LM3886 is removed and the capacitors replaced, I'd run the voltage test again and check the voltage regulators are still giving an acceptable output.
Thanks Adrian, I'm feeling like a doh-doh for making such a silly error, so I appreciate your feedback.
I think I'm on the right path to getting it corrected, and hopefully the new lm3886 will be here before the weekend.
Are the electrolytic caps you mention non-polar? They are inexpensive too and easy enough to replace, just tedious to de-solder. I'm thinking it's likely they survived if they don't really have a positive/negative orientation. I know the can-type caps with the stripe on the side did not blow their tops which surprised me.
Will report back after new lm3886 is in. If I can measure +/- 17 at that point, then I'll assume the voltage regulators are still working, and proceed with assembly.
Well. That blows. In more ways that one... 
Sorry to hear.
With sparks and smoke resulting in a cracked LM3886, I concur with your diagnosis that the LM3886 is dead. The easiest way to "desolder" it is to cut the front row of leads with a pair of end cutters. Then bend the chip away from the board and cut the back row of leads. Desolder each lead one at a time, clean up with solder wick and flux remover.
Double-check with an ohmmeter that the connections from the supply pins on the LM3886 (pins 1,5 = V+, 4 = V-) to the supply connector are still good.
Once you get the LM3886 removed, I definitely suggest checking the ±17 V regulators as they may have died as well. If you used IC sockets for U1, U3, and U4, I suggest removing the ICs and testing the ±17 V. If you don't get ±17 V, replace U5 and U6.
Once you have a working ±17 V supply, I'd plug the ICs in one at a time. Start with the THAT1200, measure the ±17 V. Then the LME49710. Measure the ±17 V. Then the OPA2277. Measure the ±17 V. Turn the power off and wait for the caps to discharge before you plug in the ICs.
If everything works to here, solder in the LM3886 and follow the final check in the design doc.
I've never tried reverse polarity on an MOD86 (and have no plans to do so), so I can't tell you how far the damage will spread. By default, I'd expect all silicon to be dead and most of everything else to have survived.
There are no electrolytic caps in the signal path to worry about. You could worry a bit about the two 1000 uF supply caps and the smaller 22 uF, 50 V caps by the LM3886. If they never got hot and don't bulge, they might be OK, but if you're ordering parts anyway, I'd throw in a replacement set. They're not terribly expensive. The peace of mind is worth it.
Tom
Sorry to hear.With sparks and smoke resulting in a cracked LM3886, I concur with your diagnosis that the LM3886 is dead. The easiest way to "desolder" it is to cut the front row of leads with a pair of end cutters. Then bend the chip away from the board and cut the back row of leads. Desolder each lead one at a time, clean up with solder wick and flux remover.
Double-check with an ohmmeter that the connections from the supply pins on the LM3886 (pins 1,5 = V+, 4 = V-) to the supply connector are still good.
Once you get the LM3886 removed, I definitely suggest checking the ±17 V regulators as they may have died as well. If you used IC sockets for U1, U3, and U4, I suggest removing the ICs and testing the ±17 V. If you don't get ±17 V, replace U5 and U6.
Once you have a working ±17 V supply, I'd plug the ICs in one at a time. Start with the THAT1200, measure the ±17 V. Then the LME49710. Measure the ±17 V. Then the OPA2277. Measure the ±17 V. Turn the power off and wait for the caps to discharge before you plug in the ICs.
If everything works to here, solder in the LM3886 and follow the final check in the design doc.
I've never tried reverse polarity on an MOD86 (and have no plans to do so), so I can't tell you how far the damage will spread. By default, I'd expect all silicon to be dead and most of everything else to have survived.
There are no electrolytic caps in the signal path to worry about. You could worry a bit about the two 1000 uF supply caps and the smaller 22 uF, 50 V caps by the LM3886. If they never got hot and don't bulge, they might be OK, but if you're ordering parts anyway, I'd throw in a replacement set. They're not terribly expensive. The peace of mind is worth it.
Tom
My build is going well. I've built and tested both the Mod86 boards and a power supply board, also a relay attenuator. Still waiting for delivery of a little 12v regulator board to supply a remote control volume control board that will control the relay attenuator, but I've finished the layout (designed using cardboard cutouts, see the photo attached). Then I can get the case fabricated (it'll be a custom aluminium and oak affair) and finally all put together. I'm using Neutrik powercon and speaking connectors and locking XLR connectors, and 2 Omron relays to allow power control via either a switch on the front or the remote control. Full power off (including shutting off power to the remote control) will be possible via a switch on the back.
(I'll edit this post to attach the layout photo when I've figured out how to upload photos....)
(I'll edit this post to attach the layout photo when I've figured out how to upload photos....)
Layout planning
I've figured out how to attach pictures now. My layout will be as per the attachment: it tries to keep the power supply, mains voltages and transformers away from the signal and aims for a reasonably short signal path from the XLRs through the volume control board and the Modulus-86s to the Speakon connectors.
I've figured out how to attach pictures now. My layout will be as per the attachment: it tries to keep the power supply, mains voltages and transformers away from the signal and aims for a reasonably short signal path from the XLRs through the volume control board and the Modulus-86s to the Speakon connectors.
Attachments
That's a good idea. It will certainly simplify the top and bottom panels as I won't need to machine convection slots in them. I've turned them round, and also moved the XLR and Speakons so that they're aligned vertically, with the Speakons in between the two Modulus-86s: that made the back feel a bit less cramped.
Tom, thanks for your reply. I had not thought of how to reach the rear pins with clippers, Folding it back after the front pins are cut is a great idea. Appreciate your support. Those are clear instructions I can follow. LM3886 should be here soon, now just hoping the voltage regulators will prove to be working, otherwise I'll be waiting on shipping again. Oh well, these are luxury problems, and I'm still having fun.
Adrian, I like your method to lay out a scale diagram like that. I should have thought of that. Your volume control looks cool too. Can't wait to see it as it progresses.
Adrian, I like your method to lay out a scale diagram like that. I should have thought of that. Your volume control looks cool too. Can't wait to see it as it progresses.
Well, desoldering isn't very pretty, but I think it worked. It's tough to clear the holes all the way, but I finally got it. The new lm3886 pins slide right in. Have not soldered it yet though.
The ohm meter showed the circuit path is still intact, and no resistance from power input on the board to the power pins at U2.
So, your guess was correct, U5 and U6 are toast. Measuring at U4 I believe I got +0 and -30V. Will replace the regulators (next week) and test again. I'm betting it will be good from there, then I can install lm3886, THAT, OPA, and LME
I bet I might be the first to have made this mistake, after all, that big Plus Sign clearly marked on the board by the power input is hard to miss.
Good tip. I found a wooden toothpick works too,....well sort of. In the end it was more patience with solder-wick that did the trick.
Hi Alex,
Try a large solder sucker. There is an added advantage that most of the heat leaves with the solder. They make cheap ones, don't buy those. A half decent one will run between $20 and $30 US. They have recoil (so fun) and will save you a lot of solderwick. You still use solderwick to clean pads up to factory fresh, but do not use any pressure or you'll pull the copper off (with the wick).
Solder "helpers", or "dental tools" are great and would include the stainless probe Mooly mentioned, but as a tool with a handle. These are often sold in sets of four. I keep all these things ready depending on the situation.
-Chris
Try a large solder sucker. There is an added advantage that most of the heat leaves with the solder. They make cheap ones, don't buy those. A half decent one will run between $20 and $30 US. They have recoil (so fun) and will save you a lot of solderwick. You still use solderwick to clean pads up to factory fresh, but do not use any pressure or you'll pull the copper off (with the wick).
Solder "helpers", or "dental tools" are great and would include the stainless probe Mooly mentioned, but as a tool with a handle. These are often sold in sets of four. I keep all these things ready depending on the situation.
-Chris
Figures. The diodes should be OK. If you didn't have any of the other ICs (OPA2277, LME49710, or THAT1200) installed, replacing U5 and U6 should do the trick.
Yeah. I'm pretty sure that's a first. Hey... Somebody had to do it at some point.
I swear by my Jonard Tools DP-200 desoldering pump. Digikey P/N: K414-ND. Mouser P/N: 801-DP-200. $24 at today's prices.
Tom
Relays sound like sh*t. Fortunately there are other options, as yourself pointed out (quoted because you can hardly remember something said 2 years ago).
I think Tom thought on designing a speaker protection module IIRC, together with a soft-start/antibump perhaps.
(I'm slowly reading ALL thread and i will finish more or less when I'm ready to buy the mod686
)
Yep. While there's no real need for a turn-on delay, I might as well build one in. It doesn't cost much extra and will make my protection board more universally applicable.
I'll be using MOSFETs for the switching. I've found a good one with very low on resistance that does not add any distortion that I can measure.
I hope to have the protection board done along with a Power-686 board (the two really need to be designed together) available before the summer kicks into high gear.
Tom
A few years ago have auditioned a linkwitz system (they were the bigger speakers) and I have NOT been impressed, although I do like open baffle.
Probably because of the harsh highs of the typical 3886 implementation and high noise of the crossovers... dunno for sure, my knowledge was much lower then. Anyway, from memory, my latest OB with software crossover had none of those issues (they had other issues of course).
I'm sure they sound much better with the modulus amps.
I do think that for low bass active with DSP is a must, but mids-hf do not seem to suffer too much from passive crossovers. I'm going to experiment with a passive XO in my next OB, which could be easily changed to active... just need more amps and a SOTA active xo