Project Complete, or It'll Do For Now, Anyway
This is a special-purpose Modulus-86 six-channel amp, built to drive the three-channels-each top baffles of Linkwitz LX-521 speakers. I'll start by thanking tomchr for designing and making the Mod-86 available in the first place; it fairly jumped off the screen at me for just this purpose. I've listened to it for a couple of weeks now. Before posting the picture and these comments I'd meant to tidy up its interior a bit, right-sizing the wire runs a bit better, etc....but when I go to do that I just end up listening to it instead, so I'll just go ahead now.
It replaces repurposed Linkwitz Pluto boards, three-3886s per side each powered by a 50VA transformer with 44,000 uF of reservoir caps. The power supply for this amp is an Antek 2222 plus tomchr's Power-686, so it's the same total reservoir capacitance but with 100VA per side. Concealed in the pic behind the front panel is a Neurochrome Intelligent Soft Start. I used XLR sockets for the inputs and outsourced the input pseudo-differentialing to adapter interconnects from the LX521 ASP.4 crossover, so in the unlikely event I ever find myself with a balanced-output crossover I'll be ready for it.
The repurposed Pluto boards didn't sound bad to me, and in a less-than-adequate listening space (an attic) Led Zeppelin doesn't need much power to make you a Def Leppard. I could call the Pluto boards 3886 maybe D not quite so R. The Mod-86s are a clear improvement. The thing I notice most is a better focused, more palpable image, and I feel the improvement is well worth the time I spent building this amp. One example: of Michael Murray's recordings of Bach organ music on Telarc, the one I cared least for was Bach at St. Bavo; the sound was so reverberant I could barely pick out the melody. But with this amp, the reverberance is still, there but the melody is easier to focus on so that this album is one I'll be listening to more.
Building lessons learned (with apologies to those who already know them):
1. Stick-on thick felt pads from the hardware store are like Dr. Scholl's (TM) for ModuShop's default plastic feet. They make it much more comfortable to slide the amp around on a shelf. Who knows, maybe the plastic feet were even designed with those in mind? Anyway, they work!
2. Get your chassis and power supply first, and don't start soldering parts onto the Mod-86 pcbs until you've used them as templates to mark where you'll drill their mounting holes. This became very obvious to me in hindsight.
3. If you can strip the threads you've tapped into the aluminum by overtightening a machine screw, chances are that same screw could have tapped those threads in the first place, saving you the cost of buying a tapping bit. And just because those threads stripped easily in a 3mm panel, that doesn't mean a thicker heat sink of the same metal can't bite off the end of your tapping bit, leaving you to tap the mounting holes for the pcbs and 3886s with the mounting screws amd a screwdriver. Actually, that wasn't so bad; it just took a bit of care and concentration: making sure the screw was vertical on starting, then advancing the screw a little and backing off a bit, then repeating until the hole was threaded to the length of the screw. Not really suited to mass production, but OK for a DIY. Pro tip: a US 7/64" hole is an excellent size for an m3 machine screw.
I'd like to thank Mike Hanson for his suggestions for working the rear panel, especially the stepped drill bit for the XLR mounting holes. It the first time I'd ever used a stepped bit. And of course, special thanks to tomchr for practically instantaneous responses when I needed points clarified or suggested substitutes for out-of-stock components in the BOM. This project was well worth doing!
That's an excellent-looking build. Thank you for sharing. I also appreciate your thoughts and subjective evaluation of the amp.
It's indeed funny that the default ModuShop feet don't have any sort of pad. It's literally just a piece of plastic. I've used felt furniture pads with them. Also some of their chassis come with feet, others don't. I'm not sure on which ones... I actually buy their anti-vibration feet. They're pretty nice. I don't care for the anti-vibration "feature" as there's nothing in the Modulus amps that is sensitive to vibration - at least within what can be expected in residential audio and shipping anyway - but I do like that the feet make the amp sit a bit proud on the shelf, which gives good airflow from underneath. The anti-vibration feet do come with some neoprene (I think) pads.
+1 for using the unpopulated boards as a template.
When tapping M3 threads I start with a 2.0 mm pilot hole. Then 2.5 mm. That's the recommended drill size for full M3 threads. I then use a tapping block to ensure that the tap is perpendicular to the heat sink. I also use A-9 aluminum cutting fluid (not to be confused with A-1 steak sauce). Always use cutting fluid. I find that gives me pretty reliable threads. Not as good as threads cut by CNC machine but certainly good enough.
If you're going to just force the screw into the heat sink without cutting threads, I suggest using a self-threading screw rather than a plain machine screw. They're intended for that exact purpose.
I'm glad to hear that your amp came together nicely. Enjoy!
Tom
It's indeed funny that the default ModuShop feet don't have any sort of pad. It's literally just a piece of plastic. I've used felt furniture pads with them. Also some of their chassis come with feet, others don't. I'm not sure on which ones... I actually buy their anti-vibration feet. They're pretty nice. I don't care for the anti-vibration "feature" as there's nothing in the Modulus amps that is sensitive to vibration - at least within what can be expected in residential audio and shipping anyway - but I do like that the feet make the amp sit a bit proud on the shelf, which gives good airflow from underneath. The anti-vibration feet do come with some neoprene (I think) pads.
+1 for using the unpopulated boards as a template.
When tapping M3 threads I start with a 2.0 mm pilot hole. Then 2.5 mm. That's the recommended drill size for full M3 threads. I then use a tapping block to ensure that the tap is perpendicular to the heat sink. I also use A-9 aluminum cutting fluid (not to be confused with A-1 steak sauce). Always use cutting fluid. I find that gives me pretty reliable threads. Not as good as threads cut by CNC machine but certainly good enough.
If you're going to just force the screw into the heat sink without cutting threads, I suggest using a self-threading screw rather than a plain machine screw. They're intended for that exact purpose.
I'm glad to hear that your amp came together nicely. Enjoy!
Tom
I am using 2U Galaxy 230 x 230 enclosures to make monoblock Mod 86's. I was planning to put the heat sink on the interior of the enclosure, but increase the top and bottom cover ventilation in the area of the heat sink.
Hi all,
I have a strange thing. The modulus-86 was running perfect but I blew up one LM3886T, yes, absolutely my own fault. I let it made contact with the chassis and it was smoked ☹. Fortunately I had a spare one.
I replaced the faulty LM3886T but, the sound did NOT came back, the strange thing is, neither does the other channel. Everything is completely dead.
Wherever I could, I am measuring same voltages at the board and LM3886T, for both sides, the same. So that part it looks ok. The LM3886T of both sides are heating up.
I could imagine I blew one channel and have to rebuild it, but not that both channels are not running anymore.
Somehow it looks like the output of both channels are shorted and therefor you don’t hear anything from the loudspeaker?? But the left-right are separated. The Power-86 seems to be ok.
Which components should I check or is it just build two new ones? I know that will be the easiest way… Normally I build tube amplifiers, like SET 300B, without any problems. This for me is something completely new. 🙂
Any tips would be nice where to search.
Thanks in advance,
Ad.
I have a strange thing. The modulus-86 was running perfect but I blew up one LM3886T, yes, absolutely my own fault. I let it made contact with the chassis and it was smoked ☹. Fortunately I had a spare one.
I replaced the faulty LM3886T but, the sound did NOT came back, the strange thing is, neither does the other channel. Everything is completely dead.
Wherever I could, I am measuring same voltages at the board and LM3886T, for both sides, the same. So that part it looks ok. The LM3886T of both sides are heating up.
I could imagine I blew one channel and have to rebuild it, but not that both channels are not running anymore.
Somehow it looks like the output of both channels are shorted and therefor you don’t hear anything from the loudspeaker?? But the left-right are separated. The Power-86 seems to be ok.
Which components should I check or is it just build two new ones? I know that will be the easiest way… Normally I build tube amplifiers, like SET 300B, without any problems. This for me is something completely new. 🙂
Any tips would be nice where to search.
Thanks in advance,
Ad.
Grounding the tab of the LM3886T with the power on tends to blow the bond wire to pin 4 (V-) internally in the chip. That renders the chip dead. If the LM3886T in the other channel was isolated from the chassis I can't imagine grounding one would kill the other, but it could have killed something in the power supply. The bridge rectifier would be the most likely victim.
I don't think you need to throw in the towel just yet, though.
1: I'd start by verifying that you get ±30 V (or so) out of the Power-86. Then verify that this ±30 V (or so) makes it to each of the two channels' power connectors. Also verify that your input signal makes it to the input connector. Apply 1 V, 400 Hz to the input of the amp and measure the AC voltage from pin 2 to pin 3 of the input connector on the Modulus-86 board. You should see the same 1 V, 400 Hz.
2: If it's all good until here, have a close look at the (Rev. 3.0) boards. By the speaker output you'll find two small test points marked "+16V" and "-16V". They're right next to D5. Measure the voltages there. You should see ±16 V (± 1 V, max). If you don't have ±16 V, the most likely cause is that the LM317L (+16 V) or LM337L (-16 V) regulators died.
If the incoming power (±30 V) and the on-board power (±16 V) are both good, we need to do some digging:
3: Start by removing the two LM4562 ICs. Measure the DC offset at the speaker output. It should be somewhat low. Typically you'll see around ±20 mV, but you might see ±100 mV there. Then apply a 1 V, 400 Hz test signal to pin 1 of the empty socket for U3. That's the socket closest to the LM3886. You should now have 2.75 V, 400 Hz at the speaker output.
4: If this works, plug in U2. Apply 1 V, 400 Hz test signal to the input of the amp. If you then measure differentially from pin 7 to pin 1 of U2 you should see the same 1 V (or maybe a smidge lower).
5: And if this works, plug in U3. The DC offset at the output of the amp should drop to the specified ±2 V (typical), ±10 mV (worst case). With 1 V, 400 Hz applied at the input you should see 10 V, 400 Hz at the output of the amp.
It should go without saying but I'll say it anyway: Don't plug/unplug components with the power on. Verify with a voltmeter that the power supply has discharged to a few volt, preferably less before messing with the circuit.
If you don't have a signal generator handy, you can use your smartphone or media player. I cover that here:
I'm headed to Seattle for the Pacific Audio Fest, so I won't be too active here the coming two weeks. Email takes priority, so if you need more help just toss me an email. Just add @neurochrome.com to my user ID here and you have my email address. I should have wifi access at most stops, but the point of the trip is to represent TCA and also to be social with friends, so please allow me time to respond.
Tom
I don't think you need to throw in the towel just yet, though.
1: I'd start by verifying that you get ±30 V (or so) out of the Power-86. Then verify that this ±30 V (or so) makes it to each of the two channels' power connectors. Also verify that your input signal makes it to the input connector. Apply 1 V, 400 Hz to the input of the amp and measure the AC voltage from pin 2 to pin 3 of the input connector on the Modulus-86 board. You should see the same 1 V, 400 Hz.
2: If it's all good until here, have a close look at the (Rev. 3.0) boards. By the speaker output you'll find two small test points marked "+16V" and "-16V". They're right next to D5. Measure the voltages there. You should see ±16 V (± 1 V, max). If you don't have ±16 V, the most likely cause is that the LM317L (+16 V) or LM337L (-16 V) regulators died.
If the incoming power (±30 V) and the on-board power (±16 V) are both good, we need to do some digging:
3: Start by removing the two LM4562 ICs. Measure the DC offset at the speaker output. It should be somewhat low. Typically you'll see around ±20 mV, but you might see ±100 mV there. Then apply a 1 V, 400 Hz test signal to pin 1 of the empty socket for U3. That's the socket closest to the LM3886. You should now have 2.75 V, 400 Hz at the speaker output.
4: If this works, plug in U2. Apply 1 V, 400 Hz test signal to the input of the amp. If you then measure differentially from pin 7 to pin 1 of U2 you should see the same 1 V (or maybe a smidge lower).
5: And if this works, plug in U3. The DC offset at the output of the amp should drop to the specified ±2 V (typical), ±10 mV (worst case). With 1 V, 400 Hz applied at the input you should see 10 V, 400 Hz at the output of the amp.
It should go without saying but I'll say it anyway: Don't plug/unplug components with the power on. Verify with a voltmeter that the power supply has discharged to a few volt, preferably less before messing with the circuit.
If you don't have a signal generator handy, you can use your smartphone or media player. I cover that here:
I'm headed to Seattle for the Pacific Audio Fest, so I won't be too active here the coming two weeks. Email takes priority, so if you need more help just toss me an email. Just add @neurochrome.com to my user ID here and you have my email address. I should have wifi access at most stops, but the point of the trip is to represent TCA and also to be social with friends, so please allow me time to respond.
Tom
This I wrote to Tom last night...
One of my latest tests was to measure the output with 1v 400Hz input... before pulling out the lm4562's.... It measured 10v. What!
Then I realized what happened...
I always use a couple modded Technics SB-F1.. to test.. They have a protection circuit.😯
Somehow both went of... Resetted them... Music meastro...😊
This never happened before in the past 30 years and they had to cope a lot of tests...
But thank you for your help, I understand more from the amp now.
Oh man, I am happy!
Enjoy your stay with yours friends
Ad.
One of my latest tests was to measure the output with 1v 400Hz input... before pulling out the lm4562's.... It measured 10v. What!
Then I realized what happened...
I always use a couple modded Technics SB-F1.. to test.. They have a protection circuit.😯
Somehow both went of... Resetted them... Music meastro...😊
This never happened before in the past 30 years and they had to cope a lot of tests...
But thank you for your help, I understand more from the amp now.
Oh man, I am happy!
Enjoy your stay with yours friends
Ad.
So did I Tom!The protection circuits probably tripped when the unfortunate LM3886 tab met with the chassis. That probably caused the LM3886 to output DC and the protection circuits did their job.
I'm glad you figured it out.
Tom
Enjoying your amp now.
Ad.
Is it possible to lower the gain a bit?
My speakers are rather sensitive and would like to have more control with the volumeknob.
Now only running with an 10k Alps and sound level already high only moving a bit...
Ad.
My speakers are rather sensitive and would like to have more control with the volumeknob.
Now only running with an 10k Alps and sound level already high only moving a bit...
Ad.
20 dB is the lowest. Sorry. The LM3886 made me do it. 🙂
You could add an attenuator on the input, though.
Tom
You could add an attenuator on the input, though.
Tom
Hi Tom,20 dB is the lowest. Sorry. The LM3886 made me do it. 🙂
You could add an attenuator on the input, though.
Tom
No problem at all, it was just a question. 🙂
You are making great stuff. Sounding great!!
It's just fine with an Alps but I was thinking to build a TVC with Seiden but they only have 23 steps and therefore probably the steps are too big,
Will figuring it somehow.
Regards,
Ad.
Yeah. That's why I stopped using attenuators years ago. At the time I had a pretty high-gain preamp with a run-of-the-mill power amp. I always wanted the volume control to be between steps 3 and 4. Volume control ICs solve this issue nicely (but sometimes at the expense of noise as is the case for the PGA2320 for example).It's just fine with an Alps but I was thinking to build a TVC with Seiden but they only have 23 steps and therefore probably the steps are too big,
In your case I'd probably add a 10-20 dB attenuator up front. That should get you onto the part of the volume control attenuator with a more reasonable step size. Alternatively, add the attenuator in the digital domain at the source or DAC (assuming you're using a digital source).
Tom
Hi Tom,Yeah. That's why I stopped using attenuators years ago. At the time I had a pretty high-gain preamp with a run-of-the-mill power amp. I always wanted the volume control to be between steps 3 and 4. Volume control ICs solve this issue nicely (but sometimes at the expense of noise as is the case for the PGA2320 for example).
Isn't it a pitty to use a 'noisy' volume ic control with such perfect quiet designs of your own?
Ad.
Yes, that's a great idea..😊😊Yeah. That's annoying. Maybe I should design something... 🙂
Tom
Ad.
My one problem is with wide dynamic range classical recordings that I have to play near the top of of my DAC’s volume range, where the steps are coarsest, to hear the quietest parts. I made my Mod-86s’ gain 25 dB instead of 20 to put the loud parts not quite so high in the DAC’s coarse increments range, but it doesn’t help much. I’m hoping Tom’s solution will let me say, “Siri, DAC volume range classical uncompressed.” 🤪
If/when I design a volume control it will have an even step size all the way through. Unless I go with a motorized pot. In that case it wouldn't have steps. It wouldn't be compatible with any of the digital assistants, though.
Tom
Tom
- Home
- Amplifiers
- Chip Amps
- Modulus-86 build thread