Modulus-86 build thread

Jump to Latest
T

twest820

Member
Joined 2009
tomchr said:
From Speedskater's comments about Starquad cables and microphone cables I figured he was concerned with mains hum coupling into the circuit. My basic point is that you'll be very hard pressed to couple anything into the MOD86 within the audio band.
Click to expand...
Thanks for clarifying. Audio band coupling'll be more conducted errors, hence much of the tutorial type stuff we went through about PSRR, CMRR, and managing different types of ground when bringing you up to speed on why the composite circuit topology was what it was at the beginning of Mod 1.0 design. Much of that's since become core marketing material which, like test and measurement capabilities, has been repeated many times and doesn't particularly need to be reiterated.

One comment I would make, though, is conducted errors aren't just an audio band thing. Common mode errors from SMPS in the hundreds of kHz to a few MHz are closely associated with their Y cap performance. A major thing I wanted to get from providing the circuit topology and layout methodology was verifying it was robust to hard switching downmod from forward SMPSes, either directly or via TID kinds of effects. The unchanged performance with the SMPS-86 goes some ways to answering those questions, so thanks for that.
 
henrylrjr

henrylrjr

Member
Joined 2012
AndrewT said:
Henry,
take the transformer's shield wire direct to chassis floor using a short wire.
6A mains fuse seems very high for a 110/120Vac supply into a pair of chipamps.
I see the transformer has 4 primary taps for a dual primary type.
Each primary has a Flow and Return (Live and Neutral). This pair must be twisted. The other primary should similarly used a twisted Flow and Return pair.
The secondary outputs are also 4 wires. These too should be arranged as two pairs, each pair being the Flow and return of one secondary.
All the loop areas at the terminations of the transformer's pairs should be kept small.
All the above is to minimise the emi emitted by the transformer and it's wiring.

The +ve, -ve and Zero Volts from the PSU to each amplifier should be arranged as a twisted triplet. Again the loop areas at the terminations should be kept small to minimise emi.

The input XLR has 3 pins and a shell. The shell and pin1 are connected directly to the Chassis at the entry hole. Pins 2 and 3 are your signal and these are taken to the amplifier using a twisted pair. It can be screened or unscreened. If screened then it is better to connect both ends of the screen to the chassis using low impedance connections at both ends. A long pig-tail is not as good.

Finally each speaker pair should be twisted and incorporate low loop area terminations.

Attending to these emi minimisation steps will virtually remove the need to be concerned about emi effects of nearby cable pairs affecting any other cable pairs.
Click to expand...

I must have missed your answer to my primary question below. I probably wasn't clear enough but that was the key question.

Bottom line is, if power wires and speaker output wires being parallel and 1 1/4" apart is not a concern
 
henrylrjr

henrylrjr

Member
Joined 2012
Thought I mentioned wiring would be done like my Mod86. Everything will be bundled in appropriate groups and if wires from the PCB speaker terminal block to speakons are long enough they would have a few twists. I will also run the transformer shield to the Power86 per Tom's pdf build doc. The Xlrs will also be wired as I described and there will probably a 6amp slo-blow fuse.

I was just looking for an opinion on power wires and speaker output wires being parallel and 1 1/4" apart...ok or bad.

Thanks,
henrylrjr
 
A

AndrewT

R.I.P.
Joined 2004
AndrewT said:
...................I see the transformer has 4 primary taps for a dual primary type.
Each primary has a Flow and Return (Live and Neutral). This pair must be twisted. The other primary should similarly used a twisted Flow and Return pair.
The secondary outputs are also 4 wires. These too should be arranged as two pairs, each pair being the Flow and return of one secondary.
All the loop areas at the terminations of the transformer's pairs should be kept small.
All the above is to minimise the emi emitted by the transformer and it's wiring.

The +ve, -ve and Zero Volts from the PSU to each amplifier should be arranged as a twisted triplet. Again the loop areas at the terminations should be kept small to minimise emi.

The input XLR has 3 pins and a shell. The shell and pin1 are connected directly to the Chassis at the entry hole. Pins 2 and 3 are your signal and these are taken to the amplifier using a twisted pair. It can be screened or unscreened. If screened then it is better to connect both ends of the screen to the chassis using low impedance connections at both ends. A long pig-tail is not as good.

Finally each speaker pair should be twisted and incorporate low loop area terminations.

Attending to these emi minimisation steps will virtually remove the need to be concerned about emi effects of nearby cable pairs affecting any other cable pairs.
Click to expand...

henrylrjr said:
...........................All wires will be twisted or bundled into appropriate groups. The schematic is intended to just show what connects where...................
Click to expand...

tomchr said:
.......................They should be tightly coupled. That doesn't necessarily mean twisted. Many transformer manufacturers, Electra-Print being one of them, actually advise against twisting the wires together.



Again: They should be tightly coupled. In particular with heavy gauge wire, you usually end up creating a larger loop area than you would just bundling the wires tightly with wire ties....................



*tightly coupled.......................
Click to expand...

henrylrjr said:
.........................Regarding twist vs bundle, with Tom's input I chose to bundle wires. The only wires I would have twisted would have been those from the PCB speaker connections to the binding posts. However, in my build, the wires were only about 1 1/4" long so twisting wasn't in the cards.The final pic is in post 2374.
Click to expand...
Here's some measurements and explanation of why close coupled can be bettered by using a twisted pair.
http://www.diyaudio.com/forums/analogue-source/129126-simplistic-njfet-riaa-824.html#post4830330
Power Strip induces 20microVolts into TwinLead, 1uV into TwistedPair
Looking at PCBs, there are loops everywhere. And wiring/cables are loops.
When you make your own twistedpair, the irregular twisting will cause inconsistent nulling of external fields. And where any cable is soldered to a connector, there will be a vulnerable loop formed by Signal & Return.

Having simple math for planar aggressor_wire and victim_loop, I use it.

I assume a PowerStrip attached to wall, 10cm behind the Audio Rack and extending a full meter left-to-right (yes, I know that is a huge Strip). I also have the TwinLead (not TwistedPair) running a meter from TurnTable to RIAA PreAmp, exactly parallel to the PowerStrip and 10cm away.

Result? A lonely wire, Hot or Return, of the PowerStrip induces 2,000uV.
The real PowerStrip, with Hot bus close to the Return bus, induces 20uV.
Into TwinLead.
Into TwistedPair, this should drop another 20X (26dB), else why use TP.

The loop area of the TwinLead is 1meter * 1mm, or 1cm*10cm, with 20uV induced.
If your construction has 1cm*1cm loops, that Power Strip will, worst case,
induce 2uV.
Here is the attachment. Avoid nearby PowerStrips.
Click to expand...
 
Last edited:
henrylrjr

henrylrjr

Member
Joined 2012
What are you builders of the Parallel86 using for a fuse. I'm using the Antek AS-3224 300VA transformer at 115 Vmains. Using the simple equation fuse = 3 X VA/Vmains I get 7.8. That seems a bit high based on what I've read in other posts.

On a slightly different fuse topic, my Aragon and McCormack have a 10 or 12 A slo-blo at the IEC but also have lower amperage fast-blo fuses on the rails. I haven't seen anyone using rail fuses. Why is that?
 
Last edited:
tomchr

tomchr

Neurochrome.com
Joined 2009
neurochrome.com
Paid Member
I use a 6.3 A slow-blow fuse with my 300 VA Antek transformer on 120 V mains. I've yet to pop the fuse.

As we discussed here recently, 3x VA/Vmains is likely overkill. 2x VA/Vmains is probably asking for trouble unless you use a soft-start circuit.

If you really want to get into it, you can look at the inrush current vs time and work out the I2t rating needed for the fuse. You then find a fuse that can handle this, which means you'll be specifying a specific brand, specific series, specific ampacity of the fuse ...... or you could just get a 6.3 A slow-blow fuse like I use and be done. :)

Tom
 
Last edited:
henrylrjr

henrylrjr

Member
Joined 2012
Thanks Tom. 6A slow-blow is what I used for the Mod86 and I'll use the same for the Parallel86.

I mentioned rail fuses for two reasons. One is my other amps have them and they are fast-blow to provide extra protection. The other reason is the Paralle86 uses an IC nolonger being made. If it craps out down the road am I left with a, useless, non-working amp, or is there another IC that could be used? I thought maybe rail fuses would provide extra, long term, protection.
 
tomchr

tomchr

Neurochrome.com
Joined 2009
neurochrome.com
Paid Member
I'm not sure I'd bother with rail fuses. The LM4780 is current limited, so as long as the device remains functional you don't have much to worry about.
With rail fuses I'd be concerned with one fuse blowing and the other remaining intact. Depending on which fuse blows that could give you DC on the output.

As far as the LM4780 availability is concerned, I suspect there'll be stock available for quite a while to come. It may not be available through the usual channels, thus easily obtainable, but surplus dealers will probably have a few in stock here and there.

Also, semiconductors have a nasty way of protecting fuses by blowing first... ;)

Tom
 
A

akapezi

Member
Joined 2006
It's all complete and I enjoyed a long listening session with this amp on the weekend. It was fantastic.

This was my first amp build. My decision to build this particular kit was completely validated, but it wasn't only the sound quality. It was the entire experience: knowing the design was thoroughly vetted and measured; receiving excellent documentation with the PCBs; and reading the supplemental posts from individuals on this thread who shared obsessive musings about hardware, layout, and cables. I know I did my own obsessive musing about hardware, layout, and cables.

Here are some design decision that were made:

Input Connection
I picked up bulk Mogami W2534 quad microphone cable from Guitar Center (US) for the connection from the female XLR chassis mount to the input of the Modulus-86 board. I would have preferred the W2549 two conductor balanced microphone cable because twisting the quad cable into two connectors and keeping everything as short as possible was a challenge.

Assembly Tools
It was mentioned that a ratcheting crimp tool would be beneficial. I used a standard crimp tool, but I could definitely understand the appeal of an upgraded crimp tool.

I should have put money into a better wire stripping tool. I used inexpensive handheld IDEAL brand wire strippers. Of the 12, 14, 16, 22 AWG wires used on this build, it only stripped the 16 AWG wire nicely. I was either nicking and pulling conductors on the smaller conductors or not cleanly removing the jacket on the larger ones.

Soldering
Check out the PACE Worldwide Youtube page for excellent soldering tutorials. The videos are from 1980 but the content was excellent. I had barely done any soldering before this project, but I learned everything I needed to know from these videos.

Chassis
I went with the Modushop.biz SLIMLINE 2U case. I did provide them with CAD files for routing the back and bottom panels. I received the case a week after the order. It was great.

I didn't mess with threaded holes on the chassis. This allows for the looser floating fastener tolerance stackup. I did use thread rolling screws electrically for the chassis grounding connections and mechanically for screwing into the heatsink.

I didn't do anything with the front panel, no LEDs or engraving. I used an illuminated IEC power inlet, so if I want to know the amp is on, I'll just peak my head around to the back the case.

Heatsink
I went with the 8" version from HeatSinkUSA.com cut to 2.5" height.


I'll have more to add later. Cheers - Curtis
 

Attachments

  • IMG_0273.jpg
    IMG_0273.jpg
    805.5 KB · Views: 598
You must log in or register to reply here.
Top Bottom