[f] AMB on Facebook

Balanced Alpha10 (α10) Build Log

For AMB projects that do not have their own forum.

Balanced Alpha10 (α10) Build Log

Postby blairfrischx » January 22nd, 2018, 3:59 pm

This post is really long, but I like to think it's a reflection of how much I have enjoyed building the preamp.

At this point, it's still a work in progress with some really minor finicky points left to solve. Hopefully in the next week or so it will be done, and I can update this post with a picture of the finished product.

Background
I built an unbalanced ɑ10 back in 2014 and used it in my main system for a few months before I took it out of commission due to DC offset drift (initially designed for gain of 5) after I heard my speakers pop a couple times. Since then, it mostly sat on my shelf of unused audio equipment - essentially just looking pretty.

In the meantime, I shifted my focus towards building a fully balanced β22, which has been working incredibly well since I sorted out the transistor orientation issues I had in the linked thread.

Once I finished the β22, I shifted my attention back up the signal stream, towards a fully balanced preamplifier to feed into the β22. The ɑ10 was the obvious choice, so I decided to take the one I had already built, reduce the gain to mitigate the DC offset drift, and build a matching ɑ10 to sit on top of it.

Design
I decided to reduce the gain for each ɑ20 board from 5 to 3, which seemed like a good compromise for the increased stability, especially considering the additional gain from the fully balanced design. It takes a hit on unbalanced output, but that was okay for me.

I took some design direction from the previous two topics I could find on builders of the fully balanced ɑ10: Pinocchio's thread and turbotoy's thread, especially turbotoy. The "Other options" section of the Alpha10 site was also invaluable, though I decided to make the right channel my bottom board instead of left channel.

I wanted my new ɑ10 to match the form factor of my β22 + σ22 (dual 16" x 12" x 3" cases), so I opted to build a single-chassis ɑ10 (16" x 12" x 4") instead of the two chassis approach. I used the Par-Metal 20 Series enclosure and custom machined panels from Front Panel Express (shown below). This means that there is not a separate power supply chassis, and the SSR and two power transformers are housed directly next to the σ22 / σ25 side of the ɑ10 backplane. The transformers are centered between the edge of the board and the edge of the enclosure, partially for aesthetics, partially because the vent holes made it problematic to mount the transformers closer to the edge of the enclosure.

Due to an apparent lack of availability for 9V+9V 10VA transformers to supply the two δ1 and two δ2 boards, I ended up ordering and using 15VA transformers for both the analog and digital supplies, the Talema 62051 and 62053. Maybe it’s overkill for the load, but I didn’t want to risk having too little power for all of the circuitry, and I had space to spare inside the enclosure.

Front Panel (design)
Rear Panel (design)
Internal Layout (planned)

Build
The building of everything in the enclosure went relatively smoothly, until I got to the wiring split for the T1_SEC supply to both ɑ10 backplane boards; I just couldn’t get four wires together into one of the small Molex KK 254 crimp terminals. So instead of doing that, I changed over to ring terminals and a small terminal strip for all of the transformer secondary wiring (shown below), and it went much better.

Overview Without Panels (sitting on top of my β22)
Front Panel (actual)
Rear Panel (actual)
Internal Layout (actual)
Internal Layout w/ Panels
Transformer Wiring

Initial setup on the first ɑ10 backplane and all of its constituent boards went really smoothly. This is pretty unsurprising, because it is essentially the ɑ10 that I had already built, just in a slightly larger enclosure. At least it proved that I didn’t break anything when swapping out R13 in the two ɑ20s to set the lower gain. I set up my remote, ran the LCDuino through all of the input/output/display menus, plugged in the σ22, set up the offset and bias for the two ɑ20 boards, and left it on in this state for about 24 hours, acting like it was its old unbalanced self. The DC offset stayed pretty stable right near the measuring capabilities of my multimeters, +/- 2 to 3mV, and the bias stayed around 66mV on both sides. I was happy.

ɑ10 right channel backplane

When I connected the second ɑ10 backplane and did initial setup on the second set of δ1 and δ2 boards (the new ones, which had never been tested before), I noticed that the “PWR” light on the δ2 LED+ bank did not light up. However, it was still able to be controlled by the LCDuino and switch inputs/outputs as needed, and it still lit up the other lights in the LED+ bank when actuated, so maybe it is just the “PWR” LED is burned out – I’m not sure. Setup on the second δ1 proceeded without any problems.

I plugged in the second σ22. It seemed to have a slightly lower output rail voltage than the first σ22 (around 11.89VDC vs 12.05VDC), which didn’t bother me too much, so I proceeded to setup on the second set of ɑ20 boards. One of them was well-behaved and held a stable offset around +/-2mV, bias around 66mV. The other ɑ20 realized it was the last board in the entire system to be installed and tested, so it decided to have problems. It seemed to have a damaged component somewhere, as I noticed during initial setup the output-to-ground measurement was around +5.9VDC, and adjustment of VR1 and VR2 did not do anything to bring it down beyond +5.3VDC. I started troubleshooting, but the ɑ20 modules are pretty cheap, so I am planning to just buy a new set of components and build another one.

ɑ10 left channel backplane

Troubleshooting of the "left" ɑ20 on the second ɑ10 backplane (left channel unbalanced, left channel inverted balanced) yielded the following:
All resistors tested at their proper value
I could not get a read on the capacitors while in the circuit, but they are installed in the proper orientation and not obviously damaged.
The BJTs and JFETs are installed in their correct position and orientation (verified against ɑ20 website and the functioning boards) and are not obviously damaged.

I feel like I need to take the BJTs and JFETs out of the circuit to properly test them. I have a multimeter with diode measuring functions, but the results while the transistors are in the circuit don't line up with what they should, even on the good boards.
If there is any more troubleshooting I can do in the meantime, I'm glad to take suggestions.

Sound
I haven’t actually made it to the point of actually plugging anything in yet.

It will be replacing a B&K Reference 50 in my office stereo setup, feeding the balanced β22 (headphones) and a PS Audio 200cx (speakers).
Last edited by blairfrischx on January 28th, 2018, 10:35 am, edited 2 times in total.

Pieterh likes this post.
blairfrischx
 
Posts: 23
Likes: 0 post
Liked in: 8 posts
Joined: February 7th, 2014, 8:11 pm
Location: Columbus, OH
Country: United States (us)

Re: Balanced Alpha10 (ɑ10) Build Log

Postby Pieterh » January 23rd, 2018, 12:17 am

Thanks you, your post is very inforrmative !
Pieterh
 
Posts: 132
Likes: 93 posts
Liked in: 54 posts
Joined: December 29th, 2015, 12:50 am
Location: Netherlands
Country: Netherlands (nl)

Re: Balanced Alpha10 (ɑ10) Build Log

Postby blairfrischx » January 24th, 2018, 5:04 am

Patiently waiting on parts to fix / replace the problematic ɑ20 board. Panels are supposed to be here tomorrow. :D


On the three functioning ɑ20 boards, the DC offset seems to be pretty stable in the +/- 5mV range between turning on the analog section and running for 24+ hours The exception to that is when I had it sitting on top of the β22 while listening to my headphones, since that puts off enough heat to throw the bottom ɑ10 all out of wack (DC offset crawls up to about 13mV on the bottom two ɑ20 modules)...so I think I won't sit it on top of the β22 when I have it all plugged in and running ;)
blairfrischx
 
Posts: 23
Likes: 0 post
Liked in: 8 posts
Joined: February 7th, 2014, 8:11 pm
Location: Columbus, OH
Country: United States (us)

Re: Balanced Alpha10 (ɑ10) Build Log

Postby blairfrischx » January 24th, 2018, 2:45 pm

blairfrischx wrote:When I connected the second ɑ10 backplane and did initial setup on the second set of δ1 and δ2 boards (the new ones, which had never been tested before), I noticed that the “PWR” light on the δ2 LED+ bank did not light up. However, it was still able to be controlled by the LCDuino and switch inputs/outputs as needed, and it still lit up the other lights in the LED+ bank when actuated, so maybe it is just the “PWR” LED is burned out – I’m not sure.


I solved this problem. δ2-2 LED+ bank is working perfectly now ;)

I swear, you miss one solder joint
blairfrischx
 
Posts: 23
Likes: 0 post
Liked in: 8 posts
Joined: February 7th, 2014, 8:11 pm
Location: Columbus, OH
Country: United States (us)

Re: Balanced Alpha10 (ɑ10) Build Log

Postby blairfrischx » January 24th, 2018, 4:06 pm

Since I don't have much to lose troubleshooting the problematic ɑ20, I started pulling the transistors out and measuring in both diode mode and ohmmeter mode to see if anything popped out as a problem.

Started with BD139 (Q5) and BD140 (Q6) - both appear to test good
Q5 is NPN, so I measured with positive lead of my multimeter on base
Q6 is PNP, so I measured with negative/common lead of my multimeter on base

Q5 measured 0.638V base(+) to collector(-) and 0.639V base(+) to emitter(-), OL on both B-C and B-E in reverse orientation
Q5 measured 5.93MOhm base(+) to collector(-) and 5.91MOhm base(+) to emitter, OL on both B-C and B-E in reverse orientation

Q6 measured 0.628V base(-) to collector(+) and 0.630V base(-) to emitter(+), OL on both B-C and B-E in reverse orientation
Q6 measured 6.02MOhm base(-) to collector(+) and 6.01MOhm base(-) to emitter(+), OL both B-C and B-E in reverse orientation

Continued with BC550C (Q3, Q7) and BC560C (Q4) - all of them seemed to be okay
Q3, Q7 are NPN, so I measured with positive lead of my multimeter on base
Q4 is PNP, so I measured with negative/common lead of my multimeter on base

Q3 measured 0.684V base(+) to collector(-) and 0.681V base(+) to emitter(-), OL on both B-C and B-E in reverse orientation
Q3 measured 6.79MOhm base(+) to collector(-) and 6.79MOhm base(+) to emitter, OL on both B-C and B-E in reverse orientation

Q7 measured 0.668V base(+) to collector(-) and 0.674V base(+) to emitter(-), OL on both B-C and B-E in reverse orientation
Q7 measured 6.74MOhm base(+) to collector(-) and 6.92MOhm base(+) to emitter, OL on both B-C and B-E in reverse orientation

Q4 measured 0.676V base(-) to collector(+) and 0.676V base(-) to emitter(+), OL on both B-C and B-E in reverse orientation
Q4 measured 7.11MOhm base(-) to collector(+) and 7.05MOhm base(-) to emitter(+), OL both B-C and B-E in reverse orientation

Finished with LSK170BL (Q1) and LSJ74B (Q2) - they also appear to be good
Q1 is N-Channel, so I measured with positive lead of my multimeter on gate
Q2 is P-Channel, so I measured with negative/common lead of my multimeter on gate

Q1 measured 0.673V gate(+) to source(-) and 0.671V gate(+) to drain(-), OL on both G-S and G-D in reverse orientation
Q1 measured 7.40MOhm gate(+) to source(-) and 7.39MOhm gate(+) to drain(-), OL on both G-S and G-D in reverse orientation
Q1 measured 35Ohm in either direction, source to drain

Q2 measured 0.699V gate(-) to source(+) and 0.700V gate(-) to drain(+), OL on both G-S and G-D in reverse orientation
Q2 measured 7.79MOhm gate(-) to source(+) and 7.81MOhm gate(-) to drain(+), OL on both G-S and G-D in reverse orientation
Q2 measured 37Ohm in either direction, source to drain

All this to say: I seem to have learned nothing, and whatever is broken, is still broken!
blairfrischx
 
Posts: 23
Likes: 0 post
Liked in: 8 posts
Joined: February 7th, 2014, 8:11 pm
Location: Columbus, OH
Country: United States (us)

Re: Balanced Alpha10 (ɑ10) Build Log

Postby amb » January 25th, 2018, 11:33 am

Maybe the problem is with the passive parts? Soldering problems?
User avatar
amb
Site Admin
 
Posts: 10378
Likes: 175 posts
Liked in: 482 posts
Joined: March 29th, 2010, 9:49 pm
Location: Sunnyvale, CA
Country: United States (us)

Re: Balanced Alpha10 (ɑ10) Build Log

Postby blairfrischx » January 25th, 2018, 12:03 pm

amb wrote:Maybe the problem is with the passive parts? Soldering problems?


My thought is, maybe when doing initial setup on the board I accidentally shorted V+ to G or V- to OUT, since the pins are so close together, and damaged a component that way.The underside of the board is getting more and more ugly as I try to desolder components, but I did notice that the middle pin of VR1 has what looks like some damage. See pictures of the board in its current state below.

Bottom
Top

I have already tested all of the resistors in the circuit; they tested at the correct values and not OL. None are obviously damaged.
I have not removed the capacitors to check their capacitance separate from the circuit, but I may do that tonight. None are obviously damaged.
That only leaves the trimpots. I'm not sure exactly how to troubleshoot them, but it seems like VR1 could be a culprit.
blairfrischx
 
Posts: 23
Likes: 0 post
Liked in: 8 posts
Joined: February 7th, 2014, 8:11 pm
Location: Columbus, OH
Country: United States (us)

Re: Balanced Alpha10 (ɑ10) Build Log

Postby amb » January 25th, 2018, 3:26 pm

To test the trimpots:

For VR1, place one DMM probe on the center pin and another on one of the outer pins. Set the DMM to ohms mode. Turn the trimpot adjustment screw and you should see the resistance increase or decrease. Turn it until it read 0Ω (or very close to it), then check the resistance from the center pin to the other outer pin. It should be about 100Ω. Now turn the trimpot screw the other way and the opposite should happen.

For VR2, note that the PCB shorts the center pin to one of the outer pins. So measure the resistance from either of those pins to the remaining outer pin. As you turn counterclockwise, the resistance should increase until it's 2KΩ at the end of travel. You then turn it clockwise and the resistance should decrease to 0Ω.

For both trimpots, after you're done testing, reset it to the fully counter clockwise position for initial setup again.
User avatar
amb
Site Admin
 
Posts: 10378
Likes: 175 posts
Liked in: 482 posts
Joined: March 29th, 2010, 9:49 pm
Location: Sunnyvale, CA
Country: United States (us)

Re: Balanced Alpha10 (α10) Build Log

Postby blairfrischx » January 26th, 2018, 4:25 am

My panels came in yesterday, so I put them on the preamp. It looks really nice - kind of deceiving, since it isn't finished yet!

Also I didn't have #4-40 flat machine screws to use with the XLR jacks and IEC inlet, so I used the pan head screws that I had available. It looks alright, but I might change it later to be more flush with the connectors.

I also noticed that the LCDuino cables are a little more strained than I would normally like, due to the offset in the boards. I set the length of those cables in my original unbalanced α10. I might change it later, but for now it works.

Please ignore the missing α20 module and the wires sticking out the back. Those wires are my test leads for OUTL, OUTR, and GND from the two α10 backplanes. They'll go away once I'm done trying to set things up.

Front w/ Top On
Front w/ Top Off
Rear w/ Top On
Rear w/ Top Off
Overhead View
Stack w/ β22 + σ22


I didn't get around to doing any troubleshooting of the problematic α20 module last night

I also realized once I had everything set up and could see the LCD backlight around the front panel cutout, that I really need to do the tape mod that AMB mentioned in turbotoy's thread:
amb wrote:On my α10, I placed a strip of black friction tape around the inside of the LCD display cutout, so that the backlight does not shine through the gap. You can't see the tape, and makes it look more "right". Just FYI.

2 people like this post.
blairfrischx
 
Posts: 23
Likes: 0 post
Liked in: 8 posts
Joined: February 7th, 2014, 8:11 pm
Location: Columbus, OH
Country: United States (us)

Re: Balanced Alpha10 (α10) Build Log

Postby blairfrischx » January 27th, 2018, 10:35 am

VR1 with trim set completely clockwise:
Center pin towards R8 = 2.3Ohm
Center pin towards R7 = 108.4Ohm

VR1 with trim set completely counter-clockwise:
Center pin towards R8 = 108.5Ohm
Center pin towards R7 = 2.3Ohm

VR2 with trim set completely clockwise:
Between outside pins = 2.112kOhm

VR2 with trim set completely counter-clockwise:
Between outside pins = 1.5Ohm

So both of those seem to check out.

Now that I had enough things removed from the circuit, I also tested C1 and C2 (should be 100uF each). From multiple samples, both tested around 108-110uF. So that's okay.

I couldn't measure C3 (47pF) or C4 (1uF) without removing them from the circuit. Also I'm not sure if my multimeter can even read 47pF, so I left those two alone.

And that brings me to.... nothing! There is nothing in the circuit which I haven't tested at this point, except for the PCB itself and the solder connections.

Since troubleshooting the PCB to check continuity on the traces is pretty tedious, I decided to reflow all of the remaining solder connections, put the transistors back in, and plug it back in to my second ɑ10 backplane to see if the problem magically disappeared. And now it seems to be working. :roll: I'm going to run it through all of the initial setup, let it sit for a while, and see how it's behaving. But it no longer has the same problem it did.
blairfrischx
 
Posts: 23
Likes: 0 post
Liked in: 8 posts
Joined: February 7th, 2014, 8:11 pm
Location: Columbus, OH
Country: United States (us)

Next

Return to Miscellaneous

Who is online

Users browsing this forum: No registered users and 8 guests

Quick Links

AMB audio DIY resources
AMB audio shop