[haejuk]

That does make sense.  I was wondering why it only had 3 contact points.  I guess I am back to the start for the power switch search.

 

[joeyjojo]

Quote:

Can anyone answer my noobie question? How much would all of the parts for 3 boards cost?

 
Including various import taxes in the UK I paid £78.21 per board a few months ago. $100/board is a very good estimate.
 
The same calculation for a sigma22 is £62.25.
 
jcx - are you saying 2x beta22 + 2x sigma22 is better or worse than 3x beta22 + 1x sigma22? The cost difference is small (£280.92 vs. £296.88) and if dual mono was better don't you think AMB's own beta22 would be of this design?
 
I'm interested because I currently have a passive ground build but have parts left over for 4 b22 channels and/or 1 s22. I don't believe adding any of them will produce an audible difference. I suspect Ti Kan uses a 3 channel because it gives him a sense of satisfaction when he imagines the currents cancelling on the supply rails

 

[cfcubed]

Think your queries WRT topologies is well-covered ground here & suggest searching instead of re-hashing the discussion.
You can search for jcx's posts in particular to locate the various discussions/posts for this.
 
After building, owning, comparing and/or hearing all but a dual-mono beta22 build (e.g. 4-ch, 3-ch, 2-ch w/sigma22, 2-ch w/simple LM-series regulation), IMO it's difficult to hear differences amongst them.  Probably because of the superior design, PSRR, etc of beta22.  I've even built/sold a nice 4-ch build & kept my "budget" beta22 after directly comparing them w/same source & cans and not being able to reliably discern a difference between them (w/my hearing & music - all caveats apply).
Quote:

    jcx - are you saying 2x beta22 + 2x sigma22 is better or worse than 3x beta22 + 1x sigma22?

 

[joeyjojo]

Thanks cfcubed, your experiences back up what I expected on the audible differences between these configurations - they are probably too small to justify $100+ expense, if they're audible at all.
 
The theoretical benefits are very real (current cancellation on supply rails with 3/4 channel primarily) but using such a config in real life is I think more for the satisfaction of knowing than anything else. This is why jcx calls it a "fad" I think.
 
I haven't searched through jcx's posts, but he seems to be saying that dual mono is somewhere between passive ground and active ground. I was just pointing out that the difference in price between dual mono and active ground is actually very small (5%) so not doing active ground and using 2 sigma22s is not justified by price. If one did as he suggests and gets a cheaper power supply that probably changes things.
 
To say using three B22 boards "is not technically justified" is simply false however (this has been done to death). However it probably is mainly for the warm feeling when you imagine those quiescent power rails. This is why AMB says that 2 channels is a world class amplifier, 3 is the icing on the cake (but for a 33% cost hike).

 

[OJNeg]

Let me get this straight. The dual mono design assumes you have a PS board for every amp board correct? This would mean 2 beta boards and 2 sigma boards. A fully balanced build means you would have 4 beta boards and 2 sigma boards, correct?

 

[cfcubed]

Quote:

Let me get this straight. The dual mono design assumes you have a PS board for every amp board correct? This would mean 2 beta boards and 2 sigma boards.

Yes if choosing to use the "customary"/"default" PS to power them (σ22).
Quote:

A fully balanced build means you would have 4 beta boards and 2 sigma boards, correct?

That's one option but a single σ22 will power a 4-ch / balanced β22 for headphone duty.

 

[OJNeg]

Quote:

Yes if choosing to use the "customary"/"default" PS to power them (σ22).
That's one option but a single σ22 will power a 4-ch / balanced β22 for headphone duty.

Doesn't a truly balanced design have to run off of separate power supplies?

 

[jcx]

"dual mono" completely isolated amplifiers for L,R channels With 4-pin "balanced" headphone connector/cabling separating R,L driver wiring completely cures any "cross contamination", common gnd coupling issues between the channels - R,L channel load currents never "mix"
the signal gnds are usually still connected in the source, so mains-sec leakage current can still flow in the signal gnd path - use split bobbin ps xmfr with lower parasitic C than toroids, heavy shield braid coax for signal IC, or balanced input with pin 1 gnd carrying the leakage current
 
"true balanced" has a few interpretations - bridged output "4-board" B22 doesn't meet the strictest definition but gives some of the advantages - common mode rejection happens but eats into dynamic range if the Vcm is at all large, and in Class A the bridged amps do draw ~constant current from the ps
 
'balanced" doesn't really say anything about number of power supplies - some versions of balanced/bridged amps mitigate some power supply effects so a "balanced" design may give some of the advantages of multiple isolated supplies
 
a "balanced" amp with bridged output "bypasses" common gnd impedance issues by not having load current flowing in the gnd
 
a "true balanced" input amp even with a single ended output, even single supply for both channels, can reject common gnd impedance error signals by virtue of its input signal common mode rejection - if it is wired correctly
each input signal diff receiver should translate the signal input for that channel to the gnd/return terminal of the headphone driver for the same channel
 
 
with its output cascode, filtered input circuitry giving ridiculously good PSRR I really don't think the B22 "needs" the cancellation of ps load current variation at the price of doubling the load - you could easily lose more to the effects of the heavier load
 
 
and as mentioned above, well executed B22 projects 2,3,4 board, 1,2,4 power supplies all can be excellent, differences can be made inaudibly small in most home listening situations

 

[joeyjojo]

Thanks jcx.
 
Quote:

the signal gnds are usually still connected in the source, so mains-sec leakage current can still flow in the signal gnd path - use split bobbin ps xmfr with lower parasitic C than toroids, heavy shield braid coax for signal IC, or balanced input with pin 1 gnd carrying the leakage current

 
Can you please explain this bit? Which transformers are you talking about?

 

[amb]

That does make sense.  I was wondering why it only had 3 contact points.  I guess I am back to the start for the power switch search.

You can use a normally-open momentary switch as a "trigger" for the ε24 power switch driver circuit, which then uses a relay to switch the power. With the ε24, the momentary switch can have low voltage and current ratings because it doesn't "see" the AC line. The relay must have contacts rated for the desired voltage and current.

 

[haejuk]

Thanks Ti.  I had found the e24, but I don't really want another board and another transformer in my build.  I am not super experienced at wiring stuff up and the voltage selector is already giving me a headache.  So I have just decided to go with a simple DPST switch for my power.
 
I do have to say that the e24 with the pcb for its transformer would make a really cool addition to any project and I can't rule out adding it in after I finish if I can wrap my head around how to wire it in with a voltage selector.

 

[jcx]

Quote:

Thanks jcx.
 
 
Can you please explain this bit? Which transformers are you talking about?

I was talking about the power supply transformers, different construction changes how well they couple or isolate from the wall power higher frequency noise - determines how much leakage current your system has in these parasitic gnd loops
 
to see part of the reasoning used in "gnd contamination" analysis see http://www.diyaudio.com/forums/diyaudio-com-articles/163575-audio-component-grounding-interconnection.html - referenced on AMB's site too
 
power entry line filter modules also add capacitive line leakage currents which the the signal gnd system has to carry - in single ended inupt amps always use the lowest resistance braid/shield RCA coax
 
 
"true balanced" input requires additional circuitry to make a differential signal receiver that rejects common mode V - audio input signal transformers are simple, pricey for better performance "full audio" bandwidth, low low frequency distortion
with the "balanced" B22 just using 4 boards with all the same positive gain any common mode V is amplified by that gain, eats into signal output dynamic range, but does reduce common mode output heard in the headphone drivers to the fraction of the gain resistor tolerance/mismatch

 

[haejuk]

I got some parts for my Beta22 over the past week, but I have a question about locking TRS jacks.  I got a switchcraft locking TRS jack, and tried plugging something into it.  It doesn't really seem to lock.  In fact it takes less force to pull the plug out than the non-locking TRS jacks on my current amps (I am not pushing the unlocking mechanism at all).  Did I get a defective jack or do I need a special connector to get it to lock?  I know this isn't the perfect place to ask, but I figure anyone looking here might have some experience with this.
 
Edit: Resolved on another forum.  Apparently "locking" TRS jacks may not lock super well sometime, so I ordered a replacement just in case.

 

[haejuk]

I just finished populating all my boards!  I am going to start casework this weekend.  When they are mostly cased up I will start all of the initial checks on them.  I am just waiting on my mini-grabber adapters so that I don't accidentally short anything.

 

[DefQon]

Is it easy to convert a 2 channel passive ground based b22 to a 3 channel active ground or even a balanced setup? (Excluding the fact that more board's for each channel and ground is required).