I just finished building an Alephj and it sounds really nice, but I am spoiled by this XA252. It beats anything I have heard or built so far. More people need to build this one, RusselC has the right idea of building multiples, and honestly if you could only build one amp the XA252 would be it. I just really enjoy the whole building process(plus who doesnt enjoy their wife asking why you need 3 preamps, 4 DACs and 4 Amps when you never listen to music anyways)
I have come across something that I have not seen in the past
Under load , there is a DC spike / instability - almost like the amp starts to oscillate
The amp is perfectly stable under the following conditions
1. no load and open input
2. no load and input signal at any level
3. load with no input signal or low level input signal
BUT as soon as I try and increase the volume , both channels behave badly - right more than left
I initially thought this was more a glitch with the output protection circuit I have - I am using solid state mosfet based protection from Mirand audio which I used in my F4 too and they have been very reliable. To test, I bypassed the circuit and I have not heard such screechy noises from a speaker before - DC was about 2V on the output. As soon as I dropped the volume , all back to normal. I would blame DC on the source but with load disconnected, I can crank the input signal to max and DC offset sits stable at a few mV
The test speakers are nothing special or a difficult load - just Bose 201s that I picked up for testing amps and didn't care if I blew them up. I had not appreciated that there was any problem as it allows me to go to 1/3 of volume which is plenty in the nearfield test space - i just happened to try to go louder and then this all started to happen
Under load , there is a DC spike / instability - almost like the amp starts to oscillate
The amp is perfectly stable under the following conditions
1. no load and open input
2. no load and input signal at any level
3. load with no input signal or low level input signal
BUT as soon as I try and increase the volume , both channels behave badly - right more than left
I initially thought this was more a glitch with the output protection circuit I have - I am using solid state mosfet based protection from Mirand audio which I used in my F4 too and they have been very reliable. To test, I bypassed the circuit and I have not heard such screechy noises from a speaker before - DC was about 2V on the output. As soon as I dropped the volume , all back to normal. I would blame DC on the source but with load disconnected, I can crank the input signal to max and DC offset sits stable at a few mV
The test speakers are nothing special or a difficult load - just Bose 201s that I picked up for testing amps and didn't care if I blew them up. I had not appreciated that there was any problem as it allows me to go to 1/3 of volume which is plenty in the nearfield test space - i just happened to try to go louder and then this all started to happen
same with and without jumper at negative input leg?
tried with any other source/preamp?
are rails stable, how's GND routing arranged?
I mean - you have plenty of things in that box ........ it's easy to make (even systematical) error when dealing with so many small sub-assemblies
tried with any other source/preamp?
are rails stable, how's GND routing arranged?
I mean - you have plenty of things in that box ........ it's easy to make (even systematical) error when dealing with so many small sub-assemblies
What???
WHAT???
That is honestly up side down and horrible mistreatment of your soul, the very fine audio equipment and your own ears.🙃
They are missing out of the gigantic mountains of absolutely fantasticly good sounding music (old and new) that you can access with a click or two.
Tidal
Spotify
Youtube
Do your mortal soul, and your ears the musical favor of a lifetime: Start subscriptions to all three right now! Enjoy! The most happy and best sounding music days EVER are really, here and now!
🎷🙂🎸
Just sayin…So Start playin…
never removed negative jumper - i can try this, neither have i tried a balanced source as i don't have one in my test space.
no - just the trusty CDP with internal volume - i typically use this for all amps but can try something else
yes, checked rails - all stable
Ground for amp is mostly through the cap multiplier - the initial rectified ground goes to the ground plane joining the caps , this is connected to the input ( Ground) of the cap multiplier and all rest of ground reference comes from cap multiplier. = = it is easy enough to bridge the grounds to test behavior
It must be simple dodo mistake that is making the output behave badly - odd that bias etc is rock solid.
depending of input signal - SE or Bal, that implies is neg. input grounded or not, and goes without saying
I was asking is there any difference having JP1 shorted vs. not shorted - that one when shorted enabling DC coupling of neg. input, removing last non-Kosher detail
Now, you can also make quicky measurement of gain of amp;
no special signal quality needed - even laptop/PC audio output is good enough for relative measurement
use 1KHz sine, set 100mV at amp input (DVM set to Vac good enough) then measure what's VAc on output
hope everything, regarding parts positioning and values, exactly per "straight MOS" schematic
I was asking is there any difference having JP1 shorted vs. not shorted - that one when shorted enabling DC coupling of neg. input, removing last non-Kosher detail
Now, you can also make quicky measurement of gain of amp;
no special signal quality needed - even laptop/PC audio output is good enough for relative measurement
use 1KHz sine, set 100mV at amp input (DVM set to Vac good enough) then measure what's VAc on output
hope everything, regarding parts positioning and values, exactly per "straight MOS" schematic
Or is it as preventive action to safe the output stage semi's and passive parts in case output Fet's are misbehaving, because of what ever reason (temperature....).
Transconductance is huge for these Fet's, when it goes wrong, it will go very fast.
it is as I wrote - Savior of our Souls - pretty much to prevent everything being melted.....in case of anything wrong with load, either due to short on output or massive clipping, clamping gates together and shutting them pucks of
I mean - amp is a bastard between a Hummingbird and Welding machine, and fuses are sloooooow
I mean - amp is a bastard between a Hummingbird and Welding machine, and fuses are sloooooow
Thanks ZM,
I've almost completed the PCB design for the full MOS version.
BOM is ready to be ordered in a couple of days.
As this amp is low wattage, meaning high risk for clipping when in the right mood; I've added a simple effective clipping indicator on the PCB :
This will again reduce risk for protection circuit to be triggered.
Power dissipation wise, this is similar as my current class AB monoblock (~110W/monoblock).
Really looking forward to first sound of this amp, for me, a first Pass like class A amplifier.
I've almost completed the PCB design for the full MOS version.
BOM is ready to be ordered in a couple of days.
As this amp is low wattage, meaning high risk for clipping when in the right mood; I've added a simple effective clipping indicator on the PCB :
Power dissipation wise, this is similar as my current class AB monoblock (~110W/monoblock).
Really looking forward to first sound of this amp, for me, a first Pass like class A amplifier.
"amp is a bastard between a Hummingbird and Welding machine" - ZM quote from above.
This will forever change how I look at the amp 😆
Still absolutely loving it!!
This will forever change how I look at the amp 😆
Still absolutely loving it!!
Before I start to populate XA252 SET with FQA36P15/FQA28N15 mosfets using +-24V rails then I want make sure that everything goes smoothly. 🤣
ZM mentioned that I can use the same PCB with mosfets when using 24V rails and no change is needed?
To test this XA252 without mosfets then I short output G and S. Then Iq pot all the way CCW which means 0 ohm or 500 ohm? Papa already gave me a lesson about that with F5m because some pots with CCW have other values.
Is there also a point to test it without Q1 and Q2? If yes then what values and where to test them?
R6 - 16k in schematic. I have 16,2k ohm.
BD140 replaced with TTA004b
BD139 replaced with TTC004b
Anything else I should know or to pay attention for?
ZM mentioned that I can use the same PCB with mosfets when using 24V rails and no change is needed?
To test this XA252 without mosfets then I short output G and S. Then Iq pot all the way CCW which means 0 ohm or 500 ohm? Papa already gave me a lesson about that with F5m because some pots with CCW have other values.
Is there also a point to test it without Q1 and Q2? If yes then what values and where to test them?
R6 - 16k in schematic. I have 16,2k ohm.
BD140 replaced with TTA004b
BD139 replaced with TTC004b
Anything else I should know or to pay attention for?
Go back a couple few posts, zenmod had me testing with board off mosfets and Q1 and Q2 removed.
I have isolated the problem to the capacitance multipliers. ZMs drek needs more current than they are happy delivering 
Bypassed to test and the amp is happy, now to build more classic CRC and forget uber filtering
Bypassed to test and the amp is happy, now to build more classic CRC and forget uber filtering
I try to answer to my question.
P2 set to max value, use DMM to measure it.
Remove Q1, Q2 now, to establish DC path for NFB, allowing FE to self-center, put short (piece of wire) connecting B to E pads of Q1 and Q2?
Then short output G and S.
Establish proper DC NFB path for servo action, simply shorting zeners between output gates - ZD3 and ZD4, then you'll have proper DC offset at output to check and sine at output.
If parts are OK, you'll have so-so close to 0V AC at output and close to red written check values on FE specific places/components (picture attached).
R19 - have values 300-400-500mV.
Is that so? 😆
pretty much no change needed for lowered rails, in that amount
don't bother with partial tests, if new build; better invest time and energy in triple checking everything, then put temporary 2A5 fast fuses in rails and power it on and set
as I said , if new build, why bother with procedure mainly intended for troubleshooting