Aleph J illustrated build guide

ok.
the input is still grounded.

You wont get any sound if the input (center post of RCA) is grounded. For future tests, just make up some RCA plugs, grounding center to ground lug. To temporarily short inputs, like when setting bias on some amps, or as here grounding input to see if amp is causing hum....with inputs grounded if it does not hum, problem is generally upstream, if hums with inputs shorted its in amp generally. then you can remove shorting plugs and proceed.

It will not play music with center post grounded...only the ground tab should be going to ground, input should go in amp....it is in "mute" status.

Russellc
 
Last edited:
You wont get any sound if the input (center post of RCA) is grounded. For future tests, just make up some RCA plugs, grounding center to ground lug. To temporarily short inputs, like when setting bias on some amps, or as here grounding input to see if amp is causing hum....with inputs grounded if it does not hum, problem is generally upstream, if hums with inputs shorted its in amp generally. then you can remove shorting plugs and proceed.

It will not play music with center post grounded...only the ground tab should be going to ground, input should go in amp....it is in "mute" status.

Russellc

sry Russelc, I dont know where my mind was at the moment i wrote the response (probarly i a bottle of mr. heineken 😀) but the grnd wasnt shorted when i wrote my response.
i will do some troubleshooting later tonight.
 
I'd exchange the FETs in Q6 and Q7, that gives you tighter matching between pairs that need to be matched. Q5 must match Q6 and Q7 must match Q8. Matching between those two pairs in not necessary.

Are you saying that with the FETs connected directly to the board they behave differently than when connected to the boards with short pieces of wire (also called flying leads)? If so, that is really odd.

Hello BobEllis. How can I match Q5/Q6 and Q7/Q8? I know how to match Q1A with Q1B, is the same method used?
TIA
 
Yes.

However, you only need match the Mosfets that share current - Meaning you may use a pair that has the same Vgs for the Current source fets (Q5, Q6) and a pair that has a different Vgs for the power output (Q7, Q8).

Yes, a matched Quad per channel, or a matched Octet for a stereo amp is more convenient, it isn't necessary.

Hello Jim, I would like to match q5/q6 and q7/q8. How might I test? Is it the same method that you showed me for Q1a and Q1b?
 
I've seen a number of photos that show Q1A & Q1B tied face to face with a tie wrap. Is this a common practice? What is the benefit? Should I do the same?
TIA,
John

About that: As I saw the Jfet tied on the build guide, I decided to do the same on my Aleph J. On the first board, I tried to tied them after soldering, so the Jfet are in contact only at their top end (on the second board, learning from my mistake, I tied them before soldering).
Will it cause any problem?
 
Stupid me!!!! This morning, I cleaned the mess left during the building of the amp and I found a forgotten capacitor.😱 It's the 0.0033µF cap which should be paralleled with the primaries of the transformer.
What's the purpose of this cap? My PSU seems to work well without it? Cancelling some noise when swicthing amp on or off?
Anymay, I'll put it in place this afternoon.

You can throw tomatoes, if you want.😀
 
cleaning garbage from mains and suppressing rectifying noises

leave it out or put in , your choice

for proper dimensioning of that part - there is a thread about gadget strictly made for that purpose - but I forgot it's name

hopefully someone will chime in , to refresh my ( faint ) memory


edit- not enough coffeine in blood - now I re-read your post ......

cap on primary side = cleaning garbage form mains ; spark decreasing caps - across (each) section of mains switch
cap on secondary side - what I wrote above
 
Last edited:
My right channel has become much weaker thank the left over time. I have tried different sources, and swapped speaker cables. The 2 channels were equal at the start.

Other than going over the bias and offset settings again, any suggestions as to what might have gone bad?