'AK Mods' SC Class A 15W Amplifier

[Deleted]

‘’the offset corrector is working hard to correct excessive DC offset’’

Hi Alex

That would be my next guess and is I said in a previous post that when I had messed up on the MPSA orientation the offset corrector still worked but only down to about 26Mv.

One things for sure

Something is not right and a measure round in the way that you’ve suggested will no doubt reveal the answer.

In some ways it’s easier to track down a problem that’s present on both channels as it only involves half the work.

Take care

[Deleted]

As I've finally had time to at least get one board populated I thought I'd jolly up the thread with some pictures.

(if anyone spots mistakes on the board do shout out)



ooh, I think my right side is more photogenic



JLH clone Wars?

[Deleted]

Hi Sean/Alex,

 I've been following this thread with interest thinking how fortunate I was
that my two builds of both the HA/Preamp and 15 Watt Power Amps were excellent working models.
Well today I finally delivered my elder sons SandyK HA/PRE and 15 Watt as almost finished models.
Almost finished only in that they still require me to construct suitable face-plates for both.




This is it  ready to go out the door.


 

[Deleted]

Well got the amps in their new home all wired up ready to go.
Mmmm!! not what I was expecting as I'd had them running full blast at home and in perfect order.
Problem..............the speakers are out-putting what I describe as motor boating without the audio.
Audio source wave files from lap top into Musiland DAC into SandyK Preamp into SandyK 15 Watt into Wharfedale Diamond 10.1 @ 6 ohms.




Did'nt like that switched off quick and pondered.
Tried the HA/Pre..............perfect on HPs so we're OK up to the Pre out asked my boy if he had any other speakers we could try and he produced these...........




FIDEK  @ 8 ohms

worked perfectly.
So Sean looks like you have a buddy here.
Alex I noted that you quoted me as not using Will's boards..........................not correct as No. 2 build has Will's nice boards installed.
This coming week I'll take my No. 1 build 15 Watt up to my son's flat and hook it up to his Wharfedale's for a try out.
As my No. 1 build has my own hand cut boards installed and working now over a year all be it into 8 ohms it will be interesting if it drives the 6 ohm speakers without a problem.
All will be revealed come mid week FB point in my own boards is before the zobel.
Regards,
Alan

[Deleted]

Hi Alan

test votages and photo's?

i've been driving my 88 dB 6ohm Sonus Fabers with no problemo.

[Crispy]

Apr 20, 2013 16:34:53 GMT drymdrum said:

Hi Sean/Alex,


Almost finished only in that they still require me to construct suitable face-plates for both.




This is it  ready to go out the door.


 

Great work Alan - are you going to use a wood front like on your amps? (with the Sandy K square word calligraphy logo)

If you fancy an aluminum front panel I may be able to get you some cheap from my brother who works in an alloy factory.
It is just a thought, but if you do just send me the dimensions and I will see what I can do.

[Deleted]

Hi Alan
Are you using Will's PCB in this build, and do you have a build that didn't use Will's PCB ?
To save rewriting, this is what I just said to Shaun in an email.
Kind Regards
Alex

I am about to suggest to Alan, that if possible he tries his own build with the Input to the Offset Corrector at the output of the Zobel , against Will’s version where it is on the wrong side, both into the speakers that are showing the problem. Both mine and Sean’s measurements show that there is only a minor change in the output voltage from pin 6 of the I.C. yet Sean's offset is quite different between loads.
This suggests a voltage drop across the Zobel, which may, or may not be using 1mm wire.
They could also try removing the 1uf in the offset corrector and replacing the 3.3uF with it, although I would first try just removing the 1UF
Kind Regards
Alex

[Deleted]

Hi Alan
In addition to trying both amps,(assuming you have both versions) Are you able to lift up one end of R32 , and connect a 220K resistor under the PCB between the LHS of C16 and the bottom of the Zobel choke ?
We need to know whether the different location of the input to the offset corrector is the problem, or it just doesn't like 4 ohm speakers with the present Offset Corrector input circuitry.

Kind Regards
Alex

[Deleted]

Hi Alex,
see under quote of my first post of this problem your comments are exactly what my thinking was around the Zobel so as soon as I can
I'll follow your suggestions.

Not sure when as today is out, this is the charity fund walk today  with Neave's Kiltwalk Nutters
so I'll be running support for MGLW who is attempting the 13 miles whilst my two boys are walking the 26 miles.
Thanks to all here at RG for your support it is much appreciated the donated fund is standing this morning at a staggering £6461 with a bundle more to come.
Here is the link to the fund at Virgin giving, using Paypal makes it easy to donate

uk.virginmoneygiving.com/fundraiser-web/fundraiser/showFundraiserProfilePage.action?userUrl=Neaveskiltwalknutters&isTeam=true

So Sean looks like you have a buddy here.
Alex I noted that you quoted me as not using Will's boards..........................not correct as No. 2 build has Will's nice boards installed.
This coming week I'll take my No. 1 build 15 Watt up to my son's flat and hook it up to his Wharfedale's for a try out.
As my No. 1 build has my own hand cut boards installed and working now over a year all be it into 8 ohms it will be interesting if it drives the 6 ohm speakers without a problem.
All will be revealed come mid week, the  FB point in my own boards is before the zobel.

Regards,
Alan

[Deleted]

Alan
Did you mean to say that the FB point with your own boards is before the Zobel ?
If you meant to say that the input to the Offset Corrector was before the Zobel , that would be surprising, as I expect you would have gone by the schematic with your own PCBs..
Regards
Alex

[Deleted]

Hi Alex,
to clarify that point here is a pic of my 15 Watt board layout I've detailed the path to and from the OSC in red,
hope it makes sense to you.

I see it as input to OSC coming from junction of R11 18K and R31 6.8R and inductor and going to R32 220K.....R33 330K...to pin 3.

Output from OSC pin 6 to R34 10K....R35 75K....R11 18K
Making it on input to zobel.
Perhaps I should edit my previous post.

Alan

 

[Deleted]

Hi Alan
I sent this as an email, as the forum went down for a while
Regards
Alex

Unless there is some other mechanism at play, then your amp should also do the same into those speakers.
We need to rectify the error on Will’s PCB for best results, even if we are unaware of any problems.
I would recommend the resistor under the PCB for all builds using Will’s PCB irrespective, as I mentioned in the thread.
If that doesn’t fix this particular problem I hope that the capacitor changes in the Offset Corrector’s input will, as it has inbuilt time constants
that may somehow be interacting with low impedance speakers.
Regards
Alex

[Deleted]

Alex,

In addition to trying both amps,(assuming you have both versions) Are you able to lift up one end of R32 , and connect a 220K resistor under the PCB between the LHS of C16 and the bottom of the Zobel choke ?

Read more: rockgrotto.proboards.com/thread/7461/ak-mods-class-15w-amplifier?page=49#ixzz2R6R3HnYG

Just for clarification, I'm not doing anything now as I'm waiting for the official fix, do you mean the top terminal of the choke? (viewing from component side)

Cheers

[Deleted]

Hi Chris
I mean the end of the choke nearest the output terminal blocks. A resistor with uncut leads with a bit of spaghetti or similar over them should be adequate.
When Will designed the PCB, he didn't realise that he had the offset corrector connected on the input side of the Zobel coil instead of the speaker side. I didn't realise it either until we tried to find out what was happening with Sean's 4 ohm speakers.
My PCBs are only partly completed, so I hadn't noticed this error. If nothing else, the DC out should end up pretty much the same whether 8 ohm or 4 ohm speakers.

Regards
Alex

[Will]

Hope we can get this sorted, as the last thing anybody wants is an amp that needs kids gloves.

This is a close up of the schematic for the offset corrector.



Would it be worth Sean lifting the incoming leg (bottom) of R32, and linking it with wire to the output? This could then rule out whether having the Zobel before or after spur to the offset corrector, is a problem. This would be a test only to check.

Done a check on the front end schematic, and it checks out to the original. However, if someone with more available time wants hi-res pictures of the schematic and layout wants copys to check against - please email me.

[Deleted]

Hi Alex,
I see now where I've gone astray on the take off point going to the OSC.
The SCHA shows that point as being before the 68R output resistor which is how I implemented it with that build HA/Pre.
Unfortunately I did not appreciate that adding the zobel took the output to the other side on the 15 Watt.
Not able to get at the new build till Tuesday however I'll have a look at my first build and alter it, that way I'll be able to say one way or the other.
Mmmmm! just had a thought perhaps I should try it first as is then with correction made.
Regards,
Alan

[Will]

I've found one other difference between Alex's original schematic and the one for my PCB. The loudspeaker -ve returns to the input gnd, not circuit ground. Signal ground is 10R above circuit. This could be be checked out by linking out the GNDLIFT resistor, which was a link on the HA/PRE circuit.

[Deleted]

Ahhhh

Yes well spotted Will.

Luckily enough that's going to be quite simple to sort out and could solve the problem

Rather than linking out the ground lift which may cause hum problems it could be good to just hook the LS OP ground directly to the circuit ground. So just moving one wire and leaving the lift in. that may be a good way of checking things out without changing toooo much.

Let’s have a think on it.

Take care

[Deleted]

Apr 21, 2013 14:53:01 GMT Will said:

I've found one other difference between Alex's original schematic and the one for my PCB. The loudspeaker -ve returns to the input gnd, not circuit ground. Signal ground is 10R above circuit. This could be be checked out by linking out the GNDLIFT resistor, which was a link on the HA/PRE circuit.

Hi Will
That's certainly worth trying too, as the earth lift is not so critical when using separate power supplies.
The original S,C, design had the LS earth terminal returning to the "Star" earth at the bottom centre of the PCB. The earth lift resistor was a later addition by S.C. but the LS earth was still at the same central point. The earth lift resistor may not be needed unless there are hum problems, and could perhaps be reinstated as a much lower value than the original 10 ohms if there were resulting minor hum problems.
Regards
Alex

P.S.
We need to remember here that Will has done a great job in presenting this and the previous preamp/HA projects for our RG members, and designing some beautiful PCBs.
Very few major published DIY projects in the magazines don't end up with Errata published in later issues. The original design caused S.C. some major headaches, and many months of development time.
The earth lift resistor was a later addition, involving a modified PCB and changes made available to the PCB makers and kit suppliers.. They got fabulously low distortion and noise figures with the original PCB, but when they combined both PCBs in the same case, the noise figures were completely unacceptable. After trying things such as shielding the transformer, they had to reluctantly put the power supplies in another case. Niggles like this can be annoying, but they also play a large part in the learning process. The lessons learned here may even result in further small improvements over my present working version. Shaun and a few others, are also trialling component type changes that may lift the bar even higher.
Please keep tuned!

[Deleted]

Apr 21, 2013 14:02:21 GMT drymdrum said:

Hi Alex,
I see now where I've gone astray on the take off point going to the OSC.
The SCHA shows that point as being before the 68R output resistor which is how I implemented it with that build HA/Pre.
Unfortunately I did not appreciate that adding the zobel took the output to the other side on the 15 Watt.
Not able to get at the new build till Tuesday however I'll have a look at my first build and alter it, that way I'll be able to say one way or the other.
Mmmmm! just had a thought perhaps I should try it first as is then with correction made.
Regards,
Alan

Alan
As the both builds are identical in this area, perhaps you could try bridging out the earth link resistor on this later one in advance of the comparison ?
You could carefully solder a small loop of wire to both ends of the 10Rs, and then if there is no problem with that one at the comparison, snip the 2 loops with side cutters to see if the problem then arises.
Regards
Alex

[Deleted]

''P.S.
We need to remember here that Will has done a great job in presenting this and the previous preamp/HA projects for our RG members, and designing some beautiful PCBs.
Very few major published DIY projects in the magazines don't end up with Errata published in later issues. The original design caused S.C. some major headaches, and many months of development time.
The earth lift resistor was a later addition, involving a modified PCB and changes made available to the PCB makers and kit suppliers.. They got fabulously low distortion and noise figures with the original PCB, but when they combined both PCBs in the same case, the noise figures were completely unacceptable. After trying things such as shielding the transformer, they had to reluctantly put the power supplies in another case. Niggles like this can be annoying, but they also play a large part in the learning process. The lessons learned here may even result in further small improvements over my present working version. Shaun and a few others, are also trialling component type changes that may lift the bar even higher.''
Please keep tuned!


Hi Alex

I could not have put it better myself.

Will's done a great job with these and the many other boards that he's produced for us here and I for one am really grateful for all of the hard work that he's put into making this build possible. I like to think that with these Group builds that we are all in this together. With each build completed adding to the experience gained and help available to others. What I have learned from past experiences is that GB can sometimes need a little Re-fettling as the build process moves along. All pretty normal IMHO.

OK just a thought on the ground lift thingy.

I’m thinking that it may be a better idea to just move the LS ground lead over to the circuit ground which has a spare connector. That’s going to remove the connection from LS ground back to the Ground lift. That way we get to keep the ground lift and don't need to make any alterations to the board. That’s a quick job and keeps us from creating another problem with hum if the GL resistor is linked out (which may or may not happen).

Just a thought

Take care

[Deleted]

I’m thinking that it may be a better idea to just move the LS ground lead over to the circuit ground which has a spare connector. That’s going to remove the connection from LS ground back to the Ground lift. That way we get to keep the ground lift and don't need to make any alterations to the board. That’s a quick job and keeps us from creating another problem with hum if the GL resistor is linked out (which may or may not happen).

Just a thought

Hi Shaun
I have stared, and stared, at an A4 size copy of the PCB artwork, and unless I am missing the obvious, I can't see how we can do that without cutting a track that is on top of the PCB, which wouldn't be a good look.
Kind Regards
Alex

[Deleted]

Hi Alex

OK maybe it's my lack of a good explanation

From what I see looking at the artwork the loudspeaker negative connection returns to the ground lift resistor and joins up to the signal ground. Then both signal and LS ground are taken to the circuit ground via the Ground Lift.

Something like this



If we disconnect the Loudspeaker ground then the return to the signal ground from the LS is taken out of the circuit.

Then moving the loudspeaker ground wire over to the signal ground's spare terminal puts it back in circuit but after the ground lift.



So then we are back to having the loudspeaker ground tied directly to the signal ground

Sorry for the quick and dirty drawings

[Deleted]

Hi Shaun
But doesn't that then put the earth lift 10 ohms in series with the 100nF (C12) in the Zobel circuit ?

Regards
Al;ex

[Deleted]

Hi Alex

rats i missed that one.

oh well looks like plan A

take care