'AK Mods' SC Class A 15W Amplifier

[Spirit]

Mar 2, 2012 3:13:48 GMT @sandyk said:

Feeling the tiniest bit overwhelmed with all the stuff I'll need to learn to put together the Class A, having not constructed the AK Class A HA Cheers,Phil

Hi Phil
What do you intend using to give you control of volume?
Regards
Alex

Hi Alex,

I'm using the volumite (TPA's volume control for the Buffalo) currently, and will likely continue to use it.
Based on www.esstech.com/PDF/digital-vs-analog-volume-control.pdf, digital volume control as implemented here should be pretty damn good


Shaun - glad to hear it's all been fairly straightforward.
Main concerns are with casework and what gets earthed to where.
My DAC isn't earthed at all currently, as far as I'm aware.

Cheers,
Phil

[Deleted]

Mar 2, 2012 14:01:43 GMT Spirit said:

Mar 2, 2012 3:13:48 GMT @sandyk said:

Hi Phil
What do you intend using to give you control of volume?
Regards
Alex

Hi Alex,

I'm using the volumite (TPA's volume control for the Buffalo) currently, and will likely continue to use it.
Based on www.esstech.com/PDF/digital-vs-analog-volume-control.pdf, digital volume control as implemented here should be pretty damn good


Shaun - glad to hear it's all been fairly straightforward.
Main concerns are with casework and what gets earthed to where.
My DAC isn't earthed at all currently, as far as I'm aware.

Cheers,
Phil

Hi Phil

i only have the PSU casework earthed.

the amp boards 0V are taken back to the psu ground points (one amp board and one PSU per channel) then back to the chassis, the chassis is then earthed to the IEC earth input.

same for both channels.

i also have a Alex's dual mono pre so the only place where each channels grounds meet for the pre and power are at their respective PSU chassis star ground points (IEC ground input)

yup even have signal and ground switching on the pre (for the super paranoid only)

I've had no issues with running the amp chassis UN earthed whatsoever. it's running as quiet as a mouse (well quieter than a mouse in reality). i have included an earth point on the amp chassis just in case but it's never been used.(but looks niiiice)

Alex may have a different way but the above worked well for me.

so it's all a big STAR man waiting in the sky.

hope that makes some sense

am i right in thinking that the Buff Volume control is errr unity gain?

worth a try but i found my DCB1 hotrod a little lifeless compared to Alex's pre. but that's just my opinion and i may be wrong.

Lets see how that spins

interesting idea though

take care

[Deleted]

Here you go Phil

Direct from Picasso’s studio

how i've done mine is as follows



Nothing tooo complex and as i said the Amp chassis is not grounded.

My PSU may help you to see how things are terminated



If you are running your Buff in an un grounded case with no issues then it should work just fine with Alex’s Class A.

Hope that helps.

Take care

[Deleted]

Another source of AD744s.
This time the better metal can version which would be best soldered in, although the leads could be formed for insertion into a quality I.C. socket.

www.ebay.com.au/itm/ws/eBayISAPI.dll?ViewItem&item=250830452698&ssPageName=ADME:L:OC:AU:1123

Attachments:

[Deleted]

Mar 3, 2012 0:04:28 GMT @sandyk said:

Another source of AD744s.
This time the better metal can version which would be best soldered in, although the leads could be formed for insertion into a quality I.C. socket.

www.ebay.com.au/itm/ws/eBayISAPI.dll?ViewItem&item=250830452698&ssPageName=ADME:L:OC:AU:1123

Hi Alex

these seem to be a pretty good price for what you get.

how much better are the metal can verity over the usual plastic types?

nice find

take care

[Deleted]

how much better are the metal can verity over the usual plastic types?

Hi Shaun
Good question. At the very least they should have less drift with temperature due to better thermal management. These appear to be a fairly new fleabay listing.
I also found 4 x 2N2920A for a reasonable price too. They usually have a little lower HFE than the complementary 2N3811A and are what Erno Borberly used years ago with 2N3811A in some of his amplifier designs.Could be useful elsewhere too, due to fairly close matching.
Regards
Alex

www.ebay.com.au/itm/ws/eBayISAPI.dll?ViewItem&item=160746512783&ssPageName=ADME:L:OC:AU:1123

[Deleted]

Mar 3, 2012 1:30:08 GMT @sandyk said:

how much better are the metal can verity over the usual plastic types?

Hi Shaun
Good question. At the very least they should have less drift with temperature due to better thermal management. These appear to be a fairly new fleabay listing.
I also found 4 x 2N2920A for a reasonable price too. They usually have a little lower HFE than the complementary 2N3811A and are what Erno Borberly used years ago with 2N3811A in some of his amplifier designs.Could be useful elsewhere too, due to fairly close matching.
Regards
Alex

www.ebay.com.au/itm/ws/eBayISAPI.dll?ViewItem&item=160746512783&ssPageName=ADME:L:OC:AU:1123

Hi Alex

those AD744's are about the same price including postage as i paid for my plastic fantastics. what you say about thermal drift make sense so the metal cans look good value.

as do the 2N292OA at that price for 4 it's pretty cheap but as you say data sheet lists HFE as 150min and 600max so a little lower.

I'm not sure how they stack up to the LS devices SQ wise but I've been pretty impressed with the 2N3811/LM394H combo i have on my front end at the moment.

i guess for wills board it's worth waiting to source the LS but for anyone building now

I'd happily recommend the Motorola's I'm using as they seem like good devices to me.


thanks for the heads up

take care

[Will]

Thanks for being understanding, gents, and thanks for the kind messages.

Those AD744AH are a nice find, Alex, metal jackets looking better than black plastic bugs if nothing else!

Could someone do me a favour and do a 'what we've learn't so far' post here, so that all the good advice can be ready for when I get at Eagle again?

So far as I can remember, we are looking at provision for fb cap, having links to bring the offset corrector into circuit (helping with fault-finding) and using to220 78/79xx regs for the offset corrector as the t092 types have been found to be a bit flakey.

In fact, an updated offeset corrector circuit would be a great benefit, and save some time and thinking this end.

Cheers,

Will

[Deleted]

So far as I can remember, we are looking at provision for fb cap, having links to bring the offset corrector into circuit (helping with fault-finding) and using to220 78/79xx regs for the offset corrector as the t092 types have been found to be a bit flakey.

Hi Will

nice to hear from you

what I've learned is pretty much as above and not to solder transistors in the wrong way around.

I'm pretty sure that those odd readings and the flaky to92 was due to my silly mistake. but i had a few advantages over the all in one board.

1) i could disconnect the offset corrector and rule that out as a problem pretty easily. but to be able to do that and having the FB cap in place made it much easier for Alex to fix my error.

2) without the Offset in place and FB in use i could fire up the front end in isolation (out side of the amp case). i just used a jaycar reg for that and believe me it made life much easier than having to go through the in out in out and shake it all about with the board.

so for the sake of a few small links i think it may be worth the extra hastle errr for you (sorry on that) and will allow people to do on the spot fault finding much more easily.it also allowed me to get a good grip (under Alex's tutoring) on how the amp works and what to look out for measurement wise. ALL GOOD as people will have this amp for a while and it helps to have the basics of how to fix it if things go all unexpected.

TBH i found that having the offset corrector in place masked the front end problem i had. i was still getting 26mV offset which is still pretty good and better than some commercial designs so it took longer to pin down the fault. after the Boss found the fault and it was fixed I'm getting stupidly low DC offset (less than 1mV on both cannels)

Alex found it pretty quickly once the offset was removed and fb cap used.

hope that makes some sense.

also I've had a measure up of output device HFE

2SA1302/2SC3218

2SA1943/2SC5200

NJWO281G/NJW0302G

and the ON NJW devices seem to be a little closer to each other HFE wise. the others seemed like much more of a lottery so if we could have a conversation on that it would be good.

my money for getting some close matches without ordering loads would be on the NJW but the measurements are from a very small sample. my experience with the above devices is also very limited.

is there any difference SQ wise between devices?

any thoughts Alex?

great to hear from you Will

take care

[Deleted]

HI Alex...Shaun,
here is the SPK PRO built up but not fired up.
Alex what diff if any will the relay coil resistance at 777R
for my relay against the specified 650R have?
I could put either (3K9 or 4K7 in stock) ( or 4K buy in) across the coil for a closer value.
Regards,
Alan

[Deleted]

Mar 3, 2012 16:52:39 GMT drymdrum said:

HI Alex...Shaun,
here is the SPK PRO built up but not fired up.
Alex what diff if any will the relay coil resistance at 777R
for my relay against the specified 650R have?
I could put either (3K9 or 4K7 in stock) ( or 4K buy in) across the coil for a closer value.
Regards,
Alan

Hi Alan

Nice work (again )

the boards looking really good.

I'm unsure about the coil relay resistance thaaaang so maybe best to see what the boss has to say on it.

nearly there

take care HFSM

[Deleted]

[Deleted]

Hi alan

does it work?

take care

[Deleted]

Alex what diff if any will the relay coil resistance at 777Rfor my relay against the specified 650R have?

Hi Alan
It will be fine as is.
Regards
Alex

[Deleted]

also I've had a measure up of output device HFE 2SA1302/2SC32182SA1943/2SC5200NJWO281G/NJW0302Gand the ON NJW devices seem to be a little closer to each other HFE wise. the others seemed like much more of a lottery so if we could have a conversation on that it would be good.my money for getting some close matches without ordering loads would be on the NJW but the measurements are from a very small sample. my experience with the above devices is also very limited.is there any difference SQ wise between devices?any thoughts Alex?

Hi Shaun
IIRC, they are alll 30MHZ devices, and some are lower capacitance than others, but I think the ONs would be better due to the chance of closer matched pairs without buying a lot of extras.
Kind Regards
Alex

[Deleted]

Mar 3, 2012 20:41:07 GMT @sandyk said:

also I've had a measure up of output device HFE 2SA1302/2SC32182SA1943/2SC5200NJWO281G/NJW0302Gand the ON NJW devices seem to be a little closer to each other HFE wise. the others seemed like much more of a lottery so if we could have a conversation on that it would be good.my money for getting some close matches without ordering loads would be on the NJW but the measurements are from a very small sample. my experience with the above devices is also very limited.is there any difference SQ wise between devices?any thoughts Alex?

Hi Shaun
IIRC, they are alll 30MHZ devices, and some are lower capacitance than others, but I think the ONs would be better due to the chance of closer matched pairs without buying a lot of extras.
Kind Regards
Alex

Hi Alex

that was my thinking on the OP device front.

and not wanting to need to buy the worlds supply just to get a few close matches.

take care

[Deleted]

Alan



This pic of your latest board really shows a PCB duffer like me very clearly how it all builds up, another amazing job!

Syd

[Deleted]

Hi All

further adventures in fine tuning

after changing my Loudspeakers to PMC FB1+ i found that my system was in need of some subtle sound balancing.

after a few conversations with Alex i decided to try changing the value of the PRE/HA OP resistor.

mmmm to be honest i was not expecting much change and I'm sure that some will say that there should be no change in the sound at all. a fair point but one that does not really match what I've experienced in my system.

so i started with the 68R standard OP resistor which I've liked but it could sound a little (and i mean just a little) sterile at times with the new speakers.

I'd put this down to my choice of Mundorf cables which can sound a little like that at times.

so firstly using a series resistor on the OP to make quick change easier i started @ 75R total

i did hear a little difference in the tonal balance but not enough to write home about.

next up @100r and well things started to get interesting

the 100R gave the SQ a useful nudge towards a more balanced sound. it sounded a little richer with a less ''spot lit'' top end.

@120r i felt that things where pushed a little to far and the treble seemed to loose a little bite. things became a little toooo Rosy IMHO.

i ended up with 100R on the OP which IMHO to my ears In my setup gave the best balance to the sound. more natural without loosing the focus and top end sparkle.

soooo

to my ears in my system changing the value of the OP resistor allowed the sound to be nudged in one direction or another.

not huge changes but useful all the same.

i know that some people will not hear a difference and some will which is fair enough.

but

for me it's been well worth the few hours and few pence to have a play. IMHO the outcomes that people will have are going to depend on all sorts of things such as cables used/partnering front ends/LS but i would encourage people to experiment.

but for me the small change of output R value has enabled a more ''natural sound''

All good then

take care

[Deleted]

Shaun,

That's a handy experiment well executed. I was asking about this recently, so when I get to this age I'll have tinker too.

Cheers!

[Deleted]

so firstly using a series resistor on the OP to make quick change easier i started @ 75R totali did hear a little difference in the tonal balance but not enough to write home about.next up @100r and well things started to get interesting the 100R gave the SQ a useful nudge towards a more balanced sound. it sounded a little richer with a less ''spot lit'' top end.@120r i felt that things where pushed a little to far and the treble seemed to loose a little bite. things became a little toooo Rosy IMHO.i ended up with 100R on the OP which IMHO to my ears In my setup gave the best balance to the sound. more natural without loosing the focus and top end sparkle.

Instead of my usual 68 or 75 ohm output resistors, which enabled driving a 75 ohm coax lead to longer distances if needed, this time around I opted for 120 ohms output resistors as they best matched my AT W1000 headphones as well, and I hadn't at the time provided separate output resistor values for P.A. driving and headphone listening.I went the reverse order to Shaun, initially going down to 75 ohms . then 100 and 120 ohms again.
The sweet point for me was 100 ohms. In my setup , the results were exactly the same as Shaun described.
Instead of paying huge amounts on designer interconnects, why not try this simple and cheap method first ? It doesn't just work with headphones.
Alex

[Deleted]

The sweet point for me was 100 ohms. In my setup , the results were exactly the same as Shaun described.Instead of paying huge amounts on designer interconnects, why not try this simple and cheap method first ? It doesn't just work with headphones.

Hi Alex

i found 100R a nice spot for my K701 also but agree with the above (almost) 100R is the sweet spot.

I'd go for tuning with what people have IC wise and then spend huge amounts of money on designer IC's ;D ;D

to be fair I've tried lots of IC between pre and power and still found that they made a difference even after setting the OP resistor to 100R.

some (VDH the second ) sounded pretty dull @68r and 100R so playing with those IC still pays off IMHO.

but that's just my take and others may not see it that way.

as i say I'm getting a more balanced sound with the IC's i have @100r so it was well worth experimenting.

it's fairly straight forward to try and nothing lost so worth doing

IMHO

take care

[Deleted]

Hi Shaun and Alex,
update on my build.
Shaun wrote 18 days ago.................regarding the SPKRPRO board

Hi alan

does it work?

take care

Well yes after almost a whole day of hair tearing.............see avatar.........found the dry joint culprit and gave it a bit of a roasting.
Relay kicks in after a good five/six seconds delay, cuts off imm' AC sense is removed and 9 volt bat applied as per instructions shuts it down quick.
Only thing 1.5 bat applied not so good at opening relay sometimes does then try again and it don't work also more likely to work on just the one polarity.
So onto the main boards now after having a senior few moments with them at first power up.
Regards,
Alan

[Deleted]

Only thing 1.5 bat applied not so good at opening relay sometimes does then try again and it don't work also more likely to work on just the one polarity.

Hi Alan
You seem to have a minor problem. Perhaps the different relay ?
Do you have the mod from S.C. with one of the bipolar 47uF capacitors removed, and 47K replacing the original 2 series 22K resistors ? (both channels)
You could also try increasing the 22K to earth to 47K.
Regards
Alex

[Deleted]

Mar 21, 2012 0:29:35 GMT drymdrum said:

Hi Shaun and Alex,
update on my build.
Shaun wrote 18 days ago.................regarding the SPKRPRO board

Hi alan

does it work?

take care

Well yes after almost a whole day of hair tearing.............see avatar.........found the dry joint culprit and gave it a bit of a roasting.
Relay kicks in after a good five/six seconds delay, cuts off imm' AC sense is removed and 9 volt bat applied as per instructions shuts it down quick.
Only thing 1.5 bat applied not so good at opening relay sometimes does then try again and it don't work also more likely to work on just the one polarity.
So onto the main boards now after having a senior few moments with them at first power up.
Regards,
Alan

Hi Alan

Nice work with the LS protection board
I'm not sure about the 1.5V as i only tried the 9V recommended. maybe the 1.5V is running a little close to the boards DC switch off voltage. your relay is switching in after 5 seconds yup that's all good and so is mine.

i think that Alex did mention a simple resister value swap to improve sensitivity but I've not found it necessary. (beat me to it Alex)

do you have it running in test mode IE a link between DC+ and one of the AC inputs? i have mine running like that permanently to avoid running AC into the amp case.

senior moments

yes I've had a few but then again to few to mention ;D ;D

well quite a few actually.

i did find the build a little more challenging than the HA/PRE but it's all part of the fun and how we learn.

I'm looking forward to sharing the SQ with you.

my amp has daily use and IMHO it's got better which could just be the components settling in.

just awesome

a small hop now HFSM



take care

[Deleted]

I'm not sure about the 1.5V as i only tried the 9V recommended. maybe the 1.5V is running a little close to the boards DC switch off voltage

Hi Shaun
Actually, in part 3 of the SC 20W Class A article,
S.C. recommended a 1.5V battery for this test.
Kind Regards
Alex