'AK Mods' SC Class A 15W Amplifier

[jeffc]

Thanks Alex,

That picture is what I've been pondering over now for a few days . Anyway, long story short I now seem to have the SPKPRO working OK, relay on after 5 sec delay and breaks on L/R with either +/- 1.6V DC battery. I have that cap removed and have reverted to using a link and a 47K to GND.

So.. problem #1 fixed.

Now back to the list

2. DC offset on either channel L = 1.6mV, R = 19.6mV (checked these and still close to this - are they that high to be much of a problem - or liveable with?)
5. Tp1/Tp2 voltages ~65mV with 500R trimmers bottomed out. With aluminium top cover in place and the amp on for ~40 min, essentially no change after lifting the lid and testing as quickly as I could..... so this is a problem)
6. Checked again and its actually the R channel heat sink (the channel with highest DC offset) that is the one that gets noticable warmer, R17 500R bias trimmer resistors not touched yet to set this

Hopefully these will be relatively simple fixes and I'll be done..... then fingers crossed for yahoooooooo....

cheers.. jeffc

[Deleted]

Feb 21, 2013 11:12:04 GMT jeffc said:

Thanks Alex,

That picture is what I've been pondering over now for a few days . Anyway, long story short I now seem to have the SPKPRO working OK, relay on after 5 sec delay and breaks on L/R with either +/- 1.6V DC battery. I have that cap removed and have reverted to using a link and a 47K to GND.

So.. problem #1 fixed.

Now back to the list

2. DC offset on either channel L = 1.6mV, R = 19.6mV (checked these and still close to this - are they that high to be much of a problem - or liveable with?)
5. Tp1/Tp2 voltages ~65mV with 500R trimmers bottomed out. With aluminium top cover in place and the amp on for ~40 min, essentially no change after lifting the lid and testing as quickly as I could..... so this is a problem)
6. Checked again and its actually the R channel heat sink (the channel with highest DC offset) that is the one that gets noticable warmer, R17 500R bias trimmer resistors not touched yet to set this

Hopefully these will be relatively simple fixes and I'll be done..... then fingers crossed for yahoooooooo....

cheers.. jeffc

Hi Jeff
Good to hear the speaker protector is now working correctly.
You appear to have a problem with the R channel, as the offset corrector should reduce DC Out to somewhere near 1mV or less.
Did you try fitting an electro in place of the CFB strap in that PCB? If not, cut the strap in the middle,
lift the cut ends up and quickly tack a 47 to 470uF electro across with the -VE to the earth side
You would be well advised to unsolder the I.C. for that channel (at least) and fit an I.C. socket.
Do you have -15V WRT earth at pin 4 , and +15 WRT earth at pin7 ?
We need to rule out the offset corrector itself. I presume you have 1K resistors fitted across the input terminals of both modules?
Measure the rersistance across R1 (15K) with the 1K fitted to the input socket.You should then see a little under 1Kohm.
What voltage do you have across the parallel 1 ohm resistors with the warmer heatsink ? Set for say 200mV initially,
and see that it doesn't go much more than 240mV when warmed up.
Kind regards
Alex

[Deleted]

Hi Will
Yes, as per the schematic.
BTW. I noticed in an earlier schematic that you drew showing the offset corrector,that you had +-24V rails. Any special reason why ?
Kind Regards
Alex

[Deleted]

Feb 21, 2013 12:21:44 GMT @sandyk said:

Feb 21, 2013 11:12:04 GMT jeffc said:

Thanks Alex,

That picture is what I've been pondering over now for a few days . Anyway, long story short I now seem to have the SPKPRO working OK, relay on after 5 sec delay and breaks on L/R with either +/- 1.6V DC battery. I have that cap removed and have reverted to using a link and a 47K to GND.

So.. problem #1 fixed.

Now back to the list

2. DC offset on either channel L = 1.6mV, R = 19.6mV (checked these and still close to this - are they that high to be much of a problem - or liveable with?)
5. Tp1/Tp2 voltages ~65mV with 500R trimmers bottomed out. With aluminium top cover in place and the amp on for ~40 min, essentially no change after lifting the lid and testing as quickly as I could..... so this is a problem)
6. Checked again and its actually the R channel heat sink (the channel with highest DC offset) that is the one that gets noticable warmer, R17 500R bias trimmer resistors not touched yet to set this

Hopefully these will be relatively simple fixes and I'll be done..... then fingers crossed for yahoooooooo....

cheers.. jeffc

Hi Jeff
Good to hear the speaker protector is now working correctly.
You appear to have a problem with the R channel, as the offset corrector should reduce DC Out to somewhere near 1mV or less.
Did you try fitting an electro in place of the CFB strap in that PCB? If not, cut the strap in the middle,
lift the cut ends up and quickly tack a 47 to 470uF electro across with the -VE to the earth side
You would be well advised to unsolder the I.C. for that channel (at least) and fit an I.C. socket.
Do you have -15V WRT earth at pin 4 , and +15 WRT earth at pin7 ?
We need to rule out the offset corrector itself. I presume you have 1K resistors fitted across the input terminals of both modules?
Measure the rersistance across R1 (15K) with the 1K fitted to the input socket.You should then see a little under 1Kohm.
What voltage do you have across the parallel 1 ohm resistors with the warmer heatsink ? Set for say 200mV initially,
and see that it doesn't go much more than 240mV when warmed up.
Kind regards
Alex

Hi Jeff

Good work on getting the speaker pro working.

thats a big part sorted.

Just to confirm what Alex has outlined above.

Pop that cap into the FB loop and disconnect the Offset corrector via those jumpers near the regs.

I’d double check that you have those regs in the right way round (yup one of my favourite tricks) just to be sure.

Getting that opamp out can be a PITB but I have developed a bodgers method of doing that without damaging the board.
1) Snip all of the legs off of the opamp. All of the legs are now free to be worked on.
2) Solder sucker out as much solder from each leg as possible.
3) Use a pair of surgical forceps to grip a leg whist heating from the solder side of the board.
4) Do any hole cleaning where necessary.
5) Works every time for me

when you resolder i'd use a socket in that spot as it just makes life soooo much easier should something go wrong.

I had the same offset problem with my build and the method outlined above by Alex made it much easier to fault find.

now i have less than 1mv both channels.

Good luck

Take care

[jeffc]

Do you have -15V WRT earth at pin 4 , and +15 WRT earth at pin7 ?
We need to rule out the offset corrector itself. I presume you have 1K resistors fitted across the input terminals of both modules?
Measure the rersistance across R1 (15K) with the 1K fitted to the input socket.You should then see a little under 1Kohm.
What voltage do you have across the parallel 1 ohm resistors with the warmer heatsink ? Set for say 200mV initially,

Hi Alex/Shaun,

Thanks....

Volatges OK on the pin4/7 of both opamps so no probs with the Vregs
Been bad and not checked DC offset with 1K resistors across input
With that fitted but no 4R-8R load on output (bad again ) , DC offset now 2.5mV, so better than 19.6mV ....
Resistance across the 15K R1 ~920R, so OK
Voltages across 1R about 188mV (R channel with initial offset issue - and warmer heat sink) and about 106mV (L channel - cooler heat sink)

So R channel might be OK for DC offset at least

However, L channel resistances you suggested I check are now wrong, will have to investigate this in more detail tonight as out of time now, hi ho hi ho its off to work I go .

cheers.. jeffc

[Deleted]

Hi Jeff

Good going on that

With the stunt speakers connected (loading the OP) I’d bet those DC offset figures are going to drop still further but looking good.

So now you need to clip a multimeter across those 1R resistors from IP to OP and measure the voltage then adjust with the trim pot R17 until you have 200mv. Your heat sinks are different temperatures because you have less bias current flowing through the LH channel.

Bet that when you have adjusted both channels to 200mv they will equal out out heat wise.

Not far now Jeff:-/

And it's cool stubby time

Cousin nearly ready

Take care

[jeffc]

Not far now Jeff:-/
And it's cool stubby time
Cousin nearly ready

Hi Shaun,

Really lookin' forward to that celebration stubby and thanks for the tips and support .

Now I know where to check bias current, hopefully adjustment to 200mV with the trimmer will be easy.
Just need to look into the L channel now that I thought was OK yesterday , hopefully its something silly and minor.
Last thing then will be front-end balancing to 0.0000000mV, now about 65mV on both channels even with lid on, although it might get quite a bit warmer inside the case with bias set higher, but I'm guessing based on advice to Alan/you from Alex a few pages back, that ultimately I'll need to remove the amp PCBs for resistor replacement to allow this to get closer to the ZERO target.

One other thing, I've got the SPKPRO currently connected to one of the +V JLH outputs and noted that this drops voltage another 0.1V compared to without it. This OK or should I leave the JLHs exclusively for the low-current amp sections and connent the SPKPRO to the JLH input side?

With some luck, you never know, I might just get sound happening from the new cousin downunder this weekend ......

cheers.. jeffc

[Deleted]

Feb 22, 2013 5:07:30 GMT jeffc said:

Not far now Jeff:-/
And it's cool stubby time
Cousin nearly ready

Hi Shaun,

Really lookin' forward to that celebration stubby and thanks for the tips and support .

Now I know where to check bias current, hopefully adjustment to 200mV with the trimmer will be easy.
Just need to look into the L channel now that I thought was OK yesterday , hopefully its something silly and minor.
Last thing then will be front-end balancing to 0.0000000mV, now about 65mV on both channels even with lid on, although it might get quite a bit warmer inside the case with bias set higher, but I'm guessing based on advice to Alan/you from Alex a few pages back, that ultimately I'll need to remove the amp PCBs for resistor replacement to allow this to get closer to the ZERO target.

One other thing, I've got the SPKPRO currently connected to one of the +V JLH outputs and noted that this drops voltage another 0.1V compared to without it. This OK or should I leave the JLHs exclusively for the low-current amp sections and connent the SPKPRO to the JLH input side?

With some luck, you never know, I might just get sound happening from the new cousin downunder this weekend ......

cheers.. jeffc

Hi Jeff
Did you replace the 2K0 resistors for the Current source transistor with 2K2 resistors ? (Q103 collector resistor)
To avoid stuffing around, it may also be worthwhile replacing the 120 ohm resistor in the VAS emitter (BC549C) with 130 ohm 1% MF resistors from Jaycar.
The SF11 diodes that I used may have had a lower forward voltage drop than most.
You shouldn't connect the Speaker Protector PCB to the JLH outputs.
Kind Regards
Alex

P.S.
150 + 1K in parallel underneath will give you 130 ohms

[jeffc]

Thank you Alex,

Answer, 2K or 2.2K, can't remember off hand sorry, will check and let you know.
I might have some 130R I think, but if not I need to go shopping at Jaycar anyway for solder wick I've run out of.
On the SPKPRO, I thought connection to a JLH might be non-ideal for various reasons, that's why I asked, but easy to change.
If I can resolve whatever is potentially amiss with the L channel...... I now have THE MAP to escape from the maze..... thanks for persisitng with me.....

cheers.. jeffc

[Spirit]

I'll be next

Will reply to your last email as soon as I get home Jeff - next Thursday.

[Deleted]

Hi Jeff

I guess that what I was trying to say (but being as clear as mud about it) was I had the same offset levels before connecting the stunt speakers. When I connected my stunts the DC offset just dropped away to almost nothing so that maybe worth checking before getting to the hair pulling stage.

if it's still 2mv then that’s not going to break anything and should be OK to try pushing sound through. After you set the OP bias just to make sure that all is well with the circuit.

The front end balance does need some work and changing resistors looks like the way to go but I’ve not changed my Diodes to SF’s so you’ll be one step ahead of me on that.
On the speaker pro

I have my SP connected before the JLH and running from DC in test mode which avoids polluting our nice clean front end supply OR RUNNING ac through the amp chasis.

One things for sure

You’ve got a lot closer to a working amp than I did with my build.

Alex had the patience of a saint with me and he fixed me up with lots of good advice. TBH when something works first time it’s great but the learning side suffers. When I had a small fault with my build I learned plenty from Alex. So a win win.

So just a small push to stubby time.

And another amp on the way down under.

Great stuff guys

Take care

[jeffc]

Hi guys,

Seems I had the input wiring to the L channel shorted somehow. Maybe at the RCA socket which I used a plastic washer on, but also a metal washer on the outside to provide a flat mounting surface rather than the bit-wonky wood surface left from my blunt spade bit. I'm guessing that with an imperfect hole, I had contact between signal-GND somewhere. Anyway, I'll see what's happening here and fix it. Good news is I now have resistances in the ballparks expected and exactly the same DC offset (2.5mV) with 1K fitted at input as the R channel........ so I now have functioning TWINS....

So as Alex suggests, replace that 120R on the emitter of BC549C (Q108) for 130R tomorrow when I get some .................. screw all back together again.. fit new RCAs ... wire... check.. double check... triple check.... balance the input transistors to 0mV... and .... test with speakers ... smile .... at last.... have that stubby.....and take a few glory pics... busy day..

OK one last push for the finish line... its been a marithon....

cheers.. jeffc

PS.. Wondering where you’ve been Phil.. amp build should be clear sailing for you, I’ve navigated most of the reefs and sand bars…

[jeffc]

Hi Shaun,

Beat me to this post and I'll take this info onboard for speaker testing and front end balancing. As you'll read from the post above, I'm almost set to go... and really appreciate all the encouragement and guidance over the past few weeks from both you and Alex, there's been a few times when I've really struggled and wondered whether I could get there or not....

cheers.. jeffc

[Deleted]

Feb 22, 2013 11:58:18 GMT jeffc said:

Hi Shaun,

Beat me to this post and I'll take this info onboard for speaker testing and front end balancing. As you'll read from the post above, I'm almost set to go... and really appreciate all the encouragement and guidance over the past few weeks from both you and Alex, there's been a few times when I've really struggled and wondered whether I could get there or not....

cheers.. jeffc

Hi Jeff

It’s getting exiting this end waiting for the new arrival.

You are on the home straight now with just a small push to cross the finishing line.

‘’there's been a few times when I've really struggled and wondered whether I could get there or not....’’

You and me both Jeff

I was at almost the same point of extreme self-doubt with my build but luckily enough I had Alex to keep me focused and give me a way of working through it.

I learned plenty and have a great amp to show for it.

but the up side is that if it does need a little attention you’ll know it like an old friend.

My build is over a year old now and always sounded and behaved impeccably so no worries on that front.

It seems like a long haul at the time but when I clapped ears on the amp the memories of the work put in just melted away.

I hope you like it as much as I do.

The best things are never easy are they?

Take care

[jeffc]

Hi Alex,

I have another circuit posted not very long ago, labelled 15W PER CHANNEL STEREO AMPLIFIER, the one in which Q112 is designated NJW0302G and Q114 is labelled NJW0218G with the output inductor network etc .... and Q103 is labelled BC559 and Q4 is labelled BC558B. In the circuit just posted, which I have a copy of too in my pile of stuff (not sure who posted it originally but Will I think), Q103 = T1 is labelled BC557 and Q104 = T4 is also labelled BC557. And I have a different circuit diagram image again with BC557 in both positions... Please tell me that whatever transistors are in these positions don't matter...........because I’ve used BC559 and BC558B....

And with 130R replacing 120R, I now get ~10mV cold on tp1-tp2, just checked again with the amp running for a while and quite warm, and its heading north, now ~15mV.. I’ll put the lid on for an hour and prey for some miracle. Wishing now I’d put a 150R in, which I was thinking of doing, to be able to drop this to whatever is going to work…

I'm now feeling sick in the stomach..... I really don't want to lift these boards AGAIN to change anything.... they are not commercial DIY types like the SPKPRO board with large lands easily amenable to parts swaps...

cheers.. jeffc

PS.. and I've just checked R3 and guess eff'n what, 200R and not 220R... spew time.....
No wonder I can't get this eff'n thing to work properly...........
OK no working amp this weekend by the looks, apparently I haven't been tortured enough yet....

[Deleted]

Jeff
Yes, the circuit I just posted was from Will. I used it mainly because it shows the offset corrector as part of the amplifier.
No, those transistors you mentioned are not critical. Just as long as you use the right polarity types. I think that Will has drawn that using the original published diagram. I will correct and repost the schematic.That's why I asked for advice of anything amiss.
The circuit from Will also had the older output transistors.
It sounds like the adjustment has gone too far, as the differences should decrease as it warms up and eventually change polarity
and increase in voltage, unless the trimpot is correctly adjusred.Did you have the same problem with the HA/preamp originally ?
We may need to check a few voltage readings if the trimpot adjustment doesnt corrrect this, but there certainly doesn't seem to be too much to worry about.Going to 150 ohms will make things much worse.Did you have in 200R or did you already have in 220R for the base resistor of Q104. ? I would have thought that where you are, you shouldn't have needed to replace the 120R with 130?
I will get back to you. In the meantime, if DC out is O.K. and the the speaker corrector is in line and the bias is less than 250mA, why don't you hook up those "sacrificial" speakers to see if you get sound ?
Regards
Alex
P.S.
The likely worst case scenario is that you may need to tack 1.5K across the 130R under the PCB to bring it back to 120 ohms, instead of trying to replace it again .
Did you have the trimpots set to 0 ohms before you started ? It should have been just a matter of winding in more resistance as it warmed up
Yes, I agree that the copper lands could be larger, but that is set by the software that Will uses.
I am not firing on all cylinders at the moment,and feel like I need to take a break, so please bear with me, or perhaps Shaun can assist here.

[jeffc]

Hi Alex,

Sorry, I'm pretty annoyed ATM, and amended my earlier post just now to reflect this.
Thanks for the info, steam is settling.. appreciated.. I just don't want to have to keep taking out the boards to change parts, which is a PITA.
Yes I used 2K2 for R4 but 200R rather than 220R for R3.
It was a crappy RCA socket, really crappy to the point of being shorted, that caused my L channel problem, better quality Jaycar ones now installed.
I just need to solder input wiring to these to speaker test....
I might do this now and see whether I get sound, which could cheer me up...

cheers.. jeffc

[Deleted]

Jeff
Leave the 200 ohms for now.
The 130 ohms by itself may be enough to provide proper adjustent if you don't get too confused by the trimpot adjustment.

It was a crappy RCA socket, really crappy to the point of being shorted, that caused my L channel problem

I suspected a s/c in that area, as it the offset should have been much higher without the 1K resistor across the input terminals.

Regards
Alex

[Deleted]

Hi jeffc,

This may help, I use this table to compare transistors:

users.tpg.com.au/gerskine/greg/PNP%20Small%20Signal%20Transistors.htm

[jeffc]

Hi again Alex,

Speakers connected....nothing else connected... amp case sitting directly on top of the PSU case, no Al lid on PSU case, put ear to speaker and there is hum, but I need to have my ear very close to the speaker before this is clear, touch the amp Al lid touching the heat sinks, and you hear this touching and if my finger is held there, hum gets a little louder. Connect crappy old test CDP, and hum is now very low, ear right on speaker cone needed to hear anything, nothing really, hit play and VERY VERY LOUD MUSIC.......... no preamp...

Feeling better....

It works......

Knowing this, I can rest easy and chip away for a little while longer to get everything optimized.

Sorry for getting a little OTT, as said last weekend, I’m close to exhaustion trying to build this amp amongst everything else expected of me at work and at home…..

cheers.. jeffc

[Will]

Circuit for the Class A PCB, with references to the manufactured pcb.



The values are taken from the circuit link provided after I asked questions about what the current values were.

These values and circuit idents (T4, etc) are reflected in the online BOM that was posted.

Any changes to values since then have not been updated as I don't know about them. Been too busy in real life to monitor the thread.

The 24V that was in the earlier circuit was a hangover from the labeling that is standard in the pcb software (there is no 20V) that's been changed in the above posted circuit.

[Deleted]

Hi Will
Due to the much colder climates where some of you guys live, the values needed to be changed to account for the considerably higher voltage drop of the SF11 and SF12
at those considerably lower ambient temperatures, in order to achieve complete balance .
Regards
Alex

[Deleted]

Feeling better....

It works......

Hi Jeff
Best news I have had all day.
Kind regards
Alex
P.S.
I think this may be the schematic you were working to ?
It was sent around the traps by email at the time. The change from 120 to 130 ohms was found necessary for a couple of U.K. members to achieve complete balance.
I think that Will may have been away at the time.

[Spirit]

Awesome work Jeff

I had a similar very loud moment with the SCHA - holding the output sockets onto the pads - when they connected I got very loud music.

[Will]

Hi Alex,

Thanks for that post. In all, I think that it will help prevent any confusion for future builders. I think most people like an 'absolute' when they build - this part, with this value goes in this hole. With the front end balancing being temperature dependent, knowing that these fixed values may need to be changed depending on local climate will help prevent frustration.

I think that a 'AK Mods Info' post is needed!

Doing a quick compare, the values in both circuit diagrams are the same, just not the 'you might need to change this' bits.

On a different topic, If you could do a definitive JLH RE post, I think a lot of people would really appreciate it (I know I would!)
Describing when you would or would not use the CL, how to make it effective at 5V (the crd), when to use 2200uF caps, when to use 4700uF, when they should both be low-esr, when to mix GP and low esr...

You are the world authority on the JLH RE, there is no denying it. So having a statement of best practice for the differing uses (digital supply, analog supply, 5V, 15V etc, etc) would be of so much use to all those who have them.