(AK) classA HA/Preamplifier.

[Will]

Mike,

Thanks for your info on caps. I was wondering if you have any suggestions for general purpose (non-low esr) caps, specifically for use on supply decoupling?

[Deleted]

Hi all

Advice needed for PCB drilling virgin.

IE drill from copper side or board side? any other advice (polite)would be great.

take care

[Will]

Nov 21, 2010 15:02:59 GMT @shaun said:

Hi all

Advice needed for PCB drilling virgin.

IE drill from copper side or board side? any other advice (polite)would be great.

take care

Use 0.8mm for normal components, 1mm for rectifiying diodes, to220-case transistor and similar.

If you can, use a drill press, and drill from the copper side,

Don't drill near children, unless you want them to learn 'new' words when you break another chuffing drill bit

[jeffc]

To make things simple (like wot I am )

they need to be for me

Thanks Will/Alex, now I've read through the thread again, hopefully that will be the last of any really naff questions.

cheers..jeffc

[Will]

Jeff, all the talk of duals, pairs, sets and the like was confusing me, and I've been living with this setup for a year or so

I do think that pictures are worth a million words sometimes, and I think that diagram will really help when it comes to talk of the different setups.

And please ask questions, it's what keeps thing interesting

[Deleted]

Nov 21, 2010 15:09:18 GMT Will said:

Nov 21, 2010 15:02:59 GMT @shaun said:

Hi all

Advice needed for PCB drilling virgin.

IE drill from copper side or board side? any other advice (polite)would be great.

take care

Use 0.8mm for normal components, 1mm for rectifiying diodes, to220-case transistor and similar.

If you can, use a drill press, and drill from the copper side,

Don't drill near children, unless you want them to learn 'new' words when you break another chuffing drill bit

Hi will

thanks for the advice the last time i tried to drill a board the hole was so big the cap fell through it.

i do have a bench drill but it's a monster so i think i am going to have to find a smaller alternative.

take care

[Deleted]

Nov 22, 2010 0:18:30 GMT @shaun said:

Nov 21, 2010 15:09:18 GMT Will said:

Use 0.8mm for normal components, 1mm for rectifiying diodes, to220-case transistor and similar.

If you can, use a drill press, and drill from the copper side,

Don't drill near children, unless you want them to learn 'new' words when you break another chuffing drill bit

Hi will

thanks for the advice the last time i tried to drill a board the hole was so big the cap fell through it.

i do have a bench drill but it's a monster so i think i am going to have to find a smaller alternative.

take care

Shaun
Something like the attached can come in handy for this kind of use with your cordless drill.
www.altronics.com.au/index.asp?area=item&id=T2305
Alex

[Deleted]

Nov 22, 2010 1:41:07 GMT @sandyk said:

Nov 22, 2010 0:18:30 GMT @shaun said:

Hi will

thanks for the advice the last time i tried to drill a board the hole was so big the cap fell through it.

i do have a bench drill but it's a monster so i think i am going to have to find a smaller alternative.

take care

Shaun
Something like the attached can come in handy for this kind of use with your cordless drill.
www.altronics.com.au/index.asp?area=item&id=T2305
Alex

Hi Alex

yes that looks like just the thing. i'll see what i can find.

thanks for the info

take care

[oohms]

Drill at high speed and you will save many drill bits

A dremel with a 'bench press' addon would work well

[Deleted]

Nov 22, 2010 16:19:09 GMT oohms said:

Drill at high speed and you will save many drill bits

A dremel with a 'bench press' addon would work well

Hi Oohms

thanks for the tip it's much appreciated. i thought I'd try Alex's chuck thingy first to See how that goes. i have some old PSU boards knocking around so will try them first to avoid tears ans tantrums if it goes wrong.

anyway I'm off to gown up ready for the psu op ;D

I'm going in

take care

[Deleted]

Nov 22, 2010 16:19:09 GMT oohms said:

Drill at high speed and you will save many drill bits

A dremel with a 'bench press' addon would work well

Agreed. But my suggestion was aimed at somebody who doesn't make PCBs, and just wishes to drill a few holes in a couple of PCBs.

[Will]

Something to give you an idea of how it should look.

[Deleted]

Nov 22, 2010 22:08:05 GMT Will said:

Something to give you an idea of how it should look.

Hi Will

looks great

[XTRProf]

Hi Will,

I still want 1 set (2 PCBs) of the SC Class A PCBs after Fran told me it can do about 16W class A into 8ohms. So confirmed that I'm still part of this GB contingent. Let us know when payment.

Thanks, Will. Thanks to Prof Fran again! Excellent!


Chong

[Deleted]

Nov 23, 2010 3:41:33 GMT XTRProf said:

Hi Will,

I still want 1 set (2 PCBs) of the SC Class A PCBs after Fran told me it can do about 16W class A into 8ohms. So confirmed that I'm still part of this GB contingent. Let us know when payment.

Thanks, Will. Thanks to Prof Fran again! Excellent!


Chong

Chong
To do what you want, you would need to use something like the original circuit (but with mods) including different O/P devices.
(MJL21193/MJL21194 OR 2SA1302/2SC3281 as in mine)
You will also need to use a very different PSU (or 2 separate PSUs), perhaps as per the original article, as due to the 1A bias per channel, ripple is very high at those currents, and otherwise S/N will be severely compromised. You may also need to fit external Zobel networks ?
Alex.

[XTRProf]

Nov 23, 2010 3:54:32 GMT @sandyk said:

Nov 23, 2010 3:41:33 GMT XTRProf said:

Hi Will,

I still want 1 set (2 PCBs) of the SC Class A PCBs after Fran told me it can do about 16W class A into 8ohms. So confirmed that I'm still part of this GB contingent. Let us know when payment.

Thanks, Will. Thanks to Prof Fran again! Excellent!


Chong

Chong
To do what you want, you would need to use something like the original circuit (but with mods) including different O/P devices.
(MJL21193/MJL21194 OR 2SA1302/2SC3281 as in mine)
You will also need to use a very different PSU (or 2 separate PSUs), perhaps as per the original article, as due to the 1A bias per channel, ripple is very high at those currents, and otherwise S/N will be severely compromised. You may also need to fit external Zobel networks ?
Alex.

Hi Alex,

Yes, I intend to do the mono version to go all out for it since the PCB is already in mono to take advantage of it. So 2 separate PSUs eventhough I may not have to change anything due to the underlying reasons. I believe Will and me are with the same idea in this PSU area.

Zobel network when most HPs are >> 8ohms and with resistor network? Do we still need a Zobel network with say your series 68 ohms? I think we may need that if we are talking of low ohmage orthodynamics or maybe yet to be materialised ribbon HPs without the series resistors. Even so, the LCD 2 is 50 ohms.

Anyway, what I'm asking is the equivalent Class A wattage at 8 ohms to get a better idea of whether this HA can let out close to 15 W class A at 8 ohms (speaker load). Let's say even though this class A HA version which you had proposed here may not have the juicier power transistors, amp or parts to do job best, we are still here only talking of driving real world HPs and not speakers. So using a real world HP load of say 300 ohms, as in the HD800, as reference, we are only drawing <0.3 W from this class A amp as explained by Fran. Even we say go to the lowest HP ohmage that presently we can find and that is about 30 ohms, correct me if I'm wrong, we are still not asking for 15W at 30 ohms. Maybe 3 W as an educated guess with reference with 300 ohms drawing 0.3 W. Sorry, I can't say what wattage at 30 ohms without the series resistor as I need to brush up on my electrical maths from the model answers given by Fran before I can give any whereabout figures. Fran, may be able to help straight away in this area as he's the expert in this type of calculation. So in no way, we are going to ask this poor HA to drive 15W class A into 8 ohms as that will never happen. This is with the exception of the mentioned orthodynamics and possible ribbon HPs without the series resistors.

Anyway, it's good to have uprated parts as the chances for destruction will be almost zero unless we short it. Let's see whether I can do one as I always like overdone things for the purpose like the Americans, especially in Hifi for the last ounce of definition. You are also using Class A for that purpose as well, right?

Thanks for the pointers. All noted as I still have that in mind.

[Deleted]

Nov 23, 2010 8:04:52 GMT XTRProf said:

Nov 23, 2010 3:54:32 GMT @sandyk said:

Chong
To do what you want, you would need to use something like the original circuit (but with mods) including different O/P devices.
(MJL21193/MJL21194 OR 2SA1302/2SC3281 as in mine)
You will also need to use a very different PSU (or 2 separate PSUs), perhaps as per the original article, as due to the 1A bias per channel, ripple is very high at those currents, and otherwise S/N will be severely compromised. You may also need to fit external Zobel networks ?
Alex.

Hi Alex,

Yes, I intend to do the mono version to go all out for it since the PCB is already in mono to take advantage of it. So 2 separate PSUs eventhough I may not have to change anything due to the underlying reasons. I believe Will and me are with the same idea in this PSU area.

Zobel network when most HPs are >> 8ohms and with resistor network? Do we still need a Zobel network with say your series 68 ohms? I think we may need that if we are talking of low ohmage orthodynamics or maybe yet to be materialised ribbon HPs without the series resistors. Even so, the LCD 2 is 50 ohms.

Anyway, what I'm asking is the equivalent Class A wattage at 8 ohms to get a better idea of whether this HA can let out close to 15 W class A at 8 ohms (speaker load). Let's say even though this class A HA version which you had proposed here may not have the juicier power transistors, amp or parts to do job best, we are here only talking of driving real HPs and not speakers. So using a real world HP load of say 300 ohms, as in the HD800, as reference, we are only drawing <0.3 W from this class A amp as explained by Fran. So in no way, we are going to ask this poor HA to drive 15W class A into 8 ohms as that will never happen. This is with the exception of the mentioned orthodynamics and possible ribbon HPs.

Anyway, it's good to have uprated parts as the chances for destruction will be almost zero unless we short it. Let's see whether I can do one as I always like overdone things for the purpose like the Americans.

Thanks for the pointers. All noted as I still have that in mind.

Chong
I am glad you cleared that up, as I got the impression you were trying to extract 15W !
The present version should meet your requirements without Zobel networks, although if you are going to try driving low impedance headphones, it may be prudent to fit 4.7ohm series resistors at the output.Don't forget too, that even a 50% increase in bias will give quite a bit more power into those headphones, even when using 120 ohm series resistors at the output.
Just beef up the heatsinks a little!

[XTRProf]

Nov 23, 2010 8:26:14 GMT @sandyk said:

Nov 23, 2010 8:04:52 GMT XTRProf said:

Hi Alex,

Yes, I intend to do the mono version to go all out for it since the PCB is already in mono to take advantage of it. So 2 separate PSUs eventhough I may not have to change anything due to the underlying reasons. I believe Will and me are with the same idea in this PSU area.

Zobel network when most HPs are >> 8ohms and with resistor network? Do we still need a Zobel network with say your series 68 ohms? I think we may need that if we are talking of low ohmage orthodynamics or maybe yet to be materialised ribbon HPs without the series resistors. Even so, the LCD 2 is 50 ohms.

Anyway, what I'm asking is the equivalent Class A wattage at 8 ohms to get a better idea of whether this HA can let out close to 15 W class A at 8 ohms (speaker load). Let's say even though this class A HA version which you had proposed here may not have the juicier power transistors, amp or parts to do job best, we are here only talking of driving real HPs and not speakers. So using a real world HP load of say 300 ohms, as in the HD800, as reference, we are only drawing <0.3 W from this class A amp as explained by Fran. So in no way, we are going to ask this poor HA to drive 15W class A into 8 ohms as that will never happen. This is with the exception of the mentioned orthodynamics and possible ribbon HPs.

Anyway, it's good to have uprated parts as the chances for destruction will be almost zero unless we short it. Let's see whether I can do one as I always like overdone things for the purpose like the Americans.

Thanks for the pointers. All noted as I still have that in mind.

Chong
I am glad you cleared that up, as I got the impression you were trying to extract 15W !
The present version should meet your requirements without Zobel networks, although if you are going to try driving low impedance headphones, it may be prudent to fit 4.7ohm series resistors at the output.Don't forget too, that even a 50% increase in bias will give quite a bit more power into those headphones, even when using 120 ohm series resistors at the output.
Just beef up the heatsinks a little!

Thanks. I will get back to you for good advice when the time is right for an uprated version. Come on, no way am I going to listen at 15 W output, right? Just the extra refinement of what a higher wattage Class A amp can do to the sound.

[Will]

Updated the BOM.

More components, and a larger array of heatsinks (epecially for Chong )

Also a little note on perhaps buying and using bc560/mpsa18 instead of the LS devices on first power up.

Hope it helps.

[Deleted]

Nov 23, 2010 21:43:34 GMT Will said:

Updated the BOM.

More components, and a larger array of heatsinks (epecially for Chong )

Also a little note on perhaps buying and using bc560/mpsa18 instead of the LS devices on first power up.

Hope it helps.

Hi Will
Just a mention that not everybody can open .xls files.
Alex

[Will]

True, but there are plenty of xls viewer programs out there.

Here it is as a pdf, as most would have a viewer for that (and it's easier to do than a .doc )

Part numbers refer to Farnell.

[Deleted]

Nov 23, 2010 21:43:34 GMT Will said:

Updated the BOM.

More components, and a larger array of heatsinks (epecially for Chong )

Also a little note on perhaps buying and using bc560/mpsa18 instead of the LS devices on first power up.

Hope it helps.

Hi Will
To make matters simpler, BC550C and BC549C will be fine for initial testing of the Current mirror instead of the LS313. Just be aware of the different pinout.
Alex

[XTRProf]

Nov 23, 2010 21:43:34 GMT Will said:

Updated the BOM.

More components, and a larger array of heatsinks (epecially for Chong )

[XTRProf]

For those without the PSU (I mean not the V regulators) + transformer, I think it will be good to recommend the PSU kit + transformer as well. Also, are we going to mass order the closely matched LS transistors here or what? Another thing is how can people have closely matched hfe 450-550 for transistors T4 BC558B,
T5 BC557B, T6, T8 BC547, T7 BC549C, T9 BC337 and
T10 BC327 when they don't have the equipment to do so. I believe many people will want the best for their kits as well.


Thanks.

[Deleted]

Nov 24, 2010 2:36:54 GMT XTRProf said:

For those without the PSU (I mean not the V regulators) + transformer, I think it will be good to recommend the PSU kit + transformer as well. Also, are we going to mass order the closely matched LS transistors here or what? Another thing is how can people have closely matched hfe 450-550 for transistors T4 BC558B,
T5 BC557B, T6, T8 BC547, T7 BC549C, T9 BC337 and
T10 BC327 when they don't have the equipment to do so. I believe many people will want the best for their kits as well.


Thanks.

Chong
The thread has all along recommended the Jaycar KC-5418 PSU kits, which are complete except for the need in this case for 2,200uF 35V electros instead of those supplied.
Due to the weight of the 18-0-18VAC 30VA toroidal transformers, it is recommended that they be obtained locally.
Although most members who will be building the amplifier will not have access to dedicated Transistor Testers, many DVMs do have a transistor test facility.
I feel sure that someone like Will perhaps , may be able to help out with suitable small signal devices for Q4 etc. if asked nicely when the Group Buy dual transistors are sent out, as the quantities needed will be quite small, and some of us buy them in bulk to get very good quantity pricing.
Alex