Will
Been here a while!
Ribena abuser!
Member since 2008
Posts: 2,164
|
Post by Will on Dec 20, 2009 15:05:36 GMT
Friedrich, I must admit to not being to concerned about having to use a DC offset protector, as I would guess that getting a large DC level at the output would be more to do with a sudden fault. The fart I got did nearly lead me to building a mute circuit, but in the end I just managed it. Must admit to being pleased now that it's gone, though!
Just been listening to the Dogman album, by Kings X. Alex mentions 'low-end dynamics' and he is not wrong. Dug Pinnock's bass line well and truly tickles your bollocks, very splendid! Deep, defined, tuneful bass that skips along. And on K701's at that.
|
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Dec 20, 2009 19:47:23 GMT
One thing to mention is that before I replaced the post-rectification caps in the PSU's and placed the 78/79l15 before the dc offset boards, I used to get a 'fart' through the headphones (equates to 1.6Vdc ) Now it's gone, which means I no longer have to disconnect the headphones before power off. I wonder how nobody had made a DC offset protector and power on delay for an headphone amp - the once I found is ε12 ("epsilon 12") is a turn-on delay muting and DC offset protection amb.org/audio/epsilon12/wikipedia about HA en.wikipedia.org/wiki/Headphone_amplifierFriedrich With my main preamp I use a Current Limited JLH which is slow start, as well as a muting/delay circuit which is used mainly for speaker operation. It senses the absence/presence of AC from the Power Supply when turned on or off, and controls 2 small relays. The turn off "fart" that Will mentioned certainly will not hurt headphones at that level with the series output resistor of 68 or 120 ohms. As Will says, incorporating the 15V regulators in the DC offset Corrector helped get rid of it. Adding all that extra stuff to a HA is not necessary unless you retain a Zobel network .The series output resistors give quite a deal of protection normally. DC out in this case is mainly a problem when first built due to errors, and anybody who knows what they are doing would check fior DC out etc. BEFORE plugging in headphones. Alex
|
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Dec 20, 2009 19:51:05 GMT
Will Funny about that ! I will follow up at a later date after fine tuning the Class A preamp at +-20V. The 15W Class A has dual PSUs and dual JLHs. Alex
|
|
Will
Been here a while!
Ribena abuser!
Member since 2008
Posts: 2,164
|
Post by Will on Feb 5, 2010 13:53:22 GMT
|
|
leo
Been here a while!
Team wtf is it?
Posts: 3,638
|
Post by leo on Feb 9, 2010 11:32:31 GMT
Finally!!! tried the 18v traffo and adjusted the balancing of the LTP
Yep, its better for sure, seems more open and solid, just need some nice servo's now
|
|
jonclancy
Been here a while!
Mr. Ripple Eater
Amateur EAGLEist
Posts: 1,131
|
Post by jonclancy on Feb 16, 2010 1:50:26 GMT
Maybe this is what we need for the RG amp - a souped up Class A... ;D
|
|
Spirit
Been here a while!
That's where I'm gonna go when I die
Posts: 1,107
|
Post by Spirit on Mar 16, 2010 0:04:07 GMT
So just how much better is this than an AK-SCHA anyway...? ;D
|
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Mar 16, 2010 0:15:50 GMT
So just how much better is this than an AK-SCHA anyway...? ;D Phil You will need to rely on Will or Leo's answers to that question. Or perhaps Syd if he makes one. Note also, that Will took it even further than the original, by using completely separate power supplies for each channel. (2 18-0-18 30VA toroidal transformers, 2 regulated PSU PCBS, and 2 JLHs.) Alex
|
|
Spirit
Been here a while!
That's where I'm gonna go when I die
Posts: 1,107
|
Post by Spirit on Mar 16, 2010 7:25:13 GMT
What are wrong with your answers Alex? I don't mind a bit of designer's bias
|
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Mar 16, 2010 7:46:02 GMT
Phil You will find my thoughts in the thread already.You will also see reports from both Will and Leo there . I will however say that I find it more 3D sounding than the SC HA with JLH. I must admit t, that I am tempted to try what Will has done with the dual PSUs. Alex
|
|
Spirit
Been here a while!
That's where I'm gonna go when I die
Posts: 1,107
|
Post by Spirit on Mar 16, 2010 9:48:50 GMT
Alright... I've re-read, and reckon it would be worth doing... probably. Leo, you said the HD650s sounded... not very good. Do you have, or can you easily get a hold of some HD600s to try with it? I would be most appreciative of your opinion of them. I know they make everything sound the same ;D, but they're all I've got atm... Hmm... slight blatant lie... they're all I've got till I crack out the blu-tack Cheers! Phil
|
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Oct 31, 2010 3:00:12 GMT
Well, I have finally got around to doing that. However, Will's version is slightly more full on, as it also uses 2 x 18-0-18VAC 30VA toroidal transformers, whereas I am still only using a single transformer for now. Alex (Click on the image, then click on the Imageshack photo for a full screen photo.)
|
|
FritzS
Been here a while!
Sound of Blue Danube
Sound of Blue Danube
Posts: 1,364
|
Post by FritzS on Oct 31, 2010 5:25:31 GMT
Just in case there is anybody else out there (apart from Leo, Alex and me ;D) who has built this amp, here is a new version of the DC offset corrector which incorporates 15v reguators, allowing you to use them with the amp boards running with 20V rails, which in my opinion, is the only way to run this amp How far is the SC Class A - all in one PCB project? Both modified amps + DC offset correctors on one PCB are optimally
|
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Oct 31, 2010 5:57:29 GMT
Just in case there is anybody else out there (apart from Leo, Alex and me ;D) who has built this amp, here is a new version of the DC offset corrector which incorporates 15v reguators, allowing you to use them with the amp boards running with 20V rails, which in my opinion, is the only way to run this amp How far is the SC Class A - all in one PCB project? Both modified amps + DC offset correctors on one PCB are optimally Friedrich That is not optimal, as the present method ensures maximum channel separation for starters. Yes, an offset corrector could have been integrated on each of the channel PCBs, but that would have necessitated a major PCB redesign. As it is, Will has incorporated all the modifications with his revised PCB design. It also enables the amplifier to be built more cheaply, and added to as finances and time permits. Alex
|
|
FritzS
Been here a while!
Sound of Blue Danube
Sound of Blue Danube
Posts: 1,364
|
Post by FritzS on Oct 31, 2010 6:27:20 GMT
Friedrich That is not optimal, as the present method ensures maximum channel separation for starters. Yes, an offset corrector could have been integrated on each of the channel PCBs, but that would have necessitated a major PCB redesign. As it is, Will has incorporated all the modifications with his revised PCB design. It also enables the amplifier to be built more cheaply, and added to as finances and time permits. Alex You could put them on one PCB and optimize the grounding too. With external wire you decrease the channel separation too. The Headphone jack have one ground for both, the PSU have one ground for both, the input have one ground for both (connected in the signal source, e.g. in the DAC). If you would optimize you must use symmetrical XLR connections, two separate PSU's and balanced HA connectors. I think both amps on one PCB with one star ground are the optimum for non balanced amp designs.
|
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Oct 31, 2010 7:06:26 GMT
Friedrich That is not optimal, as the present method ensures maximum channel separation for starters. Yes, an offset corrector could have been integrated on each of the channel PCBs, but that would have necessitated a major PCB redesign. As it is, Will has incorporated all the modifications with his revised PCB design. It also enables the amplifier to be built more cheaply, and added to as finances and time permits. Alex You could put them on one PCB and optimize the grounding too. With external wire you decrease the channel separation too. The Headphone jack have one ground for both, the PSU have one ground for both, the input have one ground for both (connected in the signal source, e.g. in the DAC). If you would optimize you must use symmetrical XLR connections, two separate PSU's and balanced HA connectors. I think both amps on one PCB with one star ground are the optimum for non balanced amp designs. Friedrich Have you any idea of how much time and resources it would take to do what you are proposing ? Wiill is not a professional designer either, and also has job and family commitments. BTW, I do not agree with what you are saying, as I believe that best result are obtained by keeping both channels completely separate, as with dual mono until the headphone socket, which ideally would not combine the earths at the socket. It should be made to work with the vast majority of commercial headphones without rewiring the headphones. Alex P.S. The Class A HA was used as a preamp at a recent listening session , and with decent material presented a very good 3D image. That was before the recent changes.
|
|
FritzS
Been here a while!
Sound of Blue Danube
Sound of Blue Danube
Posts: 1,364
|
Post by FritzS on Oct 31, 2010 7:27:22 GMT
[ Friedrich Have you any idea of how much time and resources it would take to do what you are proposing ? Wiill is not a professional designer either, and also has job and family commitments. BTW, I do not agree with what you are saying, as I believe that best result are obtained by keeping both channels completely separate, as with dual mono. Alex I know about the resource problem ... but Will say'd to me (some time ago) it's possibly to made a new common PCB, I send Will a PM now. I don't agree with you - you could not keeping both completely separate with one PSU, a three pin headphone jack and non balances sources. What's the difference between wires between two PCB's or conducting paths between on one PCB? You could put all electronic devices (PSU, amps, etc.) on one PCB, if it done right and use a high quality PCB the result is the same! I hope you understand what I mean.
|
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Oct 31, 2010 7:43:19 GMT
[ Friedrich Have you any idea of how much time and resources it would take to do what you are proposing ? Wiill is not a professional designer either, and also has job and family commitments. BTW, I do not agree with what you are saying, as I believe that best result are obtained by keeping both channels completely separate, as with dual mono. Alex Friedrich The original 15W Class A amplifier that this is based on, had THD estimated to be less than .0002% at lower powers, and that was without using the matched differential and current mirror devices that we are using. It was optimised using the Audio Precision test set. The present design will most likely have even lower distortion, a wider bandwidth, and higher S/N figures due to the JLHs and separate PSUs. If you have access to the AP test gear , then by all means try your own ideas on the subject. Otherwise, I suugest that we stick with a proven design that is an extremely high performer that will undoubtedly eat any of your beloved WNA designs for breakfast. Alex I know about the resource problem ... but Will say'd to me (some time ago) it's possibly to made a new common PCB, I send Will a PM now. I don't agree with you - you could not keeping both completely separate with one PSU, a three pin headphone jack and non balances sources. What's the difference between wires between two PCB's or conducting paths between on one PCB? You could put all electronic devices (PSU, amps, etc.) on one PCB, if it done right and use a high quality PCB the result is the same! I hope you understand what I mean. Friedrich The original 15W Class A amplifier that this is based on, had THD estimated to be less than .0002% at lower powers, and that was without using the matched differential and current mirror devices that we are using. It was optimised using the Audio Precision test set. The present design will most likely have even lower distortion, a wider bandwidth, and higher S/N figures due to the JLHs and separate PSUs. If you have access to the AP test gear , then by all means try your own ideas on the subject. Otherwise, I suugest that we stick with a proven design that is an extremely high performer that will undoubtedly eat any of your beloved WNA designs for breakfast. Alex P.S. You will be wasting your time sending Will a PM with your suggestions, as Will has already settled on 2 separate channel PCBs as per the original design, and usedf in most published amplifier designs. Jon and Will may be fiurther advanced than you believe with this project.
|
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Oct 31, 2010 7:45:25 GMT
a simple matter of using minimal 2 layer board layout like the Panda (preferably even 4 layer) with 2 ground planes Still making a new PCB design would take me (I do this professionally but not on a day to day base) about 1-2 days and mostly there is an 'oops' on it requiring slight redesign. Making multilayer boards in small quantities makes them expensive because of the tooling costs for these kind of boards (these can be around E 400 added to the board price on first order). Dual mono is generally considered to be a better solution... but.. a single board with ground planes will always have a LOWER ground (common) resistance then can be achieved with any dual mono design. This is certainly true for headphone amps and recommend this over dual mono construction. You can still use dual power supplies with such a board and enjoy it's benefits. Perhaps science and audio will not go hand in hand together insubjective tests though.. For speaker amps dual mono is arguably a better solution.
|
|
Will
Been here a while!
Ribena abuser!
Member since 2008
Posts: 2,164
|
Post by Will on Oct 31, 2010 10:11:16 GMT
Hi All,
Here's a quick state of play.
The schematic, used to create the pcb, has been created and checked. This in itself takes a fair while (to someone who does this as a hobby) I've also got the components placed on the pcb, roughly in the same positions as the original pcb. I'll be using the original layout as a guide, as it's proven to sound good.
In addition, the schematic incorporates the LTP mod, and also uses custom component models for 2 pin terminals and the linear systems dual transistors, which Jon has created for this project. I've also included the dc offset corrector to be integral on the pcb.
I have to go with one channel per pcb as I use the freeware version of eagle, as I cannot afford the $750 for the full license. This limits maximum pcb size to 100mmx80mm, which is smaller than the original class a pcb, whilst including the offset corrector pcb.
The original class A pcb has star earthing for its channel, including the ability to earth-lift between signal ground and power ground, this will also be included on the new pcb also. The original pcb also uses wire links to connect the circuit in places, which will bereplaced with pcb tracks, as this will be a double sided pcb.
As I have been working on other little projects in my spare time (both for myself and others) this pcb has slipped from being completed in October as I had hoped. However, I hope to get it finished and checked in November, so that anybody who is interested can buy the pcbs at cost price plus delivery, when the pcb is ordered. I'll also be organising a GB of the dual transistors, so that they'll be easily available.
|
|
FritzS
Been here a while!
Sound of Blue Danube
Sound of Blue Danube
Posts: 1,364
|
Post by FritzS on Oct 31, 2010 11:46:18 GMT
Alex, sorry, mea culpa, for me it's a hard thing to translate my German thoughts to English that you understand in English what I thought in German - and I mix two different things - PCB layout and ground path. Could we make a thought experiment? Look at the picture img2.imageshack.us/i/classahaupgraded50size.jpg/and put all PCB's that you see side by side to one big PCB without change anything of each layouts. Then translate all your hand crafted wiring to conducting paths between the PCB's. What is the difference now? I think nothing. The other part is the grounding path problem. In a headphone amp exist one problem that cannot be eliminate if the standard 3 pin headphone jack and asymmetric cinch connectors used - a loop would exist. The grounding from the source splitting in two path over the both stereo cinch connectors and coming together at the headphone jack (if a truely dual mono designed HA used). Only symmetrical connection (better electrically insulated too) between source and the HA and 4 pin wired headphones could solve this problem, like using of two mono block amps for speakers. Alex I hope now you know what I mean. This has noting to do with the well designed Class A 15 W PCB layout (apart from the known mods for using as HA and other BJT pin-layout). I think you use star grounding in your hand crafted wiring too.
|
|
FritzS
Been here a while!
Sound of Blue Danube
Sound of Blue Danube
Posts: 1,364
|
Post by FritzS on Oct 31, 2010 11:55:01 GMT
Hi All, Here's a quick state of play. The schematic, used to create the pcb, has been created and checked. This in itself takes a fair while (to someone who does this as a hobby) I've also got the components placed on the pcb, roughly in the same positions as the original pcb. I'll be using the original layout as a guide, as it's proven to sound good. In addition, the schematic incorporates the LTP mod, and also uses custom component models for 2 pin terminals and the linear systems dual transistors, which Jon has created for this project. I've also included the dc offset corrector to be integral on the pcb. I have to go with one channel per pcb as I use the freeware version of eagle, as I cannot afford the $750 for the full license. This limits maximum pcb size to 100mmx80mm, which is smaller than the original class a pcb, whilst including the offset corrector pcb. The original class A pcb has star earthing for its channel, including the ability to earth-lift between signal ground and power ground, this will also be included on the new pcb also. The original pcb also uses wire links to connect the circuit in places, which will bereplaced with pcb tracks, as this will be a double sided pcb. As I have been working on other little projects in my spare time (both for myself and others) this pcb has slipped from being completed in October as I had hoped. However, I hope to get it finished and checked in November, so that anybody who is interested can buy the pcbs at cost price plus delivery, when the pcb is ordered. I'll also be organising a GB of the dual transistors, so that they'll be easily available. Will this I mean - but I don't know that you have only the free version of eagle. I am interested for one set of PCB's - please let me know if ready. What I thought - you put all PCB's sibe by side to a big one and translate the hand crafted wiring to conducting paths between the PCB's. PS: Mea culpa - my english is not perfect
|
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Oct 31, 2010 12:04:10 GMT
Hi All, Here's a quick state of play. The schematic, used to create the pcb, has been created and checked. This in itself takes a fair while (to someone who does this as a hobby) I've also got the components placed on the pcb, roughly in the same positions as the original pcb. I'll be using the original layout as a guide, as it's proven to sound good. In addition, the schematic incorporates the LTP mod, and also uses custom component models for 2 pin terminals and the linear systems dual transistors, which Jon has created for this project. I've also included the dc offset corrector to be integral on the pcb. I have to go with one channel per pcb as I use the freeware version of eagle, as I cannot afford the $750 for the full license. This limits maximum pcb size to 100mmx80mm, which is smaller than the original class a pcb, whilst including the offset corrector pcb. The original class A pcb has star earthing for its channel, including the ability to earth-lift between signal ground and power ground, this will also be included on the new pcb also. The original pcb also uses wire links to connect the circuit in places, which will bereplaced with pcb tracks, as this will be a double sided pcb. As I have been working on other little projects in my spare time (both for myself and others) this pcb has slipped from being completed in October as I had hoped. However, I hope to get it finished and checked in November, so that anybody who is interested can buy the pcbs at cost price plus delivery, when the pcb is ordered. I'll also be organising a GB of the dual transistors, so that they'll be easily available. Will this I mean - but I don't know that you have only the free version of eagle. I am interested for one set of PCB's - please let me know if ready. What I thought - you put all PCB's sibe by side to a big one and translate the hand crafted wiring to conducting paths between the PCB's. PS: Mea culpa - my english is not perfect Friedrich You are just plain bone lazy ! Again, I will state that UNLESS you have access to test equipment like the Audio Precision gear, then you could end up with far worse distortion and performance than the original design. P.S. There aren't any connecting leads between the 2 main PCBs. They have separate power supply leads, separate Input cables, and separate output cables.The wiring used for the PSU supplies is 5 or 7.5A mains rated cable to ensure that the very low output impedance of the JLH is maintained. Thin 1 ounce or 2 ounce copper tracks are not adequate in that case. Unfortunately, almost ALL headphones use the tip, ring, sleeve connection, so the earths have to come together at that point. It would be nice if they were all 4 wire connected, but they aren't..
|
|
Will
Been here a while!
Ribena abuser!
Member since 2008
Posts: 2,164
|
Post by Will on Oct 31, 2010 12:45:03 GMT
I think you are thinking about something like this? www.amb.org/audio/epsilon22/Ti Kan created it so that a three channel b22 amp could be easily and quickly built up, with all the inter-board wiring sorted out. Just feed it a signal and power, and music pops out the other end. A nice idea, and it would certainly make things easier on the construction side. If I wasn't held held back by the size limitations I mentioned before, I'd certainly sort you something out. Also, don't worry about your English, it's very good!
|
|
Will
Been here a while!
Ribena abuser!
Member since 2008
Posts: 2,164
|
Post by Will on Nov 17, 2010 12:22:04 GMT
The pcb layout is pretty much done, so barring any mishaps we should be ready to buy them. The PCB's will be the same quality as the recent JLH GB, 3mm thick, 70uM copper traces and a pretty colour if required Size is 120mm x 90mm per pcb. Also, the DC offset corrector is integrated in to the circuit. You'll need two pcb's for a stereo pair. Based on 10 pcb sets, The approximate price for a set is €15 (£13 or AU$21) Delivery and Paypal to go on that. The more pcb's are ordered the cheaper they become for everyone, and you get them at cost price. Parts list to follow. If you want a set or two of these pcb's, please PM me, and I'll pop you on the list To get the very best out of this amp, you need to match the input pair and current mirror devices very tightly (for hFE and VBE) which allows the ltp mod mentioned previously in the thread to take full effect. To make this an easier task, I'm running an interest check/GB for the Linear Devices LS313/352 at diyAudio (purely because I think there maybe a better chance of hitting the minimum order quantity with the extra people over there) so please PM if interested. Linear Device GB. You'll also need JLH's to get the best out of this amp, with a second group buy here. www.diyaudio.com/forums/group-buys/173448-john-linsley-hood-ripple-eater-part-2-a.html2 boards for normal single supply, or 4 for dual (
|
|