Dear Ranchu32,
Thank you for quickly reply.
I am making this amplifier.
I will report later.
Best regards,
Hi.
This is my first post on the forum.I've "lurked" around for quite a bit.
Long story short...I've made a layout in Sprint.Thanks has to go to the previous posters who has made available their "lay" files(Thimios;Prasi;Dacz).It made the job a little bit easier.Not forgetting the huge input the OP(Ranchu32) in conjunction with Hugh Dean(AKSA) has had in component value selection.
I'm attaching the layout for perusal.Please have a look for any faults and I will correct.When the layout gurus are happy,I will upload the full Sprint "lay" file and pdf's.Don't be too harsh as this will be my first layout in quite a couple of decades and I have only been playing with Sprint for less than two weeks.
The size of the pcb is 100mmx50mm.I was looking around at pcb manufacturers online and that was the best size that would be cost effective for diy purposes.So,a bit of working backwards,and a little less surface area to work with on the pcb.
Righto..here goes.
This is my first post on the forum.I've "lurked" around for quite a bit.
Long story short...I've made a layout in Sprint.Thanks has to go to the previous posters who has made available their "lay" files(Thimios;Prasi;Dacz).It made the job a little bit easier.Not forgetting the huge input the OP(Ranchu32) in conjunction with Hugh Dean(AKSA) has had in component value selection.
I'm attaching the layout for perusal.Please have a look for any faults and I will correct.When the layout gurus are happy,I will upload the full Sprint "lay" file and pdf's.Don't be too harsh as this will be my first layout in quite a couple of decades and I have only been playing with Sprint for less than two weeks.
The size of the pcb is 100mmx50mm.I was looking around at pcb manufacturers online and that was the best size that would be cost effective for diy purposes.So,a bit of working backwards,and a little less surface area to work with on the pcb.
Righto..here goes.
Attachments
Welcome to Diyaudio Phunk67. 
I can see wanting to try a layout but given that about 3 or 4 different proven ones now exist (and developed by experienced layout artists). I'm not sure if this is the best thing for you to do here if your interest is in getting a good sounding and working PCB. As this design is pretty mature at this point, it would only make sense if your layout offered some clear advancements. It is much larger than the previous boards so that is not good. Compact amps are preferred. I don't want to discourage you but 3 months ago would have been the time to offer a layout. So there's always value in learning how to do layouts and Sprint and looks like you taught yourself nicely. Maybe apply your new found skills to an Amp that is still in early layout stages?
Regards,
X
I can see wanting to try a layout but given that about 3 or 4 different proven ones now exist (and developed by experienced layout artists). I'm not sure if this is the best thing for you to do here if your interest is in getting a good sounding and working PCB. As this design is pretty mature at this point, it would only make sense if your layout offered some clear advancements. It is much larger than the previous boards so that is not good. Compact amps are preferred. I don't want to discourage you but 3 months ago would have been the time to offer a layout. So there's always value in learning how to do layouts and Sprint and looks like you taught yourself nicely. Maybe apply your new found skills to an Amp that is still in early layout stages?
Regards,
X
Hi xrk971.
Thanks for the input.Let me explain the reasoning for posting.
I was hoping to use the layout for myself as I would outsource the pcb manufacturing to one of the online pcb houses.Cost effectively it would seem prudent to use one of the available size boards that they provide,which in my case is a 100mmx50mm board.At +-20USD for 10 boards it would be a cheap price for the specific size board.
Considering the effort involved in creating the layout,I thought it might benefit users of the forum who would maybe be interested in this layout.
Regarding size...this board has a surface area of 50 square cm.The other three boards which I have copies of has surface areas of 53,70 and 83 square cm's respectively.
You are right,i am learning still.I was hoping to get some guidance for my layout so that I can get my boards made and give back a little something to the community.
I've read through a couple of threads re: layouts.Some stuff makes sense and some are guideliines.I will have to go through some documentation re:layout rules(especially trace thicknesses and distances between tracks).I've used the other "lay" files as reference for trace thicknesses and trace gaps.
With that being said,thanks for the input thus far.
phunk
Thanks for the input.Let me explain the reasoning for posting.
I was hoping to use the layout for myself as I would outsource the pcb manufacturing to one of the online pcb houses.Cost effectively it would seem prudent to use one of the available size boards that they provide,which in my case is a 100mmx50mm board.At +-20USD for 10 boards it would be a cheap price for the specific size board.
Considering the effort involved in creating the layout,I thought it might benefit users of the forum who would maybe be interested in this layout.
Regarding size...this board has a surface area of 50 square cm.The other three boards which I have copies of has surface areas of 53,70 and 83 square cm's respectively.
You are right,i am learning still.I was hoping to get some guidance for my layout so that I can get my boards made and give back a little something to the community.
I've read through a couple of threads re: layouts.Some stuff makes sense and some are guideliines.I will have to go through some documentation re:layout rules(especially trace thicknesses and distances between tracks).I've used the other "lay" files as reference for trace thicknesses and trace gaps.
With that being said,thanks for the input thus far.
phunk
The PCB houses go by maximum extent so 50mm X 50mm is next cost size down and 100mm square is the $10 price point. The other amps may be 5cm2 bigger but the aspect ratio is smaller in sense it is squarish and the outputs are closer together allowing smaller heatsink (like the Dell Pentium CPU heatsink I am using). Obviously 50mm square is too small (unless you went all SMT). Thanks for sharing in any event. But if you want a working amp - the easy choice is send Dacz's latest Gerber to PCBway etc. I have actually ordered a set myself and waiting for them to arrive.
Hi xrk971.
Thanks for the feedback.After you posted I went looking at prices,specifically SeeedStudio.I must say I was a bit hasty as I did not check the options properly.They do have an option to order less boards(5 minimum).With options sorted,5 single layer boards at 10x10cm came to 18,90USD excluding shipping.Plus another 10USD for shipping which brings the total to +-30USD.Not too bad at all.I've got most of the parts for the amp except a couple of small caps which I've allready ordered online.I'm going for a stereo set.The rest of the pcb's will be utilised(always got a friend or two who's interested).
I'm also ordering pcb's for an AKSA 55 clone(still waiting on a couple of items).That was how I got to this thread,by noticing that AKSA(Hugh Dean) was giving help.His explanations are very concise and thorough to me(even I understand a lot of what he's saying and I've been out of the loop for a couple of decades).
Regarding the distance between the outputs: once I read your reply,it came back to me that there was a mention in the thread about the distance of the outputs in relation to each other.Silly me,off I went bulldozing to create a layout for my needs without applying the principle of double-checking all the details.
In the meantime ill be reading up a bit more regarding pcb layout,especially smd work with regards to amplifiers.
Regards
phunk
Thanks for the feedback.After you posted I went looking at prices,specifically SeeedStudio.I must say I was a bit hasty as I did not check the options properly.They do have an option to order less boards(5 minimum).With options sorted,5 single layer boards at 10x10cm came to 18,90USD excluding shipping.Plus another 10USD for shipping which brings the total to +-30USD.Not too bad at all.I've got most of the parts for the amp except a couple of small caps which I've allready ordered online.I'm going for a stereo set.The rest of the pcb's will be utilised(always got a friend or two who's interested).
I'm also ordering pcb's for an AKSA 55 clone(still waiting on a couple of items).That was how I got to this thread,by noticing that AKSA(Hugh Dean) was giving help.His explanations are very concise and thorough to me(even I understand a lot of what he's saying and I've been out of the loop for a couple of decades).
Regarding the distance between the outputs: once I read your reply,it came back to me that there was a mention in the thread about the distance of the outputs in relation to each other.Silly me,off I went bulldozing to create a layout for my needs without applying the principle of double-checking all the details.
In the meantime ill be reading up a bit more regarding pcb layout,especially smd work with regards to amplifiers.
Regards
phunk
phunk67, I'm no expert on PCB layout but to me yours looks OK. I understand your desire to come up with your own board.... this is DIY after all.
Q3 needs to be on a heatsink, either the main sink or a small standalone.
Not taking away from it at all but I found, using parts on hand, that Dacz layout was a bit tight. I had to bend a lot of resistor leads back under the resistors and my electrolytic caps couldn't sit right over there spots. If your redesigning might as well leave a bit of extra room.
Have fun,
Evan
Q3 needs to be on a heatsink, either the main sink or a small standalone.
Not taking away from it at all but I found, using parts on hand, that Dacz layout was a bit tight. I had to bend a lot of resistor leads back under the resistors and my electrolytic caps couldn't sit right over there spots. If your redesigning might as well leave a bit of extra room.
Have fun,
Evan
Check the PSRR. It's much poorer on the negative rail. If you add a cap multiplier to the -ve rail the PSRR improves dramatically. I used B2Spice.
From -35 dB to about -89 dB ( with cap multiplier) at 1Khz. Quite good all the way down to 20Hz. I'm building one and will check if this difference can make a significant difference in sound.
From -35 dB to about -89 dB ( with cap multiplier) at 1Khz. Quite good all the way down to 20Hz. I'm building one and will check if this difference can make a significant difference in sound.
Hi Hugh,
a question's been bugging me, should the amp be supplied with a cap multiplier ? after reading Ashok's comment, i thought i should post here... I do have a pair of Mr. Evil/PMI/Jason's nice CM-PSU PCB's thats eagerly awaiting some solder job.... http://www.diyaudio.com/forums/powe...lified-mrevil-pmi-capacitance-multiplier.html .
hope you are doing well!
regards
Prasi
a question's been bugging me, should the amp be supplied with a cap multiplier ? after reading Ashok's comment, i thought i should post here... I do have a pair of Mr. Evil/PMI/Jason's nice CM-PSU PCB's thats eagerly awaiting some solder job.... http://www.diyaudio.com/forums/powe...lified-mrevil-pmi-capacitance-multiplier.html .
hope you are doing well!
regards
Prasi
Last edited:
Prasi, I have never used cap multipliers with AB amps because:
1. Hum and noise is most evident with AB amps at very low signal when the amp is not working hard so even minimum filter capacitance is very effective at removing hum. These systems work best for Class A amplifiers.
2. Most cap multipliers use a resistive network to drive the gate/base. This does not reference the output voltage to a reference (like 36V from a 42V supply using a zener of voltage reference). This means that at high current the raw supply drops, and so does the 'regulated' figure moves down as a ratio, and the capacitor is modified by using the transfer function (and the series impedance) of the mosfet/bipolar, overlaying an active element in the waveform which is subtly different to a simple passive cap bank.
3. The best regulating systems use a fixed reference and to accommodate rises and falls of mains voltage you would use at least a drop of 15%. For a 42V regulation, you'd need 50V at the bridge at the input of the regulator, and this mandates a loss of headroom and efficiency, not to mention the heat dissipated by the active device!
4. More active devices in the supply rail increase unreliability and add cost, particularly since the actives must be at least as fast as the output devices! Even on my Class A circuits, I use expensive CLC passive filters. You can spend a lot of money on an inductor, certainly, but they very rarely fail and the way they remove artefacts from the supply, which are common as the harmonics of the mains frequency and switching issues from upstream, are outstanding.
5. My AB power supplies use CRC passive filters. They are very effective but the additional caps can be reduced. I also like to use this approach because the first cap sits on the bridge, removing most of the hum, while the second sits at the amplifier, giving a low impedance earth path for the speaker under drive. This confines the earth path from hum reduction device; the resistor is only small, maybe 0.15R, but it means the very high current rectifier pulses are confined to the first cap, not the second.
Hugh
1. Hum and noise is most evident with AB amps at very low signal when the amp is not working hard so even minimum filter capacitance is very effective at removing hum. These systems work best for Class A amplifiers.
2. Most cap multipliers use a resistive network to drive the gate/base. This does not reference the output voltage to a reference (like 36V from a 42V supply using a zener of voltage reference). This means that at high current the raw supply drops, and so does the 'regulated' figure moves down as a ratio, and the capacitor is modified by using the transfer function (and the series impedance) of the mosfet/bipolar, overlaying an active element in the waveform which is subtly different to a simple passive cap bank.
3. The best regulating systems use a fixed reference and to accommodate rises and falls of mains voltage you would use at least a drop of 15%. For a 42V regulation, you'd need 50V at the bridge at the input of the regulator, and this mandates a loss of headroom and efficiency, not to mention the heat dissipated by the active device!
4. More active devices in the supply rail increase unreliability and add cost, particularly since the actives must be at least as fast as the output devices! Even on my Class A circuits, I use expensive CLC passive filters. You can spend a lot of money on an inductor, certainly, but they very rarely fail and the way they remove artefacts from the supply, which are common as the harmonics of the mains frequency and switching issues from upstream, are outstanding.
5. My AB power supplies use CRC passive filters. They are very effective but the additional caps can be reduced. I also like to use this approach because the first cap sits on the bridge, removing most of the hum, while the second sits at the amplifier, giving a low impedance earth path for the speaker under drive. This confines the earth path from hum reduction device; the resistor is only small, maybe 0.15R, but it means the very high current rectifier pulses are confined to the first cap, not the second.
Hugh
Last edited:
I'm not a huge fan of cap multipliers driven from main rails feed the front-end circuitry. Most schemes introduce only a small drop in the order perhaps 1-2V to keep headroom losses to a minimum. Much more than that and the amplifier's maximum power output will be severely curtailed, and we want as much headroom as we can get to deal with transients in the music.
The problems begin when the input voltage falls below the voltage at the base of the pass transistor, because hash will shoot straight through unfiltered! Clearly we don't ever want this to happen but the prevention is to increase the Vdrop to deal with the peak ripple in the PSU under load, transient mains low-line, etc. With a typical 'compliant' PSU feeding the multiplier, that required voltage drop becomes substantial the headroom loss enormous! One also quickly realises that having gone to that extent, it is logical to to turn that cap multiplier into a series regulator and achieve "rock solid" regulation.
I think there is some merit in regulated boosted rails to feed the front-end, when separate transformer windings are specified so that there is no loss of headroom. But it is hard to justify the extra cost and complexity. The simple expedient to improve negative rail PSSR is a simple RC filter. I normally use a 10R resistor (to minimise losses) against a large cap about 1000uF. It won't be as effective as a regulator, granted, but it does provide a useful improvement with minimal cost/complexity.
The problems begin when the input voltage falls below the voltage at the base of the pass transistor, because hash will shoot straight through unfiltered! Clearly we don't ever want this to happen but the prevention is to increase the Vdrop to deal with the peak ripple in the PSU under load, transient mains low-line, etc. With a typical 'compliant' PSU feeding the multiplier, that required voltage drop becomes substantial the headroom loss enormous! One also quickly realises that having gone to that extent, it is logical to to turn that cap multiplier into a series regulator and achieve "rock solid" regulation.
I think there is some merit in regulated boosted rails to feed the front-end, when separate transformer windings are specified so that there is no loss of headroom. But it is hard to justify the extra cost and complexity. The simple expedient to improve negative rail PSSR is a simple RC filter. I normally use a 10R resistor (to minimise losses) against a large cap about 1000uF. It won't be as effective as a regulator, granted, but it does provide a useful improvement with minimal cost/complexity.
Last edited:
I design my Capacitance Multiplier using BJT (all of them). Voltage drop minimum must be 4V to avoid saturation condition. It was used in LM3886 amplifier and won chip amplifier blind test competition. Some of contestants use expensive parts (boutique parts), but my capacitance multiplier using generic parts and the LM3886 amp. using several expensive parts (not so expensive).
Using capacitance multiplier, if design properly, can increase the mid and high quality. It do not require golden ears
Using capacitance multiplier, if design properly, can increase the mid and high quality. It do not require golden ears
Dear phunk67,
I like your layout very much. The Power Transistors are at a distance at either end and make use of the heat sink properly. If you order the PCBs, Please keep two for me. Thanking you,
gannaji
The cap multiplier I was talking about is in the negative rail and only feeding the input stage from Q1 up to the VAS. Not for the complete -ve rail.
Yes it will loose a bit of voltage across it . As long as you play at normal levels this isn't a problem. If this is to be implemented for full clipping at the output, then you need a separate negative rail for this stage. That can come from a separate tapping on the transformer or a voltage multiplier on the transformer with following regulation.
While you get less unclipped power output with say a loss of 2 volts or so due to this cap multiplier/or regulator , the loss will hardly be noticeable when you play music.
What could be an issue if at all, is if you blast the amp wide open and keep clipping the output often , when the clipped behaviour would have to be looked at. I have not done that as yet. I'll attach the PSRR graph for the two supplies with three options. Normal rails. RC filter on the -ve rail and cap multiplier on the -ve rail. Meantime I'll also check on what happens at clipping !
Just for clarity, there is no cap multiplier or regulator on the main power supply rails.
Yes it will loose a bit of voltage across it . As long as you play at normal levels this isn't a problem. If this is to be implemented for full clipping at the output, then you need a separate negative rail for this stage. That can come from a separate tapping on the transformer or a voltage multiplier on the transformer with following regulation.
While you get less unclipped power output with say a loss of 2 volts or so due to this cap multiplier/or regulator , the loss will hardly be noticeable when you play music.
What could be an issue if at all, is if you blast the amp wide open and keep clipping the output often , when the clipped behaviour would have to be looked at. I have not done that as yet. I'll attach the PSRR graph for the two supplies with three options. Normal rails. RC filter on the -ve rail and cap multiplier on the -ve rail. Meantime I'll also check on what happens at clipping !
Just for clarity, there is no cap multiplier or regulator on the main power supply rails.
Ashok, I think we are all discussing the scenario where the cap multiplier feeds the front-end only. You may never have a problem at that 2V drop if the main rails are stiff enough (i.e. decent sized traffo, lots of capacitance, mains voltage doesn't fluctuate too much). But if you crank up the volume and there is enough ripple that the input voltage drops below the output, then the hum/buzz will shoot straight through the multiplier into the input stage, just as it would with a normal regulator.
Edit: Cap multipliers look very attractive in simulation when fed from perfect voltage sources. If you simulate with a realistic amount of supply line ripple and output load then the situation often isn't so rosy.
Edit2: bimo, 4V is a fair loss of headroom, especially in the -ve rail, which usually clips first!
Edit: Cap multipliers look very attractive in simulation when fed from perfect voltage sources. If you simulate with a realistic amount of supply line ripple and output load then the situation often isn't so rosy.
Edit2: bimo, 4V is a fair loss of headroom, especially in the -ve rail, which usually clips first!
Last edited:
If you need 25V PSU, then minimum input voltage for capacitance multiplier is 29V. If you use mosfet for pass current, the voltage drop can be lower because usually mosfet saturation volage much lower than BJT. I use capacitance multiplier for all amplifier, not only for VAS and input stage. But sometime, I use simple capacitance multiplier for VAS and input stage. Add capacitance multiplier (not a simple one, much better PSRR) again for all stage still increase sound quality.