Ouch indeed.. I've a $1000 solder station and years of SMD experience and will still take the mask without thinking twice: the trouble without solder mask often isn't during the board population, but afterwards during testing, mounting etc. Besides, after two years the dust will have covered all those pretty traces anyway, especially with fan cooling.
People who wish for the "all-gold" look should just keep in mind that the silkscreen can only be applied to bare board or solder mask, not directly onto the copper (gold). With all the traces, pads and copper pours on the top layer, this will imply severely restricted and cut off silkscreen, i.e. the PSU decoupling caps' silkscreen will only be semicircles.
It would therefore perhaps look better to just leave the silkscreen out altogether on these particular boards instead of the partially complete one - after all, the elite people boasting proper solder skills should have no trouble coping without a silkscreen either
A bit delayed, but here are the spectra I promised of the three amps: Mk1, Mk2 original, Mk2 upgraded. All are of the left channel, fed with 1kHz 1Vrms input sinewave into a purely resistive 8ohm load.
Apologies for the low resolution; I had to make it very small to fit within 100kB. Also note that the vertical scale for the Mk1 (top) isn't the same as the other two. I'm not going to elaborate too much about the details, people can make their own conclusions. My KSA50 clone plots are on totally different scales etc. but what was very interesting to note was that its low-power spectra is significantly better than the KSA100's. Much lower noise floor and very low harmonics. When the power is increased though, the harmonics rise sharply.
The KSA100 (commercial) may perhaps not measure as good as the KSA50 at low power, but its big advantage is that its distortion characteristics remain remarkably consistent as the output power increases - this is very different with amps with small power supplies and many class-D amps in particular where the distortion rises rapidly with output power increase. This is also an explanation for the legendary bass grip, since the amp doesn't buckle under heavy load. I actually needed to do some burst tests as well to investigate transient power response, but that will have to wait until the Mk1-Mk2 upgrade is complete.
People who wish for the "all-gold" look should just keep in mind that the silkscreen can only be applied to bare board or solder mask, not directly onto the copper (gold). With all the traces, pads and copper pours on the top layer, this will imply severely restricted and cut off silkscreen, i.e. the PSU decoupling caps' silkscreen will only be semicircles.
It would therefore perhaps look better to just leave the silkscreen out altogether on these particular boards instead of the partially complete one - after all, the elite people boasting proper solder skills should have no trouble coping without a silkscreen either
A bit delayed, but here are the spectra I promised of the three amps: Mk1, Mk2 original, Mk2 upgraded. All are of the left channel, fed with 1kHz 1Vrms input sinewave into a purely resistive 8ohm load.
Apologies for the low resolution; I had to make it very small to fit within 100kB. Also note that the vertical scale for the Mk1 (top) isn't the same as the other two. I'm not going to elaborate too much about the details, people can make their own conclusions. My KSA50 clone plots are on totally different scales etc. but what was very interesting to note was that its low-power spectra is significantly better than the KSA100's. Much lower noise floor and very low harmonics. When the power is increased though, the harmonics rise sharply.
The KSA100 (commercial) may perhaps not measure as good as the KSA50 at low power, but its big advantage is that its distortion characteristics remain remarkably consistent as the output power increases - this is very different with amps with small power supplies and many class-D amps in particular where the distortion rises rapidly with output power increase. This is also an explanation for the legendary bass grip, since the amp doesn't buckle under heavy load. I actually needed to do some burst tests as well to investigate transient power response, but that will have to wait until the Mk1-Mk2 upgrade is complete.
Banned
Joined 2002
PWatts said:
id have to say that smd soldering is my favorite. I used to hate it but i challenged my self to solder lots of smd stuff and learn how to do it. Example Take something that is totally broken that is not worth fixing and trying to remove the smd part off of the boards to learn how to.
and to say, i love smd part's.
only advice i have to share is take your time and do one at a time.
so far the briangt/jleaman SMD mini aleph's are my favorite. Color and size.
Jase
I hear you - if it was just for myself I'd have made it a full-out SMD project using 0603 components throughout except for a few 1210's for the higher power. With proper via'ed pads all the semiconductors up until the drivers could be SMD types, although alternative choices available in SOT-223 or similar might be needed.
Especially with through-hole plated boards SMD is just so much easier to remove the components with a dual-tip soldering iron, fine angled tweezers and a variable speed&temp air blower.
But for nostalgic purposes and increased spectrum of resistor types to play with through-hole is the best choice for a GB - I did fit in 1206 SMD pads for the power LED's though
Especially with through-hole plated boards SMD is just so much easier to remove the components with a dual-tip soldering iron, fine angled tweezers and a variable speed&temp air blower.
But for nostalgic purposes and increased spectrum of resistor types to play with through-hole is the best choice for a GB - I did fit in 1206 SMD pads for the power LED's though
Banned
Joined 2002
Yup i have had this soldering iron.. for about 5 years now and love it.
But just bought this blue weller off ebay for 60$ in march. The Weller 921 is ment for smd and small stuff so i now have 2.
An externally hosted image should be here but it was not working when we last tested it.
But just bought this blue weller off ebay for 60$ in march. The Weller 921 is ment for smd and small stuff so i now have 2.
Attachments
The main problem with SMT is that many parts are ONLY available in large quantities. To build a pair of mini boards boards would have had to order 500 SMT IRF input devices... and then there's the matching problem too, without a special jig to hold them its rediclous. I'll stick with through hole at least as long as its available. The surplus store that I frequent has finally gotten a rather large qiantity of SMT resistors and caps in many sizes and most values... But no semi's yet unless its an NTE.
As for the soldering comment that was meant as a tease, I just forgot the winkie face..... Sorry yous guys took it the wrong way. In reality any of you that can solder SMT could easily solder this board without any mask. I had all the semi's on and off one of the proto board three times while I was troubleshooting with no problem and no traces pulling up... I think that says alot for the quality of the board house and PWatts excellent design work. I have a 50.00 Weller station and with the smallest tip and smallest diameter silver bearing solder it works extremely well. Sometimes throwing too much $$ at something doesn't work either.
Don't know yet about the silkscreen but will later today. H-P silk- screened all their boards in either white or black print with no problem what so ever on top of some of the traces. But I also have no problem leaving the silk- screen out completely, assembly is very straightforword.
Jozua, You may want to consider building the 12B4 line stage... its the only one I've used that can get the full value back out of the Krell.
Mark
As for the soldering comment that was meant as a tease, I just forgot the winkie face.....
Sorry yous guys took it the wrong way. In reality any of you that can solder SMT could easily solder this board without any mask. I had all the semi's on and off one of the proto board three times while I was troubleshooting with no problem and no traces pulling up... I think that says alot for the quality of the board house and PWatts excellent design work. I have a 50.00 Weller station and with the smallest tip and smallest diameter silver bearing solder it works extremely well. Sometimes throwing too much $$ at something doesn't work either.Don't know yet about the silkscreen but will later today. H-P silk- screened all their boards in either white or black print with no problem what so ever on top of some of the traces. But I also have no problem leaving the silk- screen out completely, assembly is very straightforword.
Jozua, You may want to consider building the 12B4 line stage... its the only one I've used that can get the full value back out of the Krell.
Mark
Banned
Joined 2002
I like smd because for me it is easy. I like how clean the parts look and how small they are. Especially that the mini a's are so small and compact.
I was one shown and i proved it too. You can go out and buy a 1000$ soldering iron, and claim your the best soldering person. I know people that use a 5$ radio-shack soldering iron and can do just as good. It's not always the tool its how you can use it.
i only bought another one because the smd one can't heat larger things as fast and i don't want to burn it out.
I was one shown and i proved it too. You can go out and buy a 1000$ soldering iron, and claim your the best soldering person. I know people that use a 5$ radio-shack soldering iron and can do just as good. It's not always the tool its how you can use it.
i only bought another one because the smd one can't heat larger things as fast and i don't want to burn it out.
Not at all Mark. It was taken as a tease.
I'm looking forward to this build. I have too many amps around here and this may be the last amp build for a while (ya right). I think I'm gonna go 'balls to the walls' on this one... I'm even shopping for a (semi) full size vertical mill to help me make the chassis. Oh boy, here comes the endless money pit of machine tooling
...oh well.
AndrewT said:
Hi Andrew,
I've used four series diodes in order to cover two PNP plus two NPN for symmetry.
I typically put a micro-switch that selects between a LOW and a HIGH class-A bias point resistors on my PCB right at the Vbe transistor for low inductance, and use a plastic chop-stick to slip in to push the micro switch. I think having the on-transistor diodes would provide some adantages in pulling-off this two-bias point design.
Dynamic plateau bias as used on the big Krells is more complex than I need, but I like saving on Global Warming when I can.
John Curl also uses a MOSFET in his VAS stage and commented on how the MOSFET helps isolate the back EMF from inductive speakers running high currents from coupling back into the front end gain stage. I think this diff-MOSFET VAS helps the Mark II KSA/KMA 100 have superior sonics to the KSA50, and suspect many KSA50 folks will convert to the KMA100 II klone, perhaps with a 2-bias point idea like I mentioned above to save some heat.
It would be nice to find a good price on a KL-271 heatsink profile for purists who will not accept fan noise. It typically cost $65.
I agree its the person not the tools.... I've done some miracle soldering on 16 pin I.C.'s with a 150 watt soldering gun several times
.
The verticle mill is a great idea... We have one at work that I use ALOT. I think I do more of my stuff on it than the company stuff. Get a really good vise, at least a 6" capacity precision type. Don't be afraid to spend at least $300 to $500 on a vise. Without one clamping things down gets hopeless but is still sometimes the only way to hold a piece.
Mark
.The verticle mill is a great idea... We have one at work that I use ALOT. I think I do more of my stuff on it than the company stuff. Get a really good vise, at least a 6" capacity precision type. Don't be afraid to spend at least $300 to $500 on a vise. Without one clamping things down gets hopeless but is still sometimes the only way to hold a piece.
Mark
ANDYLASER,
See my post #693, its located here:
http://www.diyaudio.com/forums/showthread.php?s=&threadid=78129&perpage=10&pagenumber=70
Specs. for a suitable one is depending on what quality and size of transformer you are in the need of. you can give me a hint (are you going to use only ±52 DCV rails or are you planning to feed the LTP by a regulated voltage).
If using one xformer for both channels it will be BIG (ca 1.8-2.2kW). I plan to use one for each channel.
Regards
See my post #693, its located here:
http://www.diyaudio.com/forums/showthread.php?s=&threadid=78129&perpage=10&pagenumber=70
Specs. for a suitable one is depending on what quality and size of transformer you are in the need of. you can give me a hint (are you going to use only ±52 DCV rails or are you planning to feed the LTP by a regulated voltage).
If using one xformer for both channels it will be BIG (ca 1.8-2.2kW). I plan to use one for each channel.
Regards
"Apologies for the low resolution; I had to make it very small to fit within 100kB. Also note that the vertical scale for the Mk1 (top) isn't the same as the other two. I'm not going to elaborate too much about the details, people can make their own conclusions. My KSA50 clone plots are on totally different scales etc. but what was very interesting to note was that its low-power spectra is significantly better than the KSA100's. Much lower noise floor and very low harmonics. When the power is increased though, the harmonics rise sharply.
The KSA100 (commercial) may perhaps not measure as good as the KSA50 at low power, but its big advantage is that its distortion characteristics remain remarkably consistent as the output power increases - this is very different with amps with small power supplies and many class-D amps in particular where the distortion rises rapidly with output power increase. This is also an explanation for the legendary bass grip, since the amp doesn't buckle under heavy load. I actually needed to do some burst tests as well to investigate transient power response, but that will have to wait until the Mk1-Mk2 upgrade is complete."
----------------------------------------------------------------------------
Hi Pwatts
Does this mean that bi amping two KSA50 amps might give better results than one KSA100?
Also what power supply did you use for the KSA50 as you mention that the distortion increases with increase in power?
The KSA100 (commercial) may perhaps not measure as good as the KSA50 at low power, but its big advantage is that its distortion characteristics remain remarkably consistent as the output power increases - this is very different with amps with small power supplies and many class-D amps in particular where the distortion rises rapidly with output power increase. This is also an explanation for the legendary bass grip, since the amp doesn't buckle under heavy load. I actually needed to do some burst tests as well to investigate transient power response, but that will have to wait until the Mk1-Mk2 upgrade is complete."
----------------------------------------------------------------------------
Hi Pwatts
Does this mean that bi amping two KSA50 amps might give better results than one KSA100?
Also what power supply did you use for the KSA50 as you mention that the distortion increases with increase in power?
Harry3,
The KSA50 supply was a bit lacking, I had to work with what I was given. The caps were 55000uF per channel per rail, trafos were 500VA. Rails were a bit low at 35VDC; the trafo's were initially bought for a Stan Curtis amp. Higher voltage rails would probably improve the spectra, but the THD was still below 0.2% while pushing 30W into the load, but would have improved with the same power but with a smaller input signal and lower load impedance. The spectra of it driving 5W was remarkable and much better than the KSA100, so for an active biamped system the KSA50 should work superbly on the tweeters.
The KSA50 would IMO be a great amp for a biamped system driving the tweeters and mids. It will still struggle more with the woofers than a KSA100 driving the lot (with most speakers anyway).
I'd think a KSA50 driving the top end and a KSA00 for the woofers will be a good combination, but it will have to be evaluated to see how well they integrate with each other. A 3-way with the KSA100 driving only the woofers shouldn't have incompatibility issues as long as you account for the gain difference in mind.
I suppose the silkscreen-on-copper will depend on the board house, but it would be worth checking out before having the boards made. It would anycase perhaps be a good idea to leave out the silkscreen for the Golden Girls boards; many of the newer high-tech boards don't use it anymore besides for a logo or something - although it's mainly due to the dense component population that don't allow any space for Refdes and other silkscreen patterns.
Of course a $5 iron can do SMD with a little experience, and it's fine for a little project here and there, but use it for a minimum of 4 hours a day as part of your day job and you'll soon yearn for something better.
The assembly part is not that hard, removing is where it gets tough. For 2-lead stuff you can get away with one iron if you're very careful or preferably two irons, but I still have to see how one can remove a $500 160pin PQFP with it damaging NEITHER the chip nor the board. You also don't need a special holding jig for matching resistors; I use two sharp isolated tweezers connected to the multimeter. As long as you don't go higher than several megaohms or below 100pF it works fine for most matching/identification purposes.
If you buy from large agents you often need to buy MOQ's, but from the smaller vendors (here at least), you can buy as little as 10 pieces and often they give it for free since it's so inexpensive. Surplus stores are even cheaper, but at less than $0.01c per resistor it's hardly an issue.
For most projects I'd recommend the blue Weller stations as mentioned by jleaman. At less than $100 used on eBay they're cheap, and with it I've assembled huge boards with 0402 components and 0.5mm pin spacing chips without problem before I was given the larger station.
And yes.. the tools don't make the man - I've seen enough students screw up very expensive boards using that same station since they've no SMD solder experience. And then it's large stuff like 1206 and SOIC packages. Not good for their project evaluation too
The KSA50 supply was a bit lacking, I had to work with what I was given. The caps were 55000uF per channel per rail, trafos were 500VA. Rails were a bit low at 35VDC; the trafo's were initially bought for a Stan Curtis amp. Higher voltage rails would probably improve the spectra, but the THD was still below 0.2% while pushing 30W into the load, but would have improved with the same power but with a smaller input signal and lower load impedance. The spectra of it driving 5W was remarkable and much better than the KSA100, so for an active biamped system the KSA50 should work superbly on the tweeters.
The KSA50 would IMO be a great amp for a biamped system driving the tweeters and mids. It will still struggle more with the woofers than a KSA100 driving the lot (with most speakers anyway).
I'd think a KSA50 driving the top end and a KSA00 for the woofers will be a good combination, but it will have to be evaluated to see how well they integrate with each other. A 3-way with the KSA100 driving only the woofers shouldn't have incompatibility issues as long as you account for the gain difference in mind.
I suppose the silkscreen-on-copper will depend on the board house, but it would be worth checking out before having the boards made. It would anycase perhaps be a good idea to leave out the silkscreen for the Golden Girls boards; many of the newer high-tech boards don't use it anymore besides for a logo or something - although it's mainly due to the dense component population that don't allow any space for Refdes and other silkscreen patterns.
Of course a $5 iron can do SMD with a little experience, and it's fine for a little project here and there, but use it for a minimum of 4 hours a day as part of your day job and you'll soon yearn for something better.
The assembly part is not that hard, removing is where it gets tough. For 2-lead stuff you can get away with one iron if you're very careful or preferably two irons, but I still have to see how one can remove a $500 160pin PQFP with it damaging NEITHER the chip nor the board. You also don't need a special holding jig for matching resistors; I use two sharp isolated tweezers connected to the multimeter. As long as you don't go higher than several megaohms or below 100pF it works fine for most matching/identification purposes.
If you buy from large agents you often need to buy MOQ's, but from the smaller vendors (here at least), you can buy as little as 10 pieces and often they give it for free since it's so inexpensive. Surplus stores are even cheaper, but at less than $0.01c per resistor it's hardly an issue.
For most projects I'd recommend the blue Weller stations as mentioned by jleaman. At less than $100 used on eBay they're cheap, and with it I've assembled huge boards with 0402 components and 0.5mm pin spacing chips without problem before I was given the larger station.
And yes.. the tools don't make the man - I've seen enough students screw up very expensive boards using that same station since they've no SMD solder experience. And then it's large stuff like 1206 and SOIC packages. Not good for their project evaluation too
PWatts said:
Pierre,
That is indeed how the distortion spectrum of the KSA-100MKII was described in mid 80s test articles.
The difference between the 50 and the 100 correspond quite well with mosfet characteristics imo.
I'll skip posting German Lingo Blah Blah, here's a picture of the KSA-100 from the article.(the PAM-5 on the left sucked big time, also demonstrated by the accurate article. Objective German testing went down the drain in the 1990s ) A French article on the KSA-100MKII at the time derived at the same conclusion.
The Accuphase P500, with twice the output power, finished after the KSA-100. On certain loudspeakers, like electrostats, i found the KSA way superior.
Attachments
Electrostats and Magnepans, similar beasts?
Jacco,
Is this also true of Magnepans, which are a bit of a different beast then true electrostats? I assume magneplanars are also a bit power hungry. I wonder how the sound/dynamics comparison goes between
SS amps like the Krell vs. the Tube (Hurricanes), and the Class D (UCD modules, ICE, TriPath) with the bigger Magnepans.
Lyndon
Jacco,
Is this also true of Magnepans, which are a bit of a different beast then true electrostats? I assume magneplanars are also a bit power hungry. I wonder how the sound/dynamics comparison goes between
SS amps like the Krell vs. the Tube (Hurricanes), and the Class D (UCD modules, ICE, TriPath) with the bigger Magnepans.
Lyndon
The true strength of a class A amplifier such as the KSA-100 is in the mids and high frequencies.
The biggest plus of a class AB spaghetti wire amp like the P500 is in bass reproduction.
Magnetostats require a lot of power but are overall less demanding than electrostatic loudspeakers.
The Maggies i've heard are not at the level of (some) electrostats when it comes to holographic imaging and mid/highs.
But still very strong areas compared to many dynamic loudspeakers, plus heaps of bass power.
I think i've even heard a combo of the largest Maggies/KSA100 once, Magnepan used to be very popular in these parts.
In my experience, tube amps can be as good or better than SS, but at a cost.
For a diy type it is easier to reach heaven's gate on silicate than glass. A reason for me not to mess with tubes in power amps anymore. I could have bought a set of Jadis goldplated armored vehicles with the pecunia i've poored into solid state over the years, but diy is my main therapy.
My sentiments towards class D is more outspoken than to BGs, cold and sterile.
You prefer a short response instead of my usual jabbing ?
Yes !
btw, the loudspeakers used for auditing in the KSA-100 era were generally in the $+10K league.
All the CBS artists on this club who describe class A amplifiers such as the KSAs as fossiles/historical artifacts seem very unwilling to mention the kind of loudspeakers they have.
The biggest plus of a class AB spaghetti wire amp like the P500 is in bass reproduction.
Magnetostats require a lot of power but are overall less demanding than electrostatic loudspeakers.
The Maggies i've heard are not at the level of (some) electrostats when it comes to holographic imaging and mid/highs.
But still very strong areas compared to many dynamic loudspeakers, plus heaps of bass power.
I think i've even heard a combo of the largest Maggies/KSA100 once, Magnepan used to be very popular in these parts.
In my experience, tube amps can be as good or better than SS, but at a cost.
For a diy type it is easier to reach heaven's gate on silicate than glass. A reason for me not to mess with tubes in power amps anymore. I could have bought a set of Jadis goldplated armored vehicles with the pecunia i've poored into solid state over the years, but diy is my main therapy.
My sentiments towards class D is more outspoken than to BGs, cold and sterile.
You prefer a short response instead of my usual jabbing ?
Yes !
btw, the loudspeakers used for auditing in the KSA-100 era were generally in the $+10K league.
All the CBS artists on this club who describe class A amplifiers such as the KSAs as fossiles/historical artifacts seem very unwilling to mention the kind of loudspeakers they have.