Hi Gertjan,
especially the Vifa XT25TG30-04 is rated for nominal power (140W) and operating power (5.3W). I'd like to take this 1/26 ratio as an approx. number. So I can limit the input voltage of the UcD. Thank you.
BTW: I'm just writing some lines for a working stay in Japan next year, if things work out.
Regards, Timo
especially the Vifa XT25TG30-04 is rated for nominal power (140W) and operating power (5.3W). I'd like to take this 1/26 ratio as an approx. number. So I can limit the input voltage of the UcD. Thank you.
BTW: I'm just writing some lines for a working stay in Japan next year, if things work out.
Regards, Timo
Bruno:
Thanks for the answer to my previous questions. I'll wait for your whitepaper, in wich I hope there will be some drawings for us who still have doubts after re-reading many times your explanation about unbal to balanced connections (not your fault
)
I'm planning ahead to build the Power Supply. I wich I read before your opinion about Panasonic's. I just changed one for my cheap Sony DVD wich was allready light sounding in the midrange, as transport. I confirmed your statement.
I've searching around and I have some options on filtering caps:
*Elna Cerafine 10000uf*2
*Jensen 4 pole caps: anyone tried these?? In Reference Audiomod's page they seam to like them very much, better than BG's at lower cost (+/- 1/3x).
Any advice welcome (newbie, remember?)
I don't mind spending some more kopecs if I get more sensual midgrange and weightier bass.
Can somebody advice where to buy good, low noise, high current rectifier diodes (better in USA).
(I almost baught PS Audio's amp, I'm glad I discovered that DIY Audio is the cure for Audiophilia Nervosa
)
MR
Thanks for the answer to my previous questions. I'll wait for your whitepaper, in wich I hope there will be some drawings for us who still have doubts after re-reading many times your explanation about unbal to balanced connections (not your fault
)I'm planning ahead to build the Power Supply. I wich I read before your opinion about Panasonic's. I just changed one for my cheap Sony DVD wich was allready light sounding in the midrange, as transport. I confirmed your statement.
I've searching around and I have some options on filtering caps:
*Elna Cerafine 10000uf*2
*Jensen 4 pole caps: anyone tried these?? In Reference Audiomod's page they seam to like them very much, better than BG's at lower cost (+/- 1/3x).
Any advice welcome (newbie, remember?)
I don't mind spending some more kopecs if I get more sensual midgrange and weightier bass.
Can somebody advice where to buy good, low noise, high current rectifier diodes (better in USA).
(I almost baught PS Audio's amp, I'm glad I discovered that DIY Audio is the cure for Audiophilia Nervosa
)MR
Tiki,
I think you will find in practice that although you can try to reduce the possibilities of tweeter damage. Ultimately you will only damage them when something goes wrong, ie with normal music you shouldn't really get them into the danger zone.
In reality, when something goes wrong, it usually also means that it has a good chance of bypassing your safety system as well...
I suspect that most people burn tweeters with too little power not too much. Your most likely issue is perhaps putting DC into the amp, or perhaps a crossover which doesn't limit the low level power to the correct extent (or in like my case using a PC and having the PC software go a little mad and feed a full range signal to the tweeter output...)
Perhaps Bruno though (or others) has some clever practical tricks to limit power output?
I think you will find in practice that although you can try to reduce the possibilities of tweeter damage. Ultimately you will only damage them when something goes wrong, ie with normal music you shouldn't really get them into the danger zone.
In reality, when something goes wrong, it usually also means that it has a good chance of bypassing your safety system as well...
I suspect that most people burn tweeters with too little power not too much. Your most likely issue is perhaps putting DC into the amp, or perhaps a crossover which doesn't limit the low level power to the correct extent (or in like my case using a PC and having the PC software go a little mad and feed a full range signal to the tweeter output...)
Perhaps Bruno though (or others) has some clever practical tricks to limit power output?
Another DIY amp based on UCD180 maodules
Hi everybody,
I am posting my comments on the amp I just built using UCD180 modules with the AD8620 opamp.
I built the amp with a dual mono power supply. Each module has a 225VA, 2x18V secondaries encapsulated transformer (Multicomp brand), 2 35A bridge rectifiers and 3 4700uF Panasonic TSUP caps per rail, bypassed by a small ceramic (0.1uF) one. Since this is my first diy amp project, I decided to take it 'easy' and buy PCBs for mounting and wiring the PSU components, as well as a soft start module from Ampslab in Singapore. The wiring schema for the PSU follows the method explained by Dejan Veselinovic on his website (www.zero-distortion.com)
When I had almost finished assembling the amp, I read a message from Bruno on this thread saying that i) he did not think much of Panasonic caps, or electrolytics bypassed by a ceramic cap, ii) he recommended smaller Xformer, lower capacitance (I think he mentioned between 4700 and 10,000 uF per rail, or something like this) and that ii) the modules work best with higher voltage supplies (a comment later confirmed by Matjans). My heart sank and I thought: "Omigod! I have created a UDC monster ". After a sleepless night (ok, I'm dramatizing a bit), I steadied my nerve and decided to go ahead, wire the last few bits and a give the amp a listen.
I know that it is easy to be partial about own made things (they are your babies, after all...). And I know, because of what I do in my day job, that memory effects can bias listening tests when you are trying to compare two sets of sounds. Since the real things (actual instrumental sounds in a live situation) are far removed in place and time for most of us, it is easy to convince yourself that your systemn sounds "real", unless your memory of the real thing is very good (i.e., you listen very frequently to the sounds of real instruments). So I thought the best I could do to eval the sound of the UCD amp was to compare it with the other in-house amp. I could not switch instantly between the amps, but this is as good as it gets for assessing the amps chez-nous.
My other amp is one of the earliest class "T" commercial amps. I still like its sound. It has a nice flowing quality, good transparency and good balance when I compared with some of the similarly priced amps at the time of its purchase (few years ago). However, when I started listening to the UCD-based amp, I was able to tell which one is the better amp fairly quickly. Because of the reasons I mentioned above, I preferred to listen to the UCD amp for few days, then swap back to the "T" amp for few hrs and then do the comparison listening to the same music on the same system again (rest of system consists of DIY speakers using Jordan drivers, DIY TVC passive preamp, Marantz SA8260 cdp modded with Zapfilter output modules). The class "T" amp also has XLR input socket at the back so I could also use exactly the same cables.
What are the differences?
- Soundscape. I am not a big fan of this quality of reproduced sounds for various reasons, including those mentioned by Bruno in a previous post. However, it is a bit annoying to hear instruments that move around the speakers, or (e.g.) drums that occupy the whole space between the speakers. With the "T" amp these annoying percepts did happen, with some CDs more than with others. By contrast, with the UCD amp the sounds of various instruments are located in a more precise space around the speakers and do not move around so much. I also get a better sense of depth (but not width) in the placement of instruments with the UCD amp, and smaller perceived "size" of sounds.
- Transparency. This is very important to me. The ability to hear and follow the various instrumental sounds and voices. This is where the UCD amp really shines. With the "T" amp, things are not that bad (in fact better than with most other amps I havef listened to both at home and in other places), but sometimes I could hear the presence of sounds without being able to follow their tunes with ease.
- Timbre. This is also very important to me. Since I have more frequent experience with voice and speech than with music, I will talk about these sounds mostly. Again the UCD amp is better able to reproduce the nuances of voices, and give the impression of a real voice in space than the class "T" amp. One thing that always struck me is the fact that hifi reviewers often talk about the quality of voices (chesty, nasal, thin etc) but rarely comment about the quality of speech sounds. This is one difference that jumped out at me almost immediately when I started listening to the UCD amp. Now the sounds of /s/, /f/ started sounding right: just enough 'stridency' but not too much, when compared with real speech. The /p/, /t/ and /k/ sounds also sounded more accurate than with the "T" amp. With the latter, I had always been a bit unsatisfied about the reproduction of speech and I always guessed it might be the amp, not the speakers or other components that was responsible for this. Another difference that I heard between the amps is the better reproduction of transients (attacks and decays, the latter being the most noticeable) by the UCD amp (btw, transients affect both timbre and localization of sounds). Finally the UCD amp has a better presence and definition of bass notes. Control of bass notes seems a bit better with UCD (in my room I have pretty bad resonances at harmonics of about 30 Hz, so I do notice differences in the reproduction of bass notes easily using certain CDs).
Final comment about "involvement". Not entirely clear how to describe this quality of the sound of equipment. Other people call it PRAT. Well the class "T" amp, at least the first versions like the one I own, has frequently been described as a bit bland in this respect. Even at the low voltage that I am feeding the UCD modules with (I measured about 23.3 V per rail), the UCD amp sounds more "involving" than the class "T" amp.
(A brief comment about the Xformers. I would really like to say that I had the 2x18 V secondaries Xformers lying around the house so I used them. In fact, I simply messed up a Farnell order and purchased Xformers with lower secondaries V than I wanted...).
To conclude, I am selling my class "T" amp. I will get two more modules from Jan-Peter and bi-amp bi-amp the system. Being a tweaker, I will also try to compare the sound with smaller PSU (e.g., only one Xformer and PSU for two modules, changing some of the caps for BG, snip out the ceramic bypass cap and replace it with a polyprop cap). However, it will take some time before I can do all this. I am also in a bit of a dilemma because I purchased the UCD180 modules with the AD8620, which was a bit more expensive than the NE5532 modules, but I have no idea whether a better opamp results in a noticeably better sound. I hope Jan-Peter or somebody else will comment on the differences in the near future.
Well, sorry this message is abit verbose but I thought other forum members might be interested in detailed comments on UCD180, and especially in a comparison with a class "T" amp.
Ciao,
Goliardo
Hi everybody,
I am posting my comments on the amp I just built using UCD180 modules with the AD8620 opamp.
I built the amp with a dual mono power supply. Each module has a 225VA, 2x18V secondaries encapsulated transformer (Multicomp brand), 2 35A bridge rectifiers and 3 4700uF Panasonic TSUP caps per rail, bypassed by a small ceramic (0.1uF) one. Since this is my first diy amp project, I decided to take it 'easy' and buy PCBs for mounting and wiring the PSU components, as well as a soft start module from Ampslab in Singapore. The wiring schema for the PSU follows the method explained by Dejan Veselinovic on his website (www.zero-distortion.com)
When I had almost finished assembling the amp, I read a message from Bruno on this thread saying that i) he did not think much of Panasonic caps, or electrolytics bypassed by a ceramic cap, ii) he recommended smaller Xformer, lower capacitance (I think he mentioned between 4700 and 10,000 uF per rail, or something like this) and that ii) the modules work best with higher voltage supplies (a comment later confirmed by Matjans). My heart sank and I thought: "Omigod! I have created a UDC monster
". After a sleepless night (ok, I'm dramatizing a bit), I steadied my nerve and decided to go ahead, wire the last few bits and a give the amp a listen.I know that it is easy to be partial about own made things (they are your babies, after all...). And I know, because of what I do in my day job, that memory effects can bias listening tests when you are trying to compare two sets of sounds. Since the real things (actual instrumental sounds in a live situation) are far removed in place and time for most of us, it is easy to convince yourself that your systemn sounds "real", unless your memory of the real thing is very good (i.e., you listen very frequently to the sounds of real instruments). So I thought the best I could do to eval the sound of the UCD amp was to compare it with the other in-house amp. I could not switch instantly between the amps, but this is as good as it gets for assessing the amps chez-nous.
My other amp is one of the earliest class "T" commercial amps. I still like its sound. It has a nice flowing quality, good transparency and good balance when I compared with some of the similarly priced amps at the time of its purchase (few years ago). However, when I started listening to the UCD-based amp, I was able to tell which one is the better amp fairly quickly. Because of the reasons I mentioned above, I preferred to listen to the UCD amp for few days, then swap back to the "T" amp for few hrs and then do the comparison listening to the same music on the same system again (rest of system consists of DIY speakers using Jordan drivers, DIY TVC passive preamp, Marantz SA8260 cdp modded with Zapfilter output modules). The class "T" amp also has XLR input socket at the back so I could also use exactly the same cables.
What are the differences?
- Soundscape. I am not a big fan of this quality of reproduced sounds for various reasons, including those mentioned by Bruno in a previous post. However, it is a bit annoying to hear instruments that move around the speakers, or (e.g.) drums that occupy the whole space between the speakers. With the "T" amp these annoying percepts did happen, with some CDs more than with others. By contrast, with the UCD amp the sounds of various instruments are located in a more precise space around the speakers and do not move around so much. I also get a better sense of depth (but not width) in the placement of instruments with the UCD amp, and smaller perceived "size" of sounds.
- Transparency. This is very important to me. The ability to hear and follow the various instrumental sounds and voices. This is where the UCD amp really shines. With the "T" amp, things are not that bad (in fact better than with most other amps I havef listened to both at home and in other places), but sometimes I could hear the presence of sounds without being able to follow their tunes with ease.
- Timbre. This is also very important to me. Since I have more frequent experience with voice and speech than with music, I will talk about these sounds mostly. Again the UCD amp is better able to reproduce the nuances of voices, and give the impression of a real voice in space than the class "T" amp. One thing that always struck me is the fact that hifi reviewers often talk about the quality of voices (chesty, nasal, thin etc) but rarely comment about the quality of speech sounds. This is one difference that jumped out at me almost immediately when I started listening to the UCD amp. Now the sounds of /s/, /f/ started sounding right: just enough 'stridency' but not too much, when compared with real speech. The /p/, /t/ and /k/ sounds also sounded more accurate than with the "T" amp. With the latter, I had always been a bit unsatisfied about the reproduction of speech and I always guessed it might be the amp, not the speakers or other components that was responsible for this. Another difference that I heard between the amps is the better reproduction of transients (attacks and decays, the latter being the most noticeable) by the UCD amp (btw, transients affect both timbre and localization of sounds). Finally the UCD amp has a better presence and definition of bass notes. Control of bass notes seems a bit better with UCD (in my room I have pretty bad resonances at harmonics of about 30 Hz, so I do notice differences in the reproduction of bass notes easily using certain CDs).
Final comment about "involvement". Not entirely clear how to describe this quality of the sound of equipment. Other people call it PRAT. Well the class "T" amp, at least the first versions like the one I own, has frequently been described as a bit bland in this respect. Even at the low voltage that I am feeding the UCD modules with (I measured about 23.3 V per rail), the UCD amp sounds more "involving" than the class "T" amp.
(A brief comment about the Xformers. I would really like to say that I had the 2x18 V secondaries Xformers lying around the house so I used them. In fact, I simply messed up a Farnell order and purchased Xformers with lower secondaries V than I wanted...).
To conclude, I am selling my class "T" amp. I will get two more modules from Jan-Peter and bi-amp bi-amp the system. Being a tweaker, I will also try to compare the sound with smaller PSU (e.g., only one Xformer and PSU for two modules, changing some of the caps for BG, snip out the ceramic bypass cap and replace it with a polyprop cap). However, it will take some time before I can do all this. I am also in a bit of a dilemma because I purchased the UCD180 modules with the AD8620, which was a bit more expensive than the NE5532 modules, but I have no idea whether a better opamp results in a noticeably better sound. I hope Jan-Peter or somebody else will comment on the differences in the near future.
Well, sorry this message is abit verbose but I thought other forum members might be interested in detailed comments on UCD180, and especially in a comparison with a class "T" amp.
Ciao,
Goliardo
Goliard,
Many thanks for telling us your experience!
Loudandclear,
The NE5532 and the AD8620 can give you both 100K input impedance if you change the two input resistance from 10K to 100K. (Watch SMD 0805 case).
It seems the AD8620 is one of the best opamp.
Regards,
Jan-Peter
www.hypex.nl
Many thanks for telling us your experience!
Loudandclear,
The NE5532 and the AD8620 can give you both 100K input impedance if you change the two input resistance from 10K to 100K. (Watch SMD 0805 case).
It seems the AD8620 is one of the best opamp.
Regards,
Jan-Peter
www.hypex.nl
Oops I didn't know my email was disabled. In any case I got your message forwarded by JP so you will receive an answer shortly.tiki said:
The tweeter amplifier should have similar power as the woofer amp, simply in order to handle peaks. While average power on a tweeter may only be about 10% of that on the woofer, peak power will be the same.tiki said:
The result is that just like in a passive system (see ghemink's post), the wrong signal can damage your tweeter. White noise is the most obvious example.
Tweeter protection should be done by sensing power after the filter and shutting down the amp in case too much power is being delivered for too long. Optimally you'd use an rms detector followed by a lowpass filter tuned to have a time constant similar to the thermal time constant of the tweeter. Shortcuts are OK but you need to set the detection threshold on the safe side.
This analysis is 100% correct. The inductance goes up by 50% if the shell is closed, because the field that is normally external to the air coil is now conducted by the ferrite. This shortens the air path to the height of the coil. If the coil were still a "true air coil" this should not happen of course. What I meant to say is that this is the closest compromise between the air coil ideal and the requirement of no stray field.tiki said:
The input impedance is explicitly set by two resistors placed next to the input connector. If you change them to 100k but you keep the 5532, do not remove the coupling caps, because the 5532 is liable to develop large offsets.loudandclear said:
Also AC-wise the 5532 likes to be driven by somewhat lower impedances, so if you have a passive pre the sonic improvement from changing to 8620 is likely to be bigger than when a low-impedance preamp is used.
maxlorenz said:
I only know the Cerafines and on average have positive experiences with them.
On-semiconductor (ex motorola) and ST have a vast range of fast and soft recovery diodes. You might even consider schottky diodes. On the other hand, don't underestimate the effectiveness of a 47nF cap placed right across the diodes. You might be perfectly happy with more mundane diodes with these caps on.
Thank you, ewildgoose and Bruno, for your explanations.
The Prms-detector is a good idea (I^2*R), the determination of the thermal inertia of the loudspeakers coil will be more difficult. But who knows about the thermal conduction/convection/radiation out of the coil (which determine the thermal decay)? Just an idea: I heat up the coil while observing the dc resistance increase as a function of the temperature, after that let it cool down while observing the resistance decrease again.
Otherwise I should look for that in dedicated forums, not to bore the pants off you any longer.
Regards, Timo
The Prms-detector is a good idea (I^2*R), the determination of the thermal inertia of the loudspeakers coil will be more difficult. But who knows about the thermal conduction/convection/radiation out of the coil (which determine the thermal decay)? Just an idea: I heat up the coil while observing the dc resistance increase as a function of the temperature, after that let it cool down while observing the resistance decrease again.
Otherwise I should look for that in dedicated forums, not to bore the pants off you any longer.
Regards, Timo
Silly Questions
I know these are basic questions but I haven't done this before !
When connecting wires to the UCD Modules is it better to solder or use connectors?
What are the connectors called for the 5 flat power and speaker connections? are they best soldered on the wire or crimped?
Is the 3 pin connector a standard part that I can buy? if so what's it called.
What size are the screws that fit the threads for the heat sink?
What sort of stand off is best to screw the module to the case?
Its strange how these simple things catch out beginners
I know these are basic questions but I haven't done this before !
When connecting wires to the UCD Modules is it better to solder or use connectors?
What are the connectors called for the 5 flat power and speaker connections? are they best soldered on the wire or crimped?
Is the 3 pin connector a standard part that I can buy? if so what's it called.
What size are the screws that fit the threads for the heat sink?
What sort of stand off is best to screw the module to the case?
Its strange how these simple things catch out beginners
. the ucd180 boards have connector terminals so if i were you i'd use them. easier to work with if you want to replace parts on the boards.
. they're called lug connectors and imho they're best crimped. don't use solid wire for these connectors if you crimp them. use shielded ones, especially for the power supply pins.
. with 3-pin connector i guess you mean the amp's input connector. you can use a standard 3-pin xlr chassis connector for this. neutrik has a few with pin1 soldered directly to the connector chassis. try to use those if you can get them. use pin 2 and three for hot and cold (signal+ and signal-)
. don't use screws to connect the heatsink to another heatsink or to the case, use nuts and bolts, size m3.
. they're called lug connectors and imho they're best crimped. don't use solid wire for these connectors if you crimp them. use shielded ones, especially for the power supply pins.
. with 3-pin connector i guess you mean the amp's input connector. you can use a standard 3-pin xlr chassis connector for this. neutrik has a few with pin1 soldered directly to the connector chassis. try to use those if you can get them. use pin 2 and three for hot and cold (signal+ and signal-)
. don't use screws to connect the heatsink to another heatsink or to the case, use nuts and bolts, size m3.
Bruno Putzeys said:
Hi Bruno,
Just to make sure I do things right.
I`m planning to use XLR inputs. Some people say, pin 1 should only be connected to the chassis, not to the signal GND. Then the amplifier GND should be connected to the chassis somewhere.
What I have done now is that I have connected the UcD input GND to pin 1 and I have connected the powersupply GND to the chassis and I have also connected the GND line of the power cord cable to GND.
I will connect the poweramps to a DEQX PDC unit (digital crossover and speaker correction system) that has XLR out. According to the manufacturer of that product, pin1 of their XLR out is connected to the chassis via a 1nF capacitor (to avoid RF pickup by the shield I guess), pin 1 is not connected DC to GND. I plan to connect the power amps and PDC unit all with 3-wire power cords so that all chassis are connected via the power cord GND wire. In this way I hope to prevent that there is any offset in the GND levels of the amps and the PDC. I want to do this to reduce or avoid that a common mode signal is present on the inputs of the UCD.
Does the above make sense or do you have other advise?
Thanks and best regards
Gertjan
Hello Lars,Lars Clausen said:I'm a bit puzzled:
UcD stands for "Universal class D"
Some time in 2001 I was asked by our Class D account manager to develop "the Ultimate Class D amp".
The resulting circuit was elegant and did pretty much everything you'd ever want an amplifier to do. Good efficiency, smooth frequency response (totally indifferent to load), excellent EMC and very simple (=low cost in large quantities). Only THD was just "good" at 0.02% (all freq, power up to Pr-3dB). This was not surprising since I had pinned the design target at that value. It was obvious my employer wouldn't pay a penny more for improved THD.
As I preferred to reserve the word "Ultimate" for an amplifier having, on top of the above characteristics, less than 0.0001% THD I changed the U to mean "Universal".
Hypex' modules use this circuit and hence the name.
Cheers,
Bruno
Bruno,
There has been quite a bit of chat about power supplies for the UCD. Given that the UCD is perhaps less sensitive than some other traditional amp designs could you perhaps offer an opinion on what you would suggest as a base line supply design for one of these modules.
What I mean by that is just the basics of how extreme you would go in design, eg would you go for dual transformers, dual power supplies, or even dual power supplies with two transformers each. You gave some suggestions for sizing caps before, but just to be definitive, what kind of size would you go for, for: a) UCD 180 normal load, b) UCD400 normal load, c) UCD 400 driven to the limit into dual 4ohm sub bass systems
It's all fairly obvious stuff, I'm just basically curious to understand what you would recommend as the base line spec to get sound out of these things (and perhaps also where you would go for "over-engineered" if you were building something to drive your own system...) The point is I guess it's very easy to spend time over-engineering stuff that doesn't matter, or losing potential quality with stuff that does
Thanks for any thoughts
By the way, I found this link (posted already I think), quite useful in understanding the basic flavours of powersupply designs:
http://www.tnt-audio.com/clinica/ssps1_e.html
There has been quite a bit of chat about power supplies for the UCD. Given that the UCD is perhaps less sensitive than some other traditional amp designs could you perhaps offer an opinion on what you would suggest as a base line supply design for one of these modules.
What I mean by that is just the basics of how extreme you would go in design, eg would you go for dual transformers, dual power supplies, or even dual power supplies with two transformers each. You gave some suggestions for sizing caps before, but just to be definitive, what kind of size would you go for, for: a) UCD 180 normal load, b) UCD400 normal load, c) UCD 400 driven to the limit into dual 4ohm sub bass systems
It's all fairly obvious stuff, I'm just basically curious to understand what you would recommend as the base line spec to get sound out of these things (and perhaps also where you would go for "over-engineered" if you were building something to drive your own system...) The point is I guess it's very easy to spend time over-engineering stuff that doesn't matter, or losing potential quality with stuff that does
Thanks for any thoughts
By the way, I found this link (posted already I think), quite useful in understanding the basic flavours of powersupply designs:
http://www.tnt-audio.com/clinica/ssps1_e.html
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.