I'm baffled as to why my 5800 is having such a tough time driving my ML CLS2A's(glary mids and tizzy treble). Granted, the CLS impedance dips to less than 1 ohm @18-20kHz but I always thought the beefy power supply and 16 output devices/ch of this beast could handle the job. Can this be remedied by bias adj or some other relatively easy tweak? Appreciate any input?
Jerry
Jerry
The Martin Logan CLS went through a few changes during its lifespan and I didn't hear all of them, but to be quite frank with you I never found them acceptable. Why? Exactly what you describe--glassy, forward upper midrange and less than optimal top end.
In other words, are you sure it's the amp that's the problem?
Speaking only for myself, I wouldn't take a pair of CLSs on a silver platter. They were actively offensive to me. Yes, I know they reviewed well and some people swore by them, but to me...well, you get the idea. Don't blame the amp if it's the speaker.
Grey
In other words, are you sure it's the amp that's the problem?
Speaking only for myself, I wouldn't take a pair of CLSs on a silver platter. They were actively offensive to me. Yes, I know they reviewed well and some people swore by them, but to me...well, you get the idea. Don't blame the amp if it's the speaker.
Grey
Hi Grey
I know it's the amp because my VSP Trans/Mos 200w works very well with the CLS. A friends Hafler 9505 Transnova also was quite good. BTW, one must use tubes somewhere in the chain of amplification when running CLS'S. I use an AL MOD 2D pre. Perhaps your listening experience with the CLS was affected by incompatible power amplification. You should'nt go all solid state.
Jerry
I know it's the amp because my VSP Trans/Mos 200w works very well with the CLS. A friends Hafler 9505 Transnova also was quite good. BTW, one must use tubes somewhere in the chain of amplification when running CLS'S. I use an AL MOD 2D pre. Perhaps your listening experience with the CLS was affected by incompatible power amplification. You should'nt go all solid state.
Jerry
CLS2ASTAT said:
The first thing to know is that I adore tubes. The second thing to know is that I own tubes. The third thing to know is that I've designed with tubes. That said...no speaker should force you to listen to use tubes in order to sound tolerable.
If you're trying to use a warmish sounding amp to tame the aggressive mids and highs of the CLSs, consider an Aleph.
Grey
GRollins said:
I love tubes also. In fact I Wish I could afford a high powered tube amp to drive my CLS but I think you're missing the point. Their notoriously tough load can give many so called "high end" solid state amps a hard time. And oddly enough, just for grins we tried a Berenger Euro that my friend uses strictly as a sub amp. It did a respectible job sonically(this amp is NOT euphonic sounding), because it could deliver enough volt/amps to handle the nasty impedance of the ML. The hard aggressiveness is not a speaker issue, but rather an amp issue.
Jerry
>>BTW, one ohm speaker can design only moron
Upupa Epops, 4\2006
http://www.diyaudio.com/forums/showthread.php?postid=898944#post898944
Upupa Epops, 4\2006
http://www.diyaudio.com/forums/showthread.php?postid=898944#post898944
What would you expect from an amplifier with beefy power supply and outputs? It's maintaining a strong signal into a low impedance load. An amplifier that can't handle an impedance dip will result in a dip in overall frequency response at the frequency of the impedance dip. This is just the case of a more powerful amplifier flexing its muscle.
If we're talking about an Adcom design on a Pass circuit, there is chance that the effort to make it commercially viable for repair and resale to the average joe made the unit possibly different than Nelson did in the proto stage. Most commercial designs from Adcom used quite a bit of wire and fusing, this means the transient impedance considerations could be quite far off from Nelson's original proto design.
Thus they tend to get harsher and more confused under heavy loading than what Nelson designed into them in the first place. Just my guess and experience from taking Adcoms apart and 'hot wiring' and 'hot rodding' them. I dunno really. Only Nelson and Adcom know and both are quite reasonably not expected to speak on the subject, under that particular line of thought.
Thus they tend to get harsher and more confused under heavy loading than what Nelson designed into them in the first place. Just my guess and experience from taking Adcoms apart and 'hot wiring' and 'hot rodding' them. I dunno really. Only Nelson and Adcom know and both are quite reasonably not expected to speak on the subject, under that particular line of thought.
Thanks for your input. A hypothetical question: If this was your amp and you wanted it to drive your CLS, what would you do to make this amp compatable? Bias(how high),fusing, wiring,output device matching? I think this amp would do an admirable job with the right tweaking. Thanks again.
Jerry
Jerry
KBK's point about differences between the prototype and later production models is valid. In general, if someone designs a circuit on a contract basis, you will find that changes have been made at some point later in the game--often without even telling the designer.
As to how much bias the amp can stand, that will depend on a number of factors:
1) The SOA of the output devices
2) How much current the power supply (essentially boiling down to the transformer) can provide before the rails begin to collapse
3) How much heat the heatsinks can dissipate
Points 1 & 3 are somewhat related.
In other words, it's difficult to make a blanket statement. I expect it's safe to say that the amplifier will fail if you try to bias it to class A. The question then limits to how far into class AB can it be pushed. Given production tolerances and drift in the field it's probably safe to say that you can increase the bias by 10 to 20%. Going much farther than that would depend on detailed answers to 1-3 above.
Fusing would be determined by biasing.
Wiring is one of those things that you'll get a thousand opinions on--frequently contradictory. Since audiophile cabling gets expensive quickly you'll need to decide how much you can afford to put into it.
Output device matching is another matter entirely. You're talking about a complete rebuild at that point. To match devices you'll need a large batch to select from. That's a lot of money. Do you have some reason to believe that the devices in the amp are not up to the job?
Grey
As to how much bias the amp can stand, that will depend on a number of factors:
1) The SOA of the output devices
2) How much current the power supply (essentially boiling down to the transformer) can provide before the rails begin to collapse
3) How much heat the heatsinks can dissipate
Points 1 & 3 are somewhat related.
In other words, it's difficult to make a blanket statement. I expect it's safe to say that the amplifier will fail if you try to bias it to class A. The question then limits to how far into class AB can it be pushed. Given production tolerances and drift in the field it's probably safe to say that you can increase the bias by 10 to 20%. Going much farther than that would depend on detailed answers to 1-3 above.
Fusing would be determined by biasing.
Wiring is one of those things that you'll get a thousand opinions on--frequently contradictory. Since audiophile cabling gets expensive quickly you'll need to decide how much you can afford to put into it.
Output device matching is another matter entirely. You're talking about a complete rebuild at that point. To match devices you'll need a large batch to select from. That's a lot of money. Do you have some reason to believe that the devices in the amp are not up to the job?
Grey
I agree with both of you regarding final design compromises being chosen when this amp was released for production. As an automotive test engineering tech I see this all the time in the development of a new vehicle. I guess I was too naive to think think that it wouldn't apply to audio as well. My bad. However, since Nelson Pass showed that the proto amp was up to the task of driving ML stats, I think that my 5800 certainly has the potential of working well. I will try upping the bias first and go from there but don't plan on dumping money into this. Thanks for your input.
Jerry
Jerry
CLS2ASTAT said:
Given that this is the apparently the first version of the amp, and not
a later version, such as the 5802, which was different, I don't see
any reason why it shouldn't drive the MLs. There are a number of
presumptions:
1) No weird interconnect or speaker cables.
I mention this because occasionally you can still see the occasional
issue with unterminated Litz type wire (low inductance / high
capacitance. If there is any question about this, then a "Zobel"
network should be placed across the speaker terminals, with values
in the .1 uF + 10 ohm or so.
Also, the output Zobels on the amplifier should be checked to see
if the resistor is cooked.
The interconnects need to have good shields, and you should
confirm that all the grounds on both ends are connected - you can
use an ohmmeter for this.
2) The amplifier is properly biased. This means you need to check
the operating voltage across the Source resistors after a 1 hour
warm-up.
3) The impedance of the ML's should be checked, even if only
with a multimeter. All to often people get excited about distortion
into electrostats only to find out that there is a short circuit.
Does the amplifier sound this way at low levels, high levels, etc?
How does it sound if you put a 1 ohm power resistor in series with
the loudspeaker?
My favorite thing to do with my PERSONAL use Adcom amps (Not saying you have to do this), not one I mod for re-sale, is to slightly up the fusing that goes to the DC rails, with respect to passing current in a lower impedance and slightly more linear fashion. This cascades into 'silkier' high frequencies, in the general sense, at the same time it 'solidifies' the bottom end or transients in general. This is inexpensive to execute, as fixes go.
I also run them with bypassed speaker fusing. That is the part you should evaluate yourself, as per what kinds of risks you feel you are taking with regards to potential output section failure-if it happens, and what the costs to given speakers/amp are. In my experience output sections fail regardless if a speaker fuse is there or not, if they are accidentally shorted at the speaker cable end of things. I usually buy amplifiers that have no fusing of any kind except for a AC power fuse, finding all other designs having a distinct harshness/dirtiness due to transient limiting from fusing.
I also find that the DC rail connections in Adcoms (not all) can be quite remote from the capacitance for buffering the DC rails. On the ones I've encountered, quality, low impedance DC rail buffering as close to, if not directly ON the output section boards can create quite the sonic benefit. This can also be relatively inexpensive to execute, as mods go. About $40 in quality capacitance, thus about $10 per DC rail spent (on capacitors) and then installed. This buffering should most specifically be AFTER the fusing for the DC rails and direct run to the DC rail-output section electrical point. This is the whole point of installing these particular capacitors.
Between the two simple changes, then the spoken of bias considerations, the sonics of the amplifier can be notably increased in perceived quality, in the direction you wish to head in.
I also run them with bypassed speaker fusing. That is the part you should evaluate yourself, as per what kinds of risks you feel you are taking with regards to potential output section failure-if it happens, and what the costs to given speakers/amp are. In my experience output sections fail regardless if a speaker fuse is there or not, if they are accidentally shorted at the speaker cable end of things. I usually buy amplifiers that have no fusing of any kind except for a AC power fuse, finding all other designs having a distinct harshness/dirtiness due to transient limiting from fusing.
I also find that the DC rail connections in Adcoms (not all) can be quite remote from the capacitance for buffering the DC rails. On the ones I've encountered, quality, low impedance DC rail buffering as close to, if not directly ON the output section boards can create quite the sonic benefit. This can also be relatively inexpensive to execute, as mods go. About $40 in quality capacitance, thus about $10 per DC rail spent (on capacitors) and then installed. This buffering should most specifically be AFTER the fusing for the DC rails and direct run to the DC rail-output section electrical point. This is the whole point of installing these particular capacitors.
Between the two simple changes, then the spoken of bias considerations, the sonics of the amplifier can be notably increased in perceived quality, in the direction you wish to head in.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.