Another summary
Hi,
the experiment didn't end because with soldering and unsoldering poor AMP5 I finished to fry it
Nevertheless there are some answers.
The coils I'm using now are the result of what I learned here. T106 Molypermalloy core 60 permeability, 13 turns 16AWG wire:
They are in two dual mono new 41Hz AMP10 based on the same TA2022 like AMP5. The coils sound wonderful, slightly better than Bert's coils. However, one time, I mounted Bert's coils on one mono amp and MPP coils on the other mono amp: I couldn't notice much differences. So, Bert's coils are as good as MPP, or a better coil than Bert's one has no influence on this amp.
I believe that Bert's winding technique improves coils performance, though I don't know why: the reason is the following
On the left are T106 with same core material of Bert's T90 coils (on the right). Bigger T106 coils would sound better or at least equal, but this is not the case: they sound just similar to T80 coils supplied with AMP5 kit, far worse than Bert's coils and MPP coils.
Next step would be to try Bert's winding technique with MPP cores. But I could get only MPP toroids with high permeability and with these 13 turns are enough to get the 11uH I need. With so few turn is not possible to cross windings, or I couldn' be able.
So things left to experiment collecting suggestions from this thread are the following:
1) MPP cores with 24 permeability (suggested by Piotr). ~25 turns are possible for 11uH, so Bert's winding technique is usable.
2) Copper foil coils (with these I fried AMP5)
Kurt (Soundcheck) suggested and tried these on Charlize with good results. It is a low power amp though.
Roger (sx881663) suggested good mounting way to avoid RF and EMI here
3) There is a variant for above copper foil coils suggested by Charles (phase_accurate) in another thread
4) Finally there are shielded coils suggested by Lars Clausen in the same thread
I hope these four suggestions be useful to everybody is going to build a class-D amp and wants to try something different and report here results
Hi,
the experiment didn't end because with soldering and unsoldering poor AMP5 I finished to fry it
Nevertheless there are some answers.
The coils I'm using now are the result of what I learned here. T106 Molypermalloy core 60 permeability, 13 turns 16AWG wire:
They are in two dual mono new 41Hz AMP10 based on the same TA2022 like AMP5. The coils sound wonderful, slightly better than Bert's coils. However, one time, I mounted Bert's coils on one mono amp and MPP coils on the other mono amp: I couldn't notice much differences. So, Bert's coils are as good as MPP, or a better coil than Bert's one has no influence on this amp.
I believe that Bert's winding technique improves coils performance, though I don't know why: the reason is the following
On the left are T106 with same core material of Bert's T90 coils (on the right). Bigger T106 coils would sound better or at least equal, but this is not the case: they sound just similar to T80 coils supplied with AMP5 kit, far worse than Bert's coils and MPP coils.
Next step would be to try Bert's winding technique with MPP cores. But I could get only MPP toroids with high permeability and with these 13 turns are enough to get the 11uH I need. With so few turn is not possible to cross windings, or I couldn' be able.
So things left to experiment collecting suggestions from this thread are the following:
1) MPP cores with 24 permeability (suggested by Piotr). ~25 turns are possible for 11uH, so Bert's winding technique is usable.
2) Copper foil coils (with these I fried AMP5)
An externally hosted image should be here but it was not working when we last tested it.
Kurt (Soundcheck) suggested and tried these on Charlize with good results. It is a low power amp though.
Roger (sx881663) suggested good mounting way to avoid RF and EMI here
3) There is a variant for above copper foil coils suggested by Charles (phase_accurate) in another thread
4) Finally there are shielded coils suggested by Lars Clausen in the same thread
I hope these four suggestions be useful to everybody is going to build a class-D amp and wants to try something different and report here results
Thanks, Thomas, for all your work on the coils.
I like the high perm cores because they need so few turns. Great for low DCR and easy to wind. But they seem to saturate sooner then low perm coils - have you found that?
Different core types can have a big influence on how much RF is left in the outputs.
The AMP6 basic uses surface mount coils. Brian BWRX sent me one of his amps to listen to that had similar coils. I thought they sounded good. But maybe they don't come in large enough current rating for the big amps?
I like the high perm cores because they need so few turns. Great for low DCR and easy to wind. But they seem to saturate sooner then low perm coils - have you found that?
Different core types can have a big influence on how much RF is left in the outputs.
The AMP6 basic uses surface mount coils. Brian BWRX sent me one of his amps to listen to that had similar coils. I thought they sounded good. But maybe they don't come in large enough current rating for the big amps?
16AWG....that is interestingly thick wire.
May I ask a few questions:
A) Where did you source such material? What is the quality like?
B) What about skin effect considerations? Have you experimented with Litz wire?
It would be interesting to know what your views are on my thread
Many thanks
May I ask a few questions:
A) Where did you source such material? What is the quality like?
B) What about skin effect considerations? Have you experimented with Litz wire?
It would be interesting to know what your views are on my thread
Many thanks
#16 is quite a common gauge for air core crossover design. Having tried solid core, litz, and foil in crossovers, my instinct tells me that litz should be quite good for this kind of class D application. How much difference can be heard with only 13 windings is much easier determined using experiement than analysis. One way of trying is to get a Solen litz air core inductor and unwind the portion you need.
My 2 cents:
My Charlize, after almost a year by now, is still working great with the foil coils.
There are some threads confirming my shielding approach.
Shielding the coil will improve the quality even further.
Fixing the windings with some kind of adhesive will reduce
distortions caused by vibrations.
Cheers
Klaus
My Charlize, after almost a year by now, is still working great with the foil coils.
There are some threads confirming my shielding approach.
Shielding the coil will improve the quality even further.
Fixing the windings with some kind of adhesive will reduce
distortions caused by vibrations.
Cheers
Klaus
I'd like to ask what do you think about using a class-D output filter as a crossover filter in a bi-amp or tri-amp system.
Suppose to have an amp for the woofers and another for the tweeters. Tripath output filters are usually second order low pass with a cut at 107khz. What happens if I have a ~1000Hz LRC LP filter for the woofers amp and a ~1000Hz-107Khz pass band filter for the tweeter amp?
Advantage is to have a kind of active crossover inside the amp without losses due to passive filters.
Of course otput filter can be higher order too.
What the drawbacks?
P.S.
Klaus, I don't remember you ever described your shielding approach with your copper foil coils.
Suppose to have an amp for the woofers and another for the tweeters. Tripath output filters are usually second order low pass with a cut at 107khz. What happens if I have a ~1000Hz LRC LP filter for the woofers amp and a ~1000Hz-107Khz pass band filter for the tweeter amp?
Advantage is to have a kind of active crossover inside the amp without losses due to passive filters.
Of course otput filter can be higher order too.
What the drawbacks?
P.S.
Klaus, I don't remember you ever described your shielding approach with your copper foil coils.
thomaseliot said:
Your idea might work. I still wouldn't call it "active" filtering.
It stays passive.
IMO the best would be to have no filters in the output, or at least eliminate
the filter effects and interaction with the speaker (see UCD!)
Running the speakers active is a must to me. I am filtering in front of the amp.
You might consider passive line level filtering, you'd need much smaller parts for the filters, which will end up in lower losses and more $ on your bank account.
E.g. I am using my Charlize from 200Hz onwards. I swapped my input cap to 0,02 uF giving me a nice 6db highpass. The size of the cap and the filter effect brought
quite some improvements to the sound (and is saving me lot of money!).
However - The future are DSP based crossovers anyhow. Better concentrate on these.
( I am using a DCX2496 for the subs right now and will soon switch to brutefir - under Linux) Shielding:
I mentioned that in another thread some time ago.
I wrapped my coils with ERS paper. Others are covering them with copper. MU-metal might be the best choice. If you work with copper or MU-metal
you have to connect them to ground.
BTW: You showed above an alternative foil coil. At least from that what I see on the picture I'd say - forget it. The build quality seems to me awful. I can not imagine
that this one does any good to the sound.
Cheers
woodmancy said:
Amp5 fried and 41Hz had a new model with the same chip. The better channel separation with dual mono improves a lot performances of the amp. Then with a simple jumper they can work bridged.
Here it is:
Transformers are 300VA each 2x22VAC. Input caps are Mundorf 4.7uF and coils are MPP T106 13 turns 16AWG.
Hi Thomaseliot:
I want to do this. Any second thoughts based on your experience, separate enclosures, input caps etc? Given the original title of this thread, you must believe that this is about the best you can do with a Tripath? How about the new Truepath from 41Hz? I have a lot of respect for your opinion and a lot of faith in the Tripath (I have Charlize, Amp 3, Amp 6 basic, Amp 5, Sonic T and TweaKit 3 on the TA-10 from Audiomagus)
Keith
I want to do this. Any second thoughts based on your experience, separate enclosures, input caps etc? Given the original title of this thread, you must believe that this is about the best you can do with a Tripath? How about the new Truepath from 41Hz? I have a lot of respect for your opinion and a lot of faith in the Tripath (I have Charlize, Amp 3, Amp 6 basic, Amp 5, Sonic T and TweaKit 3 on the TA-10 from Audiomagus)
Keith
soongsc said:
Hi George,
still waiting for your enclosures to appear here
Low frequencies are very good, though I have BD-Design Singulars with FE206E that don't excel in low frequencies, but the quality is excellent.
Now I'm experimenting a bi-amp system with a front horn with FE206 and Singulars with a Fostex FW208N woofer.
An externally hosted image should be here but it was not working when we last tested it.
The woofer is driven in turn by AMP10 and UCD180, and the horn in turn by ZenV9 and AMP10. Now low frequencies are huge, something that let you feel in a doubled size room, and AMP10 are excellent in both roles.
woodmancy said:
Hi Keith,
I think with Amp10 I've done the best doable just tweaking external components. Other improvements would involve amp design, that is beyond I'm able to do.
Truepath, for sure will be my next project
It is a new design with 4 layers PCB to distribute different grounds each to its own layer. This would affect EMI/RF and distorsion. This is that kind of real improvements I waited for.I don't know any other Tripath amp (including commercial Bel Canto Evo2 genII based on the same chip) with such a complex design.
thomaseliot said:
Hi George,
still waiting for your enclosures to appear here
Low frequencies are very good, though I have BD-Design Singulars with FE206E that don't excel in low frequencies, but the quality is excellent.
Now I'm experimenting a bi-amp system with a front horn with FE206 and Singulars with a Fostex FW208N woofer.
An externally hosted image should be here but it was not working when we last tested it.
The woofer is driven in turn by AMP10 and UCD180, and the horn in turn by ZenV9 and AMP10. Now low frequencies are huge, something that let you feel in a doubled size room, and AMP10 are excellent in both roles.
Here's a link where someone posted one picture.
http://www.myav.com.tw/forum/showthread.php?s=&threadid=270772&perpage=12&pagenumber=12
One factory had no problem with samples of the bottom part, but came out not so good when we gave them a small quantity order doing the MDF power coating. I've begun to discover this to be very typical among small factories that don't have the resources to get into the technical detail. So it seems like a walnut bottom will be used where we will also be trying different types of wood on the CNC. Just means more variety.
I've also been playing around with some remote volume control and source selector to get a feel on what it takes to get rid of the preamp. Also found a source that might be able to do some SMPS to match specific requirements. It would be interesting to see how a matched SMPS will effect amplifier performance.
HI Thomas,
since I also posses amp5 I was really enjoing this thread and after doing some tweaks (input cap, PS, output cap) I also think of trying new coils.
I have to place them on top, so I will go with an OD=0,928inch. I am thinking of this mmp core from cws: CM234125 with 12 Turns of AWG 16.
Im no expert but I would prefer permeability 125 over 60, simply because then I need less turns, hence less losses
Regarding your speakers, I found a nice front loaded fostex horn with a 20 cm rear loaded Woofer design:
http://eckhorn.bonifa.net/Hornlautsprecher/BK_201_KH.php
I myselve own the BK 108 FF and I like it a lot:
http://eckhorn.bonifa.net/Hornlautsprecher/BK_108_FF.php
since I also posses amp5 I was really enjoing this thread and after doing some tweaks (input cap, PS, output cap) I also think of trying new coils.
I have to place them on top, so I will go with an OD=0,928inch. I am thinking of this mmp core from cws: CM234125 with 12 Turns of AWG 16.
Im no expert but I would prefer permeability 125 over 60, simply because then I need less turns, hence less losses
Regarding your speakers, I found a nice front loaded fostex horn with a 20 cm rear loaded Woofer design:
http://eckhorn.bonifa.net/Hornlautsprecher/BK_201_KH.php
I myselve own the BK 108 FF and I like it a lot:
http://eckhorn.bonifa.net/Hornlautsprecher/BK_108_FF.php
There must be a relationship between sensitive speakers owners and Tripath lovers.
Tempting idea to use higher permeability cores. Please keep posting.
Why not to try lower permeability too, to compare? 14 or 24. Turns wouldn't be so much.
But CWS has not these. Another source is Arnold Magnetics (with on line shop) that is the CWS supplier.
Tempting idea to use higher permeability cores. Please keep posting.
Why not to try lower permeability too, to compare? 14 or 24. Turns wouldn't be so much.
But CWS has not these. Another source is Arnold Magnetics (with on line shop) that is the CWS supplier.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.