You've the electrolytic capacitors very near the signal path (volume potentiometer), since they've very big aluminum containers they could be "injecting" HF noise in the inputs (the potentiometer)..
I would place the capacitors on the other side of the metal shield..to do it I would move the toroid a bit toward the front of the case and put the filter capacitor PCB between the IEC socket and the toroid if necessary the metal shield could be placed a bit on the right..
Does the toroid hum ?
I would place the capacitors on the other side of the metal shield..to do it I would move the toroid a bit toward the front of the case and put the filter capacitor PCB between the IEC socket and the toroid if necessary the metal shield could be placed a bit on the right..
Does the toroid hum ?
The shielding around the capacitors would just diminish the interference a bit.. I would move them away.. that actually reduces the chances they could harm the inputs..
Another important thing is how you connected the ground.. but from the picture it's impossible to see that.
Last but not least.. use a DC stopper for the toroid mains.. most of the times this mod produces differences (for the better) that are easily noticeable and it's very easy to implement..
However, the amp should be only a bit warm when running.. if the heatsink gets quickly hot or the toroids on the outputs are very warm it means it is self oscillating.. and the quality of the sound is terrible when running in these conditions..
Another thing, the input cables make 100% sure their grounds are not connected to the case and don't make them pass under the board. In theory the best is to keep them as short as possible..
Another important thing is how you connected the ground.. but from the picture it's impossible to see that.
Last but not least.. use a DC stopper for the toroid mains.. most of the times this mod produces differences (for the better) that are easily noticeable and it's very easy to implement..
However, the amp should be only a bit warm when running.. if the heatsink gets quickly hot or the toroids on the outputs are very warm it means it is self oscillating.. and the quality of the sound is terrible when running in these conditions..
Another thing, the input cables make 100% sure their grounds are not connected to the case and don't make them pass under the board. In theory the best is to keep them as short as possible..
Last edited:
From the pictures that looks like a solid build Richard.
And mr. Sgarbialan, thanks on behalf of the community for sharing your wisdom and taking the time to write a very helpful rap with your experiences. My experience is the mosfets radiate most of the switching noise and emission, hence the care of 41hz products to minimize the capacitive coupling between the fets and the heatsink by using thick ceramic blocks in between. Here lies a great opportunity for the connex modules, for they have to live with only a slab of silicone patch between the heatsink and fets.
But that is just one trick. I have a Rhode and Schwarz spectrum analizer, a helpful tool to analize where the noise is radiated most and how it is best contained. Once the noise is roaming free it is notoriously hard to get rid of it, so it's best tackled at the source. Here lies a great responsibility for the board design, where a JABOT board (just a bunch of traces) just won't pull it off.
And mr. Sgarbialan, thanks on behalf of the community for sharing your wisdom and taking the time to write a very helpful rap with your experiences. My experience is the mosfets radiate most of the switching noise and emission, hence the care of 41hz products to minimize the capacitive coupling between the fets and the heatsink by using thick ceramic blocks in between. Here lies a great opportunity for the connex modules, for they have to live with only a slab of silicone patch between the heatsink and fets.
But that is just one trick. I have a Rhode and Schwarz spectrum analizer, a helpful tool to analize where the noise is radiated most and how it is best contained. Once the noise is roaming free it is notoriously hard to get rid of it, so it's best tackled at the source. Here lies a great responsibility for the board design, where a JABOT board (just a bunch of traces) just won't pull it off.
Last edited:
Lucky you that can buy Rhode and Schwars.. Unfortunately not everybody can..
However I was able to run Cristi's amps with no problems of self oscillation so I guess also someone else can do it don't you think so?
Regarding noise emissions.. I do have an AMP7 and the mosfets are tied with a mica insulator directly to the heatsink on the back of the PCB.. is this a mistake?
Thanks, I will implement that.
It only gets slightly warm, never "hot."
They are not connected to the case at all, but they pass under the board slightly in the lower right corner. I will move them asap.
Thanks for your help, you have been very very helpful.
ATB,
Rich.
so it's not an issue of self.oscillation..
Try also the DC stopper.. you can assemble one with a couple of capacitors and diode bridges.. depending on the quality of your mains you could gain a lot..
P.S: I'm not here to teach anything to anyone.. Just trying to see if we can resolve an issue to somebody..
Last edited:
After you've determined a design shows decent emission values the need for such measures is minimized. And for that last bit of magnetic interference it helps to place the amplifier power stage in a well shielded and grounded box with all in and outgoing connections choked well with ferrite beads.
I saw you named AMP15 there in your first post, which is an almost 100% through hole design. Something nice for newbie diyers, an attempt from the owner of 41hz to make these amps more accessible, but off course this makes them less good than a good SMD design.
Then the tank caps, they can be close to the amp, just not directly facing the power stage. You can make an excellent design with these caps right there on the board itself, only then the power stage components have to have their footprints on the other side of the board, and proper laid out ground planes are necessary.
Now with that last bit of information, hold that Connex amp against some strong light and behold!
Lucky you that can buy Rhode and Schwars.. Unfortunately not everybody can..
However I was able to run Cristi's amps with no problems of self oscillation so I guess also someone else can do it don't you think so?
Regarding noise emissions.. I do have an AMP7 and the mosfets are tied with a mica insulator directly to the heatsink on the back of the PCB.. is this a mistake?
Sure you can, but if that is determined by listening impressions I still am left with some doubts...
That mosfet bit is not right, damn 41hz for not making proper documentation! You must have had that bag with mounting material in your hands thinking:"What on earth is that all for??"
Well, the idea is to place a stand-off between the mosfet body and the PCB. Their sizez are chosen to exactly match with the board stand-off height. Then with the ceramic insulators the mosfets are well away from both the board and the heatsink.
Thanks Vincent, I thought it was a good layout. The shielding needs to be copper plated, and I'll be adding some more soon......
Don't say THAT ! AHAHAH I bought two AMP7 and an AMP15 (plus two AMP10) and I never read in 41Hz forums such an assertion about sound quality.. For sure your top product, the Truepath, has a very nice multi layered pcb design with very small components on it, but comparing the sound of an AMP7 and a Truepath you actually hear the difference ? Maybe you measure it.. but with boxes a human being can buy you actually HEAR that?
I have to say I've a couple of Monitor Audio that are nicely sounding with both brands of the amps.. I don't think I would be able to distinguish them in a blind test.. however I think I would be able to easily distinguish a class AB from a T-amp
regarding the footprint of the components on the other side of the pcb could you please explain me why it is a big advantage? (and here don't take it as a tricky question, I actually don't know)
P.S: for the FETS sure I used the ceramic standoffs.. however if we're speaking about RF switching does it make so much difference 5mm away? Is it something R&S showed ?
Last edited:
If I were to perfect that lay-out by the way then place the potentiometer all the way in the back,. extend the shaft even more, keep input leads as short as possible, rotate the Connex board 180 degrees, place some thick ceramic insulators between the fets and heatsink so the emitted energy towards/from the heatsink is minimized, then you can freely place the electrolytic caps near the amplifier board, which is beneficial for their ESR value. Then place the toroidal transformer wherever you feel like, usually the best is all the way towards the front panel.
There...
There...
it is.. but in fact if the amplifier is not getting hot it means it's not self oscillating..
However I'm quite surprised about the sound quality you say it comes out of the amp..
electronic considerations apart I've a couple of monitor audio too and actually they sound very well with this amp.. it's so strange..
I see you approve some of my advices
One only point, rotating the board make sure the cable of the inputs are not near the cables going to the speakers.. probably the position of the board is already the right one.. shortening the cables by extending the potentiometer shaft instead is for sure a good advice..
Last edited:
Yes I hear that. I shall speak freely as always, also about my definite favorites of 41hz. Which are not mine by any means, I did exchange a couple of ideas at max, but usually blown in the wind by a far superior mind in this respect, the shop owner of 41hz called Jan. My input is a mere very humble input, but yet valued and appreciated...
And no, I don't own the Rhode and Schwarz, they're equipment available at the place I work. I've got lots more goodies there, like audio analizers, wfm monitors, very high frequency sweep generators, tone generators, multiple lab power supplies to just cream your pants on etc. etc. etc.
Now to cut to the chase, I hear it, but does it matter? You see, I don't really hear it with music. I can just determine a different noise level by listening to the idle white noise very close to a tweeter. From normal listening distance and around 90db/w/m efficient speakers it is hardly noticeable though, let alone with music playing.
The resistor quality of even the through hole models is superb, nice thin film resistors with tiny bodies. Then the Connex amps use thick metal film resistors, which cost about a tenth of their thin film cousins....
But I have never had a playing specimen of Connex. I would have loved to make some emission analysis and even listening impressions, but helas....
Truepath is a superb TA3020 design, absolutely. But I have a preference for the TC2001 based amps. That is very personal though, but I have always loved their sound most. Their noise performance is also extremely good. But there are so many poor TC2001 designs around too these days. Like the ones every body is on about that run terribly hot at 20% under their max voltage!! I always laugh when I see the proud presentations of ridiculously oversized heatsinks that look like the empire state building in the New York skyline just after it was built....
TK2350, TK2150 are also TC2001 based chipsets, and GREAT alternatives for TA3020. And then there are the RA-2500 and RA-0105 of Resolute Audio, and the TDA-2500 modulator design of 41hz, which is available on request. These are capable of the greatest things, but it takes a scientist to bring out the max. potential in them. They are semi-solid state chips, which means that almost any value can be tweaked to perfection in your specific design.
If only I had more time and not be busy with broadcasting, video and professional cameras all day, which is my real work....
Last edited:
I know Jan obviously only by e-mail, but he resembled me a very nice, honest and capable person.
If you wanted to put ants in my ears you were very successful..
at least you're lucky you work in the "wonderland".. having such equipments in my hand would turn me crazy..
ok, but IMHO this test tells you very little.. and however what makes (always in my opinion) a real difference is the "listening experience".. with Class AB, even some of the nicest, there's always that sensation of "electronic" sound.. thing that with Class T is reduced to almost zero, and this, for me, makes a lot of difference
being able to feel the sound with that "live" sensation that makes the speakers disappear..Cosmetics in most cases.. ppl that don't know typically associate the capabilities of the amplifiers with the size of the heatsinks.. a Class T should need a very little one since it dissipates only a very little fraction of the power it absorbs.
very uncommon on the market.. unfortunately..
Last edited:
That test tells you more than you think, if you have heard it side by side you will have to agree....What is a very consistant, distant and subtle "sssssssssssssssssssss" with the one and with the ear to the tweeter is an apparent and far less consistent "srshgsjsjsssjrrzzhshshsgshj" with the other at half a meter.... But still at least as silent as a good AB class amplifier.....
And about the heatsinks I was aiming for a specific amplifier of the brand Sure, of which I am not at all that sure.... A guy here on the forum, also from the Netherlands came by with his TK2050 Sure amplifier and it had such a heatsink on it, enough to cool a nuclear core. But it got terribly hot at 24V operating voltage!! That is not right!! Something must somehow mess up the switching in those amplifier modules pretty badly, but sadly I didn't have the time to determine what. There are a few obvious possibilities though, but I wouldn't dare to speculate until I have solved the issue. I know how tricky solving something like that can be and I don't want to raise any expectations. And the list of possibilities could extend to astronomical lengths as we know, time consuming business it is all that perfecting!
Eventually there will be an evolution and excellent T amps will be available that surpass all the existing amplifiers around now.... So far we can gladly be very happy with what we have now, certainly!
And about the heatsinks I was aiming for a specific amplifier of the brand Sure, of which I am not at all that sure.... A guy here on the forum, also from the Netherlands came by with his TK2050 Sure amplifier and it had such a heatsink on it, enough to cool a nuclear core. But it got terribly hot at 24V operating voltage!! That is not right!! Something must somehow mess up the switching in those amplifier modules pretty badly, but sadly I didn't have the time to determine what. There are a few obvious possibilities though, but I wouldn't dare to speculate until I have solved the issue. I know how tricky solving something like that can be and I don't want to raise any expectations. And the list of possibilities could extend to astronomical lengths as we know, time consuming business it is all that perfecting!
Eventually there will be an evolution and excellent T amps will be available that surpass all the existing amplifiers around now.... So far we can gladly be very happy with what we have now, certainly!
Last edited:
sorry v-bro, probably it's my mistake.. but to me this makes no difference.. if the second amplifier sounds better than the first.. as long as I don't hear the hiss from my seat.. I buy the second..
Probaly that would be beneficial at Fukushima
Again, maybe it's me.. but after Tripath was thrown out of the market, I see developments of switching technologies that are closer to class "D" more than to class "T".. If you have however notices about developments on the "T" class I would be very interested to know.
Wow,
I moved the input lead so it wasn't under the PCB at all (it was a little bit originally...) That didn't make any difference...
But I added this loop breaker circuit, here: Earthing (Grounding) Your Hi-Fi - Tricks and Techniques and it has made quite a bit of difference! The mid range is now a little tighter (how did that happen from a simple loop breaker?!)
Thanks peeps!
Any further mods would be much appreciated!
Here is how the amp looks inside now: (I could tidy it up a little, the loop breaker is fresh!!!)
I moved the input lead so it wasn't under the PCB at all (it was a little bit originally...) That didn't make any difference...
But I added this loop breaker circuit, here: Earthing (Grounding) Your Hi-Fi - Tricks and Techniques and it has made quite a bit of difference! The mid range is now a little tighter (how did that happen from a simple loop breaker?!)
Thanks peeps!
Any further mods would be much appreciated!
Here is how the amp looks inside now: (I could tidy it up a little, the loop breaker is fresh!!!)
Last edited:
- Status
- Not open for further replies.