after doing some tests it seems the Vgate supply has influence on the humm beheaviour.
1 channel psu connected: 5mv humm=100hz
2channels to psu and 1Vgate off: 10mv humm=
2 channels with psu and vgate: 15mv humm= audible
when psu and vgate on 2 channels on but 1 input cable connected the humm goes a bit(5mv) down.
Shortcutting the inputs increases the humm also(5mv).
I just leave it be for now. The unit works well for the rest of it so I'll just leave it powered on and playing.
1 channel psu connected: 5mv humm=100hz
2channels to psu and 1Vgate off: 10mv humm=
2 channels with psu and vgate: 15mv humm= audible
when psu and vgate on 2 channels on but 1 input cable connected the humm goes a bit(5mv) down.
Shortcutting the inputs increases the humm also(5mv).
I just leave it be for now. The unit works well for the rest of it so I'll just leave it powered on and playing.
I am redoing the NCD amp. into a dual mono design just to do justice to the amp. The psu's will be like the NCD's cookbook X2.
Grounding will be done exactly like the cookbook. Except no mains grounding.
I know Lars developed the amp. to be as dynamic/true as possible and this what I would like to achieve.
Next week.
Grounding will be done exactly like the cookbook. Except no mains grounding.
I know Lars developed the amp. to be as dynamic/true as possible and this what I would like to achieve.
Next week.
Ok did it and it is sounding very nice
. 1 question left, since the input RCA's are now grounded(nowhere else) I saw in the cookbook that there is a wire from the ground of the RCA's going to the ground of 1 of the PCB connectors(blue). Is this correct? There is still a bit of humm left, 100hz, but not so present(when completely silent) as before. With 1 channel connected to the input there is no humm at all. Would this ground wire solve that?
Unit is now 2x 250VA/2x40V with 4 bridge rectifiers, 1 for each winding, and 4x 10.000/63V caps. Vgate is integrated on PSU PCB's to get the exact copy of Lars his setup. Use a 10VA/2x15V toroid now for Vgate because I noticed the 6VA I had before got hot. This was a normal E shaped transf. so they always get a bit hotter.
BTW noise is no issue. Use 92Db/m speakers.
. 1 question left, since the input RCA's are now grounded(nowhere else) I saw in the cookbook that there is a wire from the ground of the RCA's going to the ground of 1 of the PCB connectors(blue). Is this correct? There is still a bit of humm left, 100hz, but not so present(when completely silent) as before. With 1 channel connected to the input there is no humm at all. Would this ground wire solve that?Unit is now 2x 250VA/2x40V with 4 bridge rectifiers, 1 for each winding, and 4x 10.000/63V caps. Vgate is integrated on PSU PCB's to get the exact copy of Lars his setup. Use a 10VA/2x15V toroid now for Vgate because I noticed the 6VA I had before got hot. This was a normal E shaped transf. so they always get a bit hotter.
BTW noise is no issue. Use 92Db/m speakers.
BWRX said:
When I disconnect the input grounds there is infinite resitance to the chassis and channels(between left and right channels), so........
I think this has to do with the NCD design. Lars used a power supply somewhere on the PCB that causes this. Either a kind of coupling through the Vgate connection. I saw the 7812 has the gound on the -V line.
I haven't had any humm in my amps for over 25 years????I mean you can say the humm is of no importance but if you put your ear on the speaker you'll hear it. To me it shows there is something fundamentely wrong in the design. The sound me be very nice indeed but humm is something that should not be there. If there is humm the design is compromised/not well designed.
ok, solved it. Isolated the RCAs and put a 1ohm/1Watt film resistor between chassis ground and RCA ground. So left and right RCA's have each each 1 ohm to ground. Humm is now so low that it has to be dead silent in the room and ear on the speaker box to hear something. Noise went also down.
Sound is already as good as the UCD's. Very hard to hear a difference. Have to give it some more time and do some more testing.
Sound is already as good as the UCD's. Very hard to hear a difference. Have to give it some more time and do some more testing.
noticed something very strange, the 45db gain from the NCD is as loud as the 35db gain from the UCD400AD(used 180gain resistor). The UCD is even a tinny bit louder. Units are fed from the same passive pre. When doing them parallel, I mean the inputs, the UCD is still a bit louder, just to see if the low input impedance from the NCD has anything to do with it. The UCD is the more qiueter 1 when there is no music. The NCD has a bit of noise left, when dead quiet in the room, the UCD's are dead silent, there just is nothing to hear. The UCD's are still very impressive, even after being on the market that long.
Bgt said:
Thanks for the updates Bert, very interesting.
I know its early days but do you think it would be worth a current UCD owner upgrading to a NCD? if theres not much difference I think I'd stick with the UCD but if its worth the extra money etc
Apart from you and Chris theres not been many more opinions yet
Cheers
Hi Bgt
Looks great!
If you want to improve the sound you should consider:
1..Using shorter wire from the PSU to the modules.
2..Using a better power cap instead of the RIFA's which are known to be lesser good with Class D amplfiers.
3..Not connecting the power electrolytis through the (steel) screws to the PCB. A silver solder joint sounds much better.
4..Is the Gate Drive also separate for the two ch's? It looks like there is only one transformer.
Best regards
Lars C
Looks great!
If you want to improve the sound you should consider:
1..Using shorter wire from the PSU to the modules.
2..Using a better power cap instead of the RIFA's which are known to be lesser good with Class D amplfiers.
3..Not connecting the power electrolytis through the (steel) screws to the PCB. A silver solder joint sounds much better.
4..Is the Gate Drive also separate for the two ch's? It looks like there is only one transformer.
Best regards
Lars C
Hi T,t. said:
Yes, pitty there are no others yet. After all the hassle Chris had with his NCD's I am happy I've mine OK now. But the UCD's are easier to assemble anyway, have better connectors(NCD's are fragile/small contact area) and need less other stuff like additional psu's. And for now Í would say, even though they sound equally fine, its not worth the money in my view.
And what I absolutely don't like is you HAVE to use them in a dual mono setup to get them right because of the humm problem otherwise. A 1 supply rail psu is no option, crosstalk suffers.
Bgt said:
This is of course not a general guideline, but only effectual in Bgt's case where he used a different wiring scheme than proposed in the cookbook. I have built many NCD amplifiers using the exact scheme proposed in the cookbook, with one supply rail. There is absolutely no hum.
With regards to crosstalk, i can more agree, a dual mono setup will always have better crosstalk performance than a stereo. And that is a general guideline
Bgt said:
Thank you for your honesty Bert, its very much appreciated
Its posts like yours which we find very helpful!
Come on the rest of you chaps, get your fingers pulled out and post us some results
Normally I would just buy a new diy amp but I've already got several different types now so have to draw a line sometime, its never took me over a week to build one and have it up and running
Lars Clausen said:Hi Bgt
Looks great!
If you want to improve the sound you should consider:
1..Using shorter wire from the PSU to the modules.
2..Using a better power cap instead of the RIFA's which are known to be lesser good with Class D amplfiers.
3..Not connecting the power electrolytis through the (steel) screws to the PCB. A silver solder joint sounds much better.
4..Is the Gate Drive also separate for the two ch's? It looks like there is only one transformer.Best regards.Lars C
1. 14cm. For me short enough.
2. test modus, normally BC's(my favourate caps, only the 10.000uF ones).
3. they are screw caps. Most soldered caps. have a high ESR@20khz. No good. Tried them all.
4. yes, 2 separate windings on the toroid.
Hi Bert
Yes BCC's should be much better (series 101 or the older 154 are very good).
I agreee completely about the screw terminal types, however you have to terminate them properly, and not let the screw sit in series with the current. The screw might have several times higher ESR than the capacitor. And steel in the signal path ....
never good!
Yes BCC's should be much better (series 101 or the older 154 are very good).
I agreee completely about the screw terminal types, however you have to terminate them properly, and not let the screw sit in series with the current. The screw might have several times higher ESR than the capacitor. And steel in the signal path ....
never good!Ok Lars, ofcourse you have no problem/humm with the 1 psu supply rail assembly, but it suffers on sound.Lars Clausen said:This is of course not a general guideline, but only effectual in Bgt's case where he used a different wiring scheme than proposed in the cookbook. I have built many NCD amplifiers using the exact scheme proposed in the cookbook, with one supply rail. There is absolutely no hum.With regards to crosstalk, i can more agree, a dual mono setup will always have better crosstalk performance than a stereo. And that is a general guideline
I want to be able to use it in a smaller housing with 1 transformer and have the goodies of a dual/mono design like soundstage/positioning. With the UCD's and Coldamp this is possible with my 1 transf./2 psu supply rails and groundlift resistors scheme. The sonic difference with a dual/mono amp. is so minimal that for me it is almost non existing and I have a more economic build and smaller housing. Yours just cannot be made like this satisfactory because of the humm.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.