Bgt said:
Utter bs, if you look at the wiring diagramm you'll see the output from the preamp, input at poweramp, going directly to the + and - input of the opamp, so where is the difference in signal????? The groundlift resistor has nothing to do with purityy of sound
I think there is a lot of misunderstanding here. What Bgt says may work very well for his RCA input setup. Bgt avoids the GND loop while using the differential - input as his signal GND. Basically seperating the signal GND and the power GND. For RCA based systems that share the same power supply with multiple amps, probably that is the best solution as the power GND level maybe contaminated by the other amps, apparently that caused issues in Bgt`s setup.
What Classd4sure is talking about in my opinion is a setup where XLR input is used and where he wants to make sure that the signal GND is tightly connected (no GND lift) to the amps GND. Actually I have GNDed my mid and tweeter amp currently this way and I have a GND loop present as both amps share the same supply. Do not notice any issues at the moment with that GND loop, however, it is of course only a short loop, local in the amp and everything is balanced input so should not be very sentive to it anyway. Still, in the future I want to go to a setup where each amp has its own supply so that there are no loops anymore, waiting for the SPS30.
Best regards
Gertjan
Ghemik: thanks for these comments. Actually you are right, the GND scheme Bgt has proposed uses the minus input (which i didn't see at first glance).
This is not optimal. I can never recommend using the minus input in a RCA setup, unless it's for phase reversal. NEVER as any kind of GND. You can short it to the GND terminal next to it, but that's all.
I would recommend Bgt to use his setup, but replace the 'minus input' with GND input of his modules. The he doesn't require the 10 Ohms resistors. And simply short the minus inputs to the input GND terminal.
This is not optimal. I can never recommend using the minus input in a RCA setup, unless it's for phase reversal. NEVER as any kind of GND. You can short it to the GND terminal next to it, but that's all.
I would recommend Bgt to use his setup, but replace the 'minus input' with GND input of his modules. The he doesn't require the 10 Ohms resistors. And simply short the minus inputs to the input GND terminal.
Lars Clausen said:
Hi Lars,
It seems Bert tried that in the past and then he got issues with channel seperation if I remember correctly. That`s why he ended up with the current setup.
The best and (always?) problem free way is one supply per amp with the supplies floating from the case. That`s why Hypex shows only dual mono setups on their site. Dual supplies does not automatically mean that the cost of the supply doubles, smaller transformers and caps can be used. So serious DIYers should probably consider this option seriously.
Gertjan
Lars Clausen said:
Even if the cost of two supplies is the same as that of one supply and assuming one supply cost as much as one amp module, then a dual mono setup would cost only 33% more than a stereo setup with one supply. If I had to make a stereo amp, I knew what I would do, I would avoid a lot of potential problems for that additional 33%
Would be really nice if someone would market an SMPS for stereo or multi channel use in which a PFC (power factor correction) module drives 2 or more independent SMPS, so that we could have regulated supplies (SMPS) independent for each amp, GNDs would be floating and seperated as each SMPS would have its own output transformer (hint to Hypex, Coldamp, NCD, ......).
Gertjan
Lars, if I do this my crosstalk figures go down more than 30Db and my input cable becomes very sensitive to EMI.Lars Clausen said:
The 10 ohm resistor is just for separating the input ground from the power ground and keep all the EMI out.
In your amp. I wired it as you suggested but there is no issue here because you already separated the power ground from the input ground with a 2.5 ohm resistor which has the same effect as my 10 ohm resistor.
Gertjan, see it this way, if you take the groundlift resistors out the speakers would take some signal from the other channels ground through the input grounds and that will give distortion/crosstalk. With the groundlift resistors the supply ground, 0V, for the speakers is always better at their own ground, 0V, terminals. They don't steal signal from the other ground, 0V. This is with a 1 transf. design. 2 separate supplies ofcourse don't have this problem. Thats what I noticed as being very different. Here you just connect the input grounds to chassis.
Bgt said:
Hi Bert,
I fully agree with your explanation. I had exactly the same thing in mind, and would basically try to avoid the same thing (currents running in those input GND cables). However, in my case, I use the modules only with a real balanced XLR input, so even if there are some currents flowing in those GND cables, my signal should not be influenced by that as the GND is not really used as a reference point for the signal.
Best regards
Gertjan
Gertjan, you got it. No problem in your setup
There is no way for me to do it differently and have all the goodies in crosstalk and distortion figures with a 1 transformer design and RCA input. Remember I also don't use a mains ground, is not obligatory in the Netherlands and I am happy for it. It saves a groundloop(NAD C272 usus also no mains ground). Furthermore my grounds are switched in my pass. preamp. So only the choosen source will pass its ground to the power amp.
There is no way for me to do it differently and have all the goodies in crosstalk and distortion figures with a 1 transformer design and RCA input. Remember I also don't use a mains ground, is not obligatory in the Netherlands and I am happy for it. It saves a groundloop(NAD C272 usus also no mains ground). Furthermore my grounds are switched in my pass. preamp. So only the choosen source will pass its ground to the power amp.
Bgt said:Gertjan, you got it. No problem in your setup
There is no way for me to do it differently and have all the goodies in crosstalk and distortion figures with a 1 transformer design and RCA input. Remember I also don't use a mains ground, is not obligatory in the Netherlands and I am happy for it. It saves a groundloop(NAD C272 usus also no mains ground). Furthermore my grounds are switched in my pass. preamp. So only the choosen source will pass its ground to the power amp.
Yes, I like XLR, should be less critical than RCA although there is many articles everywhere on the infamous pin 1 problem and some equipment does not have the pin 1 connected correctly (to the chassis). I had to modify my own DEQX PDC XLR outputs to connect them to chassis (done in the newer models). This allowed me to remove the mains earth connection between all the equipment that was needed before to have a GND reference. Now everything is floating from mains GND and power supply GND and just connected via pin 1, I think it sounds better, however, can not easily do any AB testing so maybe I`m subjective
Another remark, I assume your crosstalk problems would be less if you make a very solid (thick and short copper wire) between the signal GNDs of the two modules, in that way, most of the GND loop current would flow through that wire and GND loop induced voltages would be less. Would be interesting to see if it helps. I guess however that you do not want to experiment with this anymore as you have found a solution.
Best regards
Gertjan
Gertjan,
I always compare, A/B switching, any mod. to my ref. amp. which is the UCD400AD dualmono with dual supplies/T-Network caps. and grounding at the inputs, RCA, so I can hear/scope the differences, if present.
I tried to have 1 massive ground between the amps, but still...........crosstalk suffered. The groundlift rersitors are the best thing in this case(for me). I could lower them to 2.2 ohms like in the NCD's. When my NCD amp. is finished this week...I hope I will experiment with these resistors anyway.
I always compare, A/B switching, any mod. to my ref. amp. which is the UCD400AD dualmono with dual supplies/T-Network caps. and grounding at the inputs, RCA, so I can hear/scope the differences, if present.
I tried to have 1 massive ground between the amps, but still...........crosstalk suffered. The groundlift rersitors are the best thing in this case(for me). I could lower them to 2.2 ohms like in the NCD's. When my NCD amp. is finished this week...I hope
I will experiment with these resistors anyway.Bgt said:Gertjan,
I always compare, A/B switching, any mod. to my ref. amp. which is the UCD400AD dualmono with dual supplies/T-Network caps. and grounding at the inputs, RCA, so I can hear/scope the differences, if present.
I tried to have 1 massive ground between the amps, but still...........crosstalk suffered. The groundlift rersitors are the best thing in this case(for me). I could lower them to 2.2 ohms like in the NCD's. When my NCD amp. is finished this week...I hope
HI Bert,
Let us know how NCD works for you. MY problem is that I have active speakers, this makes it tough to compare amps. I plan to replace my old UcD180 mid/tweeter amps with UcD400AD (after modding them, some mods are finished but I have not yet listened to those amps) I`m using UcD400 standard for the woofer amps, although they are slightly modded. They sound good. My modded UcD180s also sound great, but expect of course improvement when I go to UcD400AD.
Best regards
Gertjan
while building the amp. it showed you need an extra psu line of about 18V/140ma per unit. Thats a bit of a stopper. UCD and the Coldamp. have it integrated on their pcb. You can order them though with an extra psu/cap. board which has the extra Vgate line(a bit lossy though). Ah well...you cannot have it all
If I'm not mistaken, the NCD PSU fixes that for you. A 4-wire unit (for a regular 4-wire transformer) has a voltage controller to extract the gate driver voltage from the main power rails, while the 6-wire unit has a simple bridge for a separate 15V trafo, or for a 6-wire main trafo with 5th and 6th wire being a separate 15V winding.
My ZAPs also had that function built-in, but had the option to go for a separate supply to cut down idle power.
Is that about right, Lars?
My ZAPs also had that function built-in, but had the option to go for a separate supply to cut down idle power.
Is that about right, Lars?
Hmm, that sounds interesting... Could they be simple drop-in replacements? I have a hard time seeing which op-amps are on the NCD originally, since I haven't gotten mine yet, but the LM4562 looks impressive.
Maybe I should use them for my active x-over, too. Definately beats the LC Audio x-over for price, I just don't know whether butterworth is a good alternative to Linkwitz-Riley. And then there's the issue of adjustable gain. But that's for another time, sorry for the off-topic.
Maybe I should use them for my active x-over, too. Definately beats the LC Audio x-over for price, I just don't know whether butterworth is a good alternative to Linkwitz-Riley. And then there's the issue of adjustable gain. But that's for another time, sorry for the off-topic.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.