Jump to content

and now for something completely different part 3


kevin gilmore

Recommended Posts

On 9/16/2022 at 8:48 PM, dip16amp said:

Still have to build the other three channels.

What for are the other "three" channels, you do a virtual surround setup?

Very interesting build indeed. Looks like building ship models in glass bottles. 

44 minutes ago, Thai Tao said:

I'm planning to order some board, do you guy think gold immersion and 2oz copper thickness make different in sound compare to HASL and 1oz?

No, but you can do it for the sole sake of overkill.

Link to comment
Share on other sites

4 hours ago, audiostar said:

What for are the other "three" channels, you do a virtual surround setup?

Very interesting build indeed. Looks like building ship models in glass bottles. 

No, but you can do it for the sole sake of overkill.

Only one of four channels is up and running so three more channels needed to complete the full balanced system.

 

Link to comment
Share on other sites

I completed the first balanced mono CFA2 amp, so I ran RMAA on it.
The DAC2 balanced output right channel goes in on a XLR 2 foot Mogami 2534 cable and outputs on a 1 foot XLR cable to a Benchmark ADC1 USB.
The left channel goes in a XLR 2 foot cable and a 1 foot XLR cable to the ADC1.
This compares the CFA2 balanced mono to the cables.
I had to set the ADC1 left channel to 10 dB gain to equal the CFA2 gain which shows 0.3dB lower than the left channel.
This results in the CFA2 balanced mono gain at 9.7 dB.

 

cfa2bm.jpg

benchcfa2.jpg

rmaa1.jpg

rmaa2.jpg

rmaa3.jpg

rmaa4.jpg

rmaa5.jpg

rmaa6.jpg

rmaa7.jpg

rmaa8.jpg

  • Like 3
  • Thanks 1
Link to comment
Share on other sites

hi @kevin gilmore, may i ask why there are 3 emitter follower at the output stage? I think it's just too waste silicon when only the outer most pairs do the heavy lifting. Instead can we just change to 2 emitter follower with the first one use small transistor like bc546 or mpsa06 so we can take advantage of high hfe to have high input impedance and the last one use mje15030 to do heavy work?

Edited by Thai Tao
Link to comment
Share on other sites

18 hours ago, kevin gilmore said:

no, you really need that much current gain when you are doing zero feedback. especially if you like those really low distortion numbers.  

but sure, go ahead and modify.

Hi Kelvin, 

If i understand the schematic correctly, the CFA2 will have quite high open loop gain right? The CFA3 although output stage doesn't have gain like other buffer, it still wrap around the input stage which is also high loop gain so the zero feedback here only apply for the current buffer, right?

My point is if really need that high current gain, then change the first mje15030 to pzta06 we will have current gain x5 in worst case ( 20 vs 100 ), and if we continue to the second mje15030, we will continue to get another x5 current gain and the pzta06 will still be able to handle 20ma bias. Or is the reason is we need thermal couple?

Thanks!

Link to comment
Share on other sites

no the zero feedback version does not have high loop gain. no feedback anywhere. not at the input stage and not at the output stage. first stage is a transimpedance amplifier with a output resistor that fixes the gain. output buffer is just that, voltage gain of 1 with no feedback.

the feedback version bumps the gain of the input stage and adds about 5db of feedback. or whatever you set it to.

Link to comment
Share on other sites

On 9/26/2022 at 11:50 PM, kevin gilmore said:

no the zero feedback version does not have high loop gain. no feedback anywhere. not at the input stage and not at the output stage. first stage is a transimpedance amplifier with a output resistor that fixes the gain. output buffer is just that, voltage gain of 1 with no feedback.

the feedback version bumps the gain of the input stage and adds about 5db of feedback. or whatever you set it to.

Thanks, i will study the schematic more

Link to comment
Share on other sites

  • 2 weeks later...

OK... I want to finish my CFA3. It has been too much delayed

 

I have some questions..

Recently, I killed my linear power supply by mistake.. so now I think IEC socket and fuse may be very important!(?).

How can I select the suitable IEC socket with fuse?  Is there any recommendation for it?

My CFA3 will have standard 24V GRLVs

If possible, I also want to know popular front panel turn on/off switch or push  button.

 

Second, If I want to add gain select stage (like low gain/high gain) to CFA3, What should I do?

I have googled these, but it was difficult to find.. sorry. 

 

 

Thank you so much guys.

수정됨_20221009_110435.jpg

Link to comment
Share on other sites

4 hours ago, Breaklunchdinner said:

Recently, I killed my linear power supply by mistake.. so now I think IEC socket and fuse may be very important!(?).

You can use FN281-4-06 from Schaffner. It is available in a horizontal or vertical orientation and includes a fuse holder and power switch. Take a look at the pdf spec sheet as the picture isn't correct, showing a different style. 

  • Like 1
Link to comment
Share on other sites

Hi, I want to ask for help or ideas what to try to do, I'm currently running out of ideas. :(

I've build this CFA3 dual mono setup with 2xGRLV 30V power. It is actual state after two weeks of tinkering and reworking... It is somehow orthodox implementation of balanced noise self-canceling theory now.  :)

20221006_210850.thumb.jpg.d82842da3968f1f83317224179c88062.jpg

I've decided to rework grounding scheme two weeks ago. Here is block connection diagram:

20221003_123352.thumb.jpg.fa20ae95831e6a5ceb331a20ec29d037.jpg

In short: I've one grounding point (signal reference) from which is GND connected to all amp PCBs. It is only GND connection for those PCBs, to not have ground loops.

Everything looked nice after rebuild, I've connected my test HP (Koss sporta pro, Sensitivity 103 dB/mW) to SE output and listened to amp with floating input to evaluate noise. Everything was OK and I was able to hear very small amount of distant hum probably picked from large Xformers.

I've measured bias and voltage on output 4pin XLR connector, all OK.

I've connected 3m long balanced cable which came with Susvara and wanted to set 200mV output with 1kHz sine wave test signal from DAC. I've connected DMM to cable 3.5mm jack and wondered what happens! I've measured this:

20220926_103629.thumb.jpg.b19bd29818baf9337b5f9b4c95e44f93.jpg

Wow! Disconnected that cable and measured again directly at amps XLR output connector and nothing there from this voltage! Connected XLR cable again and is there again. :(

Ok, I was thinking it is not real that voltage and connected SportaPro HP again to SE output and no noise can be heard, so I let the Hp connected and connected that XLR cable to connector and immediately I can hear that ugly noise in SportaPro HP, so it is real unf. :(

I've measured bias again and it raised from 200mV to over 300mV when XLR cable is connected to my amp, when I disconnect it, bias falls back to normal.

This noise disappear when I set volume potentiometer to its maximum volume.

It looks like there is some kind of ground loop but can't find where, all continuity/resistance measurements between all PCB GND and chassis GND seem to be OK, no loop nowhere.

This noise disappear when only one CFA3 module is powered on. It doesn't matter which one it is, but you need to disconnect GND from that powered off module, not only + and - from GRLV. 

I've tried to swap GRLVs, than CFA3 modules, powered booth CFA3 from one GRLV, then from other one GRLV, left out protector board... still nothing... :((

It seems like those CFA3 modules are fighting each other. :(

This leads me to one experiment, I've disconnected pot. signal GND (reference) from central star GND and connected it GND on CFA3 modules on signal input leads. The noise decreased/halved in intensity, but still not usable.

Next experiment: I've separated GND for each CFA3 module = I've cut GND wire connecting all GND pins on pot. so that I've only connected together two GND pins for LEFT channel to LEFT CFA3 GND and two GND pins on pot for RIGHT channel to GND on RIGHT CFA3 PCB. Next improvement! Now is the noise there only at volume pot lowest volume, exactly when impedance between pot wiper and GND is 300 ohm and less. When I turn pot to higher volume there is no noise.

I have stuck here. :( no more ideas. :(

I'm sorry for this long post and bothering, but maybe some one here can see/know what is wrong, that I've overlooked.

Thank you,

Milan

Link to comment
Share on other sites

Seems like it should work with the star ground. In your drawing, you only show a single secondary from your transformers. These should be dual secondaries (or center-tapped, depending on the GRLV version). You are getting the proper +/- dc voltages from your GRLVs? You would be seeing a ton of dc offset if this were the case, just asking.

Link to comment
Share on other sites

3 hours ago, migo said:

 

I've connected 3m long balanced cable which came with Susvara and wanted to set 200mV output with 1kHz sine wave test signal from DAC. I've connected DMM to cable 3.5mm jack and wondered what happens!

Could you please check out that XLR cable you connect your DAC and amp, how is it wired internally? Pin 2-2 and 3-3 are ok, question is about pin 1: does it connect to the shells on both sides as well or is it pin 1-1 only and does it connect pin 1-1 on both sides of the cable?

Have you tried with another XLR cable?

If the cable is right, might be your DAC implements grounding in a wrong (not AES compliant) way; in this case I'll disconnect pin 1 at the receiver's (CFA) side and only leave it connected at the senders (DAC) side. Make sure only hot and cold from pin 2 and 3 go into the amp and make sure there is no hidden connection between pin 1, cable screen, connector shell or chassis at the receiver. This should break any ground loops caused by the sender. 

@ParsYes, Toroidy have two secondaries. 

 

3 hours ago, migo said:

I've measured bias again and it raised from 200mV to over 300mV when XLR cable is connected to my amp, when I disconnect it, bias falls back to normal.

The cable or your DAC is most likely the problem. 

3 hours ago, migo said:

This noise disappear when I set volume potentiometer to its maximum volume.

Strange as well, have you tried with another pot?

Edited by audiostar
Link to comment
Share on other sites

2 hours ago, migo said:

Next experiment: I've separated GND for each CFA3 module = I've cut GND wire connecting all GND pins on pot. so that I've only connected together two GND pins for LEFT channel to LEFT CFA3 GND and two GND pins on pot for RIGHT channel to GND on RIGHT CFA3 PCB. Next improvement! Now is the noise there only at volume pot lowest volume, exactly when impedance between pot wiper and GND is 300 ohm and less. When I turn pot to higher volume there is no noise.

As you basically have a dual mono setup, try completely separating ground for both channels, and connect then each ground centrally to the chassis (via the 10R resistor as you did or a real circuit breaker). So you will have a star ground per channel which is connected to another star (chassis at IEC connector with both channels master grounds).
What you did with separating ground on the pot was the right beginning in this direction.

Edited by audiostar
  • Like 1
Link to comment
Share on other sites

I would suggest go full dual mono as you've started above, with separate pot ground for left and right channels.

Also, I would suggest not to connect the grounds as in the schematic above. With everything tied to one point, all current between amp and psu boards has to flow through that point, to which the pot grounds also happen to be connected. Not optimal, and not really what is meant with a star ground designed to avoid hum.

Instead, try connecting the left amp board ground directly to the left GRLV ground and the right amp board ground directly to right GRLV ground, and then reference only the GRLV grounds and pot grounds (not amp grounds) to chassi ground, either directly or through 10R.

  • Like 2
Link to comment
Share on other sites

Since there is bias on the inputs of the CFA3 but it's also a balanced amp, an additional thing to try is disconnecting pot ground altogether. It should in practice be happy with a differential input, as long as dac and amp chassis have a good grounds connection.

Link to comment
Share on other sites

25 minutes ago, MLA said:

Instead, try connecting the left amp board ground directly to the left GRLV ground and the right amp board ground directly to right GRLV ground, and then reference only the GRLV grounds and pot grounds (not amp grounds) to chassi ground, either directly or through 10R.

Yes, very good advice! This is exactly what I have in my two chassis CFA3 with external dual PSU. Two umbilicals carry the separate GRLV grounds to the amp boards directly. Everything connects together in the PSU case including the chassis and IEC earth.

Link to comment
Share on other sites

12 hours ago, Pars said:

Seems like it should work with the star ground. In your drawing, you only show a single secondary from your transformers. These should be dual secondaries (or center-tapped, depending on the GRLV version). You are getting the proper +/- dc voltages from your GRLVs? You would be seeing a ton of dc offset if this were the case, just asking.

Thank you Pars, xformers are dual secondaries, block drawing is simplified.

Link to comment
Share on other sites

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.

Guest
Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

×
×
  • Create New...

Important Information

By using this site, you agree to our Terms of Use.