| Search | Login/Register
   Home » Melquiades Amplifier » Planning my DSET (161 posts, 9 pages)
  Print Thread | 1st Post |  
Page 7 of 7 (161 items) Select Pages:  « First ... « 3 4 5 6 7
01-07-2019 Post does not mapped to Knowledge Tree
N-set
Gdansk, Poland
Posts 458
Joined on 01-07-2006

Post #: 151
Post ID: 25231
Reply to: 25230
EI vs toroids
Congratulations on the build!

 anthony wrote:
  Yours uses all toroids, or mostly so, which are more efficient than the EI that I use.

Why have you chosen EI vs toroids? The latter emit less.

Cheers, Jarek



Cheers,
Jarek
01-08-2019 Post does not mapped to Knowledge Tree
anthony
Posts 212
Joined on 08-18-2014

Post #: 152
Post ID: 25232
Reply to: 25231
Its about noise
 N-set wrote:
Congratulations on the build!

 anthony wrote:
  Yours uses all toroids, or mostly so, which are more efficient than the EI that I use.

Why have you chosen EI vs toroids? The latter emit less.

Cheers, Jarek

Hi Jarek,

Yes, toroids can leak less magnetic flux than EI if they are wound such that the entire core is covered, but is flux leakage a problem if you are able to keep the transformers away from sensitive circuits?  Toroids come into their own when you are forced to mount the transformer near the circuits, but otherwise EI have them covered on most fronts.

Toiroids
  • are smaller for the same power rating
  • are more efficient transformers of power than EI
  • if carefully made can have low flux leakage
  • are high bandwidth devices so more noise on the power line is transferred into the circuit
  • have no gap and will hum when there is DC on the mains


EI
  • are larger for the same power rating
  • have a gap and do not hum when you have DC on your mains (unless there is so much that the core becomes saturated)
  • are lower bandwidth devices so act as a low pass filter for mains noise (we really only need to pass 50/60Hz - the rest is unwanted)
  • leak some flux so they should be kept further separated from the sensitive circuits


Personally, I would use EI in just about every situation for audio...unless I had no option but to tuck the transformer in tight to a sensitive circuit.


Cheers,

Anthony
01-08-2019 Post does not mapped to Knowledge Tree
anthony
Posts 212
Joined on 08-18-2014

Post #: 153
Post ID: 25233
Reply to: 25232
The pin 8 question
Hi Romy,

I've been reading the 6e6p-dr datasheet and it seems as though pin 8 (a shield) is recommended to be tied to a lower negative voltage than the tube cathode.  Have you bothered or experimented with this?  This is no problem in the DHT Channel because of the cathode resistor, but in Channel D the cathode pin is connected directly to ground and I don't know how to get a potential lower than ground.
Regards.
Anthony
01-08-2019 Post does not mapped to Knowledge Tree
N-set
Gdansk, Poland
Posts 458
Joined on 01-07-2006

Post #: 154
Post ID: 25234
Reply to: 25232
Cont'ed
Anthony, thank you for the explanations. For some reasons you mention, I use C-cores. Loose coupling (pri and sec on the different legs), controlled gap and low emission. BTW toroids do have some "distributed gap" but it's small and uncontrollable.

Cheers, Jarek



Cheers,
Jarek
01-09-2019 Post does not mapped to Knowledge Tree
Romy the Cat


Boston, MA
Posts 9,579
Joined on 05-28-2004

Post #: 155
Post ID: 25235
Reply to: 25233
A good question.
Anthony, I grounded in my amp it in order the shield was not hanging there free. You certainly must NOT leave it disconnected and you do need to tie to something. Generally biasing of these electrostatic fields shields is very murky territory. There is a community of people out there who insist that it makes difference and I have seen even some gizmos that introduce bias to external shields. The shields in 6E6P is much closer to anything, so a positive or negative bias might be effective. I did not experiment with it and I have no way to predict how it might impact sound. Do not take it too far as I can see you running an active feedback loop over the 6E6P shield… :-) 
 
Let me know if you do experiment with it and what you find. I would be looking for noise and sound impact. Do not forget to burn a used tube at least 1Hr before you make any judgments. It might be hard for you as you do not have 6E6P in a full bandwidth channel and to make any concussion within a narrow bandwidth channel you need the whole system to be setup, calibrated and it sound to matured in your own perception.


"I wish I could score everything for horns." - Richard Wagner. "Our writing equipment takes part in the forming of our thoughts." - Friedrich Nietzsche
01-22-2019 Post does not mapped to Knowledge Tree
anthony
Posts 212
Joined on 08-18-2014

Post #: 156
Post ID: 25239
Reply to: 25229
Barbirolli Mahler
 Romy the Cat wrote:

If you like very slow interpretation, like I do, then you might start to listen some Barbirolli Mahler… 
 


I've found Barbirolli Mahler 6 & 9 by New Philharmonia Orchestra and Berlin Philharmonic respectively.  I'll give the records a spin when they turn up.  Also grabbed Mahler 7, Bernstein NY Philharmonic.
01-22-2019 Post does not mapped to Knowledge Tree
Romy the Cat


Boston, MA
Posts 9,579
Joined on 05-28-2004

Post #: 157
Post ID: 25241
Reply to: 25239
Yep, Barbirolli would do it...
When I was making a sarcastic joke about you begin to enjoy slow interpretations (something that any VERY good playback should do to a system owner) and referred you to the Barbirolli’s Mahler of cause I mean the exquisite  performance of 1967 of Sir Barbirolli with New Philharmonia. The “Allegro” movement is so much out of this would that I hardly believe anybody would listen any other version of Mahler 6 after listening to what Barbirolli did with it. It was told that Rostropovich pull out all his own recordings of Elgar cello concerto and never recorded it again after he heard the Jacqueline du Pre’s  take on the concerto in August 1965, ironically with the same Barbirolli… I have no idea why people keep recording the M6 after Barbirolli 1967. Whatever was possible to be expressed in that music was done by maestro Barbirolli.  




"I wish I could score everything for horns." - Richard Wagner. "Our writing equipment takes part in the forming of our thoughts." - Friedrich Nietzsche
01-31-2019 Post does not mapped to Knowledge Tree
anthony
Posts 212
Joined on 08-18-2014

Post #: 158
Post ID: 25254
Reply to: 25241
First music testing
So, the first DSET is still downstairs on the electronics bench and I am back to work for a while, but not too hard and not for too long (I hope).  I've added all the filters and checked and tested and changed some more things and decided that it was time to see if I could get any music to play.  Several crappy drivers and speakers that I had scrounged for just this purpose were wired up and I pressed the go button to stream some Tidal.


First Music Testing.jpg


It works!  I have sound.  Kind of a weird sound with the mix and match driver selection but at least it is comprehensible.

This is the front of the amplifier where it wraps around the chassis pillar for the horn stack...

Amp Front.jpg


Plenty of thought has gone into the speaker cables largely because there are plenty of them and I would prefer for them to be hidden completely from view.  Five of the six channels exit the amplifier in the well that fits around the horn stack pillar where they are nicely hidden, and the bass channel exits from the side closest to the Bass Cannons.  I've thought about different types of speaker wire from exotic to simple and in the end I have gone with simple Duelund 16GA tinned copper wire.  I've used this before with fullrange speakers and it is fine.  The big benefit for me is that it is black and the individual wires are tinned:  black to hide it and the tinning because these speaker wires will not be terminated in any way, no spades, bananas or whatever, just crushed in the binding posts of the driver one end and directly soldered to the output transformer on the other end.  Inductance and capacitance should not be an issue for such band limited amplifier channels, and even the Injection Channel that runs fullrange will run 14dB or so down.  Terminations seem to be the major issue in speaker cables, and these Duelund wires when used this way will eliminate a joint/termination each side of each binding post while proving resilient to oxidation in a tight crush joint at the transducer.

As discussed earlier I have gone for ammeters instead of voltmeters.  I can hit the mA target on all but one channel (F), but with exception of E (DHT Channel) they sure do wander about a bit, particularly the 6C33C channels.  Romy, is this normal?  I set them at about 240mA/150mA/150mA and they will go down over time and then they will come back up and past their original settings then they may wander down again.  Makes setting a stable current draw like herding cats. 
Targets 1.jpg



Here is Channel F.  It won't come over about 20-25mA.  Plus the DC Offset won't trim out less than about 0.27V (yes volts).  I have a feeling that the two problems are linked and will check to see where I have wired up the input filters for each channel.

Targets 2.jpg



Anyway, music has been playing downstairs for a couple of hours now and the only temperatures that I am concerned about are the 3 final anode resistors (15k 12W) for the first stages of A/B/C.  The hottest of those according to my thermal camera is about 140C (280F) so it might pay in the long run to parallel a pair of 30k 12W resistors even though the datasheet specifies 350C maximum and there should be decent airflow from the chassis fans at that part of the case.

The next DSET is also taking shape...

Next One.jpg



Getting closer!


01-31-2019 Post does not mapped to Knowledge Tree
Romy the Cat


Boston, MA
Posts 9,579
Joined on 05-28-2004

Post #: 159
Post ID: 25256
Reply to: 25254
So far so good.
 anthony wrote:
Plenty of thought has gone into the speaker cables largely because there are plenty of them and I would prefer for them to be hidden completely from view.  Five of the six channels exit the amplifier in the well that fits around the horn stack pillar where they are nicely hidden, and the bass channel exits from the side closest to the Bass Cannons.  I've thought about different types of speaker wire from exotic to simple and in the end I have gone with simple Duelund 16GA tinned copper wire.  I've used this before with fullrange speakers and it is fine. 

Be careful with the speaker cables. The knowledge about the cables that you developed running a full range systems is not applicable running DSET. You do not need cable that run overall well. With DSET and very limited bandwidth you might go away with very cheap wires. You will find eventually the some of the speaker cables might be better but it would need a well matured system where sound will be observed in context of the given channel loading

 anthony wrote:
As discussed earlier I have gone for ammeters instead of voltmeters.  I can hit the mA target on all but one channel (F), but with exception of E (DHT Channel) they sure do wander about a bit, particularly the 6C33C channels.  Romy, is this normal?  I set them at about 240mA/150mA/150mA and they will go down over time and then they will come back up and past their original settings then they may wander down again.  Makes setting a stable current draw like herding cats. 

Sorry, I do not understand you. The 240mA/150mA/150mA is fine. As the tubes new they will run off a lot of current but after 20-30 hour of burn they will stabilize. Read my 6C33C Survival Guide.

 anthony wrote:
Here is Channel F.  It won't come over about 20-25mA.  Plus the DC Offset won't trim out less than about 0.27V (yes volts).  I have a feeling that the two problems are linked and will check to see where I have wired up the input filters for each channel.

You will find the bug, it is there. If you need to get extra 10mA then you might change the value of bias resistors.



"I wish I could score everything for horns." - Richard Wagner. "Our writing equipment takes part in the forming of our thoughts." - Friedrich Nietzsche
02-02-2019 Post does not mapped to Knowledge Tree
anthony
Posts 212
Joined on 08-18-2014

Post #: 160
Post ID: 25274
Reply to: 25256
All fixed?
So, my two problems have been fixed, I think.

First was the issue with the 6e6p  for the HF Channel not being able to draw enough current.  I checked the bias and it was fine and the potentiometer allowed plenty of adjustment so I changed nothing in the circuit.  A new tube was swapped in and bingo! the current is back and all is good.  All that I can think is that all the time that I ran that tube without pin 8 grounded has worn it out and it may be near failure.  When I first left the amp on for several hours a couple of weeks ago now the current from that rube was fine and it slowly dropped over time.  However, the three 6c33c channels were varying as well so I did not think too much of it.  The new tube pulls the required current with the pot somewhere near the middle of its range and I will leave the amp going all day today with the hope that the current will not fall as it did with the previous tube.

My second problem was not being able to set the DC Offset to zero.  It sat at about a minimum of 230mV plus or minus 15mV regardless of the gas tubes I randomly swapped in there.  I tried half a dozen all up and decided that the issue was elsewhere.  What I did in the end was replace a 1k resistor that leads into the 100k potentiometer for trimming DC on the positive side with a 100k resistor, so 100 times larger.  The offset now trims to zero with the pot close to halfway through its travel.  Romy, I don't know if that is an error on the schematic or if I have some other issue, but the offset seems stable.
02-03-2019 Post does not mapped to Knowledge Tree
Romy the Cat


Boston, MA
Posts 9,579
Joined on 05-28-2004

Post #: 161
Post ID: 25278
Reply to: 25274
It is all good.
 anthony wrote:
First was the issue with the 6e6p  for the HF Channel not being able to draw enough current.  I checked the bias and it was fine and the potentiometer allowed plenty of adjustment so I changed nothing in the circuit.  A new tube was swapped in and bingo! the current is back and all is good. 

Hm, I did not understand that you did not start with it. It is always one tube has less current them others. In my estimation 1 our 10 tubes of this class are not useable. The irony that some (!) of them still sound fine despite lower current. The 6e5p has higher ration of not usable tubes, I would say 1 our 7 tubes. They are not defective tubes, it is just the bias setting not wide enough to cover them. You might put a bias attenuator in play, I did experiment with it and in my final design I did not do it intestinally. I discover that in most of the cases the low current tubes with lower current do sound worse. I think it has to do with cathode potency and then we drop bias too much to help the tubes. To evaluate each tube and to see if it sounds good at lower current is too much pain and it much easier to get a new tube. I paid for my 6e6p under $3 per tube, so why bother…
 anthony wrote:
All that I can think is that all the time that I ran that tube without pin 8 grounded has worn it out and it may be near failure.  When I first left the amp on for several hours a couple of weeks ago now the current from that rube was fine and it slowly dropped over time.  However, the three 6c33c channels were varying as well so I did not think too much of it.  The new tube pulls the required current with the pot somewhere near the middle of its range and I will leave the amp going all day today with the hope that the current will not fall as it did with the previous tube.

Be prepared to trash all tubes that you use during the design stage. I did some very barbarian this to them during my expedients and during my act of building stupidity. So, as you have all your operation points locked  and the amp function properly then juts put the set of the “design phase” tubes in a separate box and use all new tubes.

 anthony wrote:
My second problem was not being able to set the DC Offset to zero.  It sat at about a minimum of 230mV plus or minus 15mV regardless of the gas tubes I randomly swapped in there.  I tried half a dozen all up and decided that the issue was elsewhere.  What I did in the end was replace a 1k resistor that leads into the 100k potentiometer for trimming DC on the positive side with a 100k resistor, so 100 times larger.  The offset now trims to zero with the pot close to halfway through its travel.  Romy, I don't know if that is an error on the schematic or if I have some other issue, but the offset seems stable. 

 Nope, there is no error on the schematic, you are just the recipient of a very good design. :-) I’m not kidding, let me to explain. You can put in there 10M resistor and get rid of 420K resistor and will be able to adjust anything. The resolution of the agistment will be very low however. The 420K resistor acts as buffer that slightly decuples the rheostat from the signal. The 100K variable resistor has this value because a good feeling between the resistor turns of the voltage if balances. And the 1k resistor is a wild card joker that set all thing in place. The pression of the positive supplely is very had to estimate accurately. It balances 6 channels and the voltage of your positive supply transformer is not identical to mine. It should not be. The 1K resistor before the potentiometer is not necessary from voltage perspective but is necessary as independent adjustment tool. Your experience is a perfect illustration why the 1K resistor is there. As you can see I use the same 1K wild-card resistors earthenware.


"I wish I could score everything for horns." - Richard Wagner. "Our writing equipment takes part in the forming of our thoughts." - Friedrich Nietzsche
Page 7 of 7 (161 items) Select Pages:  « First ... « 3 4 5 6 7
Home Page  |  Last 24Hours  | Search  |  SiteMap  | Questions or Problems | Copyright Note
The content of all messages within the Forums Copyright © by authors of the posts