Rerurn to Romy the Cat's Site

Melquiades Amplifier
Topic: Formation of capacitors or the "cable memeory"

Page 1 of 1 (17 items)


Posted by Romy the Cat on 03-20-2006

I have seen it cars and have seen in electronics but I never seen it as dramatic as in the Super Melquiades. I am talking about the ability of amplifiers to be unbroke-in if they are not being used.

I have seen some amplifiers the after a couple month of not working requires good 30-40 hours of burning-in to sound good. But the Super Melquiades, for whatever reasons is very picky regarding this.  Sine the amp was born in July it was pretty much running dally for a couple hours and sometime all day long on weekends. Lately I use it 2-3 times per week, not more then r-6 hour a week. The amp sounds very purely. Interesting that in order to put it back to shape it requires 3-4 hours of containing burning the amp after 3-4 none-operational days.

I wonder why would it be so and why a single chasses Melquiades monoblocks, the were pretty much the same amplifiers did not have this behavior

Rgs,
Romy the Cat


Posted by Genn on 03-20-2006

Think about your amp as a team of players - where each tube as an individual.  The higher q-ty of performers - the higher the time to get them played a capella


Posted by Romy the Cat on 03-22-2006

I do not think Genn is close to an answer, he spent too much time to read the Lehnitsky’s brain-massage that made him to feel this way.

I am sure that there are some kinds of very pragmatic, predicable and manageable explanations for the effect that I have mentioned above. Perhaps the formation of capacitors has something to do with it or perhaps the “problem” with my cables.

The entire system wired with Dominus that is very freaky. It has not-fixed capacitance and if I believe correctly it fluctuates between 15pf and 75pf per meter, while a signal flows. Perhaps because the Dominus’ fluids take time to self-adjust the cable's electric characteristics the worsening of sound does take pleace when the system is raw or out of the "cable memeory"? It is well known how Dominus reacts to shaking and I have seen how it reacts to the current loading....
 
Perhaps it is not the amps need the "refreshing memory" but the cables? Who knows…

The Cat


Posted by Genn on 03-22-2006
Roman, 
the reformation of capacitors is an unswer in many cases. 
Those, who use BG - have to spend approximatly the same time to bring the system to the proper sound, but your mind is not SO corrupted.  You use relatively big tanks of electrolitic liquid at low discharge mode. May be you could find people, who do understand the electrolytic capacitor processes, and who could give you some in depth explanation, and you will share it.

Also - Likhnitsky has nothing to do with the statement. He is too uneducated in this area. My current reading is a book of Cyril Scott, Music: its secret influence throughout the ages. That is may be the most informative introduction to the issues, happened to appear on my way. 

Posted by Romy the Cat on 04-08-2006

As I previously told to manage six channels built around Russian tubes, and patricianly as different as 6C33C, is a pain in ass. Today, carefully examining my log file that I keep about the performance of each tube in the amp, in addition to everything else, I detected another interesting pattern: the tubes in the Super Melquiades after almost a year of working …  believe you or not but they self-tuned to each other.

The six channels were initially selected more or less accurately (by gain) and I was under impression that all channels would age gracefully and identicaly. However, now all tubes of one amp self-tuned itself and have absolutely identical gain (-+ 1/4db but there were -+3/5dB). Another amp’s tubes did the same but at +.5dB upper. It all happens despite that one of the tubes in LF channel in my left amp should output +1dB (it was how it was initially intentionally set due to the LF asymmetry of my room).

Very, interesting discovery! The tubes in the Super Melquiades kind of co-set each other (probably because they were driven with the same voltages) and they somehow self-tuned each other. In a way it is similar how female’s menstruation cycles get synchronies by some mysteries reasons. It is well know that women who stay in lesbian relationship after a few moths get theirs periods completely synchronized. Why it happens is mystery, perhaps some kind of biological clock offset the periods cycles with some kind of jitter… however, what kind “clock" works in the Super Melquiades that sel-tunes the tubes?

Go figure!

The Cat


Posted by Romy the Cat on 04-09-2006

I probably overwhelmed the readers of my site with the different aspect of the 6C33C but it is what it is. This site is my diary and reflects what is going on with me in audio. Here it comes… now I am frustrated to understand the 6C33C.

The next mystery about the 6C33C is aging of a half of the tube if the tube was used only with one heater. Presumably if only half of the tube was used then we have another half is brand spanking new. The cathode was never working, what could be better? Hm… not so fast!

Today I put the 6C33Cs that were working in the Melquiades HF and upperbass channels (1/2 tube) to my tester and measured their mutual conductance that is very objective indicator is the tubes use. So, the whole irony is that the halves that did works for this year showed up approximately 10%-15% being used up. However that halves that were idling and was not even heated showed up 35%-40% of used up. Isn’t it funny?

Rgs,
The Cat


Posted by Genn on 04-09-2006

It is pitty.  Half of the 6C33C could be not used properly - if there will be 3 additional contacts for another triode.  Then you could preserve the second triode, not applying any voltage between plate and cathode.

What you do - you use one triode with normal emmision conditions - when cathode is heated and generate field of electrons.  Simultaneosly - second triode is not heated, but due to the voltage - it is also in use.  Electrons are not generated, but pulled out of the cathode - and this leads to it's destruction.

It would be interesting to observe the life-time of so used tubes, and it's sonic performance in the lifetime.

BR,
Genn


Posted by hagtech on 04-10-2006

You could always pull the grid low on the unused triode section.  And ground the cathode.  That would prevent stripping, I believe.  Just bias it to -100V or whatever.

jh


Posted by Romy the Cat on 06-19-2006
Yes, referring to my initial post on the thread:  something in the Super Melquiades is very freaky. I did not use it for a week and this morning I tuned it on. The amp generally does not sound good for a first 10 minutes and reaches it’s stable operational quality approximately after 40 minutes. However, this time after 6 days of juts sitting there (does not happen frequently in my room) it took for the amp almost a half day of running in order to go back to it’s default quality. Isn’t it annoying?

As I said, I feel that it is most likely die to the memory in formatting of the amps large capacitors. I have seen in it in photographic flashes when the flashes after a few month of doing nothing were not able to serve correct parameters and needed to be sitting under voltage for a few hours (days in some occasions) to return back to the sable duty. In out case we deal with sound that it much more sensitive and finicky then optical density or color temperature.

So, I wonder is the caps so love to be constantly be charged in order to serve good sound then why don’t I disconnect them form anode before I shut down the amp? It would not be difficult to put a relay to lift the cap from the plate. Did anyone go to this extend?

The Cat

Posted by hagtech on 06-19-2006
 Romy the Cat wrote:
I have seen in it in photographic flashes


Wow, this is news to me.  Makes perfect sense, though. 

 he also wrote:
likely due to the memory in formatting


Yes, electrolytics are funny devices.  In fact, similar in many ways to batteries.  And batteries really require a good forming after initial construction.  Their properties can vary all over the map depending on how well the chemicals and metals are formed (meaning first charge).  You've all probably heard of the deep discharge cycle on NiCd types to erase the so-called "memory" effect.  Funny stuff going on inside these chemical/metal capacitors.  They are active devices in a way.

 and then he wrote:
disconnect them from anode


Clever.  But I believe the self leakage of large electrolytics will cause a self discharge in a matter of hours.  Worth experimentation.  Maybe a standby mode where the supply is on but the tubes are off?  This also begs the question, what is the best operating voltage for a given capacitor?  Should it be 70% of rating as suggested by reliability?  Or does perhaps 95% sound better?

jh

Posted by Romy the Cat on 06-30-2007

Interesting,

Turned today my playback - the fist time after a week of break. What a miserable sound!!! I will keep it running all weekend and I hope it will be better tomorrow. 

In the posts above we blame capacitors in the phenomena but I wonder if it might be the tubes themselves, sine SS equipment has no such a phenomena, or at least has it to the lessen  degree. (Though I do admit the I did nit live a lot with many SS amps)

If it might be so, then perhaps it would make sense to drive tube’s filaments when amps are off with ~2V, sort of a stand-by voltage…

I wonder how it will affect the tubes logetivery... and if it resolves the needs for the “amplifiers training”…

Rgs, Romy the Cat

Posted by Paul S on 07-14-2007

I think the issue here may be how long the gear is "off" between sessions. If it's going to be quite a while between sessions, then perhaps having some juice coursing through there would shorten bloom time.  It literally took several weeks of hard running to open up my long-fallow "NOS" ML2s, and now that I  have just moved again, I suppose I will have to go through all this again, too.  I can't say, however, that keeping my old filament-warmed AI plugged in shortend its bloom time versus per-session ramp-up, no matter how long it was between sessions.

I'm sure we've all heard that continuously running at least the filaments [at a lower voltage] lengthens tube lfe, although you couldn't prove this by me, either, since it seems to me that running the plates at, say, 20% would just dissipate the tubes that much more quickly.  One might well get just as good mileage by waiting a while before opening the throttle.

At least in terms of usual-listening-pattern sonics it has worked just as well for me so far to keep the tubes off until I want to listen and then just ramp the system up to operating voltages over a few minutes, +/- as Lamm does.


Best regards,
Paul S


Posted by Romy the Cat on 01-12-2008

I have been contemplating for a week if I need to drain the power caps in my new revision of my power amp. The benefits to live the voltage in there is the cap continue formatting what the amp of off and it minimize the need to break-in or to train the amps. The disadvantage is the when the amp is on then the cold tube’s anode faces the residual voltages – not good…

This post is kind of off-spin of the new Super Milq amps:

http://www.romythecat.com/Forums/ShowPost.aspx?PageIndex=5&postID=6338

With the single-stage amp build around of short-life 6E5P/6E6P and armed with polypropylene caps this question particularly important. The polypropylene caps have very low leakage and they stare voltage for very long time. So, when the single-stages amps go up then the plates see the almost full B+…

I wonder, what if I discharged the cap down to 20-30V to keep the dialectic polarized and then disconnect it without a complete discharging. This way a cold tube will see a little voltage and what the filament is heating up for the first 1.5 minutes of my soft start the 20-30V will be drain out on anode incising conductively. (I apply bias 2 second before B+). I think this way I might to get the best of the both approaches: to keep the caps charged and to protect my fragile driver tubes… I do not have a final decision though and if any “common practice” exist out there then I would like to hear about it.

The Cat

Posted by be on 01-13-2008
What about a 3 position switch?

1. Direct connection between the anode and the B+
2. Connection trough a suitable resistor for a gentle turn on/off.
3. Completely disconnected.

Just remember to turn on and off the heaters at the right times in the on-off sequence between position 3 and 2.

Posted by be on 01-13-2008
Sorry, I meant turn on the heater in position 3 in the start up sequence and turn of the heater after reached position 3 in the turn off sequence!

Posted by Romy the Cat on 01-14-2008

Yes, Be, thanks. I  have revised the Milq’s start up sequence.

With the new scenario when the amp shuts down the B+ will be sitting in the power caps. It does not hurts the tube as they are cold-closed and do not care about anode voltage.  When the amp is turning on then I has 60 seconds for filaments to warm up. It is bad as the bias is not there and in 60 seconds what the bias will arrive is insufficient bias to completely close the tube at full B+ saved in the caps (It will not be full of cause). At that time the cathode begin to heat up and the voltage on anode is too high for a given bias.  So, I will be at the same time when I begin to heat filament to shunt anode to a power resistor and at the time when the bias applied I will stop discarding the power caps. I have additional 30 seconds to continue warming up filaments and stabilizing bias until the B+ applied. At that time the power supple caps will be totally discharged and disconnected from the bleeders.

I have no idea if it is “better” solution but it sounds logical to me.  I most likely will go for it.

The caT

Posted by Romy the Cat on 05-19-2008
Hm, it was 3 days as I did not play my playback and keep my Milqs off. I tuned it on today and… it was the most horrible sound that I heard from those amps. It took 6 hours for them to continue run in order to sound more or less acceptable. I remember ML2 and VK-60 (that used the same output tube) did not demonstrate the sharp dependency to run each day. They loved to be burned-in but not so aggressively as Milq remands. Interesting….

Page 1 of 1 (17 items)