GoodSoundClub - Romy the Cat's Audio Site

In the Forum: Melquiades Amplifier
In the Thread: A different breed of 6C33C amplifier.

Post Subject: Reliability and stability of 6C33C?Posted by Romy the Cat on: 9/15/2011

Well, I would say that “something is wrong” as what you report it is not how it shall be. I had my doubts about the rational to use 6C33C in direct heating coupling but the “danger”
in direct coupling uselessly coming from the plate of the driver tube and from breaking of PS sequencing. Abyhow, I think the problems you are experiencing now are normal for new amp and exposure to new tubes. As both of them mature you shall not have those problems. Let look at what we have in the hands.

KOTriode wrote:
Totally disappointed today, while I was listening to music, the left channel make a small pop and quit. The bias meter showed about 25ma bias instead of 200ma. Test the tube on bench showed that the tube current stay about 25-50ma max. This tube is gone abrubtly instead of slow death after 31 hours of playing.

I would call it unlikely. Sure any tube can catastrophically fail and you might be unlucky in this case but I would say that it is most likely not what happened. I would rather blame your regulators and PS then the tube. From the mid 90s I use 6C33C. I use a few BAT amps, few Lamms and few Milqs, in all cases it was multi-amping and I went over many of 6C33C. I have perhaps 2-3 tubes that failed with what I call “catastrophe”. So, I fell the reliability is not the factor for 6C33C at all, in fact it is VERY reliable tube. The stability of the parameters is totally other mater – it might be anything you can imagine but still the tubes with ran way paramagnets in my view are perfectly operational.

Make sure that your PS are reliable and that you do not loose bias supply “at the mid of the night”. Then the sequencing. Here is how it must be: 1) heat the filaments for at least 2 min 2) Apply bias and lock it value. 3) Apply plate voltage. I do not know how bias your 6C33C but I presume that it might have some kind of bug that kills your 6C33C.

KOTriode wrote:
Took another brand new tube and plug into the amp with just heater power on, nothing! Turn out a brand new tube has dead heater, amazing!

Never had it I think. What I has however was that the new 6C33C have the pins covered with some kind of protective anti-corrosive shit that does not conduct electricity. It is not to mention that if you use bad tube sockets then the heater female pins might burn up very fast (not after 31 hours of playing of course). So, you need to make sure that you clean/scrub the 6C33C pins what you put need tube in. If there is no ignition on the filament then jerk the tube in the socket vaoalsntly and it will scrub the pink and you will have your filaments back. It is not DHT tube and the filaments are not fragile in this tube. It is impossible to shake this tube to the point that filaments will be damaged.

KOTriode wrote:
Take out another tube, heat the filament for an hour, and check the bias, this time it is less than 100ma, another goner.

I do not recognize it as goner. This is very normal and all that you need to do is to drive bias down until the tube gives to you your expected current. The bias amount is absolutely irrelevant for this tube. You might have initial 45V or 110V – that all is normal. As the tube get break-in the bias amount will more come to the more or less expected value. KOTriode, if you remember in my initial reaction to your design I was very skeptical in your idea of direct coupled of 6C33C. In direct coupled applications the plate of driver acts as bias of the output stage. If you are accustomed to nice DTH with stable and predictable expectation of bias then this might not work with 6C33C as this tube has one parameter that is absolutely irrelevant: bias voltage. Take a look in the tube you need to have stable gain (self-explainable), stable plate current (as it set plate impedance). The only one parameter that is not being seen by your speakers is bias. In the situation with fix bias I would like to have max control over bias and unfortunately in case of direct could you have no room to maneuver. If I am right then I have no pleasure in it but if you insist to be direct coupled then you might do some creative tricks. You might for instants to intrude with kind of supplementary bias. It means the let say 50V would come from your driver tube and another adjustable 50V would come from your supplementary bias. This way you will have option to accommodate any 6C33C. In Melquiades can play absolutely any 6C33C. In fact I select the tubes in Milqs ONLY by their gain and I never pay attention to the bias voltage. No matter what current the tube gives when it new I know that I have always reserve bias to set it to be at the expected current.

KOTriode wrote:
Check reliabilty of 6C33C on the net and found this:
I can also sell you a carton of 50 for $900, including ground shipping within the continental US. At this price they are "as is" - I will not replace any which are broken, out of spec or dead (they're a bargain anyway). Reports suggest you can expect 10-20% to be significantly out of spec. So, it looked like that these tubes left the factory untested and it is up to the buyer to sort them out. I order some " Ulyanov " tubes on Ebay, anybody have tried them yet? What label of these tubes proven to be reliable?

OK, a few things. It is possible, unlikely but possible that during your trip in Asia when you “picked up 4 these tubes for less than $50” you just was given the set of garbage tubes. Some people who use a lot of tubes trash any faulty one and some of them just put it in separate box and the box got filed up then they sell it as “set of tubes”. It is possible that you got your “winning” lottery ticket. I never do it with tubes and if I have any problematic tube I rather break it glass (to avoid temptations to”fix it”) and trash it. However I did it a lot with compression drivers when I was experimenting with them: the drivers that were un-recoverably dead I combined by pairs and sold in far countries…. So, there are a lot of assholes out there…

The comments that you quoted “10-20% to be significantly out of spec” are correct but even being significantly out of spec does not mean the they are not operational as the flexibility to set your custom bias for each out of spec tube make them all perfectly usable. I made a LOT of experiments trying to detect if the 6C33C tube that need 45V bias or 120V bias have any difference in sound. After 100Hr of burn and the bias normalizing to let say 57V bias or 80V respectfully I was not able to detect any sonic distinctions. Another thing: do not believe that anybody would sort them out – no one does it. Well, some do, Vladimir Lamm for instance burn the 6C33C for 72Hr, measures them and match then by gain. So, he used to sell his confirmed 6C33C at $80 a tube and it was 10 years ago. I am sure now he sells them for $180. The Lamm’s tube are fine for some kind of mid-west gynecologist who never saw a screwdriver as he can’t not use it in your work but absolutely no necessary for the people who comfortable to measure plate current and tube gain.

There are few types of 6C33C tubes out there. The Ulyanov tubes are the most common and perhaps 95% of them are made in Ulyanov. There are productions from other plant and by other vintages but they are no better from the perspective of reliability or stability of parameter then Ulyanov. Again if you have full bais control then you shall not have any problems with reliability and stability of 6C33C. Let presume that your “lucky” buy of 6C33C in Asia was just an accident. Hey, I promise you that it was not the last $50 you have lost in audio….

Rgs, Romy the Cat

Rerurn to Romy the Cat's Site