• WARNING: Tube/Valve amplifiers use potentially LETHAL HIGH VOLTAGES.
    Building, troubleshooting and testing of these amplifiers should only be
    performed by someone who is thoroughly familiar with
    the safety precautions around high voltages.

6C33 design comments?

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
A couple things jump out at me, one minor, one potentially fatal.

Minor: the phase splitter is not a particularly great one. I don't know why the designer chose one that's inherently unbalanced, has unequal time constants on each side, and has asymmetrical distortion side to side.

Fatal: The use of fixed bias for output tubes with a very high transconductance is a near-guarantee of output stage runaway and failure at some point.
 
Fatal: The use of fixed bias for output tubes with a very high transconductance is a near-guarantee of output stage runaway and failure at some point.

Yes, that is a risk but I don't agree that it is "a near-guarantee of output stage runaway and failure at some point" it has not been a problem for me neither in my SET with 6C33C nor with any of my OTLs. I know that it can be a problem with tubes like 6336 or 6AS7 as I have had that experience amy years ago when experimenting with OTLs.

Regards Hans
 
What do people think about this design:

Short answer:
An externally hosted image should be here but it was not working when we last tested it.


Elaboration:

The phase splitter is terrible. This sort of circuit may be acceptable for the simplest projects where fidelity isn't a prime concern. Otherwise, it is unacceptable due to its inherent imbalance, and especially the unbalanced harmonic distortion which is endemic to this entire family of phase splitters. It also is not good that the inevitably required AC balance pot is connected to the 300Vdc rail. These "pots" aren't known for their insulation. A shocking experience awaits here :hot: An LTP or cathodyne would have been the superior choices here.

This splitter stage is connected to two, separate voltage amps. Even though these are two triodes in one bottle, this does not guarantee balance between sections, even if perfectly matched out of the box, they will not stay that way over time and useage. Another source of AC imbalance and unnecessary distortion. A differential should have been used here. Possibly, inadequate grid drive for the finals.

RC coupling into fixed biased finals: combined with gNFB guarantees wicked clip behaviour. A transient overshoot will turn on the gk parasitic diode and increase negative grid bias on the finals. That excessive bias will persist until it leaks off, and while it does, the finals operate in a less linear manner. Needless to say, this does not help the sonics at all.

Fixed bias on 6C33Cs -- a big no-no, and definitely not recommended by the manufacturer. These triodes with the uber low R(p)s and high g(m)s are as delicate as germanium power transistors when it comes to DC stability. You either cathode bias, or use a DC servo, or you'll have to watch 'em like a hawk so's to be able to cut power when (not if) they go red plate.

It looks like all the RC couplings cluster their phase shifts around roughly the same turn-over frequencies, a big invitation to instability. Also, looks like the gNFB is excessive here.

There are better designs out there. I'd pass (no pun intended) on this one.
 
So that you actually gets some idea of what circuit you should build, not just how bad this one is, someone previously recommended that I try the crystal palace topology from Morgan Jones. I took a long look at it and it seems like it would work well. Just need to modify the output stage for the different tube accordingly.
 
JoshK said:
So that you actually gets some idea of what circuit you should build, not just how bad this one is, someone previously recommended that I try the crystal palace topology from Morgan Jones. I took a long look at it and it seems like it would work well. Just need to modify the output stage for the different tube accordingly.

I've been stashing up 13E1s because one day I want to build the Crystal Palace (the floor in my listening room has recently been re-enforced and I've new three-phase feeders into our plant room...) I'll have a look at MJ again tonight.

Problem I have is that this'll only be my second amp (previous was a Bevois Valley), and I just love the look of 6C33Cs. The only tube I like more is the 13E1 (which is built like a tank). 6C33s are CHEAP and I would like to build a nice amp with them too...

I'm still trying to work out how to test them on a Hickok 6000A or 539C.
 
From what I know the sound is very impresive on that pp design.
In the end you have to deal with sound not with paper grafics.I agreed with some improvements but I think that pp sch. is very good.

Take a look at these beauties after some modds. on the schematics.The sound is equal with exterior design.
 

Attachments

  • truth.jpg
    truth.jpg
    87.6 KB · Views: 1,535
Look for one with cathode bias and an input stage using properly balanced diff amps (the Crystal Palace is an excellent example of that).

Sturgeon's Law (Revelation) is universal and also applies to 6C33 designs, unfortunately.

edit: With some minimal output stage design work (different output transformer, cathode bias), you could indeed use the CP driver stage. Yes, it's a bit over the top, but it's solidly engineered. And it's a great-sounding amp.
 
Fixed bias on 6C33Cs -- a big no-no, and definitely not recommended by the manufacturer. These triodes with the uber low R(p)s and high g(m)s are as delicate as germanium power transistors when it comes to DC stability. You either cathode bias, or use a DC servo, or you'll have to watch 'em like a hawk so's to be able to cut power when (not if) they go red plate.

Seems you have got a very poor quality set of these tubes.... I cannot confirm your experience after nearly 2 years of work with them. Certainly, this is not a beginner's tube, and circuit design has to be made very carefully.

I am running 2 pairs of 6C33C-Bs in my Circlotron amp, where they are running quite near their limits - and even so I never have had any problems with runaway.... I run them with 260 V plate and about -100 to -130 V bias, depending on chosen working point. Evidently all supplies are stabilised (MOSFETs). Due to its compactness the entire amp is cooled by a temperature regulated silent fan - as are the sockets of the big tubes which are mounted with 5 mm air gap below the chassis. The amp gives about 60 W rms output power on a 8 Ohm resistor (over an output transformer, mains toroidal 115/30V, 160 VA).

The only annoying thing is that these tubes need half an eternity to reach their final working point when heating up - especially, when they are being run with low power (the amp still performs perfectly with about 50 mA quiescent current per tube only). I heat up for 5 min. prior to switch on plate voltage (also to get sufficient pressure in the mercury rectifiers). And then, 20 min. later, currents have risen almost to their final values. So certainly one day I will construct servos, but - quite a challenge with the floating supplies of the Circlotron... I can run my tubes safely up to 150 mA quiescent current (but there's no need of), with still higher values the plates are starting to glow. But I never had problems with runaway...

Also heater supply stabilization helps to reduce current drift. Minute variations in cathode temperature have a big impact on working points.

Tubes are preselected to their working point (260V, 100 mA). And - I strongly recommend this to anyone using this tube with voltages over 150V - doing a hard burn-in at low voltage and high current at full plate dissipation for several hours. In this way, any remaining oxide or dirt particles on the grid or plate, who often are the reason for flashover at higher voltages, are burning away, the vacuum is stiffened, and then in my experience the tubes are safe to any kind of use.

Yes, this is not a beginners tube....

Good luck!

Ulihttp://www.uli.de/dscn5068-kl.jpg
 
ulibub said:
Seems you have got a very poor quality set of these tubes.... I cannot confirm your experience after nearly 2 years of work with them. Certainly, this is not a beginner's tube, and circuit design has to be made very carefully.

Possibility != Inevitability

Not every set of first generation germanium power transistors poofed either. You can go over to the solid state forum and find the occasional gadfly who swears by them, and who spends considerable time and effort in tracking down NOS samples of same. Doesn't alter the fact that it can happen, and has happened for whatever reasons. Not saying that you shouldn't use 6C33C's, or that it's always wrong to fix bias, however, I would give careful consideration to what the manufacturer suggests as to operation before deciding to deviate from those suggestions. Same goes for using 6AS7s. (Over current detect and interrupt circuits would be a good idea for cases like this.)
 
nickds1 said:
Now I'm confused - this design is either excellent or rubbish, possible both at the same time (parallel universes?).

Well, it would be "rubbish" only if it were totally unlistenable due to severe distortion. You can be sure that that's not the case. However, I question his judgement over this particular project. What he's done essentially is to make a huge Lafayette-KT-92. Now, the KT-92 was originally intended as a simple, very basic, kit for the DiY n00b. In that capacity, it met the prime design criteria, and was also serviceable as an inexpensive design for less than premium applications, such as inexpensive record players like those you gave to the kids so's they'd keep their mits off Mom and Dad's stereo equipment.

If this is what you need, then go for it. It'll work, probably sound better than many a solid state "Big Box" system (which can sound surprisingly bad) but don't expect a premium performance because you're not gonna get it. If that's all you're after, then keep the expense down, use 50C5/17C5/25C5/12C5/35C5/etc. and make a KT-92 (but do spend a bit extra and include a proper DC supply that operates from a power xfmr).

If you're gonna make the effort, and go to the expense, of building something more powerful, then why bother with an inferior topology? Wastea time & $$$$ AFAIC.
 
SY said:
Look for one with cathode bias and an input stage using properly balanced diff amps (the Crystal Palace is an excellent example of that).

Thanks, I'll keep looking, but recomendations always welcome!

Sturgeon's Law (Revelation) is universal and also applies to 6C33 designs, unfortunately.
edit: With some minimal output stage design work (different output transformer, cathode bias), you could indeed use the CP driver stage. Yes, it's a bit over the top, but it's solidly engineered. And it's a great-sounding amp.

As I have a whole bunch of 13E1s, I would use those probably on a CP, but I have a healthy respect for my own inadaquacies, and I am really not sure I'm up to a CP yet (I'm still pretty new to tubes). This is not a lack of self-esteem, it's called being pragmatic... I think I'd prefer a simpler driver & PSU requirements...
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.