how much filtering can a 5Y3 take?

[phsyconoodler]

The reason I'm asking is that I have a Mesa Lonestar special on the bench and the filtering is 2-220uf caps in series,for about 110uf total.
  The amp and quite a few other ones I've worked on,eat the 5Y3 rectifier tubes for lunch.
  They last about a year or even less and then crap out and blow the fuses.
You can't use just the diode rectifier as the 'progressive linkage' only goes through the 5Y3 when the amp is set for 5 watt or 15 watt oepration.
   I know that a Sovtek 5Y3 is not a 'real' 5Y3 but even so it should not have 110uf as filtering.
I think it's a design flaw but I'm waiting for a lashing from Mesa about this.
   It seems every Lonestar I get in has the same issue.

[Dave]

Many people have brought amps to me that some "tech" had done a cap job on and then the amp wouldn't work right. Normally what I have seen is that people try to cram in 80-100 microfarads claiming that it'll give them more filtering and a tighter bottom end, but what it usually gives them is an arching 5Y3. I have used 40Mfd, but I think that is even a little much.

Dave

[Joe6v6]

Most of the tec sheets say 10 or 20uf (*input cap), I saw one that said 32uf max. I have used 40uf in this position and although I was a little worried there seemed to be no ill (short term) effects. 110uf seems way above what would be acceptable.   .    Joe

[tubenit]

32-40uf, I think is about max with a good one.

Having said that, one can always use a lower value for node A (to OT) and a higher value for node B (power tube screens).  I have done this on a Carolina Blues Special with a 6AX5 and it sounded fine to me.

With respect, Tubenit

[phsyconoodler]

Ok,I got a service update on the amp.There are a couple of resistors on the board near the filter caps that have to be removed and that's supposed to help with the 5Y3 issue this amp seems to have.Early models need to have these resistors removed.I did that and the amp is up and running again,but time will tell if it lasts.
  To me 110uf is killing these tubes.

Frankly I can't stand the sound of this amp,but I have to fix them so there you are.

[FYL]

Capacitance is only a variable, as are plate supply resistances and raw voltages. All three should be considered when designing a valve-based power supply, as they determine max transient as well as steady-state current - a 5Y3GT is spec'ed at 440 ma design center max steady-state and 2.5A max peak. If you significantly exceed these values, the valve self-destructs.

In a nutshell: low supply R and high voltage => low capacitance, say up to 20µF for a real 5Y3GT; highish supply R and medium voltages => higher capacitance, say up to 47µF or so. Anything between 16 and 33µF is usually safe.

[tubeswell]

I too have used 40uF on a 5Y3 with no discernible problems, but I wouldn't go much higher than that. 100uF is way too high. 16-20uF is the norm

[HotBluePlates]

What did those resistors do? What did they connect to?

There are a lot of factors that go into determining how much capacitance you can use with a rectifier. There are about 18 straight pages in RDH4 about selecting and using rectifier tubes, and calculating what's happening. That's probably more pages than was spent on how to design an output stage.

It's all about how much current is flowing through the rectifier, especially during current peaks. And it's a lot more current than you'd think. The RDH4 design example has a amp that draws 125mA from the power supply, and the peak current through the rectifier is nearly 400mA. Of course, RDH4 also shows that there is not a lot gained from increasing the main filter cap much above 16uF (it draws one curve on a particular graph for caps that are 16-40uF, where other curves are shown for 4uF and 8uF).

The only way to keep from popping rectifiers when you have too much capacitance is to either lower the capacitance or increase the power supply impedance, usually by putting resistors in series from the high voltage winding to the rectifier plates, in order to limit current.

If the resistors you removed could have been between the PT and the rectifier (as in when the PT windings are connected to the p.c. board), then it is probably a step in the wrong direction. Hard to know without more info.

It is noteworthy that the typical condition in the 5Y3 data sheet is for 350vdc, 20uF cap and 125mA dc output, with a supply impedance of 50 ohms per plate. If you went through the RDH4 procedure using those numbers, you'd find this is a limiting condition that would pop rectifiers if exceeded. But you could go up on the capacitance if the supply voltage were reduced, if transformer winding resistance were increased (or series resistors added), if current drawn were lower, or any combination of these.

[HotBluePlates]

Oh... and FYL's quoted 440mA steady-state corresponds to what I called "peak current". The peak I was referring to is the the peak value of current drawn during each charging pulse, and so you could call it "steady-state"... that peak is happening constantly, even if momentarily. This is as opposed to a one-off surge that the tube could handle, but not sustain repeatedly.

On another angle, I don't see what's gained by using a 5Y3 and 100uF+ of capacitance. The 5Y3 is fairly high resistance, so it's not as though slapping a big capacitor gives you good bass; the 5Y3's resistance will tend to counter any improvement by limiting the peak transient current drawn.

[phsyconoodler]

The resistors in question are connected to pin 3 of the power tubes.There are a set of 100 ohm 1 watt resistors right beside them.They are some sort of cathode bias resistors that are part of the progressive linkage design.There are no details on the linkage and it's schematic is a closely guarded mesa secret,I suspect,because it does not show it on the schematic.
  As far as I'm concerned,a good VVR kills it anyway.I do not like the way these amps were designed.

[FYL]

Yup. The 440 mA steady-state figure translates to app. 150 mA current drawn (say, a typical amp with a 6V6 PP and a few 12AX7's), while the peak value is related to the initial cap charge.

[FYL]

I do not like the way these amps were designed.

Contemporary Mesa amps are IMO a mess.

[echuta13]

AB764 Vibro-Champs and Bronco's used a 40uf cap.  It's more than a 5Y3 should have IMO, but it works.

[HotBluePlates]

Perhaps, but I can't stress enough that there are multiple factors to be considered. For example, we don't know the resistance of the PT. To find this, you should find the resistance of the high-voltage secondary, and add to that the value of the primary resistance stepped up by the turns ratio. This resistance acts to limit the current drawn by the rectifier.

In a big amp where you want excellent regulation, you keep this resistance as low as possible. But the big amp would also use a bigger rectifier that wouldn't mind the extra current. And if great regulation was needed, you'd have a fairly smallish filter cap, followed by an electronic regulator of some kind. So it still all would even out.

[FYL]

AB764 Vibro-Champs and Bronco's used a 40uf cap.  It's more than a 5Y3 should have IMO, but it works.

The 125P1B / 022772 PT shows quite a high winding R, with app. 5R pri and 340R total sec (170R per half) for a source R of app. 200R using 117V US mains.

[phsyconoodler]

If I owned the amp I would drop the filtering down or set it up to use a 5AR4.The amp sounds really stiff and lacks life.
  Some guys like them but after building my own they sound generic at best.Stiff,lack of harmonics,distortion is raspy,low wattage settings sound cheap.
  Then again if I did own it I would sell it as quickly as possible.
Yuck!