First Watt F3 compared to my Cary Tube Amplifiers(modified)

[erik2A3]

Yeah, 47uf/250V Solen were not too bad to me, either. Motor run types were also pretty good. I know we've talked about single-ended EL-84s before, with a 12ax7 in front. I just got a Leak ST20 in very good shape with just some minor work needed, but it has to be one of the most modern vintage amps I have heard. I'm sure you're familiar with it. It's among those, another being the Marantz 8B, that I have always thought were really great looking. This PP 6BQ5 Leak is punchy and fast! Quite a number of electrolytics, such as cathode bypass, but has recent metallized installed for coupling. What are you using for cathode resistor bypassing in driver/input stages? I've gone the no bypass approach at a time when I thought it was a 'purer' way to go, but I gave up significant gain and frequency response, so it seemed, by not using a bypass cap in the first stage. Honestly, even the cheapest of the cheap electrolytics sounded better than nothing. Little Cerafines worked nicely in that position, though.

[JohnA]

Thanks Mike!

Though it follows what I have heard with my ears, I am surprised someone else reports little to no audible difference between amp designs (as long as they are not overdriven).

[tube fanatic]

In my experience critical level matching is absolutely necessary when comparing different amps with a given pair of speakers. The reasons should be obvious. Apologies to Mike if he made a statement about that which I overlooked. To address some of the comments made- as far as electrolytics go, they can sound just fine, and bypassing them with low value film caps (or, dare I say it, even ceramic disc types) can be useful in preventing parasitic oscillations in some amps. Regarding choke vs. capacitor input filters, the choice can be made based on the design of the amp. For low power SETs (and SEPs) in which the current variations from zero to full output are minimal, capacitor input filters work extremely well. Generally, I use 3 caps with the input cap having the largest value, and then use somewhat smaller caps after. This allows the first cap to act as a reservoir for the others. It's a technique modeled and used by the industrial power supply company for which I did some work years ago. I may be able to get hold of the technical details from one of my former co-workers if anyone wants to see the research. It's nice to see some detailed technical discussion going on!

Maynard

[Deang]

Thanks Mike!

Though it follows what I have heard with my ears, I am surprised someone else reports little to no audible difference between amp designs (as long as they are not overdriven).

I always feel a little bad for folks who can't hear real good.

[erik2A3]

Choke input to filter designs can also work extremely well -- In SE or PP alike. One also gets the benefit of better regulation. I've used both C to input and L to input with good results in both cases.

[erik2A3]

Maynard: It's interesting that for a cap input to filter you describe using a larger value for the first capacitor. I have tended to do the opposite. I agree that the kind of filter one chooses to use is at times related to the design, and the case of capacitor input B+ supplies, I would add, as I'm sure you know, that the amount of capacitance used is often determined by the the type of rectifier -- not just in terms of solid v.s. hollow state (vacuum tube), but, in the case of the usual 5V rectifier varieties, the specs for that rectifier when used in cap input designs. For example, a GZ37 (if I remember this right! Lol!) has a max. cap input to filter rating of, I believe, something in the range of 4uf. With a of course a small amount of wiggle room beyond that usually being acceptable (though I personally have learned not to exceed the published spec by more than a few uf), we know that a robust rectifier like the common 5AR4 has an input cap value rating that is significantly higher.

Another consideration is of course the fact that capacitance values influence the rise in voltage of the HT supply. Using higher values will require often require large series resistances to knock plate supplies back down to where they need to be, where one might get by equally well in terms of ripple filtering by using just a few to several uf for the first cap going into the choke (or resistor -- my preference being a choke). One will have a a less aggressive and more immediately useful anode voltage where a giant series resistor isn't needed. After the choke, than the larger values, in my case, will hit the playing field without worry of damaging an expensive rectifier. Mullard GZ 37s are not cheap these days!

For those with capacitor interests and inclinations that the the of input to filter cap used is very much in the same boat of debate as are capacitors used for interstage coupling and loudspeaker crossovers. Oil vs poly, teflon, mylar, film types, and so forth. I've used them all.

[tube fanatic]

Maynard: It's interesting that for a cap input to filter you describe using a larger value for the first capacitor. I have tended to do the opposite. I agree that the kind of filter one chooses to use is at times related to the design, and the case of capacitor input B+ supplies, I would add, as I'm sure you know, that the amount of capacitance used is often determined by the the type of rectifier -- not just in terms of solid v.s. hollow state (vacuum tube), but, in the case of the usual 5V rectifier varieties, the specs for that rectifier when used in cap input designs. For example, a GZ37 (if I remember this right! Lol!) has a max. cap input to filter rating of, I believe, something in the range of 4uf. With a of course a small amount of wiggle room beyond that usually being acceptable (though I personally have learned not to exceed the published spec by more than a few uf), we know that a robust rectifier like the common 5AR4 has an input cap value rating that is significantly higher.

Another consideration is of course the fact that capacitance values influence the rise in voltage of the HT supply. Using higher values will require often require large series resistances to knock plate supplies back down to where they need to be, where one might get by equally well in terms of ripple filtering by using just a few to several uf for the first cap going into the choke (or resistor -- my preference being a choke). One will have a a less aggressive and more immediately useful anode voltage where a giant series resistor isn't needed. After the choke, than the larger values, in my case, will hit the playing field without worry of damaging an expensive rectifier. Mullard GZ 37s are not cheap these days!

For those with capacitor interests and inclinations that the the of input to filter cap used is very much in the same boat of debate as are capacitors used for interstage coupling and loudspeaker crossovers. Oil vs poly, teflon, mylar, film types, and so forth. I've used them all.

Hi Erik! I'll try to get hold of the technical paper on that issue. You are correct, of course, that with tube type rectifiers one has to be careful not to exceed the peak current rating of the tube which necessitates using a relatively small input cap. As you are aware, if a choke is used in the pi section filter instead of a resistor, the tube won't "see" the capacitor that follows regardless of whether it's very high in value. With a filter resistor, it's more of a concern. I'm using solid state rectifiers in most of the amps which have enormous peak current handling capability so I don't usually think about it, particularly since the power xfmr gets loaded down at the initial turn on anyway. Nice to chat with you again!

Maynard

[erik2A3]

Hey, Maynard - nice to chat with as well!

Totally agree with you -- your thinking is totally clear to me now regarding SS rectifiers related to your power supply design. The bypass you mention is useful! Some designers will bypass the high voltage secondaries to ground on the AC side just before the rectifier, as you described, as a means of dealing with HF parasitic oscillation before it gets into the B+ supply, and subsequently possibly also the audio. I have built a couple of amplifiers like that which also happened to use very large input to filter capacitance.

erik

[ClaudeJ1]

Jubs fully driven by F3s

Papa Pass is the man.

[erik2A3]

Indeed! a gifted designer, and one humble enough to share many of his own designs with DIY enthusiasts. I have found myself on a similar track as he with High efficiency Lowther drivers on an open baffle. The sound from this arrangement is astonishingly realistic. The Pass First Watt amps are real favorites for these single range drivers, and horns in general.

Bass duties for me are delivered by way of rearward, room-corner-facing, split industrial La Scala bass bins wired out-of-phase. I've never heard anything so lifelike -- though our old Klipschorns were up there too!