Anybody looked at tube perveance values?

[Old Tele man]

Has anybody looked into (investigated / calculated) the Perveance (G) values for the different vacuum tubes they use?

I ask because I have (for 6L6 and 6V6 family) and others have asked about the 5687 twin triode.

[soundmasterg]

I'm not sure what they are honestly...or what relevance they have?

Greg

[Old Tele man]

I'm not sure what they are honestly...or what relevance they have?

Greg

Here are three illustrated examples of Perveance (G):

• DIODE: Ip = G*(Vp)^(3/2)

• TRIODE: Ip = G*(Vg + Vp/µ + e)^(3/2)

• TET/PENTODE: Ik = (Ip+Is) = G*(Vg + Vs/µ1 + Vp/µ2 + e)^(3/2)

Perveance (G) is:

1) the quotient of Ip over Vp^(3/2), ie: Ip/Vp^(3/2);
2) the ratio of space-charge potential (Vp^(3/2) to beam kinetic energy (Ip);
3) the coefficient of proportionality between space-charge limited current (Ip) and effective anode voltage (Vp^(3/2);
4) a tube constant dependant solely upon physical electrode geometry (spacings, diameters, etc.);
5) expressed in "Perv" units, ie: 1 Amp/(V^(3/2)), amps-per-volt^(3/2).
6) tube constant almost NEVER published by tube manufacturers.

[murrayatuptown]

perveance Definition and Meaning adds remark that this parameter changes with tube aging. (Still nothing of practical importance, yet).


Morgan Jones' "Valve Amplifiers" book says practical significance is high perveance tubes like 5687 anodes can swing anodes closer to 0V, increasing voltage swing and efficiency. I'll take a wild guess, since 5687 was a 'computer' tube (diode logic, ENIAC?), maybe the closer it can get to 0 and plate voltage, the wider the High and Low logic levels and better noise immunity. Maybe a hollow-state analogy to rail-to-rail op amps whose output can swing much closer to the power supply rails (bipolar frequently for op amps, B+ & ground for tubes).

https://books.google.com/books?id=Qd...veance&f=false

I am not sure, but think this makes me think lower internal resistance (rp). While that may be true with 5687 compared to smaller 9-pin tubes with (roughly similar in mu, without looking it up) like 12AU7, the reason may not be solely the perveance rating (if even related). Maybe perveance rating isn't a feature, but a figure of merit.

I read somewhere else that makes them capable of small power applications (dwarf output tubes).

[Old Tele man]

I suspect the "changes with tube aging" can attributed more to cathode emissivity decline than anything else, because all of the 'published' Perveance parameters are physical spacings and geometries, which theoretically are static and non-changing.

As for the "digital" use of the high-perv 5687 tube in computers, it makes sense being able to drive/slam the tube between cutoff and saturation with minimum rise/fall time, because that allowed faster computation speeds since rise/fall times are "wasted time" in digital circuits: it's only the ONE and ZERO states that count...and, time spent transiting between them is wasted time and power: 0W = 0.0A × 5V and likewise 0W = (any)A × 0.0V.

[soundmasterg]

I have a bunch of 5687's that were in a tube powered surplus nuclear radiation detector that I got for free at a surplus place like 10 years ago. I was going to try them in a build, perhaps as a phase inverter, but I didn't get around to it yet.

Greg

[Old Tele man]

I have a bunch of 5687's that were in a tube powered surplus nuclear radiation detector that I got for free at a surplus place like 10 years ago. I was going to try them in a build, perhaps as a phase inverter, but I didn't get around to it yet.

Greg

They will be VERY easy to drive into saturation--almost too much so--but you should get almost 100% voltage swing from whatever plate voltage you use. As mentioned above, they should excel at Class-B operation, having almost no cross-over distortion to contend with.

[nickb]

I have a bunch of 5687's that were in a tube powered surplus nuclear radiation detector that I got for free at a surplus place like 10 years ago. I was going to try them in a build, perhaps as a phase inverter, but I didn't get around to it yet.

Greg

Have you seen what that are going for on the 'bay'?

[soundmasterg]

Have you seen what that are going for on the 'bay'?

I haven't looked lately but perhaps I should. These are all used ones...no new ones in the bunch, but they do all work.

Greg

[Old Tele man]

I haven't looked lately but perhaps I should. These are all used ones...no new ones in the bunch, but they do all work.

Greg

A quick gm-check should establish their usability.

[nickb]

I haven't looked lately but perhaps I should. These are all used ones...no new ones in the bunch, but they do all work.

Greg

Seems they might make a decent push pull output tube with 7.2W total plate dissipation. You should be able to get 5 to 10 watts. I see the plate resistance is very low and that a maximum of 6.6mA grid current is specified. Does that suggest that they ere typically driven with the grid positive to get the low sat voltage? The low Rp and high currents would make them switch quickly.

[soundmasterg]

A quick gm-check should establish their usability.

They all test fine on my B&K 747B tester, which is mutual conductance, but the scale is arbitrary, so I'll have to try them in a circuit to really see, which is the best way anyway. I was busy from 2010 to 2015 in EE school, and since then I've been trying to get some sort of career established to pay off my school loans, and have been trying to catch up on all the necessary things I set aside while I was in school. It will be awhile still until I have the time to build up something to use those 5687's, but they can keep company with all my other projects.

Before I can get to those, I have to finish building up a particular solid state preamp board and test it with a tube power amp on a project I was doing with R.G., rebuild/restore an AIMS Dual Twelve, a Silvertone 1484, a late 60's Baldwin solid state amp, a Trace Elliot solid state bass amp, a 1963 B15N including cabinet refurbish, a Weber JTM45 kit I got cheap that the previous owner badly built. Then there are my own designs/projects from scratch...a Vox AC100 build with another higher gain channel, a Supro Tbolt build, an 18 watt Marshall build, a Revibe build, and more. Unfortunately they are all begun but not finished....I'm bad at that sometimes.

Greg

[soundmasterg]

Seems they might make a decent push pull output tube with 7.2W total plate dissipation. You should be able to get 5 to 10 watts. I see the plate resistance is very low and that a maximum of 6.6mA grid current is specified. Does that suggest that they ere typically driven with the grid positive to get the low sat voltage? The low Rp and high currents would make them switch quickly.

I've read that the hi-fi guys really like them as phase inverters. It does look like they would make a good low power output section too.

Greg

[murrayatuptown]

Here are three illustrated examples of Perveance (G):

• DIODE: Ip = G*(Vp)^(3/2)

• TRIODE: Ip = G*(Vg + Vp/µ + e)^(3/2)

• TET/PENTODE: Ik = (Ip+Is) = G*(Vg + Vs/µ1 + Vp/µ2 + e)^(3/2)

Perveance (G) is:

1) the quotient of Ip over Vp^(3/2), ie: Ip/Vp^(3/2);
2) the ratio of space-charge potential (Vp^(3/2) to beam kinetic energy (Ip);
3) the coefficient of proportionality between space-charge limited current (Ip) and effective anode voltage (Vp^(3/2);
4) a tube constant dependant solely upon physical electrode geometry (spacings, diameters, etc.);
5) expressed in "Perv" units, ie: 1 Amp/(V^(3/2)), amps-per-volt^(3/2).
6) tube constant almost NEVER published by tube manufacturers.

Odd that 'G' was used for Perveance...G is conventionally conductance, inverse of Resistance; G=1/R.

Lower case g probably tends to indicate ac, and gm indicates the 'older' term 'mutual conductance' replaced by the 'newer' term transconductance, depending how old the book is.

My books(s?) used K for perveance. Doesn't change the limited practical usefulness of the term.

[Old Tele man]

Some old college textbooks used P, some used K, but most newer (mid-1950-60) books use G, probably (strictly MY opinion) because of it's relationship to conductance (g) which is a dynamic(*) Amps-per-Volt, and perveance (G) which is theoretically a static Amps-per-Volt^(3/2).

However, I do see that P is used by the high-power ION-drive scientists, and they've assigned "perv" as the units instead of writing out Amps-per-Volt^(3/2).



(*) gm varies as cube-root of plate current; per theory: gm = (3/2)*G^(2/3)*Ib^(1/3)